diff --git a/.travis.yml b/.travis.yml index 708d3620e8a7..ebe91939a61d 100644 --- a/.travis.yml +++ b/.travis.yml @@ -97,12 +97,12 @@ matrix: sudo: required compiler: gcc node_js: - - "6" + - "7" services: - docker before_install: - - nvm install 6 - - nvm use 6 + - nvm install 7 + - nvm use 7 - docker pull trzeci/emscripten:sdk-tag-1.35.4-64bit env: - SOLC_EMSCRIPTEN=On diff --git a/CMakeLists.txt b/CMakeLists.txt index 24bea3b3e7d6..a0b526215d58 100644 --- a/CMakeLists.txt +++ b/CMakeLists.txt @@ -8,10 +8,12 @@ include(EthPolicy) eth_policy() # project name and version should be set after cmake_policy CMP0048 -set(PROJECT_VERSION "0.4.19") +set(PROJECT_VERSION "0.4.20") project(solidity VERSION ${PROJECT_VERSION}) option(SOLC_LINK_STATIC "Link solc executable statically on supported platforms" OFF) +option(LLLC_LINK_STATIC "Link lllc executable statically on supported platforms" OFF) +option(INSTALL_LLLC "Include lllc executable in installation" OFF) # Setup cccache. include(EthCcache) @@ -43,9 +45,10 @@ configure_project(TESTS) add_subdirectory(libdevcore) add_subdirectory(libevmasm) add_subdirectory(libsolidity) -add_subdirectory(solc) +add_subdirectory(libsolc) if (NOT EMSCRIPTEN) + add_subdirectory(solc) add_subdirectory(liblll) add_subdirectory(lllc) endif() diff --git a/Changelog.md b/Changelog.md index 9a30986a6710..b2e88eca6aa7 100644 --- a/Changelog.md +++ b/Changelog.md @@ -1,3 +1,31 @@ +### 0.4.20 (2018-02-14) + +Features: + * Code Generator: Prevent non-view functions in libraries from being called + directly (as opposed to via delegatecall). + * Commandline interface: Support strict mode of assembly (disallowing jumps, + instructional opcodes, etc) with the ``--strict-assembly`` switch. + * Inline Assembly: Issue warning for using jump labels (already existed for jump instructions). + * Inline Assembly: Support some restricted tokens (return, byte, address) as identifiers in Iulia mode. + * Optimiser: Replace ``x % 2**i`` by ``x & (2**i-1)``. + * Resolver: Continue resolving references after the first error. + * Resolver: Suggest alternative identifiers if a given identifier is not found. + * SMT Checker: Take if-else branch conditions into account in the SMT encoding of the program + variables. + * Syntax Checker: Deprecate the ``var`` keyword (and mark it an error as experimental 0.5.0 feature). + * Type Checker: Allow `this.f.selector` to be a pure expression. + * Type Checker: Issue warning for using ``public`` visibility for interface functions. + * Type Checker: Limit the number of warnings raised for creating abstract contracts. + +Bugfixes: + * Error Output: Truncate huge number literals in the middle to avoid output blow-up. + * Parser: Disallow event declarations with no parameter list. + * Standard JSON: Populate the ``sourceLocation`` field in the error list. + * Standard JSON: Properly support contract and library file names containing a colon (such as URLs). + * Type Checker: Suggest the experimental ABI encoder if using ``struct``s as function parameters + (instead of an internal compiler error). + * Type Checker: Improve error message for wrong struct initialization. + ### 0.4.19 (2017-11-30) Features: diff --git a/circle.yml b/circle.yml index db685da197f9..4edc9383be92 100644 --- a/circle.yml +++ b/circle.yml @@ -1,50 +1,144 @@ version: 2 jobs: - build: + build_emscripten: docker: - image: trzeci/emscripten:sdk-tag-1.37.21-64bit steps: - checkout + - run: + name: Init submodules + command: | + git submodule update --init + - restore_cache: + name: Restore Boost build + key: &boost-cache-key emscripten-boost-{{ checksum "scripts/travis-emscripten/install_deps.sh" }}{{ checksum "scripts/travis-emscripten/build_emscripten.sh" }} + - run: + name: Bootstrap Boost + command: | + scripts/travis-emscripten/install_deps.sh + - run: + name: Build + command: | + scripts/travis-emscripten/build_emscripten.sh + - save_cache: + name: Save Boost build + key: *boost-cache-key + paths: + - boost_1_57_0 + - store_artifacts: + path: build/libsolc/soljson.js + destination: soljson.js + - run: mkdir -p workspace + - run: cp build/libsolc/soljson.js workspace/soljson.js + - run: scripts/get_version.sh > workspace/version.txt + - persist_to_workspace: + root: workspace + paths: + - soljson.js + - version.txt + test_emscripten_solcjs: + docker: + - image: trzeci/emscripten:sdk-tag-1.37.21-64bit + steps: + - checkout + - attach_workspace: + at: /tmp/workspace - run: name: Install external tests deps command: | apt-get -qq update apt-get -qy install netcat curl curl -o- https://raw.githubusercontent.com/creationix/nvm/v0.33.4/install.sh | NVM_DIR=/usr/local/nvm bash + export NVM_DIR="/usr/local/nvm" + [ -s "$NVM_DIR/nvm.sh" ] && \. "$NVM_DIR/nvm.sh" # This loads nvm + nvm --version + nvm install 6 + node --version + npm --version + - run: + name: Test solcjs + command: | + . /usr/local/nvm/nvm.sh + test/solcjsTests.sh /tmp/workspace/soljson.js $(cat /tmp/workspace/version.txt) + test_emscripten_external: + docker: + - image: trzeci/emscripten:sdk-tag-1.37.21-64bit + steps: + - checkout + - attach_workspace: + at: /tmp/workspace - run: - name: Test external tests deps + name: Install external tests deps command: | + apt-get -qq update + apt-get -qy install netcat curl + curl -o- https://raw.githubusercontent.com/creationix/nvm/v0.33.4/install.sh | NVM_DIR=/usr/local/nvm bash export NVM_DIR="/usr/local/nvm" [ -s "$NVM_DIR/nvm.sh" ] && \. "$NVM_DIR/nvm.sh" # This loads nvm nvm --version - nvm install 6 + nvm install 7 node --version npm --version + - run: + name: External tests + command: | + . /usr/local/nvm/nvm.sh + test/externalTests.sh /tmp/workspace/soljson.js + build_x86: + docker: + - image: buildpack-deps:artful + steps: + - checkout + - run: + name: Install build dependencies + command: | + apt-get -qq update + apt-get -qy install ccache cmake libboost-all-dev libz3-dev - run: name: Init submodules command: | git submodule update --init - - restore_cache: - name: Restore Boost build - key: &boost-cache-key emscripten-boost-{{ checksum "scripts/travis-emscripten/install_deps.sh" }}{{ checksum "scripts/travis-emscripten/build_emscripten.sh" }} - run: - name: Bootstrap Boost + name: Store commit hash and prerelease command: | - scripts/travis-emscripten/install_deps.sh + date -u +"nightly.%Y.%-m.%-d" > prerelease.txt + echo -n "$CIRCLE_SHA1" > commit_hash.txt + - restore_cache: + key: ccache-{{ arch }}-{{ .Branch }} + key: ccache-{{ arch }} + key: ccache - run: name: Build - command: | - scripts/travis-emscripten/build_emscripten.sh + command: ./scripts/build.sh RelWithDebInfo - save_cache: - name: Save Boost build - key: *boost-cache-key + key: ccache-{{ arch }}-{{ .Branch }} paths: - - boost_1_57_0 + - ~/.ccache - run: - name: Test - command: | - . /usr/local/nvm/nvm.sh - scripts/test_emscripten.sh + name: Commandline tests + command: test/cmdlineTests.sh + - run: mkdir -p test_results + - run: + name: Test without optimizer (exclude IPC tests) + command: build/test/soltest --logger=JUNIT,test_suite,test_results/no_opt.xml -- --no-ipc + - run: + name: Test with optimizer (exclude IPC tests) + command: build/test/soltest --logger=JUNIT,test_suite,test_results/opt.xml -- --optimize --no-ipc + - store_test_results: + path: test_results/ - store_artifacts: - path: build/solc/soljson.js - destination: soljson.js + path: build/solc/solc + destination: solc + +workflows: + version: 2 + build_all: + jobs: + - build_emscripten + - test_emscripten_solcjs: + requires: + - build_emscripten + - test_emscripten_external: + requires: + - build_emscripten + - build_x86 diff --git a/cmake/EthCompilerSettings.cmake b/cmake/EthCompilerSettings.cmake index 6d4dadeb2cfd..7199f537600d 100644 --- a/cmake/EthCompilerSettings.cmake +++ b/cmake/EthCompilerSettings.cmake @@ -93,10 +93,15 @@ if (("${CMAKE_CXX_COMPILER_ID}" MATCHES "GNU") OR ("${CMAKE_CXX_COMPILER_ID}" MA # testing on at least OS X and Ubuntu. add_compile_options(-Wno-unused-function) add_compile_options(-Wno-dangling-else) - + + if ("${CMAKE_SYSTEM_NAME}" MATCHES "Darwin") + # Set stack size to 16MB - by default Apple's clang defines a stack size of 8MB, some tests require more. + set(CMAKE_EXE_LINKER_FLAGS "${CMAKE_EXE_LINKER_FLAGS} -Wl,-stack_size -Wl,0x1000000") + endif() + # Some Linux-specific Clang settings. We don't want these for OS X. if ("${CMAKE_SYSTEM_NAME}" MATCHES "Linux") - + # TODO - Is this even necessary? Why? # See http://stackoverflow.com/questions/19774778/when-is-it-necessary-to-use-use-the-flag-stdlib-libstdc. add_compile_options(-stdlib=libstdc++) @@ -115,7 +120,7 @@ if (("${CMAKE_CXX_COMPILER_ID}" MATCHES "GNU") OR ("${CMAKE_CXX_COMPILER_ID}" MA endif() if (EMSCRIPTEN) - # Do emit a separate memory initialiser file + # Do not emit a separate memory initialiser file set(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} --memory-init-file 0") # Leave only exported symbols as public and agressively remove others set(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} -fdata-sections -ffunction-sections -Wl,--gc-sections -fvisibility=hidden") @@ -133,6 +138,8 @@ if (("${CMAKE_CXX_COMPILER_ID}" MATCHES "GNU") OR ("${CMAKE_CXX_COMPILER_ID}" MA set(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} -s ALLOW_MEMORY_GROWTH=1") # Disable eval() set(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} -s NO_DYNAMIC_EXECUTION=1") + # Disable greedy exception catcher + set(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} -s NODEJS_CATCH_EXIT=0") add_definitions(-DETH_EMSCRIPTEN=1) endif() endif() diff --git a/docs/abi-spec.rst b/docs/abi-spec.rst index c93ce25bd8aa..8095a3b7b4b5 100644 --- a/docs/abi-spec.rst +++ b/docs/abi-spec.rst @@ -40,21 +40,21 @@ The following elementary types exist: - ``int``: two's complement signed integer type of ``M`` bits, ``0 < M <= 256``, ``M % 8 == 0``. -- ``address``: equivalent to ``uint160``, except for the assumed interpretation and language typing. +- ``address``: equivalent to ``uint160``, except for the assumed interpretation and language typing. For computing the function selector, ``address`` is used. -- ``uint``, ``int``: synonyms for ``uint256``, ``int256`` respectively (this shorthand not to be used for computing the function selector). +- ``uint``, ``int``: synonyms for ``uint256``, ``int256`` respectively. For computing the function selector, ``uint256`` and ``int256`` have to be used. -- ``bool``: equivalent to ``uint8`` restricted to the values 0 and 1 +- ``bool``: equivalent to ``uint8`` restricted to the values 0 and 1. For computing the function selector, ``bool`` is used. - ``fixedx``: signed fixed-point decimal number of ``M`` bits, ``8 <= M <= 256``, ``M % 8 ==0``, and ``0 < N <= 80``, which denotes the value ``v`` as ``v / (10 ** N)``. - ``ufixedx``: unsigned variant of ``fixedx``. -- ``fixed``, ``ufixed``: synonyms for ``fixed128x19``, ``ufixed128x19`` respectively (this shorthand not to be used for computing the function selector). +- ``fixed``, ``ufixed``: synonyms for ``fixed128x19``, ``ufixed128x19`` respectively. For computing the function selector, ``fixed128x19`` and ``ufixed128x19`` have to be used. - ``bytes``: binary type of ``M`` bytes, ``0 < M <= 32``. -- ``function``: equivalent to ``bytes24``: an address, followed by a function selector +- ``function``: an address (20 bytes) folled by a function selector (4 bytes). Encoded identical to ``bytes24``. The following (fixed-size) array type exists: @@ -187,12 +187,12 @@ Given the contract: :: - pragma solidity ^0.4.0; + pragma solidity ^0.4.16; contract Foo { - function bar(bytes3[2] xy) {} - function baz(uint32 x, bool y) returns (bool r) { r = x > 32 || y; } - function sam(bytes name, bool z, uint[] data) {} + function bar(bytes3[2]) public pure {} + function baz(uint32 x, bool y) public pure returns (bool r) { r = x > 32 || y; } + function sam(bytes, bool, uint[]) public pure {} } @@ -288,6 +288,8 @@ In effect, a log entry using this ABI is described as: - ``topics[n]``: ``EVENT_INDEXED_ARGS[n - 1]`` (``EVENT_INDEXED_ARGS`` is the series of ``EVENT_ARGS`` that are indexed); - ``data``: ``abi_serialise(EVENT_NON_INDEXED_ARGS)`` (``EVENT_NON_INDEXED_ARGS`` is the series of ``EVENT_ARGS`` that are not indexed, ``abi_serialise`` is the ABI serialisation function used for returning a series of typed values from a function, as described above). +For all fixed-length Solidity types, the ``EVENT_INDEXED_ARGS`` array contains the 32-byte encoded value directly. However, for *types of dynamic length*, which include ``string``, ``bytes``, and arrays, ``EVENT_INDEXED_ARGS`` will contain the *Keccak hash* of the encoded value, rather than the encoded value directly. This allows applications to efficiently query for values of dynamic-length types (by setting the hash of the encoded value as the topic), but leaves applications unable to decode indexed values they have not queried for. For dynamic-length types, application developers face a trade-off between fast search for predetermined values (if the argument is indexed) and legibility of arbitrary values (which requires that the arguments not be indexed). Developers may overcome this tradeoff and achieve both efficient search and arbitrary legibility by defining events with two arguments — one indexed, one not — intended to hold the same value. + JSON ==== @@ -333,10 +335,10 @@ For example, pragma solidity ^0.4.0; contract Test { - function Test(){ b = 0x12345678901234567890123456789012; } + function Test() public { b = 0x12345678901234567890123456789012; } event Event(uint indexed a, bytes32 b); event Event2(uint indexed a, bytes32 b); - function foo(uint a) { Event(a, b); } + function foo(uint a) public { Event(a, b); } bytes32 b; } @@ -377,10 +379,14 @@ As an example, the code :: + pragma solidity ^0.4.19; + pragma experimental ABIEncoderV2; + contract Test { struct S { uint a; uint[] b; T[] c; } struct T { uint x; uint y; } - function f(S s, T t, uint a) { } + function f(S s, T t, uint a) public { } + function g() public returns (S s, T t, uint a) {} } would result in the JSON: diff --git a/docs/assembly.rst b/docs/assembly.rst index c233985baacb..afc44d665924 100644 --- a/docs/assembly.rst +++ b/docs/assembly.rst @@ -23,7 +23,7 @@ arising when writing manual assembly by the following features: * functional-style opcodes: ``mul(1, add(2, 3))`` instead of ``push1 3 push1 2 add push1 1 mul`` * assembly-local variables: ``let x := add(2, 3) let y := mload(0x40) x := add(x, y)`` -* access to external variables: ``function f(uint x) { assembly { x := sub(x, 1) } }`` +* access to external variables: ``function f(uint x) public { assembly { x := sub(x, 1) } }`` * labels: ``let x := 10 repeat: x := sub(x, 1) jumpi(repeat, eq(x, 0))`` * loops: ``for { let i := 0 } lt(i, x) { i := add(i, 1) } { y := mul(2, y) }`` * if statements: ``if slt(x, 0) { x := sub(0, x) }`` @@ -54,7 +54,7 @@ idea is that assembly libraries will be used to enhance the language in such way pragma solidity ^0.4.0; library GetCode { - function at(address _addr) returns (bytes o_code) { + function at(address _addr) public view returns (bytes o_code) { assembly { // retrieve the size of the code, this needs assembly let size := extcodesize(_addr) @@ -78,12 +78,12 @@ you really know what you are doing. .. code:: - pragma solidity ^0.4.12; + pragma solidity ^0.4.16; library VectorSum { // This function is less efficient because the optimizer currently fails to // remove the bounds checks in array access. - function sumSolidity(uint[] _data) returns (uint o_sum) { + function sumSolidity(uint[] _data) public view returns (uint o_sum) { for (uint i = 0; i < _data.length; ++i) o_sum += _data[i]; } @@ -91,7 +91,7 @@ you really know what you are doing. // We know that we only access the array in bounds, so we can avoid the check. // 0x20 needs to be added to an array because the first slot contains the // array length. - function sumAsm(uint[] _data) returns (uint o_sum) { + function sumAsm(uint[] _data) public view returns (uint o_sum) { for (uint i = 0; i < _data.length; ++i) { assembly { o_sum := add(o_sum, mload(add(add(_data, 0x20), mul(i, 0x20)))) @@ -100,7 +100,7 @@ you really know what you are doing. } // Same as above, but accomplish the entire code within inline assembly. - function sumPureAsm(uint[] _data) returns (uint o_sum) { + function sumPureAsm(uint[] _data) public view returns (uint o_sum) { assembly { // Load the length (first 32 bytes) let len := mload(_data) @@ -388,7 +388,7 @@ changes during the call, and thus references to local variables will be wrong. contract C { uint b; - function f(uint x) returns (uint r) { + function f(uint x) public returns (uint r) { assembly { r := mul(x, sload(b_slot)) // ignore the offset, we know it is zero } @@ -447,31 +447,6 @@ will have a wrong impression about the stack height at label ``two``: three: } -This problem can be fixed by manually adjusting the stack height for the -assembler - you can provide a stack height delta that is added -to the stack height just prior to the label. -Note that you will not have to care about these things if you just use -loops and assembly-level functions. - -As an example how this can be done in extreme cases, please see the following. - -.. code:: - - { - let x := 8 - jump(two) - 0 // This code is unreachable but will adjust the stack height correctly - one: - x := 9 // Now x can be accessed properly. - jump(three) - pop // Similar negative correction. - two: - 7 // push something onto the stack - jump(one) - three: - pop // We have to pop the manually pushed value here again. - } - Declaring Assembly-Local Variables ---------------------------------- @@ -484,10 +459,10 @@ be just ``0``, but it can also be a complex functional-style expression. .. code:: - pragma solidity ^0.4.0; + pragma solidity ^0.4.16; contract C { - function f(uint x) returns (uint b) { + function f(uint x) public view returns (uint b) { assembly { let v := add(x, 1) mstore(0x80, v) @@ -599,7 +574,7 @@ Simply leave the initialization and post-iteration parts empty. x := add(x, mload(i)) i := add(i, 0x20) } - } + } Functions --------- @@ -735,10 +710,10 @@ Example: We will follow an example compilation from Solidity to desugared assembly. We consider the runtime bytecode of the following Solidity program:: - pragma solidity ^0.4.0; + pragma solidity ^0.4.16; contract C { - function f(uint x) returns (uint y) { + function f(uint x) public pure returns (uint y) { y = 1; for (uint i = 0; i < x; i++) y = 2 * y; diff --git a/docs/bugs_by_version.json b/docs/bugs_by_version.json index 3a8ff9a10ee8..5a4c9e29e88f 100644 --- a/docs/bugs_by_version.json +++ b/docs/bugs_by_version.json @@ -414,6 +414,10 @@ ], "released": "2016-09-17" }, + "0.4.20": { + "bugs": [], + "released": "2018-02-14" + }, "0.4.3": { "bugs": [ "ZeroFunctionSelector", diff --git a/docs/common-patterns.rst b/docs/common-patterns.rst index 52319be0d550..7e09f53468de 100644 --- a/docs/common-patterns.rst +++ b/docs/common-patterns.rst @@ -36,12 +36,12 @@ become the new richest. mapping (address => uint) pendingWithdrawals; - function WithdrawalContract() payable { + function WithdrawalContract() public payable { richest = msg.sender; mostSent = msg.value; } - function becomeRichest() payable returns (bool) { + function becomeRichest() public payable returns (bool) { if (msg.value > mostSent) { pendingWithdrawals[richest] += msg.value; richest = msg.sender; @@ -52,7 +52,7 @@ become the new richest. } } - function withdraw() { + function withdraw() public { uint amount = pendingWithdrawals[msg.sender]; // Remember to zero the pending refund before // sending to prevent re-entrancy attacks @@ -71,12 +71,12 @@ This is as opposed to the more intuitive sending pattern: address public richest; uint public mostSent; - function SendContract() payable { + function SendContract() public payable { richest = msg.sender; mostSent = msg.value; } - function becomeRichest() payable returns (bool) { + function becomeRichest() public payable returns (bool) { if (msg.value > mostSent) { // This line can cause problems (explained below). richest.transfer(msg.value); @@ -157,6 +157,7 @@ restrictions highly readable. /// Make `_newOwner` the new owner of this /// contract. function changeOwner(address _newOwner) + public onlyBy(owner) { owner = _newOwner; @@ -171,6 +172,7 @@ restrictions highly readable. /// May only be called 6 weeks after /// the contract has been created. function disown() + public onlyBy(owner) onlyAfter(creationTime + 6 weeks) { @@ -191,6 +193,7 @@ restrictions highly readable. } function forceOwnerChange(address _newOwner) + public costs(200 ether) { owner = _newOwner; @@ -310,6 +313,7 @@ function finishes. // Order of the modifiers matters here! function bid() + public payable timedTransitions atStage(Stages.AcceptingBlindedBids) @@ -318,6 +322,7 @@ function finishes. } function reveal() + public timedTransitions atStage(Stages.RevealBids) { @@ -332,6 +337,7 @@ function finishes. } function g() + public timedTransitions atStage(Stages.AnotherStage) transitionNext @@ -339,6 +345,7 @@ function finishes. } function h() + public timedTransitions atStage(Stages.AreWeDoneYet) transitionNext @@ -346,6 +353,7 @@ function finishes. } function i() + public timedTransitions atStage(Stages.Finished) { diff --git a/docs/contracts.rst b/docs/contracts.rst index 2b0956bb53c6..afc32b169291 100644 --- a/docs/contracts.rst +++ b/docs/contracts.rst @@ -40,7 +40,7 @@ This means that cyclic creation dependencies are impossible. :: - pragma solidity ^0.4.0; + pragma solidity ^0.4.16; contract OwnedToken { // TokenCreator is a contract type that is defined below. @@ -52,7 +52,7 @@ This means that cyclic creation dependencies are impossible. // This is the constructor which registers the // creator and the assigned name. - function OwnedToken(bytes32 _name) { + function OwnedToken(bytes32 _name) public { // State variables are accessed via their name // and not via e.g. this.owner. This also applies // to functions and especially in the constructors, @@ -67,7 +67,7 @@ This means that cyclic creation dependencies are impossible. name = _name; } - function changeName(bytes32 newName) { + function changeName(bytes32 newName) public { // Only the creator can alter the name -- // the comparison is possible since contracts // are implicitly convertible to addresses. @@ -75,7 +75,7 @@ This means that cyclic creation dependencies are impossible. name = newName; } - function transfer(address newOwner) { + function transfer(address newOwner) public { // Only the current owner can transfer the token. if (msg.sender != owner) return; // We also want to ask the creator if the transfer @@ -90,25 +90,27 @@ This means that cyclic creation dependencies are impossible. contract TokenCreator { function createToken(bytes32 name) + public returns (OwnedToken tokenAddress) { // Create a new Token contract and return its address. // From the JavaScript side, the return type is simply - // "address", as this is the closest type available in + // `address`, as this is the closest type available in // the ABI. return new OwnedToken(name); } - function changeName(OwnedToken tokenAddress, bytes32 name) { - // Again, the external type of "tokenAddress" is - // simply "address". + function changeName(OwnedToken tokenAddress, bytes32 name) public { + // Again, the external type of `tokenAddress` is + // simply `address`. tokenAddress.changeName(name); } - function isTokenTransferOK( - address currentOwner, - address newOwner - ) returns (bool ok) { + function isTokenTransferOK(address currentOwner, address newOwner) + public + view + returns (bool ok) + { // Check some arbitrary condition. address tokenAddress = msg.sender; return (keccak256(newOwner) & 0xff) == (bytes20(tokenAddress) & 0xff); @@ -171,10 +173,10 @@ return parameter list for functions. :: - pragma solidity ^0.4.0; + pragma solidity ^0.4.16; contract C { - function f(uint a) private returns (uint b) { return a + 1; } + function f(uint a) private pure returns (uint b) { return a + 1; } function setData(uint a) internal { data = a; } uint public data; } @@ -193,23 +195,23 @@ In the following example, ``D``, can call ``c.getData()`` to retrieve the value uint private data; function f(uint a) private returns(uint b) { return a + 1; } - function setData(uint a) { data = a; } + function setData(uint a) public { data = a; } function getData() public returns(uint) { return data; } function compute(uint a, uint b) internal returns (uint) { return a+b; } } contract D { - function readData() { + function readData() public { C c = new C(); - uint local = c.f(7); // error: member "f" is not visible + uint local = c.f(7); // error: member `f` is not visible c.setData(3); local = c.getData(); - local = c.compute(3, 5); // error: member "compute" is not visible + local = c.compute(3, 5); // error: member `compute` is not visible } } contract E is C { - function g() { + function g() public { C c = new C(); uint val = compute(3, 5); // access to internal member (from derived to parent contract) } @@ -238,7 +240,7 @@ be done at declaration. contract Caller { C c = new C(); - function f() { + function f() public { uint local = c.data(); } } @@ -254,7 +256,7 @@ it is evaluated as a state variable. If it is accessed externally contract C { uint public data; - function x() { + function x() public { data = 3; // internal access uint val = this.data(); // external access } @@ -277,7 +279,7 @@ The next example is a bit more complex: It will generate a function of the following form:: - function data(uint arg1, bool arg2, uint arg3) returns (uint a, bytes3 b) { + function data(uint arg1, bool arg2, uint arg3) public returns (uint a, bytes3 b) { a = data[arg1][arg2][arg3].a; b = data[arg1][arg2][arg3].b; } @@ -302,13 +304,13 @@ inheritable properties of contracts and may be overridden by derived contracts. pragma solidity ^0.4.11; contract owned { - function owned() { owner = msg.sender; } + function owned() public { owner = msg.sender; } address owner; // This contract only defines a modifier but does not use - // it - it will be used in derived contracts. + // it: it will be used in derived contracts. // The function body is inserted where the special symbol - // "_;" in the definition of a modifier appears. + // `_;` in the definition of a modifier appears. // This means that if the owner calls this function, the // function is executed and otherwise, an exception is // thrown. @@ -319,11 +321,11 @@ inheritable properties of contracts and may be overridden by derived contracts. } contract mortal is owned { - // This contract inherits the "onlyOwner"-modifier from - // "owned" and applies it to the "close"-function, which - // causes that calls to "close" only have an effect if + // This contract inherits the `onlyOwner` modifier from + // `owned` and applies it to the `close` function, which + // causes that calls to `close` only have an effect if // they are made by the stored owner. - function close() onlyOwner { + function close() public onlyOwner { selfdestruct(owner); } } @@ -341,16 +343,16 @@ inheritable properties of contracts and may be overridden by derived contracts. mapping (address => bool) registeredAddresses; uint price; - function Register(uint initialPrice) { price = initialPrice; } + function Register(uint initialPrice) public { price = initialPrice; } // It is important to also provide the - // "payable" keyword here, otherwise the function will + // `payable` keyword here, otherwise the function will // automatically reject all Ether sent to it. - function register() payable costs(price) { + function register() public payable costs(price) { registeredAddresses[msg.sender] = true; } - function changePrice(uint _price) onlyOwner { + function changePrice(uint _price) public onlyOwner { price = _price; } } @@ -365,10 +367,10 @@ inheritable properties of contracts and may be overridden by derived contracts. } /// This function is protected by a mutex, which means that - /// reentrant calls from within msg.sender.call cannot call f again. + /// reentrant calls from within `msg.sender.call` cannot call `f` again. /// The `return 7` statement assigns 7 to the return value but still /// executes the statement `locked = false` in the modifier. - function f() noReentrancy returns (uint) { + function f() public noReentrancy returns (uint) { require(msg.sender.call()); return 7; } @@ -426,19 +428,27 @@ value types and strings. bytes32 constant myHash = keccak256("abc"); } +.. index:: ! functions + +.. _functions: + +********* +Functions +********* + +.. index:: ! view function, function;view .. _view-functions: -************** View Functions -************** +============== Functions can be declared ``view`` in which case they promise not to modify the state. The following statements are considered modifying the state: #. Writing to state variables. -#. :ref:`Emitting events. `. +#. :ref:`Emitting events `. #. :ref:`Creating other contracts `. #. Using ``selfdestruct``. #. Sending Ether via calls. @@ -451,7 +461,7 @@ The following statements are considered modifying the state: pragma solidity ^0.4.16; contract C { - function f(uint a, uint b) view returns (uint) { + function f(uint a, uint b) public view returns (uint) { return a * (b + 42) + now; } } @@ -465,11 +475,12 @@ The following statements are considered modifying the state: .. warning:: The compiler does not enforce yet that a ``view`` method is not modifying state. +.. index:: ! pure function, function;pure + .. _pure-functions: -************** Pure Functions -************** +============== Functions can be declared ``pure`` in which case they promise not to read from or modify the state. @@ -486,7 +497,7 @@ In addition to the list of state modifying statements explained above, the follo pragma solidity ^0.4.16; contract C { - function f(uint a, uint b) pure returns (uint) { + function f(uint a, uint b) public pure returns (uint) { return a * (b + 42); } } @@ -498,9 +509,8 @@ In addition to the list of state modifying statements explained above, the follo .. _fallback-function: -***************** Fallback Function -***************** +================= A contract can have exactly one unnamed function. This function cannot have arguments and cannot return anything. @@ -551,9 +561,9 @@ Please ensure you test your fallback function thoroughly to ensure the execution // This function is called for all messages sent to // this contract (there is no other function). // Sending Ether to this contract will cause an exception, - // because the fallback function does not have the "payable" + // because the fallback function does not have the `payable` // modifier. - function() { x = 1; } + function() public { x = 1; } uint x; } @@ -561,11 +571,11 @@ Please ensure you test your fallback function thoroughly to ensure the execution // This contract keeps all Ether sent to it with no way // to get it back. contract Sink { - function() payable { } + function() public payable { } } contract Caller { - function callTest(Test test) { + function callTest(Test test) public { test.call(0xabcdef01); // hash does not exist // results in test.x becoming == 1. @@ -577,6 +587,85 @@ Please ensure you test your fallback function thoroughly to ensure the execution } } +.. index:: ! overload + +.. _overload-function: + +Function Overloading +==================== + +A Contract can have multiple functions of the same name but with different arguments. +This also applies to inherited functions. The following example shows overloading of the +``f`` function in the scope of contract ``A``. + +:: + + pragma solidity ^0.4.16; + + contract A { + function f(uint _in) public pure returns (uint out) { + out = 1; + } + + function f(uint _in, bytes32 _key) public pure returns (uint out) { + out = 2; + } + } + +Overloaded functions are also present in the external interface. It is an error if two +externally visible functions differ by their Solidity types but not by their external types. + +:: + + // This will not compile + pragma solidity ^0.4.16; + + contract A { + function f(B _in) public pure returns (B out) { + out = _in; + } + + function f(address _in) public pure returns (address out) { + out = _in; + } + } + + contract B { + } + + +Both ``f`` function overloads above end up accepting the address type for the ABI although +they are considered different inside Solidity. + +Overload resolution and Argument matching +----------------------------------------- + +Overloaded functions are selected by matching the function declarations in the current scope +to the arguments supplied in the function call. Functions are selected as overload candidates +if all arguments can be implicitly converted to the expected types. If there is not exactly one +candidate, resolution fails. + +.. note:: + Return parameters are not taken into account for overload resolution. + +:: + + pragma solidity ^0.4.16; + + contract A { + function f(uint8 _in) public pure returns (uint8 out) { + out = _in; + } + + function f(uint256 _in) public pure returns (uint256 out) { + out = _in; + } + } + +Calling ``f(50)`` would create a type error since ``250`` can be implicitly converted both to ``uint8`` +and ``uint256`` types. On another hand ``f(256)`` would resolve to ``f(uint256)`` overload as ``256`` cannot be implicitly +converted to ``uint8``. + .. index:: ! event .. _events: @@ -634,7 +723,7 @@ All non-indexed arguments will be stored in the data part of the log. uint _value ); - function deposit(bytes32 _id) payable { + function deposit(bytes32 _id) public payable { // Any call to this function (even deeply nested) can // be detected from the JavaScript API by filtering // for `Deposit` to be called. @@ -655,7 +744,7 @@ The use in the JavaScript API would be as follows: // watch for changes event.watch(function(error, result){ // result will contain various information - // including the argumets given to the Deposit + // including the argumets given to the `Deposit` // call. if (!error) console.log(result); @@ -683,7 +772,7 @@ as topics. The event call above can be performed in the same way as pragma solidity ^0.4.10; contract C { - function f() { + function f() public payable { bytes32 _id = 0x420042; log3( bytes32(msg.value), @@ -727,14 +816,14 @@ Details are given in the following example. :: - pragma solidity ^0.4.0; + pragma solidity ^0.4.16; contract owned { function owned() { owner = msg.sender; } address owner; } - // Use "is" to derive from another contract. Derived + // Use `is` to derive from another contract. Derived // contracts can access all non-private members including // internal functions and state variables. These cannot be // accessed externally via `this`, though. @@ -749,20 +838,20 @@ Details are given in the following example. // without body. If a contract does not implement all // functions it can only be used as an interface. contract Config { - function lookup(uint id) returns (address adr); + function lookup(uint id) public returns (address adr); } contract NameReg { - function register(bytes32 name); - function unregister(); + function register(bytes32 name) public; + function unregister() public; } - // Multiple inheritance is possible. Note that "owned" is - // also a base class of "mortal", yet there is only a single - // instance of "owned" (as for virtual inheritance in C++). + // Multiple inheritance is possible. Note that `owned` is + // also a base class of `mortal`, yet there is only a single + // instance of `owned` (as for virtual inheritance in C++). contract named is owned, mortal { function named(bytes32 name) { - Config config = Config(0xd5f9d8d94886e70b06e474c3fb14fd43e2f23970); + Config config = Config(0xD5f9D8D94886E70b06E474c3fB14Fd43E2f23970); NameReg(config.lookup(1)).register(name); } @@ -771,9 +860,9 @@ Details are given in the following example. // types of output parameters, that causes an error. // Both local and message-based function calls take these overrides // into account. - function kill() { + function kill() public { if (msg.sender == owner) { - Config config = Config(0xd5f9d8d94886e70b06e474c3fb14fd43e2f23970); + Config config = Config(0xD5f9D8D94886E70b06E474c3fB14Fd43E2f23970); NameReg(config.lookup(1)).unregister(); // It is still possible to call a specific // overridden function. @@ -786,11 +875,11 @@ Details are given in the following example. // provided in the header (or modifier-invocation-style at // the constructor of the derived contract (see below)). contract PriceFeed is owned, mortal, named("GoldFeed") { - function updateInfo(uint newInfo) { + function updateInfo(uint newInfo) public { if (msg.sender == owner) info = newInfo; } - function get() constant returns(uint r) { return info; } + function get() public view returns(uint r) { return info; } uint info; } @@ -802,22 +891,22 @@ seen in the following example:: pragma solidity ^0.4.0; contract owned { - function owned() { owner = msg.sender; } + function owned() public { owner = msg.sender; } address owner; } contract mortal is owned { - function kill() { + function kill() public { if (msg.sender == owner) selfdestruct(owner); } } contract Base1 is mortal { - function kill() { /* do cleanup 1 */ mortal.kill(); } + function kill() public { /* do cleanup 1 */ mortal.kill(); } } contract Base2 is mortal { - function kill() { /* do cleanup 2 */ mortal.kill(); } + function kill() public { /* do cleanup 2 */ mortal.kill(); } } contract Final is Base1, Base2 { @@ -831,34 +920,34 @@ derived override, but this function will bypass pragma solidity ^0.4.0; contract owned { - function owned() { owner = msg.sender; } + function owned() public { owner = msg.sender; } address owner; } contract mortal is owned { - function kill() { + function kill() public { if (msg.sender == owner) selfdestruct(owner); } } contract Base1 is mortal { - function kill() { /* do cleanup 1 */ super.kill(); } + function kill() public { /* do cleanup 1 */ super.kill(); } } contract Base2 is mortal { - function kill() { /* do cleanup 2 */ super.kill(); } + function kill() public { /* do cleanup 2 */ super.kill(); } } - contract Final is Base2, Base1 { + contract Final is Base1, Base2 { } -If ``Base1`` calls a function of ``super``, it does not simply -call this function on one of its base contracts. Rather, it +If ``Base2`` calls a function of ``super``, it does not simply +call this function on one of its base contracts. Rather, it calls this function on the next base contract in the final -inheritance graph, so it will call ``Base2.kill()`` (note that +inheritance graph, so it will call ``Base1.kill()`` (note that the final inheritance sequence is -- starting with the most -derived contract: Final, Base1, Base2, mortal, owned). +derived contract: Final, Base2, Base1, mortal, owned). The actual function that is called when using super is not known in the context of the class where it is used, although its type is known. This is similar for ordinary @@ -876,11 +965,11 @@ the base constructors. This can be done in two ways:: contract Base { uint x; - function Base(uint _x) { x = _x; } + function Base(uint _x) public { x = _x; } } contract Derived is Base(7) { - function Derived(uint _y) Base(_y * _y) { + function Derived(uint _y) Base(_y * _y) public { } } @@ -945,7 +1034,7 @@ Contract functions can lack an implementation as in the following example (note pragma solidity ^0.4.0; contract Feline { - function utterance() returns (bytes32); + function utterance() public returns (bytes32); } Such contracts cannot be compiled (even if they contain @@ -955,11 +1044,11 @@ but they can be used as base contracts:: pragma solidity ^0.4.0; contract Feline { - function utterance() returns (bytes32); + function utterance() public returns (bytes32); } contract Cat is Feline { - function utterance() returns (bytes32) { return "miaow"; } + function utterance() public returns (bytes32) { return "miaow"; } } If a contract inherits from an abstract contract and does not implement all non-implemented functions by overriding, it will itself be abstract. @@ -990,7 +1079,7 @@ Interfaces are denoted by their own keyword: pragma solidity ^0.4.11; interface Token { - function transfer(address recipient, uint amount); + function transfer(address recipient, uint amount) public; } Contracts can inherit interfaces as they would inherit other contracts. @@ -1011,7 +1100,11 @@ is executed in the context of the calling contract, i.e. ``this`` points to the calling contract, and especially the storage from the calling contract can be accessed. As a library is an isolated piece of source code, it can only access state variables of the calling contract if they are explicitly supplied (it -would have no way to name them, otherwise). +would have no way to name them, otherwise). Library functions can only be +called directly (i.e. without the use of ``DELEGATECALL``) if they do not modify +the state (i.e. if they are ``view`` or ``pure`` functions), +because libraries are assumed to be stateless. In particular, it is +not possible to destroy a library unless Solidity's type system is circumvented. Libraries can be seen as implicit base contracts of the contracts that use them. They will not be explicitly visible in the inheritance hierarchy, but calls @@ -1022,7 +1115,7 @@ if the library were a base contract. Of course, calls to internal functions use the internal calling convention, which means that all internal types can be passed and memory types will be passed by reference and not copied. To realize this in the EVM, code of internal library functions -and all functions called from therein will be pulled into the calling +and all functions called from therein will at compile time be pulled into the calling contract, and a regular ``JUMP`` call will be used instead of a ``DELEGATECALL``. .. index:: using for, set @@ -1033,7 +1126,7 @@ more advanced example to implement a set). :: - pragma solidity ^0.4.11; + pragma solidity ^0.4.16; library Set { // We define a new struct datatype that will be used to @@ -1044,9 +1137,10 @@ more advanced example to implement a set). // reference" and thus only its storage address and not // its contents is passed as part of the call. This is a // special feature of library functions. It is idiomatic - // to call the first parameter 'self', if the function can + // to call the first parameter `self`, if the function can // be seen as a method of that object. function insert(Data storage self, uint value) + public returns (bool) { if (self.flags[value]) @@ -1056,6 +1150,7 @@ more advanced example to implement a set). } function remove(Data storage self, uint value) + public returns (bool) { if (!self.flags[value]) @@ -1065,6 +1160,8 @@ more advanced example to implement a set). } function contains(Data storage self, uint value) + public + view returns (bool) { return self.flags[value]; @@ -1074,7 +1171,7 @@ more advanced example to implement a set). contract C { Set.Data knownValues; - function register(uint value) { + function register(uint value) public { // The library functions can be called without a // specific instance of the library, since the // "instance" will be the current contract. @@ -1084,7 +1181,7 @@ more advanced example to implement a set). } Of course, you do not have to follow this way to use -libraries - they can also be used without defining struct +libraries: they can also be used without defining struct data types. Functions also work without any storage reference parameters, and they can have multiple storage reference parameters and in any position. @@ -1103,19 +1200,19 @@ custom types without the overhead of external function calls: :: - pragma solidity ^0.4.0; + pragma solidity ^0.4.16; library BigInt { struct bigint { uint[] limbs; } - function fromUint(uint x) internal returns (bigint r) { + function fromUint(uint x) internal pure returns (bigint r) { r.limbs = new uint[](1); r.limbs[0] = x; } - function add(bigint _a, bigint _b) internal returns (bigint r) { + function add(bigint _a, bigint _b) internal pure returns (bigint r) { r.limbs = new uint[](max(_a.limbs.length, _b.limbs.length)); uint carry = 0; for (uint i = 0; i < r.limbs.length; ++i) { @@ -1137,11 +1234,11 @@ custom types without the overhead of external function calls: } } - function limb(bigint _a, uint _limb) internal returns (uint) { + function limb(bigint _a, uint _limb) internal pure returns (uint) { return _limb < _a.limbs.length ? _a.limbs[_limb] : 0; } - function max(uint a, uint b) private returns (uint) { + function max(uint a, uint b) private pure returns (uint) { return a > b ? a : b; } } @@ -1149,7 +1246,7 @@ custom types without the overhead of external function calls: contract C { using BigInt for BigInt.bigint; - function f() { + function f() public pure { var x = BigInt.fromUint(7); var y = BigInt.fromUint(uint(-1)); var z = x.add(y); @@ -1175,6 +1272,30 @@ Restrictions for libraries in comparison to contracts: (These might be lifted at a later point.) +Call Protection For Libraries +============================= + +As mentioned in the introduction, if a library's code is executed +using a ``CALL`` instead of a ``DELEGATECALL`` or ``CALLCODE``, +it will revert unless a ``view`` or ``pure`` function is called. + +The EVM does not provide a direct way for a contract to detect +whether it was called using ``CALL`` or not, but a contract +can use the ``ADDRESS`` opcode to find out "where" it is +currently running. The generated code compares this address +to the address used at construction time to determine the mode +of calling. + +More specifically, the runtime code of a library always starts +with a push instruction, which is a zero of 20 bytes at +compilation time. When the deploy code runs, this constant +is replaced in memory by the current address and this +modified code is stored in the contract. At runtime, +this causes the deploy time address to be the first +constant to be pushed onto the stack and the dispatcher +code compares the current address against this constant +for any non-view and non-pure function. + .. index:: ! using for, library .. _using-for: @@ -1207,13 +1328,14 @@ available without having to add further code. Let us rewrite the set example from the :ref:`libraries` in this way:: - pragma solidity ^0.4.11; + pragma solidity ^0.4.16; // This is the same code as before, just without comments library Set { struct Data { mapping(uint => bool) flags; } function insert(Data storage self, uint value) + public returns (bool) { if (self.flags[value]) @@ -1223,6 +1345,7 @@ Let us rewrite the set example from the } function remove(Data storage self, uint value) + public returns (bool) { if (!self.flags[value]) @@ -1232,6 +1355,8 @@ Let us rewrite the set example from the } function contains(Data storage self, uint value) + public + view returns (bool) { return self.flags[value]; @@ -1242,21 +1367,25 @@ Let us rewrite the set example from the using Set for Set.Data; // this is the crucial change Set.Data knownValues; - function register(uint value) { + function register(uint value) public { // Here, all variables of type Set.Data have // corresponding member functions. // The following function call is identical to - // Set.insert(knownValues, value) + // `Set.insert(knownValues, value)` require(knownValues.insert(value)); } } It is also possible to extend elementary types in that way:: - pragma solidity ^0.4.0; + pragma solidity ^0.4.16; library Search { - function indexOf(uint[] storage self, uint value) returns (uint) { + function indexOf(uint[] storage self, uint value) + public + view + returns (uint) + { for (uint i = 0; i < self.length; i++) if (self[i] == value) return i; return uint(-1); @@ -1267,11 +1396,11 @@ It is also possible to extend elementary types in that way:: using Search for uint[]; uint[] data; - function append(uint value) { + function append(uint value) public { data.push(value); } - function replace(uint _old, uint _new) { + function replace(uint _old, uint _new) public { // This performs the library function call uint index = data.indexOf(_old); if (index == uint(-1)) diff --git a/docs/contributing.rst b/docs/contributing.rst index 0f7c3e7266da..a5efba8bf789 100644 --- a/docs/contributing.rst +++ b/docs/contributing.rst @@ -50,8 +50,10 @@ and instead, ``git rebase`` your branch. Additionally, if you are writing a new feature, please ensure you write appropriate Boost test cases and place them under ``test/``. -However, if you are making a larger change, please consult with the Gitter -channel, first. +However, if you are making a larger change, please consult with the `Solidity Development Gitter channel +`_ (different from the one mentioned above, this on is +focused on compiler and language development instead of language use) first. + Finally, please make sure you respect the `coding standards `_ diff --git a/docs/control-structures.rst b/docs/control-structures.rst index bcb597cf1dee..1a8ee25b458f 100644 --- a/docs/control-structures.rst +++ b/docs/control-structures.rst @@ -20,10 +20,10 @@ For example, suppose we want our contract to accept one kind of external calls with two integers, we would write something like:: - pragma solidity ^0.4.0; + pragma solidity ^0.4.16; contract Simple { - function taker(uint _a, uint _b) { + function taker(uint _a, uint _b) public pure { // do something with _a and _b. } } @@ -36,10 +36,14 @@ The output parameters can be declared with the same syntax after the the sum and the product of the two given integers, then we would write:: - pragma solidity ^0.4.0; + pragma solidity ^0.4.16; contract Simple { - function arithmetics(uint _a, uint _b) returns (uint o_sum, uint o_product) { + function arithmetics(uint _a, uint _b) + public + pure + returns (uint o_sum, uint o_product) + { o_sum = _a + _b; o_product = _a * _b; } @@ -95,11 +99,11 @@ Internal Function Calls Functions of the current contract can be called directly ("internally"), also recursively, as seen in this nonsensical example:: - pragma solidity ^0.4.0; + pragma solidity ^0.4.16; contract C { - function g(uint a) returns (uint ret) { return f(); } - function f() returns (uint ret) { return g(7) + f(); } + function g(uint a) public pure returns (uint ret) { return f(); } + function f() internal pure returns (uint ret) { return g(7) + f(); } } These function calls are translated into simple jumps inside the EVM. This has @@ -125,13 +129,13 @@ the gas can be specified with special options ``.value()`` and ``.gas()``, respe pragma solidity ^0.4.0; contract InfoFeed { - function info() payable returns (uint ret) { return 42; } + function info() public payable returns (uint ret) { return 42; } } contract Consumer { InfoFeed feed; - function setFeed(address addr) { feed = InfoFeed(addr); } - function callFeed() { feed.info.value(10).gas(800)(); } + function setFeed(address addr) public { feed = InfoFeed(addr); } + function callFeed() public { feed.info.value(10).gas(800)(); } } The modifier ``payable`` has to be used for ``info``, because otherwise, the `.value()` @@ -180,11 +184,11 @@ parameters from the function declaration, but can be in arbitrary order. pragma solidity ^0.4.0; contract C { - function f(uint key, uint value) { + function f(uint key, uint value) public { // ... } - function g() { + function g() public { // named arguments f({value: 2, key: 3}); } @@ -198,11 +202,11 @@ Those parameters will still be present on the stack, but they are inaccessible. :: - pragma solidity ^0.4.0; + pragma solidity ^0.4.16; contract C { // omitted name for parameter - function func(uint k, uint) returns(uint) { + function func(uint k, uint) public pure returns(uint) { return k; } } @@ -225,7 +229,7 @@ creation-dependencies are not possible. contract D { uint x; - function D(uint a) payable { + function D(uint a) public payable { x = a; } } @@ -233,11 +237,11 @@ creation-dependencies are not possible. contract C { D d = new D(4); // will be executed as part of C's constructor - function createD(uint arg) { + function createD(uint arg) public { D newD = new D(arg); } - function createAndEndowD(uint arg, uint amount) payable { + function createAndEndowD(uint arg, uint amount) public payable { // Send ether along with the creation D newD = (new D).value(amount)(arg); } @@ -270,16 +274,16 @@ Destructuring Assignments and Returning Multiple Values Solidity internally allows tuple types, i.e. a list of objects of potentially different types whose size is a constant at compile-time. Those tuples can be used to return multiple values at the same time and also assign them to multiple variables (or LValues in general) at the same time:: - pragma solidity ^0.4.0; + pragma solidity ^0.4.16; contract C { uint[] data; - function f() returns (uint, bool, uint) { + function f() public pure returns (uint, bool, uint) { return (7, true, 2); } - function g() { + function g() public { // Declares and assigns the variables. Specifying the type explicitly is not possible. var (x, b, y) = f(); // Assigns to a pre-existing variable. @@ -291,6 +295,7 @@ Solidity internally allows tuple types, i.e. a list of objects of potentially di // the rest of the values are discarded. (data.length,) = f(); // Sets the length to 7 // The same can be done on the left side. + // If the tuple begins in an empty component, the beginning values are discarded. (,data[3]) = f(); // Sets data[3] to 2 // Components can only be left out at the left-hand-side of assignments, with // one exception: @@ -326,10 +331,10 @@ As a result, the following code is illegal and cause the compiler to throw an er // This will not compile - pragma solidity ^0.4.0; + pragma solidity ^0.4.16; contract ScopingErrors { - function scoping() { + function scoping() public { uint i = 0; while (i++ < 1) { @@ -341,7 +346,7 @@ As a result, the following code is illegal and cause the compiler to throw an er } } - function minimalScoping() { + function minimalScoping() public { { uint same2 = 0; } @@ -351,7 +356,7 @@ As a result, the following code is illegal and cause the compiler to throw an er } } - function forLoopScoping() { + function forLoopScoping() public { for (uint same3 = 0; same3 < 1; same3++) { } @@ -364,9 +369,9 @@ In addition to this, if a variable is declared, it will be initialized at the be As a result, the following code is legal, despite being poorly written:: pragma solidity ^0.4.0; - + contract C { - function foo() returns (uint) { + function foo() public pure returns (uint) { // baz is implicitly initialized as 0 uint bar = 5; if (true) { @@ -412,7 +417,7 @@ and how ``assert`` can be used for internal error checking:: pragma solidity ^0.4.0; contract Sharer { - function sendHalf(address addr) payable returns (uint balance) { + function sendHalf(address addr) public payable returns (uint balance) { require(msg.value % 2 == 0); // Only allow even numbers uint balanceBeforeTransfer = this.balance; addr.transfer(msg.value / 2); diff --git a/docs/frequently-asked-questions.rst b/docs/frequently-asked-questions.rst index 5c427c692ec3..a6bead29fd6f 100644 --- a/docs/frequently-asked-questions.rst +++ b/docs/frequently-asked-questions.rst @@ -39,12 +39,13 @@ This is just the bytecode "data" sent along with the request. Is there a decompiler available? ================================ -There is no decompiler to Solidity. This is in principle possible -to some degree, but for example variable names will be lost and -great effort will be necessary to make it look similar to -the original source code. +There is no exact decompiler to Solidity, but +`Porosity `_ is close. +Because some information like variable names, comments, and +source code formatting is lost in the compilation process, +it is not possible to completely recover the original source code. -Bytecode can be decompiled to opcodes, a service that is provided by +Bytecode can be disassembled to opcodes, a service that is provided by several blockchain explorers. Contracts on the blockchain should have their original source @@ -111,10 +112,10 @@ array in the return statement. Pretty cool, huh? Example:: - pragma solidity ^0.4.0; + pragma solidity ^0.4.16; contract C { - function f() returns (uint8[5]) { + function f() public pure returns (uint8[5]) { string[4] memory adaArr = ["This", "is", "an", "array"]; return ([1, 2, 3, 4, 5]); } @@ -190,11 +191,11 @@ you should always convert it to a ``bytes`` first:: contract C { string s; - function append(byte c) { + function append(byte c) public { bytes(s).push(c); } - function set(uint i, byte c) { + function set(uint i, byte c) public { bytes(s)[i] = c; } } @@ -232,12 +233,14 @@ situation. If you do not want to throw, you can return a pair:: - pragma solidity ^0.4.0; + pragma solidity ^0.4.16; contract C { uint[] counters; function getCounter(uint index) + public + view returns (uint counter, bool error) { if (index >= counters.length) return (0, true); @@ -245,7 +248,7 @@ If you do not want to throw, you can return a pair:: return (counters[index], false); } - function checkCounter(uint index) { + function checkCounter(uint index) public view { var (counter, error) = getCounter(index); if (error) { // ... @@ -316,11 +319,11 @@ Example:: uint[] data1; uint[] data2; - function appendOne() { + function appendOne() public { append(data1); } - function appendTwo() { + function appendTwo() public { append(data2); } @@ -349,7 +352,7 @@ be created in memory, although it will be created in storage:: uint someVariable; uint[] data; - function f() { + function f() public { uint[] x; x.push(2); data = x; @@ -375,7 +378,7 @@ The correct way to do this is the following:: uint someVariable; uint[] data; - function f() { + function f() public { uint[] x = data; x.push(2); } @@ -431,14 +434,14 @@ What happens to a ``struct``'s mapping when copying over a ``struct``? This is a very interesting question. Suppose that we have a contract field set up like such:: - struct user { + struct User { mapping(string => string) comments; } - function somefunction { - user user1; + function somefunction public { + User user1; user1.comments["Hello"] = "World"; - user user2 = user1; + User user2 = user1; } In this case, the mapping of the struct being copied over into the userList is ignored as there is no "list of mapped keys". @@ -456,13 +459,13 @@ In this example:: pragma solidity ^0.4.0; contract B { - function B() payable {} + function B() public payable {} } contract A { address child; - function test() { + function test() public { child = (new B).value(10)(); //construct a new B with 10 wei } } @@ -501,17 +504,17 @@ Can a contract pass an array (static size) or string or ``bytes`` (dynamic size) Sure. Take care that if you cross the memory / storage boundary, independent copies will be created:: - pragma solidity ^0.4.0; + pragma solidity ^0.4.16; contract C { uint[20] x; - function f() { + function f() public { g(x); h(x); } - function g(uint[20] y) internal { + function g(uint[20] y) internal pure { y[2] = 3; } diff --git a/docs/grammar.txt b/docs/grammar.txt index ce3fd3ad5683..e700c9465050 100644 --- a/docs/grammar.txt +++ b/docs/grammar.txt @@ -27,14 +27,19 @@ ModifierInvocation = Identifier ( '(' ExpressionList? ')' )? FunctionDefinition = 'function' Identifier? ParameterList ( ModifierInvocation | StateMutability | 'external' | 'public' | 'internal' | 'private' )* ( 'returns' ParameterList )? ( ';' | Block ) -EventDefinition = 'event' Identifier IndexedParameterList 'anonymous'? ';' +EventDefinition = 'event' Identifier EventParameterList 'anonymous'? ';' EnumValue = Identifier EnumDefinition = 'enum' Identifier '{' EnumValue? (',' EnumValue)* '}' -IndexedParameterList = '(' ( TypeName 'indexed'? Identifier? (',' TypeName 'indexed'? Identifier?)* )? ')' -ParameterList = '(' ( TypeName Identifier? (',' TypeName Identifier?)* )? ')' -TypeNameList = '(' ( TypeName (',' TypeName )* )? ')' +ParameterList = '(' ( Parameter (',' Parameter)* )? ')' +Parameter = TypeName StorageLocation? Identifier? + +EventParameterList = '(' ( EventParameter (',' EventParameter )* )? ')' +EventParameter = TypeName 'indexed'? Identifier? + +FunctionTypeParameterList = '(' ( FunctionTypeParameter (',' FunctionTypeParameter )* )? ')' +FunctionTypeParameter = TypeName StorageLocation? // semantic restriction: mappings and structs (recursively) containing mappings // are not allowed in argument lists @@ -50,8 +55,8 @@ UserDefinedTypeName = Identifier ( '.' Identifier )* Mapping = 'mapping' '(' ElementaryTypeName '=>' TypeName ')' ArrayTypeName = TypeName '[' Expression? ']' -FunctionTypeName = 'function' TypeNameList ( 'internal' | 'external' | StateMutability )* - ( 'returns' TypeNameList )? +FunctionTypeName = 'function' FunctionTypeParameterList ( 'internal' | 'external' | StateMutability )* + ( 'returns' FunctionTypeParameterList )? StorageLocation = 'memory' | 'storage' StateMutability = 'pure' | 'constant' | 'view' | 'payable' diff --git a/docs/index.rst b/docs/index.rst index 3c617d36ba49..3df0af3c6c76 100644 --- a/docs/index.rst +++ b/docs/index.rst @@ -21,6 +21,14 @@ crowdfunding, blind auctions, multi-signature wallets and more. `Remix `_ (it can take a while to load, please be patient). +.. warning:: + Since software is written by humans, it can have bugs. Thus, also + smart contracts should be created following well-known best-practices in + software development. This includes code review, testing, audits and correctness proofs. + Also note that users are sometimes more confident in code than its authors. + Finally, blockchains have their own things to watch out for, so please take + a look at the section :ref:`security_considerations`. + Translations ------------ @@ -70,7 +78,7 @@ Available Solidity Integrations Configurable Solidty linter for Atom using Solium as a base. * `Solium `_ - A commandline linter for Solidity which strictly follows the rules prescribed by the `Solidity Style Guide `_. + Linter to identify and fix style and security issues in Solidity. * `Solhint `_ Solidity linter that provides security, style guide and best practice rules for smart contract validation. diff --git a/docs/installing-solidity.rst b/docs/installing-solidity.rst index b660cf027e5f..8f30f19972bd 100644 --- a/docs/installing-solidity.rst +++ b/docs/installing-solidity.rst @@ -2,9 +2,9 @@ .. _installing-solidity: -################### -Installing Solidity -################### +################################ +Installing the Solidity Compiler +################################ Versioning ========== @@ -18,30 +18,38 @@ will use the latest release. Remix ===== -If you just want to try Solidity for small contracts, you -can try `Remix `_ -which does not need any installation. If you want to use it -without connection to the Internet, you can go to -https://github.com/ethereum/browser-solidity/tree/gh-pages and -download the .ZIP file as explained on that page. +*We recommend Remix for small contracts and for quickly learning Solidity.* + +`Access Remix online `_, you don't need to install anything. +If you want to use it without connection to the Internet, go to +https://github.com/ethereum/browser-solidity/tree/gh-pages and download the .ZIP file as +explained on that page. + +Further options on this page detail installing commandline Solidity compiler software +on your computer. Choose a commandline compiler if you are working on a larger contract +or if you require more compilation options. npm / Node.js ============= -This is probably the most portable and most convenient way to install Solidity locally. +Use `npm` for a convenient and portable way to install `solcjs`, a Solidity compiler. The +`solcjs` program has less features than all options further down this page. Our +`Using the compiler ` documentation assumes you are using +the full-featured compiler, `solc`. So if you install `solcjs` from `npm` then you will +stop reading the documentation here and then continue to , -A platform-independent JavaScript library is provided by compiling the C++ source -into JavaScript using Emscripten. It can be used in projects directly (such as Remix). +Note: The `solc-js ` project is derived from the C++ +`solc` by using Emscripten. `solc-js` can be used in JavaScript projects directly (such as Remix). Please refer to the `solc-js `_ repository for instructions. -It also contains a commandline tool called `solcjs`, which can be installed via npm: - .. code:: bash npm install -g solc .. note:: + The commandline is named `solcjs`. + The comandline options of `solcjs` are not compatible with `solc` and tools (such as `geth`) expecting the behaviour of `solc` will not work with `solcjs`. @@ -63,7 +71,7 @@ output directories. Binary Packages =============== -Binary packages of Solidity available at +Binary packages of Solidity are available at `solidity/releases `_. We also have PPAs for Ubuntu. For the latest stable version. diff --git a/docs/introduction-to-smart-contracts.rst b/docs/introduction-to-smart-contracts.rst index aedc0c09f668..6425dcaabf73 100644 --- a/docs/introduction-to-smart-contracts.rst +++ b/docs/introduction-to-smart-contracts.rst @@ -21,11 +21,11 @@ Storage contract SimpleStorage { uint storedData; - function set(uint x) { + function set(uint x) public { storedData = x; } - function get() constant returns (uint) { + function get() public constant returns (uint) { return storedData; } } @@ -94,16 +94,16 @@ registering with username and password - all you need is an Ethereum keypair. // This is the constructor whose code is // run only when the contract is created. - function Coin() { + function Coin() public { minter = msg.sender; } - function mint(address receiver, uint amount) { + function mint(address receiver, uint amount) public { if (msg.sender != minter) return; balances[receiver] += amount; } - function send(address receiver, uint amount) { + function send(address receiver, uint amount) public { if (balances[msg.sender] < amount) return; balances[msg.sender] -= amount; balances[receiver] += amount; @@ -145,7 +145,7 @@ like this one. The :ref:`getter function` created by the ``pub is a bit more complex in this case. It roughly looks like the following:: - function balances(address _account) returns (uint) { + function balances(address _account) public view returns (uint) { return balances[_account]; } diff --git a/docs/julia.rst b/docs/julia.rst index 309e6b369d16..9e961a9ddd43 100644 --- a/docs/julia.rst +++ b/docs/julia.rst @@ -14,6 +14,13 @@ It can already be used for "inline assembly" inside Solidity and future versions of the Solidity compiler will even use JULIA as intermediate language. It should also be easy to build high-level optimizer stages for JULIA. +.. note:: + + Note that the flavour used for "inline assembly" does not have types + (everything is ``u256``) and the built-in functions are identical + to the EVM opcodes. Please resort to the inline assembly documentation + for details. + The core components of JULIA are functions, blocks, variables, literals, for-loops, if-statements, switch-statements, expressions and assignments to variables. diff --git a/docs/metadata.rst b/docs/metadata.rst index dbde87e86f80..5e37219e1de5 100644 --- a/docs/metadata.rst +++ b/docs/metadata.rst @@ -131,6 +131,8 @@ user interface for the contract. Furthermore, Mist can use the userdoc to display a confirmation message to the user whenever they interact with the contract. +Additional information about Ethereum Natural Specification (NatSpec) can be found `here `_. + Usage for Source Code Verification ================================== diff --git a/docs/security-considerations.rst b/docs/security-considerations.rst index 337a3d3fb089..49fd7ea4eefd 100644 --- a/docs/security-considerations.rst +++ b/docs/security-considerations.rst @@ -62,17 +62,34 @@ complete contract): /// Mapping of ether shares of the contract. mapping(address => uint) shares; /// Withdraw your share. - function withdraw() { + function withdraw() public { if (msg.sender.send(shares[msg.sender])) shares[msg.sender] = 0; } } The problem is not too serious here because of the limited gas as part -of ``send``, but it still exposes a weakness: Ether transfer always -includes code execution, so the recipient could be a contract that calls +of ``send``, but it still exposes a weakness: Ether transfer can always +include code execution, so the recipient could be a contract that calls back into ``withdraw``. This would let it get multiple refunds and -basically retrieve all the Ether in the contract. +basically retrieve all the Ether in the contract. In particular, the +following contract will allow an attacker to refund multiple times +as it uses ``call`` which forwards all remaining gas by default: + +:: + + pragma solidity ^0.4.0; + + // THIS CONTRACT CONTAINS A BUG - DO NOT USE + contract Fund { + /// Mapping of ether shares of the contract. + mapping(address => uint) shares; + /// Withdraw your share. + function withdraw() public { + if (msg.sender.call.value(shares[msg.sender])()) + shares[msg.sender] = 0; + } + } To avoid re-entrancy, you can use the Checks-Effects-Interactions pattern as outlined further below: @@ -85,7 +102,7 @@ outlined further below: /// Mapping of ether shares of the contract. mapping(address => uint) shares; /// Withdraw your share. - function withdraw() { + function withdraw() public { var share = shares[msg.sender]; shares[msg.sender] = 0; msg.sender.transfer(share); @@ -113,7 +130,7 @@ Sending and Receiving Ether - Neither contracts nor "external accounts" are currently able to prevent that someone sends them Ether. Contracts can react on and reject a regular transfer, but there are ways to move Ether without creating a message call. One way is to simply "mine to" - the contract address and the second way is using ``selfdestruct(x)``. + the contract address and the second way is using ``selfdestruct(x)``. - If a contract receives Ether (without a function being called), the fallback function is executed. If it does not have a fallback function, the Ether will be rejected (by throwing an exception). @@ -169,11 +186,11 @@ Never use tx.origin for authorization. Let's say you have a wallet contract like contract TxUserWallet { address owner; - function TxUserWallet() { + function TxUserWallet() public { owner = msg.sender; } - function transferTo(address dest, uint amount) { + function transferTo(address dest, uint amount) public { require(tx.origin == owner); dest.transfer(amount); } @@ -186,17 +203,17 @@ Now someone tricks you into sending ether to the address of this attack wallet: pragma solidity ^0.4.11; interface TxUserWallet { - function transferTo(address dest, uint amount); + function transferTo(address dest, uint amount) public; } contract TxAttackWallet { address owner; - function TxAttackWallet() { + function TxAttackWallet() public { owner = msg.sender; } - function() { + function() public { TxUserWallet(msg.sender).transferTo(owner, msg.sender.balance); } } diff --git a/docs/solidity-by-example.rst b/docs/solidity-by-example.rst index 9489665e9adf..b663083cdd5c 100644 --- a/docs/solidity-by-example.rst +++ b/docs/solidity-by-example.rst @@ -36,7 +36,7 @@ of votes. :: - pragma solidity ^0.4.11; + pragma solidity ^0.4.16; /// @title Voting with delegation. contract Ballot { @@ -66,7 +66,7 @@ of votes. Proposal[] public proposals; /// Create a new ballot to choose one of `proposalNames`. - function Ballot(bytes32[] proposalNames) { + function Ballot(bytes32[] proposalNames) public { chairperson = msg.sender; voters[chairperson].weight = 1; @@ -86,7 +86,7 @@ of votes. // Give `voter` the right to vote on this ballot. // May only be called by `chairperson`. - function giveRightToVote(address voter) { + function giveRightToVote(address voter) public { // If the argument of `require` evaluates to `false`, // it terminates and reverts all changes to // the state and to Ether balances. It is often @@ -99,7 +99,7 @@ of votes. } /// Delegate your vote to the voter `to`. - function delegate(address to) { + function delegate(address to) public { // assigns reference Voter storage sender = voters[msg.sender]; require(!sender.voted); @@ -140,7 +140,7 @@ of votes. /// Give your vote (including votes delegated to you) /// to proposal `proposals[proposal].name`. - function vote(uint proposal) { + function vote(uint proposal) public { Voter storage sender = voters[msg.sender]; require(!sender.voted); sender.voted = true; @@ -154,7 +154,7 @@ of votes. /// @dev Computes the winning proposal taking all /// previous votes into account. - function winningProposal() constant + function winningProposal() public view returns (uint winningProposal) { uint winningVoteCount = 0; @@ -169,7 +169,7 @@ of votes. // Calls winningProposal() function to get the index // of the winner contained in the proposals array and then // returns the name of the winner - function winnerName() constant + function winnerName() public view returns (bytes32 winnerName) { winnerName = proposals[winningProposal()].name; @@ -248,7 +248,7 @@ activate themselves. function SimpleAuction( uint _biddingTime, address _beneficiary - ) { + ) public { beneficiary = _beneficiary; auctionEnd = now + _biddingTime; } @@ -257,7 +257,7 @@ activate themselves. /// together with this transaction. /// The value will only be refunded if the /// auction is not won. - function bid() payable { + function bid() public payable { // No arguments are necessary, all // information is already part of // the transaction. The keyword payable @@ -286,7 +286,7 @@ activate themselves. } /// Withdraw a bid that was overbid. - function withdraw() returns (bool) { + function withdraw() public returns (bool) { uint amount = pendingReturns[msg.sender]; if (amount > 0) { // It is important to set this to zero because the recipient @@ -305,7 +305,7 @@ activate themselves. /// End the auction and send the highest bid /// to the beneficiary. - function auctionEnd() { + function auctionEnd() public { // It is a good guideline to structure functions that interact // with other contracts (i.e. they call functions or send Ether) // into three phases: @@ -405,7 +405,7 @@ high or low invalid bids. uint _biddingTime, uint _revealTime, address _beneficiary - ) { + ) public { beneficiary = _beneficiary; biddingEnd = now + _biddingTime; revealEnd = biddingEnd + _revealTime; @@ -421,6 +421,7 @@ high or low invalid bids. /// still make the required deposit. The same address can /// place multiple bids. function bid(bytes32 _blindedBid) + public payable onlyBefore(biddingEnd) { @@ -438,6 +439,7 @@ high or low invalid bids. bool[] _fake, bytes32[] _secret ) + public onlyAfter(biddingEnd) onlyBefore(revealEnd) { @@ -487,7 +489,7 @@ high or low invalid bids. } /// Withdraw a bid that was overbid. - function withdraw() { + function withdraw() public { uint amount = pendingReturns[msg.sender]; if (amount > 0) { // It is important to set this to zero because the recipient @@ -503,6 +505,7 @@ high or low invalid bids. /// End the auction and send the highest bid /// to the beneficiary. function auctionEnd() + public onlyAfter(revealEnd) { require(!ended); @@ -533,7 +536,7 @@ Safe Remote Purchase // Ensure that `msg.value` is an even number. // Division will truncate if it is an odd number. // Check via multiplication that it wasn't an odd number. - function Purchase() payable { + function Purchase() public payable { seller = msg.sender; value = msg.value / 2; require((2 * value) == msg.value); @@ -567,6 +570,7 @@ Safe Remote Purchase /// Can only be called by the seller before /// the contract is locked. function abort() + public onlySeller inState(State.Created) { @@ -580,6 +584,7 @@ Safe Remote Purchase /// The ether will be locked until confirmReceived /// is called. function confirmPurchase() + public inState(State.Created) condition(msg.value == (2 * value)) payable @@ -592,6 +597,7 @@ Safe Remote Purchase /// Confirm that you (the buyer) received the item. /// This will release the locked ether. function confirmReceived() + public onlyBuyer inState(State.Locked) { diff --git a/docs/structure-of-a-contract.rst b/docs/structure-of-a-contract.rst index 0b5548004c33..a9a7ed526040 100644 --- a/docs/structure-of-a-contract.rst +++ b/docs/structure-of-a-contract.rst @@ -8,7 +8,7 @@ Structure of a Contract Contracts in Solidity are similar to classes in object-oriented languages. Each contract can contain declarations of :ref:`structure-state-variables`, :ref:`structure-functions`, -:ref:`structure-function-modifiers`, :ref:`structure-events`, :ref:`structure-structs-types` and :ref:`structure-enum-types`. +:ref:`structure-function-modifiers`, :ref:`structure-events`, :ref:`structure-struct-types` and :ref:`structure-enum-types`. Furthermore, contracts can inherit from other contracts. .. _structure-state-variables: @@ -43,7 +43,7 @@ Functions are the executable units of code within a contract. pragma solidity ^0.4.0; contract SimpleAuction { - function bid() payable { // Function + function bid() public payable { // Function // ... } } @@ -72,7 +72,7 @@ Function modifiers can be used to amend the semantics of functions in a declarat _; } - function abort() onlySeller { // Modifier usage + function abort() public onlySeller { // Modifier usage // ... } } @@ -91,7 +91,7 @@ Events are convenience interfaces with the EVM logging facilities. contract SimpleAuction { event HighestBidIncreased(address bidder, uint amount); // Event - function bid() payable { + function bid() public payable { // ... HighestBidIncreased(msg.sender, msg.value); // Triggering event } @@ -100,9 +100,9 @@ Events are convenience interfaces with the EVM logging facilities. See :ref:`events` in contracts section for information on how events are declared and can be used from within a dapp. -.. _structure-structs-types: +.. _structure-struct-types: -Structs Types +Struct Types ============= Structs are custom defined types that can group several variables (see diff --git a/docs/style-guide.rst b/docs/style-guide.rst index a438b3d0abe7..4c0d44f05c03 100644 --- a/docs/style-guide.rst +++ b/docs/style-guide.rst @@ -86,17 +86,17 @@ Blank lines may be omitted between groups of related one-liners (such as stub fu Yes:: contract A { - function spam(); - function ham(); + function spam() public; + function ham() public; } contract B is A { - function spam() { + function spam() public { ... } - function ham() { + function ham() public { ... } } @@ -104,10 +104,10 @@ Yes:: No:: contract A { - function spam() { + function spam() public { ... } - function ham() { + function ham() public { ... } } @@ -169,26 +169,26 @@ Within a grouping, place the ``constant`` functions last. Yes:: contract A { - function A() { + function A() public { ... } - - function() { + + function() public { ... } - + // External functions // ... - + // External functions that are constant // ... - + // Public functions // ... - + // Internal functions // ... - + // Private functions // ... } @@ -196,7 +196,7 @@ Yes:: No:: contract A { - + // External functions // ... @@ -206,16 +206,16 @@ No:: // Public functions // ... - function A() { + function A() public { ... } - - function() { + + function() public { ... } // Internal functions - // ... + // ... } Whitespace in Expressions @@ -235,17 +235,17 @@ No:: Exception:: - function singleLine() { spam(); } + function singleLine() public { spam(); } Immediately before a comma, semicolon: Yes:: - function spam(uint i, Coin coin); + function spam(uint i, Coin coin) public; No:: - function spam(uint i , Coin coin) ; + function spam(uint i , Coin coin) public ; More than one space around an assignment or other operator to align with another: @@ -266,13 +266,13 @@ Don't include a whitespace in the fallback function: Yes:: - function() { + function() public { ... } No:: - - function () { + + function () public { ... } @@ -395,30 +395,30 @@ The opening brace should be preceeded by a single space. Yes:: - function increment(uint x) returns (uint) { + function increment(uint x) public pure returns (uint) { return x + 1; } - function increment(uint x) public onlyowner returns (uint) { + function increment(uint x) public pure onlyowner returns (uint) { return x + 1; } No:: - function increment(uint x) returns (uint) + function increment(uint x) public pure returns (uint) { return x + 1; } - function increment(uint x) returns (uint){ + function increment(uint x) public pure returns (uint){ return x + 1; } - function increment(uint x) returns (uint) { + function increment(uint x) public pure returns (uint) { return x + 1; } - function increment(uint x) returns (uint) { + function increment(uint x) public pure returns (uint) { return x + 1;} The visibility modifiers for a function should come before any custom @@ -450,14 +450,16 @@ Yes:: address d, address e, address f - ) { + ) + public + { doSomething(); } No:: function thisFunctionHasLotsOfArguments(address a, address b, address c, - address d, address e, address f) { + address d, address e, address f) public { doSomething(); } @@ -466,7 +468,7 @@ No:: address c, address d, address e, - address f) { + address f) public { doSomething(); } @@ -476,12 +478,12 @@ No:: address c, address d, address e, - address f) { + address f) public { doSomething(); } If a long function declaration has modifiers, then each modifier should be -dropped to it's own line. +dropped to its own line. Yes:: @@ -542,6 +544,7 @@ Yes:: B(param1) C(param2, param3) D(param4) + public { // do something with param5 } @@ -554,6 +557,7 @@ No:: B(param1) C(param2, param3) D(param4) + public { // do something with param5 } @@ -563,7 +567,8 @@ No:: function A(uint param1, uint param2, uint param3, uint param4, uint param5) B(param1) C(param2, param3) - D(param4) { + D(param4) + public { // do something with param5 } } @@ -572,7 +577,7 @@ When declaring short functions with a single statement, it is permissible to do Permissible:: - function shortFunction() { doSomething(); } + function shortFunction() public { doSomething(); } These guidelines for function declarations are intended to improve readability. Authors should use their best judgement as this guide does not try to cover all @@ -699,6 +704,12 @@ Contract and Library Names Contracts and libraries should be named using the CapWords style. Examples: ``SimpleToken``, ``SmartBank``, ``CertificateHashRepository``, ``Player``. +Struct Names +========================== + +Structs should be named using the CapWords style. Examples: ``MyCoin``, ``Position``, ``PositionXY``. + + Event Names =========== @@ -708,7 +719,7 @@ Events should be named using the CapWords style. Examples: ``Deposit``, ``Transf Function Names ============== -Functions should use mixedCase. Examples: ``getBalance``, ``transfer``, ``verifyOwner``, ``addMember``, ``changeOwner``. +Functions other than constructors should use mixedCase. Examples: ``getBalance``, ``transfer``, ``verifyOwner``, ``addMember``, ``changeOwner``. Function Argument Names @@ -739,6 +750,12 @@ Modifier Names Use mixedCase. Examples: ``onlyBy``, ``onlyAfter``, ``onlyDuringThePreSale``. +Enums +===== + +Enums, in the style of simple type declarations, should be named using the CapWords style. Examples: ``TokenGroup``, ``Frame``, ``HashStyle``, ``CharacterLocation``. + + Avoiding Naming Collisions ========================== diff --git a/docs/types.rst b/docs/types.rst index c716b95eac04..55eaa69ae9ea 100644 --- a/docs/types.rst +++ b/docs/types.rst @@ -79,8 +79,8 @@ Fixed Point Numbers Fixed point numbers are not fully supported by Solidity yet. They can be declared, but cannot be assigned to or from. -``fixed`` / ``ufixed``: Signed and unsigned fixed point number of various sizes. Keywords ``ufixedMxN`` and ``fixedMxN``, where ``M`` represent the number of bits taken by -the type and ``N`` represent how many decimal points are available. ``M`` must be divisible by 8 and goes from 8 to 256 bits. ``N`` must be between 0 and 80, inclusive. +``fixed`` / ``ufixed``: Signed and unsigned fixed point number of various sizes. Keywords ``ufixedMxN`` and ``fixedMxN``, where ``M`` represents the number of bits taken by +the type and ``N`` represents how many decimal points are available. ``M`` must be divisible by 8 and goes from 8 to 256 bits. ``N`` must be between 0 and 80, inclusive. ``ufixed`` and ``fixed`` are aliases for ``ufixed128x19`` and ``fixed128x19``, respectively. Operators: @@ -331,14 +331,14 @@ check the value ranges at runtime and a failure causes an exception. Enums need :: - pragma solidity ^0.4.0; + pragma solidity ^0.4.16; contract test { enum ActionChoices { GoLeft, GoRight, GoStraight, SitStill } ActionChoices choice; ActionChoices constant defaultChoice = ActionChoices.GoStraight; - function setGoStraight() { + function setGoStraight() public { choice = ActionChoices.GoStraight; } @@ -347,11 +347,11 @@ check the value ranges at runtime and a failure causes an exception. Enums need // for all matters external to Solidity. The integer type used is just // large enough to hold all enum values, i.e. if you have more values, // `uint16` will be used and so on. - function getChoice() returns (ActionChoices) { + function getChoice() public view returns (ActionChoices) { return choice; } - function getDefaultChoice() returns (uint) { + function getDefaultChoice() public pure returns (uint) { return uint(defaultChoice); } } @@ -409,23 +409,24 @@ just use ``f``, if you want to use its external form, use ``this.f``. Additionally, public (or external) functions also have a special member called ``selector``, which returns the :ref:`ABI function selector `:: - pragma solidity ^0.4.0; + pragma solidity ^0.4.16; contract Selector { - function f() returns (bytes4) { + function f() public view returns (bytes4) { return this.f.selector; } } Example that shows how to use internal function types:: - pragma solidity ^0.4.5; + pragma solidity ^0.4.16; library ArrayUtils { // internal functions can be used in internal library functions because // they will be part of the same code context - function map(uint[] memory self, function (uint) returns (uint) f) + function map(uint[] memory self, function (uint) pure returns (uint) f) internal + pure returns (uint[] memory r) { r = new uint[](self.length); @@ -435,9 +436,10 @@ Example that shows how to use internal function types:: } function reduce( uint[] memory self, - function (uint, uint) returns (uint) f + function (uint, uint) pure returns (uint) f ) internal + pure returns (uint r) { r = self[0]; @@ -445,23 +447,23 @@ Example that shows how to use internal function types:: r = f(r, self[i]); } } - function range(uint length) internal returns (uint[] memory r) { + function range(uint length) internal pure returns (uint[] memory r) { r = new uint[](length); for (uint i = 0; i < r.length; i++) { r[i] = i; } } } - + contract Pyramid { using ArrayUtils for *; - function pyramid(uint l) returns (uint) { + function pyramid(uint l) public pure returns (uint) { return ArrayUtils.range(l).map(square).reduce(sum); } - function square(uint x) internal returns (uint) { + function square(uint x) internal pure returns (uint) { return x * x; } - function sum(uint x, uint y) internal returns (uint) { + function sum(uint x, uint y) internal pure returns (uint) { return x + y; } } @@ -477,11 +479,11 @@ Another example that uses external function types:: } Request[] requests; event NewRequest(uint); - function query(bytes data, function(bytes memory) external callback) { + function query(bytes data, function(bytes memory) external callback) public { requests.push(Request(data, callback)); NewRequest(requests.length - 1); } - function reply(uint requestID, bytes response) { + function reply(uint requestID, bytes response) public { // Here goes the check that the reply comes from a trusted source requests[requestID].callback(response); } @@ -492,7 +494,7 @@ Another example that uses external function types:: function buySomething() { oracle.query("USD", this.oracleResponse); } - function oracleResponse(bytes response) { + function oracleResponse(bytes response) public { require(msg.sender == address(oracle)); // Use the data } @@ -543,7 +545,7 @@ memory-stored reference type do not create a copy. uint[] x; // the data location of x is storage // the data location of memoryArray is memory - function f(uint[] memoryArray) { + function f(uint[] memoryArray) public { x = memoryArray; // works, copies the whole array to storage var y = x; // works, assigns a pointer, data location of y is storage y[7]; // fine, returns the 8th element @@ -560,7 +562,7 @@ memory-stored reference type do not create a copy. } function g(uint[] storage storageArray) internal {} - function h(uint[] memoryArray) {} + function h(uint[] memoryArray) public {} } Summary @@ -620,10 +622,10 @@ the ``.length`` member. :: - pragma solidity ^0.4.0; + pragma solidity ^0.4.16; contract C { - function f(uint len) { + function f(uint len) public pure { uint[] memory a = new uint[](7); bytes memory b = new bytes(len); // Here we have a.length == 7 and b.length == len @@ -641,13 +643,13 @@ assigned to a variable right away. :: - pragma solidity ^0.4.0; + pragma solidity ^0.4.16; contract C { - function f() { + function f() public pure { g([uint(1), 2, 3]); } - function g(uint[3] _data) { + function g(uint[3] _data) public pure { // ... } } @@ -667,7 +669,7 @@ possible: pragma solidity ^0.4.0; contract C { - function f() { + function f() public { // The next line creates a type error because uint[3] memory // cannot be converted to uint[] memory. uint[] x = [uint(1), 3, 4]; @@ -703,7 +705,7 @@ Members :: - pragma solidity ^0.4.0; + pragma solidity ^0.4.16; contract ArrayContract { uint[2**20] m_aLotOfIntegers; @@ -712,23 +714,23 @@ Members bool[2][] m_pairsOfFlags; // newPairs is stored in memory - the default for function arguments - function setAllFlagPairs(bool[2][] newPairs) { + function setAllFlagPairs(bool[2][] newPairs) public { // assignment to a storage array replaces the complete array m_pairsOfFlags = newPairs; } - function setFlagPair(uint index, bool flagA, bool flagB) { + function setFlagPair(uint index, bool flagA, bool flagB) public { // access to a non-existing index will throw an exception m_pairsOfFlags[index][0] = flagA; m_pairsOfFlags[index][1] = flagB; } - function changeFlagArraySize(uint newSize) { + function changeFlagArraySize(uint newSize) public { // if the new size is smaller, removed array elements will be cleared m_pairsOfFlags.length = newSize; } - function clear() { + function clear() public { // these clear the arrays completely delete m_pairsOfFlags; delete m_aLotOfIntegers; @@ -738,20 +740,20 @@ Members bytes m_byteData; - function byteArrays(bytes data) { + function byteArrays(bytes data) public { // byte arrays ("bytes") are different as they are stored without padding, // but can be treated identical to "uint8[]" m_byteData = data; m_byteData.length += 7; - m_byteData[3] = 8; + m_byteData[3] = byte(8); delete m_byteData[2]; } - function addFlag(bool[2] flag) returns (uint) { + function addFlag(bool[2] flag) public returns (uint) { return m_pairsOfFlags.push(flag); } - function createMemoryArray(uint size) returns (bytes) { + function createMemoryArray(uint size) public pure returns (bytes) { // Dynamic memory arrays are created using `new`: uint[2][] memory arrayOfPairs = new uint[2][](size); // Create a dynamic byte array: @@ -795,13 +797,13 @@ shown in the following example: uint numCampaigns; mapping (uint => Campaign) campaigns; - function newCampaign(address beneficiary, uint goal) returns (uint campaignID) { + function newCampaign(address beneficiary, uint goal) public returns (uint campaignID) { campaignID = numCampaigns++; // campaignID is return variable // Creates new struct and saves in storage. We leave out the mapping type. campaigns[campaignID] = Campaign(beneficiary, goal, 0, 0); } - function contribute(uint campaignID) payable { + function contribute(uint campaignID) public payable { Campaign storage c = campaigns[campaignID]; // Creates a new temporary memory struct, initialised with the given values // and copies it over to storage. @@ -810,7 +812,7 @@ shown in the following example: c.amount += msg.value; } - function checkGoalReached(uint campaignID) returns (bool reached) { + function checkGoalReached(uint campaignID) public returns (bool reached) { Campaign storage c = campaigns[campaignID]; if (c.amount < c.fundingGoal) return false; @@ -872,13 +874,13 @@ for each ``_KeyType``, recursively. contract MappingExample { mapping(address => uint) public balances; - function update(uint newBalance) { + function update(uint newBalance) public { balances[msg.sender] = newBalance; } } contract MappingUser { - function f() returns (uint) { + function f() public returns (uint) { MappingExample m = new MappingExample(); m.update(100); return m.balances(this); @@ -897,7 +899,7 @@ Operators Involving LValues If ``a`` is an LValue (i.e. a variable or something that can be assigned to), the following operators are available as shorthands: -``a += e`` is equivalent to ``a = a + e``. The operators ``-=``, ``*=``, ``/=``, ``%=``, ``a |=``, ``&=`` and ``^=`` are defined accordingly. ``a++`` and ``a--`` are equivalent to ``a += 1`` / ``a -= 1`` but the expression itself still has the previous value of ``a``. In contrast, ``--a`` and ``++a`` have the same effect on ``a`` but return the value after the change. +``a += e`` is equivalent to ``a = a + e``. The operators ``-=``, ``*=``, ``/=``, ``%=``, ``|=``, ``&=`` and ``^=`` are defined accordingly. ``a++`` and ``a--`` are equivalent to ``a += 1`` / ``a -= 1`` but the expression itself still has the previous value of ``a``. In contrast, ``--a`` and ``++a`` have the same effect on ``a`` but return the value after the change. delete ------ @@ -916,11 +918,11 @@ It is important to note that ``delete a`` really behaves like an assignment to ` uint data; uint[] dataArray; - function f() { + function f() public { uint x = data; delete x; // sets x to 0, does not affect data delete data; // sets data to 0, does not affect x which still holds a copy - uint[] y = dataArray; + uint[] storage y = dataArray; delete dataArray; // this sets dataArray.length to zero, but as uint[] is a complex object, also // y is affected which is an alias to the storage object // On the other hand: "delete y" is not valid, as assignments to local variables diff --git a/docs/units-and-global-variables.rst b/docs/units-and-global-variables.rst index 8261bdde711f..ce58cf56fece 100644 --- a/docs/units-and-global-variables.rst +++ b/docs/units-and-global-variables.rst @@ -34,7 +34,7 @@ library has to be updated by an external oracle. These suffixes cannot be applied to variables. If you want to interpret some input variable in e.g. days, you can do it in the following way:: - function f(uint start, uint daysAfter) { + function f(uint start, uint daysAfter) public { if (now >= start + daysAfter * 1 days) { // ... } diff --git a/docs/using-the-compiler.rst b/docs/using-the-compiler.rst index c12750c8231d..42cc807aa33c 100644 --- a/docs/using-the-compiler.rst +++ b/docs/using-the-compiler.rst @@ -147,11 +147,11 @@ Input Description }, // Enable the abi and opcodes output of MyContract defined in file def. "def": { - "MyContract": [ "abi", "evm.opcodes" ] + "MyContract": [ "abi", "evm.bytecode.opcodes" ] }, // Enable the source map output of every single contract. "*": { - "*": [ "evm.sourceMap" ] + "*": [ "evm.bytecode.sourceMap" ] }, // Enable the legacy AST output of every single file. "*": { @@ -177,7 +177,8 @@ Output Description start: 0, end: 100 ], - // Mandatory: Error type, such as "TypeError", "InternalCompilerError", "Exception", etc + // Mandatory: Error type, such as "TypeError", "InternalCompilerError", "Exception", etc. + // See below for complete list of types. type: "TypeError", // Mandatory: Component where the error originated, such as "general", "ewasm", etc. component: "general", @@ -273,3 +274,21 @@ Output Description } } } + + +Error types +~~~~~~~~~~~ + +1. ``JSONError``: JSON input doesn't conform to the required format, e.g. input is not a JSON object, the language is not supported, etc. +2. ``IOError``: IO and import processing errors, such as unresolvable URL or hash mismatch in supplied sources. +3. ``ParserError``: Source code doesn't conform to the language rules. +4. ``DocstringParsingError``: The NatSpec tags in the comment block cannot be parsed. +5. ``SyntaxError``: Syntactical error, such as ``continue`` is used outside of a ``for`` loop. +6. ``DeclarationError``: Invalid, unresolvable or clashing identifier names. e.g. ``Identifier not found`` +7. ``TypeError``: Error within the type system, such as invalid type conversions, invalid assignments, etc. +8. ``UnimplementedFeatureError``: Feature is not supported by the compiler, but is expected to be supported in future versions. +9. ``InternalCompilerError``: Internal bug triggered in the compiler - this should be reported as an issue. +10. ``Exception``: Unknown failure during compilation - this should be reported as an issue. +11. ``CompilerError``: Invalid use of the compiler stack - this should be reported as an issue. +12. ``FatalError``: Fatal error not processed correctly - this should be reported as an issue. +13. ``Warning``: A warning, which didn't stop the compilation, but should be addressed if possible. diff --git a/libdevcore/CommonData.h b/libdevcore/CommonData.h index e76a09491133..e410af5cc700 100644 --- a/libdevcore/CommonData.h +++ b/libdevcore/CommonData.h @@ -155,6 +155,14 @@ inline std::string formatNumber(bigint const& _value) return _value.str(); } +inline std::string formatNumber(u256 const& _value) +{ + if (_value > 0x1000000) + return toHex(toCompactBigEndian(_value), 2, HexPrefix::Add); + else + return _value.str(); +} + inline std::string toCompactHexWithPrefix(u256 val) { std::ostringstream ret; @@ -183,6 +191,12 @@ template std::vector& operator+=(std::vector& _a, U con _a.push_back(i); return _a; } +/// Concatenate the contents of a container onto a vector, move variant. +template std::vector& operator+=(std::vector& _a, U&& _b) +{ + std::move(_b.begin(), _b.end(), std::back_inserter(_a)); + return _a; +} /// Concatenate the contents of a container onto a set template std::set& operator+=(std::set& _a, U const& _b) { @@ -197,6 +211,17 @@ inline std::vector operator+(std::vector const& _a, std::vector const& ret += _b; return ret; } +/// Concatenate two vectors of elements, moving them. +template +inline std::vector operator+(std::vector&& _a, std::vector&& _b) +{ + std::vector ret(std::move(_a)); + if (&_a == &_b) + ret += ret; + else + ret += std::move(_b); + return ret; +} template bool contains(T const& _t, V const& _v) diff --git a/libdevcore/StringUtils.cpp b/libdevcore/StringUtils.cpp new file mode 100644 index 000000000000..2ff86bd56c32 --- /dev/null +++ b/libdevcore/StringUtils.cpp @@ -0,0 +1,101 @@ +/* + This file is part of solidity. + + solidity is free software: you can redistribute it and/or modify + it under the terms of the GNU General Public License as published by + the Free Software Foundation, either version 3 of the License, or + (at your option) any later version. + + solidity is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + GNU General Public License for more details. + + You should have received a copy of the GNU General Public License + along with solidity. If not, see . +*/ +/** @file StringUtils.h + * @author Balajiganapathi S + * @date 2017 + * + * String routines + */ + +#include "StringUtils.h" +#include +#include +#include + +using namespace std; +using namespace dev; + +bool dev::stringWithinDistance(string const& _str1, string const& _str2, size_t _maxDistance) +{ + if (_str1 == _str2) + return true; + + size_t n1 = _str1.size(); + size_t n2 = _str2.size(); + size_t distance = stringDistance(_str1, _str2); + + // if distance is not greater than _maxDistance, and distance is strictly less than length of both names, they can be considered similar + // this is to avoid irrelevant suggestions + return distance <= _maxDistance && distance < n1 && distance < n2; +} + +size_t dev::stringDistance(string const& _str1, string const& _str2) +{ + size_t n1 = _str1.size(); + size_t n2 = _str2.size(); + // Optimize by storing only last 2 rows and current row. So first index is considered modulo 3 + // This is a two-dimensional array of size 3 x (n2 + 1). + vector dp(3 * (n2 + 1)); + + // In this dp formulation of Damerau–Levenshtein distance we are assuming that the strings are 1-based to make base case storage easier. + // So index accesser to _name1 and _name2 have to be adjusted accordingly + for (size_t i1 = 0; i1 <= n1; ++i1) + for (size_t i2 = 0; i2 <= n2; ++i2) + { + size_t x = 0; + if (min(i1, i2) == 0) // base case + x = max(i1, i2); + else + { + size_t left = dp[(i1 - 1) % 3 + i2 * 3]; + size_t up = dp[(i1 % 3) + (i2 - 1) * 3]; + size_t upleft = dp[((i1 - 1) % 3) + (i2 - 1) * 3]; + // deletion and insertion + x = min(left + 1, up + 1); + if (_str1[i1-1] == _str2[i2-1]) + // same chars, can skip + x = min(x, upleft); + else + // different chars so try substitution + x = min(x, upleft + 1); + + // transposing + if (i1 > 1 && i2 > 1 && _str1[i1 - 1] == _str2[i2 - 2] && _str1[i1 - 2] == _str2[i2 - 1]) + x = min(x, dp[((i1 - 2) % 3) + (i2 - 2) * 3] + 1); + } + dp[(i1 % 3) + i2 * 3] = x; + } + + return dp[(n1 % 3) + n2 * 3]; +} + +string dev::quotedAlternativesList(vector const& suggestions) +{ + if (suggestions.empty()) + return ""; + if (suggestions.size() == 1) + return "\"" + suggestions.front() + "\""; + + string choices = "\"" + suggestions.front() + "\""; + for (size_t i = 1; i + 1 < suggestions.size(); ++i) + choices += ", \"" + suggestions[i] + "\""; + + choices += " or \"" + suggestions.back() + "\""; + + return choices; +} + diff --git a/libdevcore/StringUtils.h b/libdevcore/StringUtils.h new file mode 100644 index 000000000000..acd93e326ec0 --- /dev/null +++ b/libdevcore/StringUtils.h @@ -0,0 +1,39 @@ +/* + This file is part of solidity. + + solidity is free software: you can redistribute it and/or modify + it under the terms of the GNU General Public License as published by + the Free Software Foundation, either version 3 of the License, or + (at your option) any later version. + + solidity is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + GNU General Public License for more details. + + You should have received a copy of the GNU General Public License + along with solidity. If not, see . +*/ +/** @file StringUtils.h + * @author Balajiganapathi S + * @date 2017 + * + * String routines + */ + +#pragma once + +#include +#include + +namespace dev +{ + +// Calculates the Damerau–Levenshtein distance between _str1 and _str2 and returns true if that distance is not greater than _maxDistance +bool stringWithinDistance(std::string const& _str1, std::string const& _str2, size_t _maxDistance); +// Calculates the Damerau–Levenshtein distance between _str1 and _str2 +size_t stringDistance(std::string const& _str1, std::string const& _str2); +// Return a string having elements of suggestions as quoted, alternative suggestions. e.g. "a", "b" or "c" +std::string quotedAlternativesList(std::vector const& suggestions); + +} diff --git a/libevmasm/Assembly.cpp b/libevmasm/Assembly.cpp index 5fab24e191c4..b9fedf26c40e 100644 --- a/libevmasm/Assembly.cpp +++ b/libevmasm/Assembly.cpp @@ -283,6 +283,11 @@ Json::Value Assembly::assemblyJSON(StringMap const& _sourceCodes) const createJsonValue("PUSHLIB", i.location().start, i.location().end, m_libraries.at(h256(i.data()))) ); break; + case PushDeployTimeAddress: + collection.append( + createJsonValue("PUSHDEPLOYADDRESS", i.location().start, i.location().end) + ); + break; case Tag: collection.append( createJsonValue("tag", i.location().start, i.location().end, string(i.data()))); @@ -590,6 +595,10 @@ LinkerObject const& Assembly::assemble() const ret.linkReferences[ret.bytecode.size()] = m_libraries.at(i.data()); ret.bytecode.resize(ret.bytecode.size() + 20); break; + case PushDeployTimeAddress: + ret.bytecode.push_back(byte(Instruction::PUSH20)); + ret.bytecode.resize(ret.bytecode.size() + 20); + break; case Tag: assertThrow(i.data() != 0, AssemblyException, "Invalid tag position."); assertThrow(i.splitForeignPushTag().first == size_t(-1), AssemblyException, "Foreign tag."); diff --git a/libevmasm/AssemblyItem.cpp b/libevmasm/AssemblyItem.cpp index 649630217bda..5af618ff545a 100644 --- a/libevmasm/AssemblyItem.cpp +++ b/libevmasm/AssemblyItem.cpp @@ -68,6 +68,7 @@ unsigned AssemblyItem::bytesRequired(unsigned _addressLength) const case PushSub: return 1 + _addressLength; case PushLibraryAddress: + case PushDeployTimeAddress: return 1 + 20; default: break; @@ -97,6 +98,7 @@ int AssemblyItem::returnValues() const case PushSubSize: case PushProgramSize: case PushLibraryAddress: + case PushDeployTimeAddress: return 1; case Tag: return 0; @@ -119,6 +121,7 @@ bool AssemblyItem::canBeFunctional() const case PushSubSize: case PushProgramSize: case PushLibraryAddress: + case PushDeployTimeAddress: return true; case Tag: return false; @@ -190,6 +193,9 @@ string AssemblyItem::toAssemblyText() const case PushLibraryAddress: text = string("linkerSymbol(\"") + toHex(data()) + string("\")"); break; + case PushDeployTimeAddress: + text = string("deployTimeAddress()"); + break; case UndefinedItem: assertThrow(false, AssemblyException, "Invalid assembly item."); break; @@ -252,6 +258,9 @@ ostream& dev::eth::operator<<(ostream& _out, AssemblyItem const& _item) _out << " PushLibraryAddress " << hash.substr(0, 8) + "..." + hash.substr(hash.length() - 8); break; } + case PushDeployTimeAddress: + _out << " PushDeployTimeAddress"; + break; case UndefinedItem: _out << " ???"; break; diff --git a/libevmasm/AssemblyItem.h b/libevmasm/AssemblyItem.h index d38db927ca66..5319a2b6a4ca 100644 --- a/libevmasm/AssemblyItem.h +++ b/libevmasm/AssemblyItem.h @@ -46,7 +46,8 @@ enum AssemblyItemType { PushProgramSize, Tag, PushData, - PushLibraryAddress ///< Push a currently unknown address of another (library) contract. + PushLibraryAddress, ///< Push a currently unknown address of another (library) contract. + PushDeployTimeAddress ///< Push an address to be filled at deploy time. Should not be touched by the optimizer. }; class Assembly; diff --git a/libevmasm/ExpressionClasses.cpp b/libevmasm/ExpressionClasses.cpp index fc283b0bcef7..69b33ec5f8ec 100644 --- a/libevmasm/ExpressionClasses.cpp +++ b/libevmasm/ExpressionClasses.cpp @@ -181,7 +181,7 @@ string ExpressionClasses::fullDAGToString(ExpressionClasses::Id _id) const return str.str(); } -ExpressionClasses::Id ExpressionClasses::tryToSimplify(Expression const& _expr, bool _secondRun) +ExpressionClasses::Id ExpressionClasses::tryToSimplify(Expression const& _expr) { static Rules rules; @@ -202,14 +202,7 @@ ExpressionClasses::Id ExpressionClasses::tryToSimplify(Expression const& _expr, //cout << "with rule " << match->first.toString() << endl; //ExpressionTemplate t(match->second()); //cout << "to " << match->second().toString() << endl; - return rebuildExpression(ExpressionTemplate(match->second(), _expr.item->location())); - } - - if (!_secondRun && _expr.arguments.size() == 2 && SemanticInformation::isCommutativeOperation(*_expr.item)) - { - Expression expr = _expr; - swap(expr.arguments[0], expr.arguments[1]); - return tryToSimplify(expr, true); + return rebuildExpression(ExpressionTemplate(match->action(), _expr.item->location())); } return -1; diff --git a/libevmasm/ExpressionClasses.h b/libevmasm/ExpressionClasses.h index 6b426e971b17..df8082f93775 100644 --- a/libevmasm/ExpressionClasses.h +++ b/libevmasm/ExpressionClasses.h @@ -108,8 +108,7 @@ class ExpressionClasses private: /// Tries to simplify the given expression. /// @returns its class if it possible or Id(-1) otherwise. - /// @param _secondRun is set to true for the second run where arguments of commutative expressions are reversed - Id tryToSimplify(Expression const& _expr, bool _secondRun = false); + Id tryToSimplify(Expression const& _expr); /// Rebuilds an expression from a (matched) pattern. Id rebuildExpression(ExpressionTemplate const& _template); diff --git a/libevmasm/GasMeter.cpp b/libevmasm/GasMeter.cpp index dad952bc3920..543f1cbc3e32 100644 --- a/libevmasm/GasMeter.cpp +++ b/libevmasm/GasMeter.cpp @@ -52,6 +52,7 @@ GasMeter::GasConsumption GasMeter::estimateMax(AssemblyItem const& _item, bool _ case PushSubSize: case PushProgramSize: case PushLibraryAddress: + case PushDeployTimeAddress: gas = runGas(Instruction::PUSH1); break; case Tag: diff --git a/libevmasm/RuleList.h b/libevmasm/RuleList.h new file mode 100644 index 000000000000..da522cec8a7e --- /dev/null +++ b/libevmasm/RuleList.h @@ -0,0 +1,266 @@ +/* + This file is part of solidity. + + solidity is free software: you can redistribute it and/or modify + it under the terms of the GNU General Public License as published by + the Free Software Foundation, either version 3 of the License, or + (at your option) any later version. + + solidity is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + GNU General Public License for more details. + + You should have received a copy of the GNU General Public License + along with solidity. If not, see . +*/ +/** + * @date 2018 + * Templatized list of simplification rules. + */ + +#pragma once + +#include +#include + +#include +#include + +#include + +namespace dev +{ +namespace solidity +{ + +template S divWorkaround(S const& _a, S const& _b) +{ + return (S)(bigint(_a) / bigint(_b)); +} + +template S modWorkaround(S const& _a, S const& _b) +{ + return (S)(bigint(_a) % bigint(_b)); +} + +/// @returns a list of simplification rules given certain match placeholders. +/// A, B and C should represent constants, X and Y arbitrary expressions. +/// The simplifications should neven change the order of evaluation of +/// arbitrary operations. +template +std::vector> simplificationRuleList( + Pattern A, + Pattern B, + Pattern C, + Pattern X, + Pattern Y +) +{ + std::vector> rules; + rules += std::vector>{ + // arithmetics on constants + {{Instruction::ADD, {A, B}}, [=]{ return A.d() + B.d(); }, false}, + {{Instruction::MUL, {A, B}}, [=]{ return A.d() * B.d(); }, false}, + {{Instruction::SUB, {A, B}}, [=]{ return A.d() - B.d(); }, false}, + {{Instruction::DIV, {A, B}}, [=]{ return B.d() == 0 ? 0 : divWorkaround(A.d(), B.d()); }, false}, + {{Instruction::SDIV, {A, B}}, [=]{ return B.d() == 0 ? 0 : s2u(divWorkaround(u2s(A.d()), u2s(B.d()))); }, false}, + {{Instruction::MOD, {A, B}}, [=]{ return B.d() == 0 ? 0 : modWorkaround(A.d(), B.d()); }, false}, + {{Instruction::SMOD, {A, B}}, [=]{ return B.d() == 0 ? 0 : s2u(modWorkaround(u2s(A.d()), u2s(B.d()))); }, false}, + {{Instruction::EXP, {A, B}}, [=]{ return u256(boost::multiprecision::powm(bigint(A.d()), bigint(B.d()), bigint(1) << 256)); }, false}, + {{Instruction::NOT, {A}}, [=]{ return ~A.d(); }, false}, + {{Instruction::LT, {A, B}}, [=]() -> u256 { return A.d() < B.d() ? 1 : 0; }, false}, + {{Instruction::GT, {A, B}}, [=]() -> u256 { return A.d() > B.d() ? 1 : 0; }, false}, + {{Instruction::SLT, {A, B}}, [=]() -> u256 { return u2s(A.d()) < u2s(B.d()) ? 1 : 0; }, false}, + {{Instruction::SGT, {A, B}}, [=]() -> u256 { return u2s(A.d()) > u2s(B.d()) ? 1 : 0; }, false}, + {{Instruction::EQ, {A, B}}, [=]() -> u256 { return A.d() == B.d() ? 1 : 0; }, false}, + {{Instruction::ISZERO, {A}}, [=]() -> u256 { return A.d() == 0 ? 1 : 0; }, false}, + {{Instruction::AND, {A, B}}, [=]{ return A.d() & B.d(); }, false}, + {{Instruction::OR, {A, B}}, [=]{ return A.d() | B.d(); }, false}, + {{Instruction::XOR, {A, B}}, [=]{ return A.d() ^ B.d(); }, false}, + {{Instruction::BYTE, {A, B}}, [=]{ return A.d() >= 32 ? 0 : (B.d() >> unsigned(8 * (31 - A.d()))) & 0xff; }, false}, + {{Instruction::ADDMOD, {A, B, C}}, [=]{ return C.d() == 0 ? 0 : u256((bigint(A.d()) + bigint(B.d())) % C.d()); }, false}, + {{Instruction::MULMOD, {A, B, C}}, [=]{ return C.d() == 0 ? 0 : u256((bigint(A.d()) * bigint(B.d())) % C.d()); }, false}, + {{Instruction::MULMOD, {A, B, C}}, [=]{ return A.d() * B.d(); }, false}, + {{Instruction::SIGNEXTEND, {A, B}}, [=]() -> u256 { + if (A.d() >= 31) + return B.d(); + unsigned testBit = unsigned(A.d()) * 8 + 7; + u256 mask = (u256(1) << testBit) - 1; + return u256(boost::multiprecision::bit_test(B.d(), testBit) ? B.d() | ~mask : B.d() & mask); + }, false}, + + // invariants involving known constants + {{Instruction::ADD, {X, 0}}, [=]{ return X; }, false}, + {{Instruction::ADD, {0, X}}, [=]{ return X; }, false}, + {{Instruction::SUB, {X, 0}}, [=]{ return X; }, false}, + {{Instruction::MUL, {X, 0}}, [=]{ return u256(0); }, true}, + {{Instruction::MUL, {0, X}}, [=]{ return u256(0); }, true}, + {{Instruction::MUL, {X, 1}}, [=]{ return X; }, false}, + {{Instruction::MUL, {1, X}}, [=]{ return X; }, false}, + {{Instruction::MUL, {X, u256(-1)}}, [=]() -> Pattern { return {Instruction::SUB, {0, X}}; }, false}, + {{Instruction::MUL, {u256(-1), X}}, [=]() -> Pattern { return {Instruction::SUB, {0, X}}; }, false}, + {{Instruction::DIV, {X, 0}}, [=]{ return u256(0); }, true}, + {{Instruction::DIV, {0, X}}, [=]{ return u256(0); }, true}, + {{Instruction::DIV, {X, 1}}, [=]{ return X; }, false}, + {{Instruction::SDIV, {X, 0}}, [=]{ return u256(0); }, true}, + {{Instruction::SDIV, {0, X}}, [=]{ return u256(0); }, true}, + {{Instruction::SDIV, {X, 1}}, [=]{ return X; }, false}, + {{Instruction::AND, {X, ~u256(0)}}, [=]{ return X; }, false}, + {{Instruction::AND, {~u256(0), X}}, [=]{ return X; }, false}, + {{Instruction::AND, {X, 0}}, [=]{ return u256(0); }, true}, + {{Instruction::AND, {0, X}}, [=]{ return u256(0); }, true}, + {{Instruction::OR, {X, 0}}, [=]{ return X; }, false}, + {{Instruction::OR, {0, X}}, [=]{ return X; }, false}, + {{Instruction::OR, {X, ~u256(0)}}, [=]{ return ~u256(0); }, true}, + {{Instruction::OR, {~u256(0), X}}, [=]{ return ~u256(0); }, true}, + {{Instruction::XOR, {X, 0}}, [=]{ return X; }, false}, + {{Instruction::XOR, {0, X}}, [=]{ return X; }, false}, + {{Instruction::MOD, {X, 0}}, [=]{ return u256(0); }, true}, + {{Instruction::MOD, {0, X}}, [=]{ return u256(0); }, true}, + {{Instruction::EQ, {X, 0}}, [=]() -> Pattern { return {Instruction::ISZERO, {X}}; }, false }, + {{Instruction::EQ, {0, X}}, [=]() -> Pattern { return {Instruction::ISZERO, {X}}; }, false }, + + // operations involving an expression and itself + {{Instruction::AND, {X, X}}, [=]{ return X; }, true}, + {{Instruction::OR, {X, X}}, [=]{ return X; }, true}, + {{Instruction::XOR, {X, X}}, [=]{ return u256(0); }, true}, + {{Instruction::SUB, {X, X}}, [=]{ return u256(0); }, true}, + {{Instruction::EQ, {X, X}}, [=]{ return u256(1); }, true}, + {{Instruction::LT, {X, X}}, [=]{ return u256(0); }, true}, + {{Instruction::SLT, {X, X}}, [=]{ return u256(0); }, true}, + {{Instruction::GT, {X, X}}, [=]{ return u256(0); }, true}, + {{Instruction::SGT, {X, X}}, [=]{ return u256(0); }, true}, + {{Instruction::MOD, {X, X}}, [=]{ return u256(0); }, true}, + + // logical instruction combinations + {{Instruction::NOT, {{Instruction::NOT, {X}}}}, [=]{ return X; }, false}, + {{Instruction::XOR, {X, {Instruction::XOR, {X, Y}}}}, [=]{ return Y; }, true}, + {{Instruction::XOR, {X, {Instruction::XOR, {Y, X}}}}, [=]{ return Y; }, true}, + {{Instruction::XOR, {{Instruction::XOR, {X, Y}}, X}}, [=]{ return Y; }, true}, + {{Instruction::XOR, {{Instruction::XOR, {Y, X}}, X}}, [=]{ return Y; }, true}, + {{Instruction::OR, {X, {Instruction::AND, {X, Y}}}}, [=]{ return X; }, true}, + {{Instruction::OR, {X, {Instruction::AND, {Y, X}}}}, [=]{ return X; }, true}, + {{Instruction::OR, {{Instruction::AND, {X, Y}}, X}}, [=]{ return X; }, true}, + {{Instruction::OR, {{Instruction::AND, {Y, X}}, X}}, [=]{ return X; }, true}, + {{Instruction::AND, {X, {Instruction::OR, {X, Y}}}}, [=]{ return X; }, true}, + {{Instruction::AND, {X, {Instruction::OR, {Y, X}}}}, [=]{ return X; }, true}, + {{Instruction::AND, {{Instruction::OR, {X, Y}}, X}}, [=]{ return X; }, true}, + {{Instruction::AND, {{Instruction::OR, {Y, X}}, X}}, [=]{ return X; }, true}, + {{Instruction::AND, {X, {Instruction::NOT, {X}}}}, [=]{ return u256(0); }, true}, + {{Instruction::AND, {{Instruction::NOT, {X}}, X}}, [=]{ return u256(0); }, true}, + {{Instruction::OR, {X, {Instruction::NOT, {X}}}}, [=]{ return ~u256(0); }, true}, + {{Instruction::OR, {{Instruction::NOT, {X}}, X}}, [=]{ return ~u256(0); }, true}, + }; + + // Replace MOD X, with AND X, - 1 + for (size_t i = 0; i < 256; ++i) + { + u256 value = u256(1) << i; + rules.push_back({ + {Instruction::MOD, {X, value}}, + [=]() -> Pattern { return {Instruction::AND, {X, value - 1}}; }, + false + }); + } + + // Double negation of opcodes with boolean result + for (auto const& op: std::vector{ + Instruction::EQ, + Instruction::LT, + Instruction::SLT, + Instruction::GT, + Instruction::SGT + }) + rules.push_back({ + {Instruction::ISZERO, {{Instruction::ISZERO, {{op, {X, Y}}}}}}, + [=]() -> Pattern { return {op, {X, Y}}; }, + false + }); + + rules.push_back({ + {Instruction::ISZERO, {{Instruction::ISZERO, {{Instruction::ISZERO, {X}}}}}}, + [=]() -> Pattern { return {Instruction::ISZERO, {X}}; }, + false + }); + + rules.push_back({ + {Instruction::ISZERO, {{Instruction::XOR, {X, Y}}}}, + [=]() -> Pattern { return { Instruction::EQ, {X, Y} }; }, + false + }); + + // Associative operations + for (auto const& opFun: std::vector>>{ + {Instruction::ADD, std::plus()}, + {Instruction::MUL, std::multiplies()}, + {Instruction::AND, std::bit_and()}, + {Instruction::OR, std::bit_or()}, + {Instruction::XOR, std::bit_xor()} + }) + { + auto op = opFun.first; + auto fun = opFun.second; + // Moving constants to the outside, order matters here - we first add rules + // for constants and then for non-constants. + // xa can be (X, A) or (A, X) + for (auto xa: {std::vector{X, A}, std::vector{A, X}}) + { + rules += std::vector>{{ + // (X+A)+B -> X+(A+B) + {op, {{op, xa}, B}}, + [=]() -> Pattern { return {op, {X, fun(A.d(), B.d())}}; }, + false + }, { + // (X+A)+Y -> (X+Y)+A + {op, {{op, xa}, Y}}, + [=]() -> Pattern { return {op, {{op, {X, Y}}, A}}; }, + false + }, { + // B+(X+A) -> X+(A+B) + {op, {B, {op, xa}}}, + [=]() -> Pattern { return {op, {X, fun(A.d(), B.d())}}; }, + false + }, { + // Y+(X+A) -> (Y+X)+A + {op, {Y, {op, xa}}}, + [=]() -> Pattern { return {op, {{op, {Y, X}}, A}}; }, + false + }}; + } + } + + // move constants across subtractions + rules += std::vector>{ + { + // X - A -> X + (-A) + {Instruction::SUB, {X, A}}, + [=]() -> Pattern { return {Instruction::ADD, {X, 0 - A.d()}}; }, + false + }, { + // (X + A) - Y -> (X - Y) + A + {Instruction::SUB, {{Instruction::ADD, {X, A}}, Y}}, + [=]() -> Pattern { return {Instruction::ADD, {{Instruction::SUB, {X, Y}}, A}}; }, + false + }, { + // (A + X) - Y -> (X - Y) + A + {Instruction::SUB, {{Instruction::ADD, {A, X}}, Y}}, + [=]() -> Pattern { return {Instruction::ADD, {{Instruction::SUB, {X, Y}}, A}}; }, + false + }, { + // X - (Y + A) -> (X - Y) + (-A) + {Instruction::SUB, {X, {Instruction::ADD, {Y, A}}}}, + [=]() -> Pattern { return {Instruction::ADD, {{Instruction::SUB, {X, Y}}, 0 - A.d()}}; }, + false + }, { + // X - (A + Y) -> (X - Y) + (-A) + {Instruction::SUB, {X, {Instruction::ADD, {A, Y}}}}, + [=]() -> Pattern { return {Instruction::ADD, {{Instruction::SUB, {X, Y}}, 0 - A.d()}}; }, + false + } + }; + return rules; +} + +} +} diff --git a/libevmasm/SemanticInformation.cpp b/libevmasm/SemanticInformation.cpp index 61a6ccda4461..03870f7cab51 100644 --- a/libevmasm/SemanticInformation.cpp +++ b/libevmasm/SemanticInformation.cpp @@ -35,6 +35,7 @@ bool SemanticInformation::breaksCSEAnalysisBlock(AssemblyItem const& _item) default: case UndefinedItem: case Tag: + case PushDeployTimeAddress: return true; case Push: case PushString: @@ -153,6 +154,31 @@ bool SemanticInformation::isDeterministic(AssemblyItem const& _item) } } +bool SemanticInformation::movable(Instruction _instruction) +{ + // These are not really functional. + if (isDupInstruction(_instruction) || isSwapInstruction(_instruction)) + return false; + InstructionInfo info = instructionInfo(_instruction); + if (info.sideEffects) + return false; + switch (_instruction) + { + case Instruction::KECCAK256: + case Instruction::BALANCE: + case Instruction::EXTCODESIZE: + case Instruction::RETURNDATASIZE: + case Instruction::SLOAD: + case Instruction::PC: + case Instruction::MSIZE: + case Instruction::GAS: + return false; + default: + return true; + } + return true; +} + bool SemanticInformation::invalidatesMemory(Instruction _instruction) { switch (_instruction) diff --git a/libevmasm/SemanticInformation.h b/libevmasm/SemanticInformation.h index e5ea7c180459..83656252c637 100644 --- a/libevmasm/SemanticInformation.h +++ b/libevmasm/SemanticInformation.h @@ -49,6 +49,10 @@ struct SemanticInformation /// @returns false if the value put on the stack by _item depends on anything else than /// the information in the current block header, memory, storage or stack. static bool isDeterministic(AssemblyItem const& _item); + /// @returns true if the instruction can be moved or copied (together with its arguments) + /// without altering the semantics. This means it cannot depend on storage or memory, + /// cannot have any side-effects, but it can depend on a call-constant state of the blockchain. + static bool movable(solidity::Instruction _instruction); /// @returns true if the given instruction modifies memory. static bool invalidatesMemory(solidity::Instruction _instruction); /// @returns true if the given instruction modifies storage (even indirectly). diff --git a/libevmasm/SimplificationRule.h b/libevmasm/SimplificationRule.h new file mode 100644 index 000000000000..7b4dea6875e3 --- /dev/null +++ b/libevmasm/SimplificationRule.h @@ -0,0 +1,45 @@ +/* + This file is part of solidity. + + solidity is free software: you can redistribute it and/or modify + it under the terms of the GNU General Public License as published by + the Free Software Foundation, either version 3 of the License, or + (at your option) any later version. + + solidity is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + GNU General Public License for more details. + + You should have received a copy of the GNU General Public License + along with solidity. If not, see . +*/ +/** + * Expression simplification pattern. + */ + +#pragma once + +#include + +namespace dev +{ +namespace solidity +{ + +/** + * Rule that contains a pattern, an action that can be applied + * after the pattern has matched and a bool that indicates + * whether the action would remove something from the expression + * than is not a constant literal. + */ +template +struct SimplificationRule +{ + Pattern pattern; + std::function action; + bool removesNonConstants; +}; + +} +} diff --git a/libevmasm/SimplificationRules.cpp b/libevmasm/SimplificationRules.cpp index e6c51f95690c..53a5f9fc5974 100644 --- a/libevmasm/SimplificationRules.cpp +++ b/libevmasm/SimplificationRules.cpp @@ -23,7 +23,6 @@ #include #include -#include #include #include #include @@ -31,12 +30,14 @@ #include #include +#include + using namespace std; using namespace dev; using namespace dev::eth; -pair > const* Rules::findFirstMatch( +SimplificationRule const* Rules::findFirstMatch( Expression const& _expr, ExpressionClasses const& _classes ) @@ -46,32 +47,22 @@ pair > const* Rules::findFirstMatch( assertThrow(_expr.item, OptimizerException, ""); for (auto const& rule: m_rules[byte(_expr.item->instruction())]) { - if (rule.first.matches(_expr, _classes)) + if (rule.pattern.matches(_expr, _classes)) return &rule; resetMatchGroups(); } return nullptr; } -void Rules::addRules(std::vector > > const& _rules) +void Rules::addRules(std::vector> const& _rules) { for (auto const& r: _rules) addRule(r); } -void Rules::addRule(std::pair > const& _rule) +void Rules::addRule(SimplificationRule const& _rule) { - m_rules[byte(_rule.first.instruction())].push_back(_rule); -} - -template S divWorkaround(S const& _a, S const& _b) -{ - return (S)(bigint(_a) / bigint(_b)); -} - -template S modWorkaround(S const& _a, S const& _b) -{ - return (S)(bigint(_a) % bigint(_b)); + m_rules[byte(_rule.pattern.instruction())].push_back(_rule); } Rules::Rules() @@ -84,165 +75,13 @@ Rules::Rules() // Anything. Pattern X; Pattern Y; - Pattern Z; A.setMatchGroup(1, m_matchGroups); B.setMatchGroup(2, m_matchGroups); C.setMatchGroup(3, m_matchGroups); X.setMatchGroup(4, m_matchGroups); Y.setMatchGroup(5, m_matchGroups); - Z.setMatchGroup(6, m_matchGroups); - - addRules(vector>>{ - // arithmetics on constants - {{Instruction::ADD, {A, B}}, [=]{ return A.d() + B.d(); }}, - {{Instruction::MUL, {A, B}}, [=]{ return A.d() * B.d(); }}, - {{Instruction::SUB, {A, B}}, [=]{ return A.d() - B.d(); }}, - {{Instruction::DIV, {A, B}}, [=]{ return B.d() == 0 ? 0 : divWorkaround(A.d(), B.d()); }}, - {{Instruction::SDIV, {A, B}}, [=]{ return B.d() == 0 ? 0 : s2u(divWorkaround(u2s(A.d()), u2s(B.d()))); }}, - {{Instruction::MOD, {A, B}}, [=]{ return B.d() == 0 ? 0 : modWorkaround(A.d(), B.d()); }}, - {{Instruction::SMOD, {A, B}}, [=]{ return B.d() == 0 ? 0 : s2u(modWorkaround(u2s(A.d()), u2s(B.d()))); }}, - {{Instruction::EXP, {A, B}}, [=]{ return u256(boost::multiprecision::powm(bigint(A.d()), bigint(B.d()), bigint(1) << 256)); }}, - {{Instruction::NOT, {A}}, [=]{ return ~A.d(); }}, - {{Instruction::LT, {A, B}}, [=]() -> u256 { return A.d() < B.d() ? 1 : 0; }}, - {{Instruction::GT, {A, B}}, [=]() -> u256 { return A.d() > B.d() ? 1 : 0; }}, - {{Instruction::SLT, {A, B}}, [=]() -> u256 { return u2s(A.d()) < u2s(B.d()) ? 1 : 0; }}, - {{Instruction::SGT, {A, B}}, [=]() -> u256 { return u2s(A.d()) > u2s(B.d()) ? 1 : 0; }}, - {{Instruction::EQ, {A, B}}, [=]() -> u256 { return A.d() == B.d() ? 1 : 0; }}, - {{Instruction::ISZERO, {A}}, [=]() -> u256 { return A.d() == 0 ? 1 : 0; }}, - {{Instruction::AND, {A, B}}, [=]{ return A.d() & B.d(); }}, - {{Instruction::OR, {A, B}}, [=]{ return A.d() | B.d(); }}, - {{Instruction::XOR, {A, B}}, [=]{ return A.d() ^ B.d(); }}, - {{Instruction::BYTE, {A, B}}, [=]{ return A.d() >= 32 ? 0 : (B.d() >> unsigned(8 * (31 - A.d()))) & 0xff; }}, - {{Instruction::ADDMOD, {A, B, C}}, [=]{ return C.d() == 0 ? 0 : u256((bigint(A.d()) + bigint(B.d())) % C.d()); }}, - {{Instruction::MULMOD, {A, B, C}}, [=]{ return C.d() == 0 ? 0 : u256((bigint(A.d()) * bigint(B.d())) % C.d()); }}, - {{Instruction::MULMOD, {A, B, C}}, [=]{ return A.d() * B.d(); }}, - {{Instruction::SIGNEXTEND, {A, B}}, [=]() -> u256 { - if (A.d() >= 31) - return B.d(); - unsigned testBit = unsigned(A.d()) * 8 + 7; - u256 mask = (u256(1) << testBit) - 1; - return u256(boost::multiprecision::bit_test(B.d(), testBit) ? B.d() | ~mask : B.d() & mask); - }}, - - // invariants involving known constants (commutative instructions will be checked with swapped operants too) - {{Instruction::ADD, {X, 0}}, [=]{ return X; }}, - {{Instruction::SUB, {X, 0}}, [=]{ return X; }}, - {{Instruction::MUL, {X, 0}}, [=]{ return u256(0); }}, - {{Instruction::MUL, {X, 1}}, [=]{ return X; }}, - {{Instruction::DIV, {X, 0}}, [=]{ return u256(0); }}, - {{Instruction::DIV, {0, X}}, [=]{ return u256(0); }}, - {{Instruction::DIV, {X, 1}}, [=]{ return X; }}, - {{Instruction::SDIV, {X, 0}}, [=]{ return u256(0); }}, - {{Instruction::SDIV, {0, X}}, [=]{ return u256(0); }}, - {{Instruction::SDIV, {X, 1}}, [=]{ return X; }}, - {{Instruction::AND, {X, ~u256(0)}}, [=]{ return X; }}, - {{Instruction::AND, {X, 0}}, [=]{ return u256(0); }}, - {{Instruction::OR, {X, 0}}, [=]{ return X; }}, - {{Instruction::OR, {X, ~u256(0)}}, [=]{ return ~u256(0); }}, - {{Instruction::XOR, {X, 0}}, [=]{ return X; }}, - {{Instruction::MOD, {X, 0}}, [=]{ return u256(0); }}, - {{Instruction::MOD, {0, X}}, [=]{ return u256(0); }}, - {{Instruction::EQ, {X, 0}}, [=]() -> Pattern { return {Instruction::ISZERO, {X}}; } }, - - // operations involving an expression and itself - {{Instruction::AND, {X, X}}, [=]{ return X; }}, - {{Instruction::OR, {X, X}}, [=]{ return X; }}, - {{Instruction::XOR, {X, X}}, [=]{ return u256(0); }}, - {{Instruction::SUB, {X, X}}, [=]{ return u256(0); }}, - {{Instruction::EQ, {X, X}}, [=]{ return u256(1); }}, - {{Instruction::LT, {X, X}}, [=]{ return u256(0); }}, - {{Instruction::SLT, {X, X}}, [=]{ return u256(0); }}, - {{Instruction::GT, {X, X}}, [=]{ return u256(0); }}, - {{Instruction::SGT, {X, X}}, [=]{ return u256(0); }}, - {{Instruction::MOD, {X, X}}, [=]{ return u256(0); }}, - - // logical instruction combinations - {{Instruction::NOT, {{Instruction::NOT, {X}}}}, [=]{ return X; }}, - {{Instruction::XOR, {{{X}, {Instruction::XOR, {X, Y}}}}}, [=]{ return Y; }}, - {{Instruction::OR, {{{X}, {Instruction::AND, {X, Y}}}}}, [=]{ return X; }}, - {{Instruction::AND, {{{X}, {Instruction::OR, {X, Y}}}}}, [=]{ return X; }}, - {{Instruction::AND, {{{X}, {Instruction::NOT, {X}}}}}, [=]{ return u256(0); }}, - {{Instruction::OR, {{{X}, {Instruction::NOT, {X}}}}}, [=]{ return ~u256(0); }}, - }); - - // Double negation of opcodes with binary result - for (auto const& op: vector{ - Instruction::EQ, - Instruction::LT, - Instruction::SLT, - Instruction::GT, - Instruction::SGT - }) - addRule({ - {Instruction::ISZERO, {{Instruction::ISZERO, {{op, {X, Y}}}}}}, - [=]() -> Pattern { return {op, {X, Y}}; } - }); - - addRule({ - {Instruction::ISZERO, {{Instruction::ISZERO, {{Instruction::ISZERO, {X}}}}}}, - [=]() -> Pattern { return {Instruction::ISZERO, {X}}; } - }); - - addRule({ - {Instruction::ISZERO, {{Instruction::XOR, {X, Y}}}}, - [=]() -> Pattern { return { Instruction::EQ, {X, Y} }; } - }); - - // Associative operations - for (auto const& opFun: vector>>{ - {Instruction::ADD, plus()}, - {Instruction::MUL, multiplies()}, - {Instruction::AND, bit_and()}, - {Instruction::OR, bit_or()}, - {Instruction::XOR, bit_xor()} - }) - { - auto op = opFun.first; - auto fun = opFun.second; - // Moving constants to the outside, order matters here! - // we need actions that return expressions (or patterns?) here, and we need also reversed rules - // (X+A)+B -> X+(A+B) - addRules(vector>>{{ - {op, {{op, {X, A}}, B}}, - [=]() -> Pattern { return {op, {X, fun(A.d(), B.d())}}; } - }, { - // X+(Y+A) -> (X+Y)+A - {op, {{op, {X, A}}, Y}}, - [=]() -> Pattern { return {op, {{op, {X, Y}}, A}}; } - }, { - // For now, we still need explicit commutativity for the inner pattern - {op, {{op, {A, X}}, B}}, - [=]() -> Pattern { return {op, {X, fun(A.d(), B.d())}}; } - }, { - {op, {{op, {A, X}}, Y}}, - [=]() -> Pattern { return {op, {{op, {X, Y}}, A}}; } - }}); - } - // move constants across subtractions - addRules(vector>>{ - { - // X - A -> X + (-A) - {Instruction::SUB, {X, A}}, - [=]() -> Pattern { return {Instruction::ADD, {X, 0 - A.d()}}; } - }, { - // (X + A) - Y -> (X - Y) + A - {Instruction::SUB, {{Instruction::ADD, {X, A}}, Y}}, - [=]() -> Pattern { return {Instruction::ADD, {{Instruction::SUB, {X, Y}}, A}}; } - }, { - // (A + X) - Y -> (X - Y) + A - {Instruction::SUB, {{Instruction::ADD, {A, X}}, Y}}, - [=]() -> Pattern { return {Instruction::ADD, {{Instruction::SUB, {X, Y}}, A}}; } - }, { - // X - (Y + A) -> (X - Y) + (-A) - {Instruction::SUB, {X, {Instruction::ADD, {Y, A}}}}, - [=]() -> Pattern { return {Instruction::ADD, {{Instruction::SUB, {X, Y}}, 0 - A.d()}}; } - }, { - // X - (A + Y) -> (X - Y) + (-A) - {Instruction::SUB, {X, {Instruction::ADD, {A, Y}}}}, - [=]() -> Pattern { return {Instruction::ADD, {{Instruction::SUB, {X, Y}}, 0 - A.d()}}; } - } - }); + addRules(simplificationRuleList(A, B, C, X, Y)); } Pattern::Pattern(Instruction _instruction, std::vector const& _arguments): diff --git a/libevmasm/SimplificationRules.h b/libevmasm/SimplificationRules.h index a4da5849c178..53f7e59545a9 100644 --- a/libevmasm/SimplificationRules.h +++ b/libevmasm/SimplificationRules.h @@ -24,6 +24,7 @@ #pragma once #include +#include #include #include @@ -47,19 +48,21 @@ class Rules: public boost::noncopyable /// @returns a pointer to the first matching pattern and sets the match /// groups accordingly. - std::pair> const* findFirstMatch( + SimplificationRule const* findFirstMatch( Expression const& _expr, ExpressionClasses const& _classes ); private: - void addRules(std::vector>> const& _rules); - void addRule(std::pair> const& _rule); + void addRules(std::vector> const& _rules); + void addRule(SimplificationRule const& _rule); void resetMatchGroups() { m_matchGroups.clear(); } std::map m_matchGroups; - std::vector>> m_rules[256]; + /// Pattern to match, replacement to be applied and flag indicating whether + /// the replacement might remove some elements (except constants). + std::vector> m_rules[256]; }; /** diff --git a/libjulia/ASTDataForward.h b/libjulia/ASTDataForward.h new file mode 100644 index 000000000000..143b9c46afaf --- /dev/null +++ b/libjulia/ASTDataForward.h @@ -0,0 +1,54 @@ +/* + This file is part of solidity. + + solidity is free software: you can redistribute it and/or modify + it under the terms of the GNU General Public License as published by + the Free Software Foundation, either version 3 of the License, or + (at your option) any later version. + + solidity is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + GNU General Public License for more details. + + You should have received a copy of the GNU General Public License + along with solidity. If not, see . +*/ +/** + * @date 2017 + * Pull in some identifiers from the solidity::assembly namespace. + */ + +#pragma once + +#include + +namespace dev +{ +namespace julia +{ + +using Instruction = solidity::assembly::Instruction; +using Literal = solidity::assembly::Literal; +using Label = solidity::assembly::Label; +using StackAssignment = solidity::assembly::StackAssignment; +using Identifier = solidity::assembly::Identifier; +using Assignment = solidity::assembly::Assignment; +using VariableDeclaration = solidity::assembly::VariableDeclaration; +using FunctionalInstruction = solidity::assembly::FunctionalInstruction; +using FunctionDefinition = solidity::assembly::FunctionDefinition; +using FunctionCall = solidity::assembly::FunctionCall; +using If = solidity::assembly::If; +using Case = solidity::assembly::Case; +using Switch = solidity::assembly::Switch; +using ForLoop = solidity::assembly::ForLoop; +using ExpressionStatement = solidity::assembly::ExpressionStatement; +using Block = solidity::assembly::Block; + +using TypedName = solidity::assembly::TypedName; + +using Expression = boost::variant; +using Statement = boost::variant; + +} +} diff --git a/libjulia/backends/evm/EVMAssembly.cpp b/libjulia/backends/evm/EVMAssembly.cpp index 1d499b200be7..07ad05c9c093 100644 --- a/libjulia/backends/evm/EVMAssembly.cpp +++ b/libjulia/backends/evm/EVMAssembly.cpp @@ -26,7 +26,7 @@ using namespace std; using namespace dev; -using namespace julia; +using namespace dev::julia; namespace { diff --git a/libjulia/backends/evm/EVMCodeTransform.cpp b/libjulia/backends/evm/EVMCodeTransform.cpp index 13d9d01171de..0c7365fbae2a 100644 --- a/libjulia/backends/evm/EVMCodeTransform.cpp +++ b/libjulia/backends/evm/EVMCodeTransform.cpp @@ -31,7 +31,8 @@ using namespace std; using namespace dev; using namespace dev::julia; using namespace dev::solidity; -using namespace dev::solidity::assembly; + +using Scope = dev::solidity::assembly::Scope; void CodeTransform::operator()(VariableDeclaration const& _varDecl) { @@ -76,6 +77,13 @@ void CodeTransform::operator()(StackAssignment const& _assignment) checkStackHeight(&_assignment); } +void CodeTransform::operator()(ExpressionStatement const& _statement) +{ + m_assembly.setSourceLocation(_statement.location); + boost::apply_visitor(*this, _statement.expression); + checkStackHeight(&_statement); +} + void CodeTransform::operator()(Label const& _label) { m_assembly.setSourceLocation(_label.location); @@ -124,11 +132,11 @@ void CodeTransform::operator()(FunctionCall const& _call) void CodeTransform::operator()(FunctionalInstruction const& _instruction) { if (m_evm15 && ( - _instruction.instruction.instruction == solidity::Instruction::JUMP || - _instruction.instruction.instruction == solidity::Instruction::JUMPI + _instruction.instruction == solidity::Instruction::JUMP || + _instruction.instruction == solidity::Instruction::JUMPI )) { - bool const isJumpI = _instruction.instruction.instruction == solidity::Instruction::JUMPI; + bool const isJumpI = _instruction.instruction == solidity::Instruction::JUMPI; if (isJumpI) { solAssert(_instruction.arguments.size() == 2, ""); @@ -149,7 +157,8 @@ void CodeTransform::operator()(FunctionalInstruction const& _instruction) { for (auto const& arg: _instruction.arguments | boost::adaptors::reversed) visitExpression(arg); - (*this)(_instruction.instruction); + m_assembly.setSourceLocation(_instruction.location); + m_assembly.appendInstruction(_instruction.instruction); } checkStackHeight(&_instruction); } @@ -289,7 +298,7 @@ void CodeTransform::operator()(FunctionDefinition const& _function) solAssert(m_info.scopes.at(&_function.body), ""); Scope* varScope = m_info.scopes.at(m_info.virtualBlocks.at(&_function).get()).get(); solAssert(varScope, ""); - for (auto const& v: _function.arguments | boost::adaptors::reversed) + for (auto const& v: _function.parameters | boost::adaptors::reversed) { auto& var = boost::get(varScope->identifiers.at(v.name)); m_context->variableStackHeights[&var] = height++; @@ -302,7 +311,7 @@ void CodeTransform::operator()(FunctionDefinition const& _function) if (m_evm15) { m_assembly.appendJumpTo(afterFunction, -stackHeightBefore); - m_assembly.appendBeginsub(functionEntryID(_function.name, function), _function.arguments.size()); + m_assembly.appendBeginsub(functionEntryID(_function.name, function), _function.parameters.size()); } else { @@ -311,7 +320,7 @@ void CodeTransform::operator()(FunctionDefinition const& _function) } m_stackAdjustment += localStackAdjustment; - for (auto const& v: _function.returns) + for (auto const& v: _function.returnVariables) { auto& var = boost::get(varScope->identifiers.at(v.name)); m_context->variableStackHeights[&var] = height++; @@ -341,9 +350,9 @@ void CodeTransform::operator()(FunctionDefinition const& _function) // modified parallel to the actual stack. vector stackLayout; if (!m_evm15) - stackLayout.push_back(_function.returns.size()); // Move return label to the top - stackLayout += vector(_function.arguments.size(), -1); // discard all arguments - for (size_t i = 0; i < _function.returns.size(); ++i) + stackLayout.push_back(_function.returnVariables.size()); // Move return label to the top + stackLayout += vector(_function.parameters.size(), -1); // discard all arguments + for (size_t i = 0; i < _function.returnVariables.size(); ++i) stackLayout.push_back(i); // Move return values down, but keep order. solAssert(stackLayout.size() <= 17, "Stack too deep"); @@ -363,9 +372,9 @@ void CodeTransform::operator()(FunctionDefinition const& _function) } if (m_evm15) - m_assembly.appendReturnsub(_function.returns.size(), stackHeightBefore); + m_assembly.appendReturnsub(_function.returnVariables.size(), stackHeightBefore); else - m_assembly.appendJump(stackHeightBefore - _function.returns.size()); + m_assembly.appendJump(stackHeightBefore - _function.returnVariables.size()); m_stackAdjustment -= localStackAdjustment; m_assembly.appendLabel(afterFunction); checkStackHeight(&_function); @@ -458,7 +467,7 @@ AbstractAssembly::LabelID CodeTransform::functionEntryID(string const& _name, Sc return m_context->functionEntryIDs[&_function]; } -void CodeTransform::visitExpression(Statement const& _expression) +void CodeTransform::visitExpression(Expression const& _expression) { int height = m_assembly.stackHeight(); boost::apply_visitor(*this, _expression); diff --git a/libjulia/backends/evm/EVMCodeTransform.h b/libjulia/backends/evm/EVMCodeTransform.h index 387720a20308..0f2aaf955765 100644 --- a/libjulia/backends/evm/EVMCodeTransform.h +++ b/libjulia/backends/evm/EVMCodeTransform.h @@ -20,8 +20,9 @@ #include +#include + #include -#include #include #include @@ -95,38 +96,39 @@ class CodeTransform: public boost::static_visitor<> {} public: - void operator()(solidity::assembly::Instruction const& _instruction); - void operator()(solidity::assembly::Literal const& _literal); - void operator()(solidity::assembly::Identifier const& _identifier); - void operator()(solidity::assembly::FunctionalInstruction const& _instr); - void operator()(solidity::assembly::FunctionCall const&); - void operator()(solidity::assembly::Label const& _label); - void operator()(solidity::assembly::StackAssignment const& _assignment); - void operator()(solidity::assembly::Assignment const& _assignment); - void operator()(solidity::assembly::VariableDeclaration const& _varDecl); - void operator()(solidity::assembly::If const& _if); - void operator()(solidity::assembly::Switch const& _switch); - void operator()(solidity::assembly::FunctionDefinition const&); - void operator()(solidity::assembly::ForLoop const&); - void operator()(solidity::assembly::Block const& _block); + void operator()(Instruction const& _instruction); + void operator()(Literal const& _literal); + void operator()(Identifier const& _identifier); + void operator()(FunctionalInstruction const& _instr); + void operator()(FunctionCall const&); + void operator()(ExpressionStatement const& _statement); + void operator()(Label const& _label); + void operator()(StackAssignment const& _assignment); + void operator()(Assignment const& _assignment); + void operator()(VariableDeclaration const& _varDecl); + void operator()(If const& _if); + void operator()(Switch const& _switch); + void operator()(FunctionDefinition const&); + void operator()(ForLoop const&); + void operator()(Block const& _block); private: - AbstractAssembly::LabelID labelFromIdentifier(solidity::assembly::Identifier const& _identifier); + AbstractAssembly::LabelID labelFromIdentifier(Identifier const& _identifier); /// @returns the label ID corresponding to the given label, allocating a new one if /// necessary. AbstractAssembly::LabelID labelID(solidity::assembly::Scope::Label const& _label); AbstractAssembly::LabelID functionEntryID(std::string const& _name, solidity::assembly::Scope::Function const& _function); /// Generates code for an expression that is supposed to return a single value. - void visitExpression(solidity::assembly::Statement const& _expression); + void visitExpression(Expression const& _expression); - void visitStatements(std::vector const& _statements); + void visitStatements(std::vector const& _statements); /// Pops all variables declared in the block and checks that the stack height is equal /// to @a _blackStartStackHeight. - void finalizeBlock(solidity::assembly::Block const& _block, int _blockStartStackHeight); + void finalizeBlock(Block const& _block, int _blockStartStackHeight); - void generateMultiAssignment(std::vector const& _variableNames); - void generateAssignment(solidity::assembly::Identifier const& _variableName); + void generateMultiAssignment(std::vector const& _variableNames); + void generateAssignment(Identifier const& _variableName); /// Determines the stack height difference to the given variables. Throws /// if it is not yet in scope or the height difference is too large. Returns diff --git a/libjulia/optimiser/ASTCopier.cpp b/libjulia/optimiser/ASTCopier.cpp new file mode 100644 index 000000000000..5c47be64b99f --- /dev/null +++ b/libjulia/optimiser/ASTCopier.cpp @@ -0,0 +1,185 @@ +/* + This file is part of solidity. + + solidity is free software: you can redistribute it and/or modify + it under the terms of the GNU General Public License as published by + the Free Software Foundation, either version 3 of the License, or + (at your option) any later version. + + solidity is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + GNU General Public License for more details. + + You should have received a copy of the GNU General Public License + along with solidity. If not, see . +*/ +/** + * Creates an independent copy of an AST, renaming identifiers to be unique. + */ + +#include + +#include + +#include + +#include + +using namespace std; +using namespace dev; +using namespace dev::julia; + + +Statement ASTCopier::operator()(Instruction const&) +{ + solAssert(false, "Invalid operation."); + return {}; +} + +Statement ASTCopier::operator()(ExpressionStatement const& _statement) +{ + return ExpressionStatement{ _statement.location, translate(_statement.expression) }; +} + +Statement ASTCopier::operator()(VariableDeclaration const& _varDecl) +{ + return VariableDeclaration{ + _varDecl.location, + translateVector(_varDecl.variables), + translate(_varDecl.value) + }; +} + +Statement ASTCopier::operator()(Assignment const& _assignment) +{ + return Assignment{ + _assignment.location, + translateVector(_assignment.variableNames), + translate(_assignment.value) + }; +} + +Statement ASTCopier::operator()(StackAssignment const&) +{ + solAssert(false, "Invalid operation."); + return {}; +} + +Statement ASTCopier::operator()(Label const&) +{ + solAssert(false, "Invalid operation."); + return {}; +} + +Expression ASTCopier::operator()(FunctionCall const& _call) +{ + return FunctionCall{ + _call.location, + translate(_call.functionName), + translateVector(_call.arguments) + }; +} + +Expression ASTCopier::operator()(FunctionalInstruction const& _instruction) +{ + return FunctionalInstruction{ + _instruction.location, + _instruction.instruction, + translateVector(_instruction.arguments) + }; +} + +Expression ASTCopier::operator()(Identifier const& _identifier) +{ + return Identifier{_identifier.location, translateIdentifier(_identifier.name)}; +} + +Expression ASTCopier::operator()(Literal const& _literal) +{ + return translate(_literal); +} + +Statement ASTCopier::operator()(If const& _if) +{ + return If{_if.location, translate(_if.condition), translate(_if.body)}; +} + +Statement ASTCopier::operator()(Switch const& _switch) +{ + return Switch{_switch.location, translate(_switch.expression), translateVector(_switch.cases)}; +} + +Statement ASTCopier::operator()(FunctionDefinition const& _function) +{ + string translatedName = translateIdentifier(_function.name); + + enterFunction(_function); + ScopeGuard g([&]() { this->leaveFunction(_function); }); + + return FunctionDefinition{ + _function.location, + move(translatedName), + translateVector(_function.parameters), + translateVector(_function.returnVariables), + translate(_function.body) + }; +} + +Statement ASTCopier::operator()(ForLoop const& _forLoop) +{ + enterScope(_forLoop.pre); + ScopeGuard g([&]() { this->leaveScope(_forLoop.pre); }); + + return ForLoop{ + _forLoop.location, + translate(_forLoop.pre), + translate(_forLoop.condition), + translate(_forLoop.post), + translate(_forLoop.body) + }; +} + +Statement ASTCopier::operator ()(Block const& _block) +{ + return translate(_block); +} + +Expression ASTCopier::translate(Expression const& _expression) +{ + return _expression.apply_visitor(static_cast(*this)); +} + +Statement ASTCopier::translate(Statement const& _statement) +{ + return _statement.apply_visitor(static_cast(*this)); +} + +Block ASTCopier::translate(Block const& _block) +{ + enterScope(_block); + ScopeGuard g([&]() { this->leaveScope(_block); }); + + return Block{_block.location, translateVector(_block.statements)}; +} + +Case ASTCopier::translate(Case const& _case) +{ + return Case{_case.location, translate(_case.value), translate(_case.body)}; +} + +Identifier ASTCopier::translate(Identifier const& _identifier) +{ + return Identifier{_identifier.location, translateIdentifier(_identifier.name)}; +} + +Literal ASTCopier::translate(Literal const& _literal) +{ + return _literal; +} + +TypedName ASTCopier::translate(TypedName const& _typedName) +{ + return TypedName{_typedName.location, translateIdentifier(_typedName.name), _typedName.type}; +} + diff --git a/libjulia/optimiser/ASTCopier.h b/libjulia/optimiser/ASTCopier.h new file mode 100644 index 000000000000..36a1ced51ca0 --- /dev/null +++ b/libjulia/optimiser/ASTCopier.h @@ -0,0 +1,121 @@ +/* + This file is part of solidity. + + solidity is free software: you can redistribute it and/or modify + it under the terms of the GNU General Public License as published by + the Free Software Foundation, either version 3 of the License, or + (at your option) any later version. + + solidity is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + GNU General Public License for more details. + + You should have received a copy of the GNU General Public License + along with solidity. If not, see . +*/ +/** + * Creates an independent copy of an AST, renaming identifiers to be unique. + */ + +#pragma once + +#include + +#include +#include + +#include +#include +#include + +namespace dev +{ +namespace julia +{ + +class ExpressionCopier: public boost::static_visitor +{ +public: + virtual Expression operator()(Literal const& _literal) = 0; + virtual Expression operator()(Identifier const& _identifier) = 0; + virtual Expression operator()(FunctionalInstruction const& _instr) = 0; + virtual Expression operator()(FunctionCall const&) = 0; +}; + +class StatementCopier: public boost::static_visitor +{ +public: + virtual Statement operator()(ExpressionStatement const& _statement) = 0; + virtual Statement operator()(Instruction const& _instruction) = 0; + virtual Statement operator()(Label const& _label) = 0; + virtual Statement operator()(StackAssignment const& _assignment) = 0; + virtual Statement operator()(Assignment const& _assignment) = 0; + virtual Statement operator()(VariableDeclaration const& _varDecl) = 0; + virtual Statement operator()(If const& _if) = 0; + virtual Statement operator()(Switch const& _switch) = 0; + virtual Statement operator()(FunctionDefinition const&) = 0; + virtual Statement operator()(ForLoop const&) = 0; + virtual Statement operator()(Block const& _block) = 0; +}; + +/** + * Creates a copy of a iulia AST potentially replacing identifier names. + * Base class to be extended. + */ +class ASTCopier: public ExpressionCopier, public StatementCopier +{ +public: + virtual Expression operator()(Literal const& _literal) override; + virtual Statement operator()(Instruction const& _instruction) override; + virtual Expression operator()(Identifier const& _identifier) override; + virtual Expression operator()(FunctionalInstruction const& _instr) override; + virtual Expression operator()(FunctionCall const&) override; + virtual Statement operator()(ExpressionStatement const& _statement) override; + virtual Statement operator()(Label const& _label) override; + virtual Statement operator()(StackAssignment const& _assignment) override; + virtual Statement operator()(Assignment const& _assignment) override; + virtual Statement operator()(VariableDeclaration const& _varDecl) override; + virtual Statement operator()(If const& _if) override; + virtual Statement operator()(Switch const& _switch) override; + virtual Statement operator()(FunctionDefinition const&) override; + virtual Statement operator()(ForLoop const&) override; + virtual Statement operator()(Block const& _block) override; + + virtual Expression translate(Expression const& _expression); + virtual Statement translate(Statement const& _statement); + +protected: + template + std::vector translateVector(std::vector const& _values); + + template + std::shared_ptr translate(std::shared_ptr const& _v) + { + return _v ? std::make_shared(translate(*_v)) : nullptr; + } + Block translate(Block const& _block); + Case translate(Case const& _case); + Identifier translate(Identifier const& _identifier); + Literal translate(Literal const& _literal); + TypedName translate(TypedName const& _typedName); + + virtual void enterScope(Block const&) { } + virtual void leaveScope(Block const&) { } + virtual void enterFunction(FunctionDefinition const&) { } + virtual void leaveFunction(FunctionDefinition const&) { } + virtual std::string translateIdentifier(std::string const& _name) { return _name; } +}; + +template +std::vector ASTCopier::translateVector(std::vector const& _values) +{ + std::vector translated; + for (auto const& v: _values) + translated.emplace_back(translate(v)); + return translated; +} + + +} +} diff --git a/libjulia/optimiser/ASTWalker.cpp b/libjulia/optimiser/ASTWalker.cpp new file mode 100644 index 000000000000..0344498492d4 --- /dev/null +++ b/libjulia/optimiser/ASTWalker.cpp @@ -0,0 +1,159 @@ +/* + This file is part of solidity. + + solidity is free software: you can redistribute it and/or modify + it under the terms of the GNU General Public License as published by + the Free Software Foundation, either version 3 of the License, or + (at your option) any later version. + + solidity is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + GNU General Public License for more details. + + You should have received a copy of the GNU General Public License + along with solidity. If not, see . +*/ +/** + * Generic AST walker. + */ + +#include + +#include + +#include + +#include + +using namespace std; +using namespace dev; +using namespace dev::julia; +using namespace dev::solidity; + + +void ASTWalker::operator()(FunctionalInstruction const& _instr) +{ + walkVector(_instr.arguments | boost::adaptors::reversed); +} + +void ASTWalker::operator()(FunctionCall const& _funCall) +{ + walkVector(_funCall.arguments | boost::adaptors::reversed); +} + +void ASTWalker::operator()(ExpressionStatement const& _statement) +{ + visit(_statement.expression); +} + +void ASTWalker::operator()(Assignment const& _assignment) +{ + for (auto const& name: _assignment.variableNames) + (*this)(name); + visit(*_assignment.value); +} + +void ASTWalker::operator()(VariableDeclaration const& _varDecl) +{ + if (_varDecl.value) + visit(*_varDecl.value); +} + +void ASTWalker::operator()(If const& _if) +{ + visit(*_if.condition); + (*this)(_if.body); +} + +void ASTWalker::operator()(Switch const& _switch) +{ + visit(*_switch.expression); + for (auto const& _case: _switch.cases) + { + if (_case.value) + (*this)(*_case.value); + (*this)(_case.body); + } +} + +void ASTWalker::operator()(FunctionDefinition const& _fun) +{ + (*this)(_fun.body); +} + +void ASTWalker::operator()(ForLoop const& _for) +{ + (*this)(_for.pre); + visit(*_for.condition); + (*this)(_for.body); + (*this)(_for.post); +} + +void ASTWalker::operator()(Block const& _block) +{ + walkVector(_block.statements); +} + +void ASTModifier::operator()(FunctionalInstruction& _instr) +{ + walkVector(_instr.arguments | boost::adaptors::reversed); +} + +void ASTModifier::operator()(FunctionCall& _funCall) +{ + walkVector(_funCall.arguments | boost::adaptors::reversed); +} + +void ASTModifier::operator()(ExpressionStatement& _statement) +{ + visit(_statement.expression); +} + +void ASTModifier::operator()(Assignment& _assignment) +{ + for (auto& name: _assignment.variableNames) + (*this)(name); + visit(*_assignment.value); +} + +void ASTModifier::operator()(VariableDeclaration& _varDecl) +{ + if (_varDecl.value) + visit(*_varDecl.value); +} + +void ASTModifier::operator()(If& _if) +{ + visit(*_if.condition); + (*this)(_if.body); +} + +void ASTModifier::operator()(Switch& _switch) +{ + visit(*_switch.expression); + for (auto& _case: _switch.cases) + { + if (_case.value) + (*this)(*_case.value); + (*this)(_case.body); + } +} + +void ASTModifier::operator()(FunctionDefinition& _fun) +{ + (*this)(_fun.body); +} + +void ASTModifier::operator()(ForLoop& _for) +{ + (*this)(_for.pre); + visit(*_for.condition); + (*this)(_for.post); + (*this)(_for.body); +} + +void ASTModifier::operator()(Block& _block) +{ + walkVector(_block.statements); +} diff --git a/libjulia/optimiser/ASTWalker.h b/libjulia/optimiser/ASTWalker.h new file mode 100644 index 000000000000..9789138133c3 --- /dev/null +++ b/libjulia/optimiser/ASTWalker.h @@ -0,0 +1,120 @@ +/* + This file is part of solidity. + + solidity is free software: you can redistribute it and/or modify + it under the terms of the GNU General Public License as published by + the Free Software Foundation, either version 3 of the License, or + (at your option) any later version. + + solidity is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + GNU General Public License for more details. + + You should have received a copy of the GNU General Public License + along with solidity. If not, see . +*/ +/** + * Generic AST walker. + */ + +#pragma once + +#include + +#include + +#include +#include + +#include +#include +#include + +namespace dev +{ +namespace julia +{ + +/** + * Generic AST walker. + */ +class ASTWalker: public boost::static_visitor<> +{ +public: + virtual void operator()(Literal const&) {} + virtual void operator()(Instruction const&) { solAssert(false, ""); } + virtual void operator()(Identifier const&) {} + virtual void operator()(FunctionalInstruction const& _instr); + virtual void operator()(FunctionCall const& _funCall); + virtual void operator()(ExpressionStatement const& _statement); + virtual void operator()(Label const&) { solAssert(false, ""); } + virtual void operator()(StackAssignment const&) { solAssert(false, ""); } + virtual void operator()(Assignment const& _assignment); + virtual void operator()(VariableDeclaration const& _varDecl); + virtual void operator()(If const& _if); + virtual void operator()(Switch const& _switch); + virtual void operator()(FunctionDefinition const&); + virtual void operator()(ForLoop const&); + virtual void operator()(Block const& _block); + + virtual void visit(Statement const& _st) + { + boost::apply_visitor(*this, _st); + } + virtual void visit(Expression const& _e) + { + boost::apply_visitor(*this, _e); + } + +protected: + template + void walkVector(T const& _statements) + { + for (auto const& statement: _statements) + visit(statement); + } +}; + +/** + * Generic AST modifier (i.e. non-const version of ASTWalker). + */ +class ASTModifier: public boost::static_visitor<> +{ +public: + virtual void operator()(Literal&) {} + virtual void operator()(Instruction&) { solAssert(false, ""); } + virtual void operator()(Identifier&) {} + virtual void operator()(FunctionalInstruction& _instr); + virtual void operator()(FunctionCall& _funCall); + virtual void operator()(ExpressionStatement& _statement); + virtual void operator()(Label&) { solAssert(false, ""); } + virtual void operator()(StackAssignment&) { solAssert(false, ""); } + virtual void operator()(Assignment& _assignment); + virtual void operator()(VariableDeclaration& _varDecl); + virtual void operator()(If& _if); + virtual void operator()(Switch& _switch); + virtual void operator()(FunctionDefinition&); + virtual void operator()(ForLoop&); + virtual void operator()(Block& _block); + + virtual void visit(Statement& _st) + { + boost::apply_visitor(*this, _st); + } + virtual void visit(Expression& _e) + { + boost::apply_visitor(*this, _e); + } + +protected: + template + void walkVector(T&& _statements) + { + for (auto& st: _statements) + visit(st); + } +}; + +} +} diff --git a/libjulia/optimiser/DataFlowAnalyzer.cpp b/libjulia/optimiser/DataFlowAnalyzer.cpp new file mode 100644 index 000000000000..5665339303a9 --- /dev/null +++ b/libjulia/optimiser/DataFlowAnalyzer.cpp @@ -0,0 +1,193 @@ +/*( + This file is part of solidity. + + solidity is free software: you can redistribute it and/or modify + it under the terms of the GNU General Public License as published by + the Free Software Foundation, either version 3 of the License, or + (at your option) any later version. + + solidity is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + GNU General Public License for more details. + + You should have received a copy of the GNU General Public License + along with solidity. If not, see . +*/ +/** + * Base class to perform data flaw analysis during AST walks. + * Tracks assignments and is used as base class for both Rematerialiser and + * Common Subexpression Eliminator. + */ + +#include + +#include + +#include + +#include + +#include + +#include + +using namespace std; +using namespace dev; +using namespace dev::julia; + +void DataFlowAnalyzer::operator()(Assignment& _assignment) +{ + set names; + for (auto const& var: _assignment.variableNames) + names.insert(var.name); + solAssert(_assignment.value, ""); + visit(*_assignment.value); + handleAssignment(names, _assignment.value.get()); +} + +void DataFlowAnalyzer::operator()(VariableDeclaration& _varDecl) +{ + set names; + for (auto const& var: _varDecl.variables) + names.insert(var.name); + m_variableScopes.back().variables += names; + if (_varDecl.value) + visit(*_varDecl.value); + handleAssignment(names, _varDecl.value.get()); +} + +void DataFlowAnalyzer::operator()(If& _if) +{ + ASTModifier::operator()(_if); + + Assignments assignments; + assignments(_if.body); + clearValues(assignments.names()); +} + +void DataFlowAnalyzer::operator()(Switch& _switch) +{ + visit(*_switch.expression); + set assignedVariables; + for (auto& _case: _switch.cases) + { + (*this)(_case.body); + Assignments assignments; + assignments(_case.body); + assignedVariables += assignments.names(); + // This is a little too destructive, we could retain the old values. + clearValues(assignments.names()); + } + clearValues(assignedVariables); +} + +void DataFlowAnalyzer::operator()(FunctionDefinition& _fun) +{ + m_variableScopes.emplace_back(true); + for (auto const& parameter: _fun.parameters) + m_variableScopes.back().variables.insert(parameter.name); + for (auto const& var: _fun.returnVariables) + m_variableScopes.back().variables.insert(var.name); + ASTModifier::operator()(_fun); + m_variableScopes.pop_back(); +} + +void DataFlowAnalyzer::operator()(ForLoop& _for) +{ + // Special scope handling of the pre block. + m_variableScopes.emplace_back(false); + for (auto& statement: _for.pre.statements) + visit(statement); + + Assignments assignments; + assignments(_for.body); + assignments(_for.post); + clearValues(assignments.names()); + + visit(*_for.condition); + (*this)(_for.body); + (*this)(_for.post); + + clearValues(assignments.names()); + + m_variableScopes.pop_back(); +} + +void DataFlowAnalyzer::operator()(Block& _block) +{ + size_t numScopes = m_variableScopes.size(); + m_variableScopes.emplace_back(false); + ASTModifier::operator()(_block); + m_variableScopes.pop_back(); + solAssert(numScopes == m_variableScopes.size(), ""); +} + +void DataFlowAnalyzer::handleAssignment(set const& _variables, Expression* _value) +{ + clearValues(_variables); + + MovableChecker movableChecker; + if (_value) + movableChecker.visit(*_value); + if (_variables.size() == 1) + { + string const& name = *_variables.begin(); + // Expression has to be movable and cannot contain a reference + // to the variable that will be assigned to. + if (_value && movableChecker.movable() && !movableChecker.referencedVariables().count(name)) + m_value[name] = _value; + } + + auto const& referencedVariables = movableChecker.referencedVariables(); + for (auto const& name: _variables) + { + m_references[name] = referencedVariables; + for (auto const& ref: referencedVariables) + m_referencedBy[ref].insert(name); + } +} + +void DataFlowAnalyzer::clearValues(set const& _variables) +{ + // All variables that reference variables to be cleared also have to be + // cleared, but not recursively, since only the value of the original + // variables changes. Example: + // let a := 1 + // let b := a + // let c := b + // let a := 2 + // add(b, c) + // In the last line, we can replace c by b, but not b by a. + // + // This cannot be easily tested since the substitutions will be done + // one by one on the fly, and the last line will just be add(1, 1) + + set variables = _variables; + // Clear variables that reference variables to be cleared. + for (auto const& name: variables) + for (auto const& ref: m_referencedBy[name]) + variables.insert(ref); + + // Clear the value and update the reference relation. + for (auto const& name: variables) + m_value.erase(name); + for (auto const& name: variables) + { + for (auto const& ref: m_references[name]) + m_referencedBy[ref].erase(name); + m_references[name].clear(); + } +} + +bool DataFlowAnalyzer::inScope(string const& _variableName) const +{ + for (auto const& scope: m_variableScopes | boost::adaptors::reversed) + { + if (scope.variables.count(_variableName)) + return true; + if (scope.isFunction) + return false; + } + return false; +} diff --git a/libjulia/optimiser/DataFlowAnalyzer.h b/libjulia/optimiser/DataFlowAnalyzer.h new file mode 100644 index 000000000000..4cb3d4cd0d54 --- /dev/null +++ b/libjulia/optimiser/DataFlowAnalyzer.h @@ -0,0 +1,84 @@ +/* + This file is part of solidity. + + solidity is free software: you can redistribute it and/or modify + it under the terms of the GNU General Public License as published by + the Free Software Foundation, either version 3 of the License, or + (at your option) any later version. + + solidity is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + GNU General Public License for more details. + + You should have received a copy of the GNU General Public License + along with solidity. If not, see . +*/ +/** + * Base class to perform data flow analysis during AST walks. + * Tracks assignments and is used as base class for both Rematerialiser and + * Common Subexpression Eliminator. + */ + +#pragma once + +#include + +#include +#include +#include + +namespace dev +{ +namespace julia +{ + +/** + * Base class to perform data flow analysis during AST walks. + * Tracks assignments and is used as base class for both Rematerialiser and + * Common Subexpression Eliminator. + * + * Prerequisite: Disambiguator + */ +class DataFlowAnalyzer: public ASTModifier +{ +public: + using ASTModifier::operator(); + virtual void operator()(Assignment& _assignment) override; + virtual void operator()(VariableDeclaration& _varDecl) override; + virtual void operator()(If& _if) override; + virtual void operator()(Switch& _switch) override; + virtual void operator()(FunctionDefinition&) override; + virtual void operator()(ForLoop&) override; + virtual void operator()(Block& _block) override; + +protected: + /// Registers the assignment. + void handleAssignment(std::set const& _names, Expression* _value); + + /// Clears information about the valuse assigned to the given variables, + /// for example at points where control flow is merged. + void clearValues(std::set const& _names); + + /// Returns true iff the variable is in scope. + bool inScope(std::string const& _variableName) const; + + /// Current values of variables, always movable. + std::map m_value; + /// m_references[a].contains(b) <=> the current expression assigned to a references b + std::map> m_references; + /// m_referencedBy[b].contains(a) <=> the current expression assigned to a references b + std::map> m_referencedBy; + + struct Scope + { + explicit Scope(bool _isFunction): isFunction(_isFunction) {} + std::set variables; + bool isFunction; + }; + /// List of scopes. + std::vector m_variableScopes; +}; + +} +} diff --git a/libjulia/optimiser/Disambiguator.cpp b/libjulia/optimiser/Disambiguator.cpp new file mode 100644 index 000000000000..df2984e50b3a --- /dev/null +++ b/libjulia/optimiser/Disambiguator.cpp @@ -0,0 +1,85 @@ +/* + This file is part of solidity. + + solidity is free software: you can redistribute it and/or modify + it under the terms of the GNU General Public License as published by + the Free Software Foundation, either version 3 of the License, or + (at your option) any later version. + + solidity is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + GNU General Public License for more details. + + You should have received a copy of the GNU General Public License + along with solidity. If not, see . +*/ +/** + * Optimiser component that makes all identifiers unique. + */ + +#include + +#include +#include + +#include + +using namespace std; +using namespace dev; +using namespace dev::julia; +using namespace dev::solidity; + +using Scope = dev::solidity::assembly::Scope; + +string Disambiguator::translateIdentifier(string const& _originalName) +{ + solAssert(!m_scopes.empty() && m_scopes.back(), ""); + Scope::Identifier const* id = m_scopes.back()->lookup(_originalName); + solAssert(id, ""); + if (!m_translations.count(id)) + { + string translated = _originalName; + size_t suffix = 0; + while (m_usedNames.count(translated)) + { + suffix++; + translated = _originalName + "_" + std::to_string(suffix); + } + m_usedNames.insert(translated); + m_translations[id] = translated; + } + return m_translations.at(id); +} + +void Disambiguator::enterScope(Block const& _block) +{ + enterScopeInternal(*m_info.scopes.at(&_block)); +} + +void Disambiguator::leaveScope(Block const& _block) +{ + leaveScopeInternal(*m_info.scopes.at(&_block)); +} + +void Disambiguator::enterFunction(FunctionDefinition const& _function) +{ + enterScopeInternal(*m_info.scopes.at(m_info.virtualBlocks.at(&_function).get())); +} + +void Disambiguator::leaveFunction(FunctionDefinition const& _function) +{ + leaveScopeInternal(*m_info.scopes.at(m_info.virtualBlocks.at(&_function).get())); +} + +void Disambiguator::enterScopeInternal(Scope& _scope) +{ + m_scopes.push_back(&_scope); +} + +void Disambiguator::leaveScopeInternal(Scope& _scope) +{ + solAssert(!m_scopes.empty(), ""); + solAssert(m_scopes.back() == &_scope, ""); + m_scopes.pop_back(); +} diff --git a/libjulia/optimiser/Disambiguator.h b/libjulia/optimiser/Disambiguator.h new file mode 100644 index 000000000000..18ffd157b713 --- /dev/null +++ b/libjulia/optimiser/Disambiguator.h @@ -0,0 +1,67 @@ +/* + This file is part of solidity. + + solidity is free software: you can redistribute it and/or modify + it under the terms of the GNU General Public License as published by + the Free Software Foundation, either version 3 of the License, or + (at your option) any later version. + + solidity is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + GNU General Public License for more details. + + You should have received a copy of the GNU General Public License + along with solidity. If not, see . +*/ +/** + * Optimiser component that makes all identifiers unique. + */ + +#pragma once + +#include + +#include + +#include + +#include +#include + +#include + +namespace dev +{ +namespace julia +{ + +/** + * Creates a copy of a iulia AST replacing all identifiers by unique names. + */ +class Disambiguator: public ASTCopier +{ +public: + Disambiguator(solidity::assembly::AsmAnalysisInfo const& _analysisInfo): + m_info(_analysisInfo) + {} + +protected: + virtual void enterScope(Block const& _block) override; + virtual void leaveScope(Block const& _block) override; + virtual void enterFunction(FunctionDefinition const& _function) override; + virtual void leaveFunction(FunctionDefinition const& _function) override; + virtual std::string translateIdentifier(std::string const& _name) override; + + void enterScopeInternal(solidity::assembly::Scope& _scope); + void leaveScopeInternal(solidity::assembly::Scope& _scope); + + solidity::assembly::AsmAnalysisInfo const& m_info; + + std::vector m_scopes; + std::map m_translations; + std::set m_usedNames; +}; + +} +} diff --git a/libjulia/optimiser/ExpressionInliner.cpp b/libjulia/optimiser/ExpressionInliner.cpp new file mode 100644 index 000000000000..450769fdf20b --- /dev/null +++ b/libjulia/optimiser/ExpressionInliner.cpp @@ -0,0 +1,74 @@ +/* + This file is part of solidity. + + solidity is free software: you can redistribute it and/or modify + it under the terms of the GNU General Public License as published by + the Free Software Foundation, either version 3 of the License, or + (at your option) any later version. + + solidity is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + GNU General Public License for more details. + + You should have received a copy of the GNU General Public License + along with solidity. If not, see . +*/ +/** + * Optimiser component that performs function inlining inside expressions. + */ + +#include + +#include +#include +#include + +#include + +#include + +using namespace std; +using namespace dev; +using namespace dev::julia; +using namespace dev::solidity; + +void ExpressionInliner::run() +{ + InlinableExpressionFunctionFinder funFinder; + funFinder(m_block); + m_inlinableFunctions = funFinder.inlinableFunctions(); + + (*this)(m_block); +} + + +void ExpressionInliner::operator()(FunctionDefinition& _fun) +{ + ASTModifier::operator()(_fun); +} + +void ExpressionInliner::visit(Expression& _expression) +{ + ASTModifier::visit(_expression); + if (_expression.type() == typeid(FunctionCall)) + { + FunctionCall& funCall = boost::get(_expression); + + bool movable = boost::algorithm::all_of( + funCall.arguments, + [=](Expression const& _arg) { return MovableChecker(_arg).movable(); } + ); + if (m_inlinableFunctions.count(funCall.functionName.name) && movable) + { + FunctionDefinition const& fun = *m_inlinableFunctions.at(funCall.functionName.name); + map substitutions; + for (size_t i = 0; i < fun.parameters.size(); ++i) + substitutions[fun.parameters[i].name] = &funCall.arguments[i]; + _expression = Substitution(substitutions).translate(*boost::get(fun.body.statements.front()).value); + + // TODO Add metric! This metric should perform well on a pair of functions who + // call each other recursively. + } + } +} diff --git a/libjulia/optimiser/ExpressionInliner.h b/libjulia/optimiser/ExpressionInliner.h new file mode 100644 index 000000000000..10d7659cab84 --- /dev/null +++ b/libjulia/optimiser/ExpressionInliner.h @@ -0,0 +1,74 @@ +/* + This file is part of solidity. + + solidity is free software: you can redistribute it and/or modify + it under the terms of the GNU General Public License as published by + the Free Software Foundation, either version 3 of the License, or + (at your option) any later version. + + solidity is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + GNU General Public License for more details. + + You should have received a copy of the GNU General Public License + along with solidity. If not, see . +*/ +/** + * Optimiser component that performs function inlining. + */ +#pragma once + +#include + +#include + +#include + +#include +#include + +#include + +namespace dev +{ +namespace julia +{ + +/** + * Optimiser component that modifies an AST in place, inlining functions that can be + * inlined inside functional expressions, i.e. functions that + * - return a single value + * - have a body like r := + * - neither reference themselves nor r in the right hand side + * + * Furthermore, the arguments of the function call cannot have any side-effects. + * + * This component can only be used on sources with unique names. + */ +class ExpressionInliner: public ASTModifier +{ +public: + ExpressionInliner(Block& _block): + m_block(_block) + {} + + void run(); + + using ASTModifier::operator(); + virtual void operator()(FunctionDefinition& _fun) override; + + virtual void visit(Expression& _expression) override; + +private: + std::map m_inlinableFunctions; + std::map m_varReplacements; + /// Set of functions we are currently visiting inside. + std::set m_currentFunctions; + + Block& m_block; +}; + + +} +} diff --git a/libjulia/optimiser/ExpressionSimplifier.cpp b/libjulia/optimiser/ExpressionSimplifier.cpp new file mode 100644 index 000000000000..3d471cb3fd58 --- /dev/null +++ b/libjulia/optimiser/ExpressionSimplifier.cpp @@ -0,0 +1,50 @@ +/* + This file is part of solidity. + + solidity is free software: you can redistribute it and/or modify + it under the terms of the GNU General Public License as published by + the Free Software Foundation, either version 3 of the License, or + (at your option) any later version. + + solidity is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + GNU General Public License for more details. + + You should have received a copy of the GNU General Public License + along with solidity. If not, see . +*/ +/** + * Optimiser component that uses the simplification rules to simplify expressions. + */ + +#include + +#include +#include + +#include + +#include + +#include + +using namespace std; +using namespace dev; +using namespace dev::julia; +using namespace dev::solidity; + + +void ExpressionSimplifier::visit(Expression& _expression) +{ + ASTModifier::visit(_expression); + while (auto match = SimplificationRules::findFirstMatch(_expression)) + { + // Do not apply the rule if it removes non-constant parts of the expression. + // TODO: The check could actually be less strict than "movable". + // We only require "Does not cause side-effects". + if (match->removesNonConstants && !MovableChecker(_expression).movable()) + return; + _expression = match->action().toExpression(locationOf(_expression)); + } +} diff --git a/libjulia/optimiser/ExpressionSimplifier.h b/libjulia/optimiser/ExpressionSimplifier.h new file mode 100644 index 000000000000..8a9e0e200a00 --- /dev/null +++ b/libjulia/optimiser/ExpressionSimplifier.h @@ -0,0 +1,45 @@ +/* + This file is part of solidity. + + solidity is free software: you can redistribute it and/or modify + it under the terms of the GNU General Public License as published by + the Free Software Foundation, either version 3 of the License, or + (at your option) any later version. + + solidity is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + GNU General Public License for more details. + + You should have received a copy of the GNU General Public License + along with solidity. If not, see . +*/ +/** + * Optimiser component that uses the simplification rules to simplify expressions. + */ + +#pragma once + +#include + +#include + +namespace dev +{ +namespace julia +{ + +/** + * Applies simplification rules to all expressions. + */ +class ExpressionSimplifier: public ASTModifier +{ +public: + using ASTModifier::operator(); + virtual void visit(Expression& _expression); + +private: +}; + +} +} diff --git a/libjulia/optimiser/FunctionGrouper.cpp b/libjulia/optimiser/FunctionGrouper.cpp new file mode 100644 index 000000000000..cc40bc46b1e4 --- /dev/null +++ b/libjulia/optimiser/FunctionGrouper.cpp @@ -0,0 +1,49 @@ +/* + This file is part of solidity. + + solidity is free software: you can redistribute it and/or modify + it under the terms of the GNU General Public License as published by + the Free Software Foundation, either version 3 of the License, or + (at your option) any later version. + + solidity is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + GNU General Public License for more details. + + You should have received a copy of the GNU General Public License + along with solidity. If not, see . +*/ +/** + * Optimiser component that changes the code of a block so that all non-function definition + * statements are moved to a block of their own followed by all function definitions. + */ + +#include + +#include + +#include + +#include + +using namespace std; +using namespace dev; +using namespace dev::julia; +using namespace dev::solidity; + + +void FunctionGrouper::operator()(Block& _block) +{ + vector reordered; + reordered.emplace_back(Block{_block.location, {}}); + + for (auto&& statement: _block.statements) + { + if (statement.type() == typeid(FunctionDefinition)) + reordered.emplace_back(std::move(statement)); + else + boost::get(reordered.front()).statements.emplace_back(std::move(statement)); + } + _block.statements = std::move(reordered); +} diff --git a/libjulia/optimiser/FunctionGrouper.h b/libjulia/optimiser/FunctionGrouper.h new file mode 100644 index 000000000000..64a7131865b9 --- /dev/null +++ b/libjulia/optimiser/FunctionGrouper.h @@ -0,0 +1,46 @@ +/* + This file is part of solidity. + + solidity is free software: you can redistribute it and/or modify + it under the terms of the GNU General Public License as published by + the Free Software Foundation, either version 3 of the License, or + (at your option) any later version. + + solidity is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + GNU General Public License for more details. + + You should have received a copy of the GNU General Public License + along with solidity. If not, see . +*/ +/** + * Optimiser component that changes the code of a black so that all non-function definition + * instructions are moved to a block of their own followed by all function definitions. + */ + +#pragma once + +#include + +namespace dev +{ +namespace julia +{ + +/** + * Moves all instructions in a block into a new block at the start of the block, followed by + * all function definitions. + * + * After this step, a block is of the form + * { { I...} F... } + * Where I are (non-function-definition) instructions and F are function definitions. + */ +class FunctionGrouper +{ +public: + void operator()(Block& _block); +}; + +} +} diff --git a/libjulia/optimiser/FunctionHoister.cpp b/libjulia/optimiser/FunctionHoister.cpp new file mode 100644 index 000000000000..63f6edb983ce --- /dev/null +++ b/libjulia/optimiser/FunctionHoister.cpp @@ -0,0 +1,53 @@ +/* + This file is part of solidity. + + solidity is free software: you can redistribute it and/or modify + it under the terms of the GNU General Public License as published by + the Free Software Foundation, either version 3 of the License, or + (at your option) any later version. + + solidity is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + GNU General Public License for more details. + + You should have received a copy of the GNU General Public License + along with solidity. If not, see . +*/ +/** + * Optimiser component that changes the code so that it consists of a block starting with + * a single block followed only by function definitions and with no functions defined + * anywhere else. + */ + +#include +#include + +#include + +#include + +#include + +using namespace std; +using namespace dev; +using namespace dev::julia; +using namespace dev::solidity; + +void FunctionHoister::operator()(Block& _block) +{ + bool topLevel = m_isTopLevel; + m_isTopLevel = false; + for (auto&& statement: _block.statements) + { + boost::apply_visitor(*this, statement); + if (statement.type() == typeid(FunctionDefinition)) + { + m_functions.emplace_back(std::move(statement)); + statement = Block{_block.location, {}}; + } + } + removeEmptyBlocks(_block); + if (topLevel) + _block.statements += std::move(m_functions); +} diff --git a/libjulia/optimiser/FunctionHoister.h b/libjulia/optimiser/FunctionHoister.h new file mode 100644 index 000000000000..f9f8bce0a1db --- /dev/null +++ b/libjulia/optimiser/FunctionHoister.h @@ -0,0 +1,52 @@ +/* + This file is part of solidity. + + solidity is free software: you can redistribute it and/or modify + it under the terms of the GNU General Public License as published by + the Free Software Foundation, either version 3 of the License, or + (at your option) any later version. + + solidity is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + GNU General Public License for more details. + + You should have received a copy of the GNU General Public License + along with solidity. If not, see . +*/ +/** + * Optimiser component that changes the code so that all function definitions are at the top + * level block. + */ + +#pragma once + +#include + +#include + +namespace dev +{ +namespace julia +{ + +/** + * Moves all functions to the top-level scope. + * Applying this transformation to source code that has ambiguous identifiers might + * lead to invalid code. + * + * Prerequisites: Disambiguator + */ +class FunctionHoister: public ASTModifier +{ +public: + using ASTModifier::operator(); + virtual void operator()(Block& _block); + +private: + bool m_isTopLevel = true; + std::vector m_functions; +}; + +} +} diff --git a/libjulia/optimiser/InlinableExpressionFunctionFinder.cpp b/libjulia/optimiser/InlinableExpressionFunctionFinder.cpp new file mode 100644 index 000000000000..2097e091a580 --- /dev/null +++ b/libjulia/optimiser/InlinableExpressionFunctionFinder.cpp @@ -0,0 +1,70 @@ +/* + This file is part of solidity. + + solidity is free software: you can redistribute it and/or modify + it under the terms of the GNU General Public License as published by + the Free Software Foundation, either version 3 of the License, or + (at your option) any later version. + + solidity is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + GNU General Public License for more details. + + You should have received a copy of the GNU General Public License + along with solidity. If not, see . +*/ +/** + * Optimiser component that identifies functions to be inlined. + */ + +#include + +#include + +#include + +using namespace std; +using namespace dev; +using namespace dev::julia; + +void InlinableExpressionFunctionFinder::operator()(Identifier const& _identifier) +{ + checkAllowed(_identifier.name); + ASTWalker::operator()(_identifier); +} + +void InlinableExpressionFunctionFinder::operator()(FunctionCall const& _funCall) +{ + checkAllowed(_funCall.functionName.name); + ASTWalker::operator()(_funCall); +} + +void InlinableExpressionFunctionFinder::operator()(FunctionDefinition const& _function) +{ + if (_function.returnVariables.size() == 1 && _function.body.statements.size() == 1) + { + string const& retVariable = _function.returnVariables.front().name; + Statement const& bodyStatement = _function.body.statements.front(); + if (bodyStatement.type() == typeid(Assignment)) + { + Assignment const& assignment = boost::get(bodyStatement); + if (assignment.variableNames.size() == 1 && assignment.variableNames.front().name == retVariable) + { + // FIXME: use code size metric here + + // We cannot overwrite previous settings, because this function definition + // would not be valid here if we were searching inside a functionally inlinable + // function body. + solAssert(m_disallowedIdentifiers.empty() && !m_foundDisallowedIdentifier, ""); + m_disallowedIdentifiers = set{retVariable, _function.name}; + boost::apply_visitor(*this, *assignment.value); + if (!m_foundDisallowedIdentifier) + m_inlinableFunctions[_function.name] = &_function; + m_disallowedIdentifiers.clear(); + m_foundDisallowedIdentifier = false; + } + } + } + ASTWalker::operator()(_function.body); +} diff --git a/libjulia/optimiser/InlinableExpressionFunctionFinder.h b/libjulia/optimiser/InlinableExpressionFunctionFinder.h new file mode 100644 index 000000000000..36cb557a3ea4 --- /dev/null +++ b/libjulia/optimiser/InlinableExpressionFunctionFinder.h @@ -0,0 +1,71 @@ +/* + This file is part of solidity. + + solidity is free software: you can redistribute it and/or modify + it under the terms of the GNU General Public License as published by + the Free Software Foundation, either version 3 of the License, or + (at your option) any later version. + + solidity is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + GNU General Public License for more details. + + You should have received a copy of the GNU General Public License + along with solidity. If not, see . +*/ +/** + * Optimiser component that identifies functions to be inlined. + */ + +#pragma once + +#include +#include + +#include + +#include + +namespace dev +{ +namespace julia +{ + +/** + * Optimiser component that finds functions that can be + * inlined inside functional expressions, i.e. functions that + * - have a single return parameter r + * - have a body like r := + * - neither reference themselves nor r in the right hand side + * + * This component can only be used on sources with unique names. + */ +class InlinableExpressionFunctionFinder: public ASTWalker +{ +public: + + std::map const& inlinableFunctions() const + { + return m_inlinableFunctions; + } + + using ASTWalker::operator(); + virtual void operator()(Identifier const& _identifier) override; + virtual void operator()(FunctionCall const& _funCall) override; + virtual void operator()(FunctionDefinition const& _function) override; + +private: + void checkAllowed(std::string const& _name) + { + if (m_disallowedIdentifiers.count(_name)) + m_foundDisallowedIdentifier = true; + } + + bool m_foundDisallowedIdentifier = false; + std::set m_disallowedIdentifiers; + std::map m_inlinableFunctions; +}; + +} +} diff --git a/libjulia/optimiser/Metrics.cpp b/libjulia/optimiser/Metrics.cpp new file mode 100644 index 000000000000..eaa1494f5e66 --- /dev/null +++ b/libjulia/optimiser/Metrics.cpp @@ -0,0 +1,52 @@ +/*( + This file is part of solidity. + + solidity is free software: you can redistribute it and/or modify + it under the terms of the GNU General Public License as published by + the Free Software Foundation, either version 3 of the License, or + (at your option) any later version. + + solidity is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + GNU General Public License for more details. + + You should have received a copy of the GNU General Public License + along with solidity. If not, see . +*/ +/** +* Module providing metrics for the optimizer. +*/ + +#include + +#include + +using namespace dev; +using namespace dev::julia; + +size_t CodeSize::codeSize(Statement const& _statement) +{ + CodeSize cs; + cs.visit(_statement); + return cs.m_size; +} + +size_t CodeSize::codeSize(Expression const& _expression) +{ + CodeSize cs; + cs.visit(_expression); + return cs.m_size; +} + +void CodeSize::visit(Statement const& _statement) +{ + ++m_size; + ASTWalker::visit(_statement); +} + +void CodeSize::visit(Expression const& _expression) +{ + ++m_size; + ASTWalker::visit(_expression); +} diff --git a/libjulia/optimiser/Metrics.h b/libjulia/optimiser/Metrics.h new file mode 100644 index 000000000000..ff058fc39cc0 --- /dev/null +++ b/libjulia/optimiser/Metrics.h @@ -0,0 +1,49 @@ +/* + This file is part of solidity. + + solidity is free software: you can redistribute it and/or modify + it under the terms of the GNU General Public License as published by + the Free Software Foundation, either version 3 of the License, or + (at your option) any later version. + + solidity is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + GNU General Public License for more details. + + You should have received a copy of the GNU General Public License + along with solidity. If not, see . +*/ +/** + * Module providing metrics for the optimizer. + */ + +#pragma once + +#include + +namespace dev +{ +namespace julia +{ + +class CodeSize: public ASTWalker +{ +public: + /// Returns a metric for the code size of an AST element. + /// More specifically, it returns the number of AST nodes. + static size_t codeSize(Statement const& _statement); + /// Returns a metric for the code size of an AST element. + /// More specifically, it returns the number of AST nodes. + static size_t codeSize(Expression const& _expression); + +private: + virtual void visit(Statement const& _statement) override; + virtual void visit(Expression const& _expression) override; + +private: + size_t m_size = 0; +}; + +} +} diff --git a/libjulia/optimiser/NameCollector.cpp b/libjulia/optimiser/NameCollector.cpp new file mode 100644 index 000000000000..510ee2897266 --- /dev/null +++ b/libjulia/optimiser/NameCollector.cpp @@ -0,0 +1,75 @@ +/* + This file is part of solidity. + + solidity is free software: you can redistribute it and/or modify + it under the terms of the GNU General Public License as published by + the Free Software Foundation, either version 3 of the License, or + (at your option) any later version. + + solidity is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + GNU General Public License for more details. + + You should have received a copy of the GNU General Public License + along with solidity. If not, see . +*/ +/** + * Specific AST walker that collects all defined names. + */ + +#include + +#include + +using namespace std; +using namespace dev; +using namespace dev::julia; + +void NameCollector::operator()(VariableDeclaration const& _varDecl) +{ + for (auto const& var: _varDecl.variables) + m_names.insert(var.name); +} + +void NameCollector::operator ()(FunctionDefinition const& _funDef) +{ + m_names.insert(_funDef.name); + m_functions[_funDef.name] = &_funDef; + for (auto const arg: _funDef.parameters) + m_names.insert(arg.name); + for (auto const ret: _funDef.returnVariables) + m_names.insert(ret.name); + ASTWalker::operator ()(_funDef); +} + +void ReferencesCounter::operator()(Identifier const& _identifier) +{ + ++m_references[_identifier.name]; +} + +void ReferencesCounter::operator()(FunctionCall const& _funCall) +{ + ++m_references[_funCall.functionName.name]; + ASTWalker::operator()(_funCall); +} + +map ReferencesCounter::countReferences(Block const& _block) +{ + ReferencesCounter counter; + counter(_block); + return counter.references(); +} + +map ReferencesCounter::countReferences(Expression const& _expression) +{ + ReferencesCounter counter; + counter.visit(_expression); + return counter.references(); +} + +void Assignments::operator()(Assignment const& _assignment) +{ + for (auto const& var: _assignment.variableNames) + m_names.insert(var.name); +} diff --git a/libjulia/optimiser/NameCollector.h b/libjulia/optimiser/NameCollector.h new file mode 100644 index 000000000000..2d4a1d4bfac5 --- /dev/null +++ b/libjulia/optimiser/NameCollector.h @@ -0,0 +1,84 @@ +/* + This file is part of solidity. + + solidity is free software: you can redistribute it and/or modify + it under the terms of the GNU General Public License as published by + the Free Software Foundation, either version 3 of the License, or + (at your option) any later version. + + solidity is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + GNU General Public License for more details. + + You should have received a copy of the GNU General Public License + along with solidity. If not, see . +*/ +/** + * Specific AST walkers that collect facts about identifiers and definitions. + */ + +#pragma once + +#include + +#include +#include +#include + +namespace dev +{ +namespace julia +{ + +/** + * Specific AST walker that collects all defined names. + */ +class NameCollector: public ASTWalker +{ +public: + using ASTWalker::operator (); + virtual void operator()(VariableDeclaration const& _varDecl) override; + virtual void operator()(FunctionDefinition const& _funDef) override; + + std::set const& names() const { return m_names; } + std::map const& functions() const { return m_functions; } +private: + std::set m_names; + std::map m_functions; +}; + +/** + * Specific AST walker that counts all references to all declarations. + */ +class ReferencesCounter: public ASTWalker +{ +public: + using ASTWalker::operator (); + virtual void operator()(Identifier const& _identifier); + virtual void operator()(FunctionCall const& _funCall); + + static std::map countReferences(Block const& _block); + static std::map countReferences(Expression const& _expression); + + std::map const& references() const { return m_references; } +private: + std::map m_references; +}; + +/** + * Specific AST walker that finds all variables that are assigned to. + */ +class Assignments: public ASTWalker +{ +public: + using ASTWalker::operator (); + virtual void operator()(Assignment const& _assignment) override; + + std::set const& names() const { return m_names; } +private: + std::set m_names; +}; + +} +} diff --git a/libjulia/optimiser/README.md b/libjulia/optimiser/README.md new file mode 100644 index 000000000000..e71344404fa6 --- /dev/null +++ b/libjulia/optimiser/README.md @@ -0,0 +1,89 @@ +## IULIA Optimiser + +The iulia optimiser consists of several stages and components that all transform +the AST in a semantically equivalent way. The goal is to end up either with code +that is shorter or at least only marginally longer but will allow further +optimisation steps. + +The optimiser currently follows a purely greedy strategy and does not do any +backtracking. + +## Disambiguator + +The disambiguator takes an AST and returns a fresh copy where all identifiers have +names unique to the input AST. This is a prerequisite for all other optimiser stages. +One of the benefits is that identifier lookup does not need to take scopes into account +and we can basically ignore the result of the analysis phase. + +All subsequent stages have the property that all names stay unique. This means if +a new identifier needs to be introduced, a new unique name is generated. + +## Function Hoister + +The function hoister moves all function definitions to the topmost block. This is +a semantically equivalent transformation as long as it is performed after the +disambiguation stage. The reason is that moving a definition upwards cannot decrease +its visibility and it is impossible to reference variables defined in a different function. + +The benefit of this stage is that function definitions can be lookup up more easily. + +## Function Grouper + +The function grouper has to be applied after the disambiguator and the function hoister. +Its effect is that all topmost elements that are not function definitions are moved +into a single block which is the first satement of the root block. + +After this step, a program has the following normal form: + + { I F... } + +Where I is a block that does not contain any function definitions (not even recursively) +and F is a list of function definitions such that no function contains a function definition. + +## Functional Inliner + +The functional inliner depends on the disambiguator, the function hoister and function grouper. +It performs function inlining such that the result of the inlining is an expression. This can +only be done if the body of the function to be inlined has the form ``{ r := E }`` where ``r`` +is the single return value of the function, ``E`` is an expression and all arguments in the +function call are so-called movable expressions. A movable expression is either a literal, a +variable or a function call (or EVM opcode) which does not have side-effects and also does not +depend on any side-effects. + +As an example, neither ``mload`` nor ``mstore`` would be allowed. + +## Full Function Inliner + +## Rematerialisation + +The rematerialisation stage tries to replace variable references by the expression that +was last assigned to the variable. This is of course only beneficial if this expression +is comparatively cheap to evaluate. Furthermore, it is only semantically equivalent if +the value of the expression did not change between the point of assignment and the +point of use. The main benefit of this stage is that it can save stack slots if it +leads to a variable being eliminated completely (see below), but it can also +save a DUP opcode on the EVM if the expression is very cheap. + +The algorithm only allows movable expressions (see above for a definition) in this case. +Expressions that contain other variables are also disallowed if one of those variables +have been assigned to in the meantime. This is also not applied to variables where +assignment and use span across loops and conditionals. + +## Unused Definition Pruner + +If a variable or function is not referenced, it is removed from the code. +If there are two assignments to a variable where the first one is a movable expression +and the variable is not used between the two assignments (and the second is not inside +a loop or conditional, the first one is not inside), the first assignment is removed. + + +## Function Unifier + +## Expression Simplifier + +This step can only be applied for the EVM-flavoured dialect of iulia. It applies +simple rules like ``x + 0 == x`` to simplify expressions. + +## Ineffective Statement Remover + +This step removes statements that have no side-effects. diff --git a/libjulia/optimiser/Rematerialiser.cpp b/libjulia/optimiser/Rematerialiser.cpp new file mode 100644 index 000000000000..eaa75e339e9d --- /dev/null +++ b/libjulia/optimiser/Rematerialiser.cpp @@ -0,0 +1,54 @@ +/*( + This file is part of solidity. + + solidity is free software: you can redistribute it and/or modify + it under the terms of the GNU General Public License as published by + the Free Software Foundation, either version 3 of the License, or + (at your option) any later version. + + solidity is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + GNU General Public License for more details. + + You should have received a copy of the GNU General Public License + along with solidity. If not, see . +*/ +/** + * Optimisation stage that replaces variables by their most recently assigned expressions. + */ + +#include + +#include +#include + +#include + +using namespace std; +using namespace dev; +using namespace dev::julia; + +void Rematerialiser::visit(Expression& _e) +{ + if (_e.type() == typeid(Identifier)) + { + Identifier& identifier = boost::get(_e); + if (m_value.count(identifier.name)) + { + string name = identifier.name; + bool expressionValid = true; + for (auto const& ref: m_references[name]) + if (!inScope(ref)) + { + expressionValid = false; + break; + } + solAssert(m_value.at(name), ""); + auto const& value = *m_value.at(name); + if (expressionValid && CodeSize::codeSize(value) <= 7) + _e = (ASTCopier{}).translate(value); + } + } + DataFlowAnalyzer::visit(_e); +} diff --git a/libjulia/optimiser/Rematerialiser.h b/libjulia/optimiser/Rematerialiser.h new file mode 100644 index 000000000000..60dbfada56f2 --- /dev/null +++ b/libjulia/optimiser/Rematerialiser.h @@ -0,0 +1,48 @@ +/* + This file is part of solidity. + + solidity is free software: you can redistribute it and/or modify + it under the terms of the GNU General Public License as published by + the Free Software Foundation, either version 3 of the License, or + (at your option) any later version. + + solidity is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + GNU General Public License for more details. + + You should have received a copy of the GNU General Public License + along with solidity. If not, see . +*/ +/** + * Optimisation stage that replaces variables by their most recently assigned expressions. + */ + +#pragma once + +#include + +#include +#include +#include + +namespace dev +{ +namespace julia +{ + +/** + * Optimisation stage that replaces variables by their most recently assigned expressions. + * + * Prerequisite: Disambiguator + */ +class Rematerialiser: public DataFlowAnalyzer +{ +protected: + using ASTModifier::visit; + virtual void visit(Expression& _e) override; + +}; + +} +} diff --git a/libjulia/optimiser/Semantics.cpp b/libjulia/optimiser/Semantics.cpp new file mode 100644 index 000000000000..92728c46b6ff --- /dev/null +++ b/libjulia/optimiser/Semantics.cpp @@ -0,0 +1,60 @@ +/*( + This file is part of solidity. + + solidity is free software: you can redistribute it and/or modify + it under the terms of the GNU General Public License as published by + the Free Software Foundation, either version 3 of the License, or + (at your option) any later version. + + solidity is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + GNU General Public License for more details. + + You should have received a copy of the GNU General Public License + along with solidity. If not, see . +*/ +/** + * Specific AST walkers that collect semantical facts. + */ + +#include + +#include + +#include + +#include + +using namespace std; +using namespace dev; +using namespace dev::julia; + +MovableChecker::MovableChecker(Expression const& _expression) +{ + visit(_expression); +} + +void MovableChecker::operator()(Identifier const& _identifier) +{ + ASTWalker::operator()(_identifier); + m_variableReferences.insert(_identifier.name); +} + +void MovableChecker::operator()(FunctionalInstruction const& _instr) +{ + if (!eth::SemanticInformation::movable(_instr.instruction)) + m_movable = false; + else + ASTWalker::operator()(_instr); +} + +void MovableChecker::operator()(FunctionCall const&) +{ + m_movable = false; +} + +void MovableChecker::visit(Statement const&) +{ + solAssert(false, "Movability for statement requested."); +} diff --git a/libjulia/optimiser/Semantics.h b/libjulia/optimiser/Semantics.h new file mode 100644 index 000000000000..6df5f01a8681 --- /dev/null +++ b/libjulia/optimiser/Semantics.h @@ -0,0 +1,62 @@ +/* + This file is part of solidity. + + solidity is free software: you can redistribute it and/or modify + it under the terms of the GNU General Public License as published by + the Free Software Foundation, either version 3 of the License, or + (at your option) any later version. + + solidity is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + GNU General Public License for more details. + + You should have received a copy of the GNU General Public License + along with solidity. If not, see . +*/ +/** + * Specific AST walkers that collect semantical facts. + */ + +#pragma once + +#include + +#include +#include +#include + +namespace dev +{ +namespace julia +{ + +/** + * Specific AST walker that determines whether an expression is movable. + */ +class MovableChecker: public ASTWalker +{ +public: + MovableChecker() = default; + explicit MovableChecker(Expression const& _expression); + + virtual void operator()(Identifier const& _identifier) override; + virtual void operator()(FunctionalInstruction const& _functionalInstruction) override; + virtual void operator()(FunctionCall const& _functionCall) override; + + /// Disallow visiting anything apart from Expressions (this throws). + virtual void visit(Statement const&) override; + using ASTWalker::visit; + + bool movable() const { return m_movable; } + std::set const& referencedVariables() const { return m_variableReferences; } + +private: + /// Which variables the current expression references. + std::set m_variableReferences; + /// Is the current expression movable or not. + bool m_movable = true; +}; + +} +} diff --git a/libjulia/optimiser/SimplificationRules.cpp b/libjulia/optimiser/SimplificationRules.cpp new file mode 100644 index 000000000000..e5dbb0591a68 --- /dev/null +++ b/libjulia/optimiser/SimplificationRules.cpp @@ -0,0 +1,182 @@ +/* + This file is part of solidity. + + solidity is free software: you can redistribute it and/or modify + it under the terms of the GNU General Public License as published by + the Free Software Foundation, either version 3 of the License, or + (at your option) any later version. + + solidity is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + GNU General Public License for more details. + + You should have received a copy of the GNU General Public License + along with solidity. If not, see . +*/ +/** + * Module for applying replacement rules against Expressions. + */ + +#include + +#include +#include +#include +#include + +#include + +#include + +using namespace std; +using namespace dev; +using namespace dev::julia; + + +SimplificationRule const* SimplificationRules::findFirstMatch(Expression const& _expr) +{ + if (_expr.type() != typeid(FunctionalInstruction)) + return nullptr; + + static SimplificationRules rules; + + FunctionalInstruction const& instruction = boost::get(_expr); + for (auto const& rule: rules.m_rules[byte(instruction.instruction)]) + { + rules.resetMatchGroups(); + if (rule.pattern.matches(_expr)) + return &rule; + } + return nullptr; +} + +void SimplificationRules::addRules(vector> const& _rules) +{ + for (auto const& r: _rules) + addRule(r); +} + +void SimplificationRules::addRule(SimplificationRule const& _rule) +{ + m_rules[byte(_rule.pattern.instruction())].push_back(_rule); +} + +SimplificationRules::SimplificationRules() +{ + // Multiple occurences of one of these inside one rule must match the same equivalence class. + // Constants. + Pattern A(PatternKind::Constant); + Pattern B(PatternKind::Constant); + Pattern C(PatternKind::Constant); + // Anything. + Pattern X; + Pattern Y; + A.setMatchGroup(1, m_matchGroups); + B.setMatchGroup(2, m_matchGroups); + C.setMatchGroup(3, m_matchGroups); + X.setMatchGroup(4, m_matchGroups); + Y.setMatchGroup(5, m_matchGroups); + + addRules(simplificationRuleList(A, B, C, X, Y)); +} + +Pattern::Pattern(solidity::Instruction _instruction, vector const& _arguments): + m_kind(PatternKind::Operation), + m_instruction(_instruction), + m_arguments(_arguments) +{ +} + +void Pattern::setMatchGroup(unsigned _group, map& _matchGroups) +{ + m_matchGroup = _group; + m_matchGroups = &_matchGroups; +} + +bool Pattern::matches(Expression const& _expr) const +{ + if (m_kind == PatternKind::Constant) + { + if (_expr.type() != typeid(Literal)) + return false; + Literal const& literal = boost::get(_expr); + if (literal.kind != assembly::LiteralKind::Number) + return false; + if (m_data && *m_data != u256(literal.value)) + return false; + assertThrow(m_arguments.empty(), OptimizerException, ""); + } + else if (m_kind == PatternKind::Operation) + { + if (_expr.type() != typeid(FunctionalInstruction)) + return false; + FunctionalInstruction const& instr = boost::get(_expr); + if (m_instruction != instr.instruction) + return false; + assertThrow(m_arguments.size() == instr.arguments.size(), OptimizerException, ""); + for (size_t i = 0; i < m_arguments.size(); ++i) + if (!m_arguments[i].matches(instr.arguments.at(i))) + return false; + } + else + { + assertThrow(m_arguments.empty(), OptimizerException, ""); + } + // We support matching multiple expressions that require the same value + // based on identical ASTs, which have to be movable. + if (m_matchGroup) + { + if (m_matchGroups->count(m_matchGroup)) + { + Expression const* firstMatch = (*m_matchGroups)[m_matchGroup]; + assertThrow(firstMatch, OptimizerException, "Match set but to null."); + return + SyntacticalEqualityChecker::equal(*firstMatch, _expr) && + MovableChecker(_expr).movable(); + } + else + (*m_matchGroups)[m_matchGroup] = &_expr; + } + return true; +} + +solidity::Instruction Pattern::instruction() const +{ + assertThrow(m_kind == PatternKind::Operation, OptimizerException, ""); + return m_instruction; +} + +Expression Pattern::toExpression(SourceLocation const& _location) const +{ + if (matchGroup()) + return ASTCopier().translate(matchGroupValue()); + if (m_kind == PatternKind::Constant) + { + assertThrow(m_data, OptimizerException, "No match group and no constant value given."); + return Literal{_location, assembly::LiteralKind::Number, formatNumber(*m_data), ""}; + } + else if (m_kind == PatternKind::Operation) + { + vector arguments; + for (auto const& arg: m_arguments) + arguments.emplace_back(arg.toExpression(_location)); + return FunctionalInstruction{_location, m_instruction, std::move(arguments)}; + } + assertThrow(false, OptimizerException, "Pattern of kind 'any', but no match group."); +} + +u256 Pattern::d() const +{ + Literal const& literal = boost::get(matchGroupValue()); + assertThrow(literal.kind == assembly::LiteralKind::Number, OptimizerException, ""); + return u256(literal.value); +} + +Expression const& Pattern::matchGroupValue() const +{ + assertThrow(m_matchGroup > 0, OptimizerException, ""); + assertThrow(!!m_matchGroups, OptimizerException, ""); + assertThrow((*m_matchGroups)[m_matchGroup], OptimizerException, ""); + return *(*m_matchGroups)[m_matchGroup]; +} diff --git a/libjulia/optimiser/SimplificationRules.h b/libjulia/optimiser/SimplificationRules.h new file mode 100644 index 000000000000..68b640b1d7ca --- /dev/null +++ b/libjulia/optimiser/SimplificationRules.h @@ -0,0 +1,117 @@ +/* + This file is part of solidity. + + solidity is free software: you can redistribute it and/or modify + it under the terms of the GNU General Public License as published by + the Free Software Foundation, either version 3 of the License, or + (at your option) any later version. + + solidity is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + GNU General Public License for more details. + + You should have received a copy of the GNU General Public License + along with solidity. If not, see . +*/ +/** + * Module for applying replacement rules against Expressions. + */ + +#pragma once + +#include +#include + +#include + +#include + +#include + +#include +#include + +namespace dev +{ +namespace julia +{ + +class Pattern; + +/** + * Container for all simplification rules. + */ +class SimplificationRules: public boost::noncopyable +{ +public: + SimplificationRules(); + + /// @returns a pointer to the first matching pattern and sets the match + /// groups accordingly. + static SimplificationRule const* findFirstMatch(Expression const& _expr); + +private: + void addRules(std::vector> const& _rules); + void addRule(SimplificationRule const& _rule); + + void resetMatchGroups() { m_matchGroups.clear(); } + + std::map m_matchGroups; + std::vector> m_rules[256]; +}; + +enum class PatternKind +{ + Operation, + Constant, + Any +}; + +/** + * Pattern to match against an expression. + * Also stores matched expressions to retrieve them later, for constructing new expressions using + * ExpressionTemplate. + */ +class Pattern +{ +public: + /// Matches any expression. + Pattern(PatternKind _kind = PatternKind::Any): m_kind(_kind) {} + // Matches a specific constant value. + Pattern(unsigned _value): Pattern(u256(_value)) {} + // Matches a specific constant value. + Pattern(u256 const& _value): m_kind(PatternKind::Constant), m_data(std::make_shared(_value)) {} + // Matches a given instruction with given arguments + Pattern(solidity::Instruction _instruction, std::vector const& _arguments = {}); + /// Sets this pattern to be part of the match group with the identifier @a _group. + /// Inside one rule, all patterns in the same match group have to match expressions from the + /// same expression equivalence class. + void setMatchGroup(unsigned _group, std::map& _matchGroups); + unsigned matchGroup() const { return m_matchGroup; } + bool matches(Expression const& _expr) const; + + std::vector arguments() const { return m_arguments; } + + /// @returns the data of the matched expression if this pattern is part of a match group. + u256 d() const; + + solidity::Instruction instruction() const; + + /// Turns this pattern into an actual expression. Should only be called + /// for patterns resulting from an action, i.e. with match groups assigned. + Expression toExpression(SourceLocation const& _location) const; + +private: + Expression const& matchGroupValue() const; + + PatternKind m_kind = PatternKind::Any; + solidity::Instruction m_instruction; ///< Only valid if m_kind is Operation + std::shared_ptr m_data; ///< Only valid if m_kind is Constant + std::vector m_arguments; + unsigned m_matchGroup = 0; + std::map* m_matchGroups = nullptr; +}; + +} +} diff --git a/libjulia/optimiser/Substitution.cpp b/libjulia/optimiser/Substitution.cpp new file mode 100644 index 000000000000..668b6cb644e5 --- /dev/null +++ b/libjulia/optimiser/Substitution.cpp @@ -0,0 +1,39 @@ +/* + This file is part of solidity. + + solidity is free software: you can redistribute it and/or modify + it under the terms of the GNU General Public License as published by + the Free Software Foundation, either version 3 of the License, or + (at your option) any later version. + + solidity is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + GNU General Public License for more details. + + You should have received a copy of the GNU General Public License + along with solidity. If not, see . +*/ +/** + * Specific AST copier that replaces certain identifiers with expressions. + */ + +#include + +#include + +using namespace std; +using namespace dev; +using namespace dev::julia; + +Expression Substitution::translate(Expression const& _expression) +{ + if (_expression.type() == typeid(Identifier)) + { + string const& name = boost::get(_expression).name; + if (m_substitutions.count(name)) + // No recursive substitution + return ASTCopier().translate(*m_substitutions.at(name)); + } + return ASTCopier::translate(_expression); +} diff --git a/libjulia/optimiser/Substitution.h b/libjulia/optimiser/Substitution.h new file mode 100644 index 000000000000..313a08d7884d --- /dev/null +++ b/libjulia/optimiser/Substitution.h @@ -0,0 +1,50 @@ +/* + This file is part of solidity. + + solidity is free software: you can redistribute it and/or modify + it under the terms of the GNU General Public License as published by + the Free Software Foundation, either version 3 of the License, or + (at your option) any later version. + + solidity is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + GNU General Public License for more details. + + You should have received a copy of the GNU General Public License + along with solidity. If not, see . +*/ +/** + * Specific AST copier that replaces certain identifiers with expressions. + */ + +#pragma once + +#include + +#include +#include +#include + +namespace dev +{ +namespace julia +{ + +/** + * Specific AST copier that replaces certain identifiers with expressions. + */ +class Substitution: public ASTCopier +{ +public: + Substitution(std::map const& _substitutions): + m_substitutions(_substitutions) + {} + virtual Expression translate(Expression const& _expression) override; + +private: + std::map const& m_substitutions; +}; + +} +} diff --git a/libjulia/optimiser/SyntacticalEquality.cpp b/libjulia/optimiser/SyntacticalEquality.cpp new file mode 100644 index 000000000000..2b90b0911198 --- /dev/null +++ b/libjulia/optimiser/SyntacticalEquality.cpp @@ -0,0 +1,75 @@ +/*( + This file is part of solidity. + + solidity is free software: you can redistribute it and/or modify + it under the terms of the GNU General Public License as published by + the Free Software Foundation, either version 3 of the License, or + (at your option) any later version. + + solidity is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + GNU General Public License for more details. + + You should have received a copy of the GNU General Public License + along with solidity. If not, see . +*/ +/** + * Component that can compare ASTs for equality on a syntactic basis. + */ + +#include + +#include +#include + +#include + +using namespace std; +using namespace dev; +using namespace dev::julia; + +bool SyntacticalEqualityChecker::equal(Expression const& _e1, Expression const& _e2) +{ + if (_e1.type() != _e2.type()) + return false; + + // TODO This should be replaced by some kind of AST walker as soon as it gets + // more complex. + if (_e1.type() == typeid(FunctionalInstruction)) + { + auto const& e1 = boost::get(_e1); + auto const& e2 = boost::get(_e2); + return + e1.instruction == e2.instruction && + equalVector(e1.arguments, e2.arguments); + } + else if (_e1.type() == typeid(FunctionCall)) + { + auto const& e1 = boost::get(_e1); + auto const& e2 = boost::get(_e2); + return + equal(e1.functionName, e2.functionName) && + equalVector(e1.arguments, e2.arguments); + } + else if (_e1.type() == typeid(Identifier)) + return boost::get(_e1).name == boost::get(_e2).name; + else if (_e1.type() == typeid(Literal)) + { + auto const& e1 = boost::get(_e1); + auto const& e2 = boost::get(_e2); + return e1.kind == e2.kind && e1.value == e2.value && e1.type == e2.type; + } + else + { + solAssert(false, "Invlid expression"); + } + return false; +} + +bool SyntacticalEqualityChecker::equalVector(vector const& _e1, vector const& _e2) +{ + return _e1.size() == _e2.size() && + std::equal(begin(_e1), end(_e1), begin(_e2), SyntacticalEqualityChecker::equal); + +} diff --git a/libjulia/optimiser/SyntacticalEquality.h b/libjulia/optimiser/SyntacticalEquality.h new file mode 100644 index 000000000000..b7c09330ff7d --- /dev/null +++ b/libjulia/optimiser/SyntacticalEquality.h @@ -0,0 +1,50 @@ +/* + This file is part of solidity. + + solidity is free software: you can redistribute it and/or modify + it under the terms of the GNU General Public License as published by + the Free Software Foundation, either version 3 of the License, or + (at your option) any later version. + + solidity is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + GNU General Public License for more details. + + You should have received a copy of the GNU General Public License + along with solidity. If not, see . +*/ +/** + * Component that can compare ASTs for equality on a syntactic basis. + */ + +#pragma once + +#include + +#include + +namespace dev +{ +namespace julia +{ + +/** + * Component that can compare ASTs for equality on a syntactic basis. + * Ignores source locations but requires exact matches otherwise. + * + * TODO: Only implemented for Expressions for now. + * A future version might also recognize renamed variables and thus could be used to + * remove duplicate functions. + */ +class SyntacticalEqualityChecker +{ +public: + static bool equal(Expression const& _e1, Expression const& _e2); + +protected: + static bool equalVector(std::vector const& _e1, std::vector const& _e2); +}; + +} +} diff --git a/libjulia/optimiser/UnusedPruner.cpp b/libjulia/optimiser/UnusedPruner.cpp new file mode 100644 index 000000000000..500384314512 --- /dev/null +++ b/libjulia/optimiser/UnusedPruner.cpp @@ -0,0 +1,116 @@ +/*( + This file is part of solidity. + + solidity is free software: you can redistribute it and/or modify + it under the terms of the GNU General Public License as published by + the Free Software Foundation, either version 3 of the License, or + (at your option) any later version. + + solidity is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + GNU General Public License for more details. + + You should have received a copy of the GNU General Public License + along with solidity. If not, see . +*/ +/** + * Optimisation stage that removes unused variables and functions. + */ + +#include + +#include +#include +#include + +#include + +#include + +using namespace std; +using namespace dev; +using namespace dev::julia; + +UnusedPruner::UnusedPruner(Block& _ast) +{ + ReferencesCounter counter; + counter(_ast); + + m_references = counter.references(); +} + +void UnusedPruner::operator()(Block& _block) +{ + for (auto&& statement: _block.statements) + if (statement.type() == typeid(FunctionDefinition)) + { + FunctionDefinition& funDef = boost::get(statement); + if (!used(funDef.name)) + { + subtractReferences(ReferencesCounter::countReferences(funDef.body)); + statement = Block{std::move(funDef.location), {}}; + } + } + else if (statement.type() == typeid(VariableDeclaration)) + { + VariableDeclaration& varDecl = boost::get(statement); + // Multi-variable declarations are special. We can only remove it + // if all vairables are unused and the right-hand-side is either + // movable or it return a single value. In the latter case, we + // replace `let a := f()` by `pop(f())` (in pure IULIA, this will be + // `drop(f())`). + if (boost::algorithm::none_of( + varDecl.variables, + [=](TypedName const& _typedName) { return used(_typedName.name); } + )) + { + if (!varDecl.value) + statement = Block{std::move(varDecl.location), {}}; + else if (MovableChecker(*varDecl.value).movable()) + { + subtractReferences(ReferencesCounter::countReferences(*varDecl.value)); + statement = Block{std::move(varDecl.location), {}}; + } + else if (varDecl.variables.size() == 1) + // In pure IULIA, this should be replaced by a function call to `drop` + // instead of `pop`. + statement = ExpressionStatement{varDecl.location, FunctionalInstruction{ + varDecl.location, + solidity::Instruction::POP, + {*std::move(varDecl.value)} + }}; + } + } + + removeEmptyBlocks(_block); + + ASTModifier::operator()(_block); +} + +void UnusedPruner::runUntilStabilised(Block& _ast) +{ + while (true) + { + UnusedPruner pruner(_ast); + pruner(_ast); + if (!pruner.shouldRunAgain()) + return; + } +} + +bool UnusedPruner::used(string const& _name) const +{ + return m_references.count(_name) && m_references.at(_name) > 0; +} + +void UnusedPruner::subtractReferences(map const& _subtrahend) +{ + for (auto const& ref: _subtrahend) + { + solAssert(m_references.count(ref.first), ""); + solAssert(m_references.at(ref.first) >= ref.second, ""); + m_references[ref.first] -= ref.second; + m_shouldRunAgain = true; + } +} diff --git a/libjulia/optimiser/UnusedPruner.h b/libjulia/optimiser/UnusedPruner.h new file mode 100644 index 000000000000..73e8de7c690d --- /dev/null +++ b/libjulia/optimiser/UnusedPruner.h @@ -0,0 +1,67 @@ +/* + This file is part of solidity. + + solidity is free software: you can redistribute it and/or modify + it under the terms of the GNU General Public License as published by + the Free Software Foundation, either version 3 of the License, or + (at your option) any later version. + + solidity is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + GNU General Public License for more details. + + You should have received a copy of the GNU General Public License + along with solidity. If not, see . +*/ +/** + * Optimisation stage that removes unused variables and functions. + */ + +#pragma once + +#include + +#include +#include +#include + +namespace dev +{ +namespace julia +{ + +/** + * Optimisation stage that removes unused variables and functions. + * + * TODO: Also remove intermediate variable assignments from movable expressions + * which are not referenced until after the next assignment to the same variable. + * + * Note that this does not remove circular references. + * + * Prerequisite: Disambiguator + */ +class UnusedPruner: public ASTModifier +{ +public: + explicit UnusedPruner(Block& _ast); + + using ASTModifier::operator(); + virtual void operator()(Block& _block) override; + + // @returns true iff the code changed in the previous run. + bool shouldRunAgain() const { return m_shouldRunAgain; } + + // Run the pruner until the code does not change anymore. + static void runUntilStabilised(Block& _ast); + +private: + bool used(std::string const& _name) const; + void subtractReferences(std::map const& _subtrahend); + + bool m_shouldRunAgain = false; + std::map m_references; +}; + +} +} diff --git a/libjulia/optimiser/Utilities.cpp b/libjulia/optimiser/Utilities.cpp new file mode 100644 index 000000000000..ff108b895f06 --- /dev/null +++ b/libjulia/optimiser/Utilities.cpp @@ -0,0 +1,39 @@ +/*( + This file is part of solidity. + + solidity is free software: you can redistribute it and/or modify + it under the terms of the GNU General Public License as published by + the Free Software Foundation, either version 3 of the License, or + (at your option) any later version. + + solidity is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + GNU General Public License for more details. + + You should have received a copy of the GNU General Public License + along with solidity. If not, see . +*/ +/** + * Some useful snippets for the optimiser. + */ + +#include + +#include + +#include + +#include + +using namespace std; +using namespace dev; +using namespace dev::julia; + +void dev::julia::removeEmptyBlocks(Block& _block) +{ + auto isEmptyBlock = [](Statement const& _st) -> bool { + return _st.type() == typeid(Block) && boost::get(_st).statements.empty(); + }; + boost::range::remove_erase_if(_block.statements, isEmptyBlock); +} diff --git a/libjulia/optimiser/Utilities.h b/libjulia/optimiser/Utilities.h new file mode 100644 index 000000000000..e3b4b087aa5d --- /dev/null +++ b/libjulia/optimiser/Utilities.h @@ -0,0 +1,39 @@ +/* + This file is part of solidity. + + solidity is free software: you can redistribute it and/or modify + it under the terms of the GNU General Public License as published by + the Free Software Foundation, either version 3 of the License, or + (at your option) any later version. + + solidity is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + GNU General Public License for more details. + + You should have received a copy of the GNU General Public License + along with solidity. If not, see . +*/ +/** + * Small useful snippets for the optimiser. + */ + +#pragma once + +#include + +#include + +namespace dev +{ +namespace julia +{ + +struct IuliaException: virtual Exception {}; +struct OptimizerException: virtual IuliaException {}; + +/// Removes statements that are just empty blocks (non-recursive). +void removeEmptyBlocks(Block& _block); + +} +} diff --git a/libsolc/CMakeLists.txt b/libsolc/CMakeLists.txt new file mode 100644 index 000000000000..e67583dd2e96 --- /dev/null +++ b/libsolc/CMakeLists.txt @@ -0,0 +1,9 @@ +if (EMSCRIPTEN) + set(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} -s EXPORTED_FUNCTIONS='[\"_compileJSON\",\"_license\",\"_version\",\"_compileJSONMulti\",\"_compileJSONCallback\",\"_compileStandard\"]' -s RESERVED_FUNCTION_POINTERS=20") + add_executable(soljson libsolc.cpp) + target_link_libraries(soljson PRIVATE solidity) +else() + add_library(libsolc libsolc.cpp) + set_target_properties(libsolc PROPERTIES OUTPUT_NAME solc) + target_link_libraries(libsolc PRIVATE solidity) +endif() diff --git a/solc/jsonCompiler.cpp b/libsolc/libsolc.cpp similarity index 99% rename from solc/jsonCompiler.cpp rename to libsolc/libsolc.cpp index 23feaa2ae49f..3a6e15215d3d 100644 --- a/solc/jsonCompiler.cpp +++ b/libsolc/libsolc.cpp @@ -20,7 +20,7 @@ * JSON interface for the solidity compiler to be used from Javascript. */ -#include +#include #include #include #include diff --git a/solc/jsonCompiler.h b/libsolc/libsolc.h similarity index 100% rename from solc/jsonCompiler.h rename to libsolc/libsolc.h diff --git a/libsolidity/analysis/ConstantEvaluator.cpp b/libsolidity/analysis/ConstantEvaluator.cpp index 4d546e68d015..83f37f47414d 100644 --- a/libsolidity/analysis/ConstantEvaluator.cpp +++ b/libsolidity/analysis/ConstantEvaluator.cpp @@ -28,51 +28,42 @@ using namespace std; using namespace dev; using namespace dev::solidity; -/// FIXME: this is pretty much a copy of TypeChecker::endVisit(BinaryOperation) void ConstantEvaluator::endVisit(UnaryOperation const& _operation) { - TypePointer const& subType = _operation.subExpression().annotation().type; - if (!dynamic_cast(subType.get())) - m_errorReporter.fatalTypeError(_operation.subExpression().location(), "Invalid constant expression."); - TypePointer t = subType->unaryOperatorResult(_operation.getOperator()); - _operation.annotation().type = t; + auto sub = type(_operation.subExpression()); + if (sub) + setType(_operation, sub->unaryOperatorResult(_operation.getOperator())); } -/// FIXME: this is pretty much a copy of TypeChecker::endVisit(BinaryOperation) void ConstantEvaluator::endVisit(BinaryOperation const& _operation) { - TypePointer const& leftType = _operation.leftExpression().annotation().type; - TypePointer const& rightType = _operation.rightExpression().annotation().type; - if (!dynamic_cast(leftType.get())) - m_errorReporter.fatalTypeError(_operation.leftExpression().location(), "Invalid constant expression."); - if (!dynamic_cast(rightType.get())) - m_errorReporter.fatalTypeError(_operation.rightExpression().location(), "Invalid constant expression."); - TypePointer commonType = leftType->binaryOperatorResult(_operation.getOperator(), rightType); - if (!commonType) + auto left = type(_operation.leftExpression()); + auto right = type(_operation.rightExpression()); + if (left && right) { - m_errorReporter.typeError( - _operation.location(), - "Operator " + - string(Token::toString(_operation.getOperator())) + - " not compatible with types " + - leftType->toString() + - " and " + - rightType->toString() + auto commonType = left->binaryOperatorResult(_operation.getOperator(), right); + if (!commonType) + m_errorReporter.fatalTypeError( + _operation.location(), + "Operator " + + string(Token::toString(_operation.getOperator())) + + " not compatible with types " + + left->toString() + + " and " + + right->toString() + ); + setType( + _operation, + Token::isCompareOp(_operation.getOperator()) ? + make_shared() : + commonType ); - commonType = leftType; } - _operation.annotation().commonType = commonType; - _operation.annotation().type = - Token::isCompareOp(_operation.getOperator()) ? - make_shared() : - commonType; } void ConstantEvaluator::endVisit(Literal const& _literal) { - _literal.annotation().type = Type::forLiteral(_literal); - if (!_literal.annotation().type) - m_errorReporter.fatalTypeError(_literal.location(), "Invalid literal value."); + setType(_literal, Type::forLiteral(_literal)); } void ConstantEvaluator::endVisit(Identifier const& _identifier) @@ -81,18 +72,34 @@ void ConstantEvaluator::endVisit(Identifier const& _identifier) if (!variableDeclaration) return; if (!variableDeclaration->isConstant()) - m_errorReporter.fatalTypeError(_identifier.location(), "Identifier must be declared constant."); + return; - ASTPointer value = variableDeclaration->value(); + ASTPointer const& value = variableDeclaration->value(); if (!value) - m_errorReporter.fatalTypeError(_identifier.location(), "Constant identifier declaration must have a constant value."); - - if (!value->annotation().type) + return; + else if (!m_types->count(value.get())) { if (m_depth > 32) m_errorReporter.fatalTypeError(_identifier.location(), "Cyclic constant definition (or maximum recursion depth exhausted)."); - ConstantEvaluator e(*value, m_errorReporter, m_depth + 1); + ConstantEvaluator(m_errorReporter, m_depth + 1, m_types).evaluate(*value); } - _identifier.annotation().type = value->annotation().type; + setType(_identifier, type(*value)); +} + +void ConstantEvaluator::setType(ASTNode const& _node, TypePointer const& _type) +{ + if (_type && _type->category() == Type::Category::RationalNumber) + (*m_types)[&_node] = _type; +} + +TypePointer ConstantEvaluator::type(ASTNode const& _node) +{ + return (*m_types)[&_node]; +} + +TypePointer ConstantEvaluator::evaluate(Expression const& _expr) +{ + _expr.accept(*this); + return type(_expr); } diff --git a/libsolidity/analysis/ConstantEvaluator.h b/libsolidity/analysis/ConstantEvaluator.h index 6725d610f972..77a357b65e28 100644 --- a/libsolidity/analysis/ConstantEvaluator.h +++ b/libsolidity/analysis/ConstantEvaluator.h @@ -38,22 +38,32 @@ class TypeChecker; class ConstantEvaluator: private ASTConstVisitor { public: - ConstantEvaluator(Expression const& _expr, ErrorReporter& _errorReporter, size_t _newDepth = 0): + ConstantEvaluator( + ErrorReporter& _errorReporter, + size_t _newDepth = 0, + std::shared_ptr> _types = std::make_shared>() + ): m_errorReporter(_errorReporter), - m_depth(_newDepth) + m_depth(_newDepth), + m_types(_types) { - _expr.accept(*this); } + TypePointer evaluate(Expression const& _expr); + private: virtual void endVisit(BinaryOperation const& _operation); virtual void endVisit(UnaryOperation const& _operation); virtual void endVisit(Literal const& _literal); virtual void endVisit(Identifier const& _identifier); + void setType(ASTNode const& _node, TypePointer const& _type); + TypePointer type(ASTNode const& _node); + ErrorReporter& m_errorReporter; /// Current recursion depth. - size_t m_depth; + size_t m_depth = 0; + std::shared_ptr> m_types; }; } diff --git a/libsolidity/analysis/DeclarationContainer.cpp b/libsolidity/analysis/DeclarationContainer.cpp index b33c856859c9..7508ad9e1ddd 100644 --- a/libsolidity/analysis/DeclarationContainer.cpp +++ b/libsolidity/analysis/DeclarationContainer.cpp @@ -23,6 +23,7 @@ #include #include #include +#include using namespace std; using namespace dev; @@ -105,7 +106,7 @@ bool DeclarationContainer::registerDeclaration( return true; } -std::vector DeclarationContainer::resolveName(ASTString const& _name, bool _recursive) const +vector DeclarationContainer::resolveName(ASTString const& _name, bool _recursive) const { solAssert(!_name.empty(), "Attempt to resolve empty name."); auto result = m_declarations.find(_name); @@ -115,3 +116,22 @@ std::vector DeclarationContainer::resolveName(ASTString cons return m_enclosingContainer->resolveName(_name, true); return vector({}); } + +vector DeclarationContainer::similarNames(ASTString const& _name) const +{ + static size_t const MAXIMUM_EDIT_DISTANCE = 2; + + vector similar; + + for (auto const& declaration: m_declarations) + { + string const& declarationName = declaration.first; + if (stringWithinDistance(_name, declarationName, MAXIMUM_EDIT_DISTANCE)) + similar.push_back(declarationName); + } + + if (m_enclosingContainer) + similar += m_enclosingContainer->similarNames(_name); + + return similar; +} diff --git a/libsolidity/analysis/DeclarationContainer.h b/libsolidity/analysis/DeclarationContainer.h index 301998b71333..f9b1bda4b003 100644 --- a/libsolidity/analysis/DeclarationContainer.h +++ b/libsolidity/analysis/DeclarationContainer.h @@ -58,6 +58,10 @@ class DeclarationContainer /// @returns whether declaration is valid, and if not also returns previous declaration. Declaration const* conflictingDeclaration(Declaration const& _declaration, ASTString const* _name = nullptr) const; + /// @returns existing declaration names similar to @a _name. + /// Searches this and all parent containers. + std::vector similarNames(ASTString const& _name) const; + private: ASTNode const* m_enclosingNode; DeclarationContainer const* m_enclosingContainer; diff --git a/libsolidity/analysis/NameAndTypeResolver.cpp b/libsolidity/analysis/NameAndTypeResolver.cpp index 5d0106935296..5e4d414be3b9 100644 --- a/libsolidity/analysis/NameAndTypeResolver.cpp +++ b/libsolidity/analysis/NameAndTypeResolver.cpp @@ -25,6 +25,7 @@ #include #include #include +#include #include @@ -425,6 +426,11 @@ vector<_T const*> NameAndTypeResolver::cThreeMerge(list>& _toMer return result; } +string NameAndTypeResolver::similarNameSuggestions(ASTString const& _name) const +{ + return quotedAlternativesList(m_currentScope->similarNames(_name)); +} + DeclarationRegistrationHelper::DeclarationRegistrationHelper( map>& _scopes, ASTNode& _astRoot, diff --git a/libsolidity/analysis/NameAndTypeResolver.h b/libsolidity/analysis/NameAndTypeResolver.h index d83697cdab6f..9aea07abf187 100644 --- a/libsolidity/analysis/NameAndTypeResolver.h +++ b/libsolidity/analysis/NameAndTypeResolver.h @@ -93,6 +93,9 @@ class NameAndTypeResolver: private boost::noncopyable /// Generate and store warnings about variables that are named like instructions. void warnVariablesNamedLikeInstructions(); + /// @returns a list of similar identifiers in the current and enclosing scopes. May return empty string if no suggestions. + std::string similarNameSuggestions(ASTString const& _name) const; + private: /// Internal version of @a resolveNamesAndTypes (called from there) throws exceptions on fatal errors. bool resolveNamesAndTypesInternal(ASTNode& _node, bool _resolveInsideCode = true); diff --git a/libsolidity/analysis/ReferencesResolver.cpp b/libsolidity/analysis/ReferencesResolver.cpp index f22c95cc1878..0bb5e3fef869 100644 --- a/libsolidity/analysis/ReferencesResolver.cpp +++ b/libsolidity/analysis/ReferencesResolver.cpp @@ -47,7 +47,13 @@ bool ReferencesResolver::visit(Identifier const& _identifier) { auto declarations = m_resolver.nameFromCurrentScope(_identifier.name()); if (declarations.empty()) - fatalDeclarationError(_identifier.location(), "Undeclared identifier."); + { + string suggestions = m_resolver.similarNameSuggestions(_identifier.name()); + string errorMessage = + "Undeclared identifier." + + (suggestions.empty()? "": " Did you mean " + std::move(suggestions) + "?"); + declarationError(_identifier.location(), errorMessage); + } else if (declarations.size() == 1) _identifier.annotation().referencedDeclaration = declarations.front(); else @@ -90,7 +96,10 @@ void ReferencesResolver::endVisit(UserDefinedTypeName const& _typeName) { Declaration const* declaration = m_resolver.pathFromCurrentScope(_typeName.namePath()); if (!declaration) - fatalDeclarationError(_typeName.location(), "Identifier not found or not unique."); + { + declarationError(_typeName.location(), "Identifier not found or not unique."); + return; + } _typeName.annotation().referencedDeclaration = declaration; @@ -101,7 +110,7 @@ void ReferencesResolver::endVisit(UserDefinedTypeName const& _typeName) else if (ContractDefinition const* contract = dynamic_cast(declaration)) _typeName.annotation().type = make_shared(*contract); else - fatalTypeError(_typeName.location(), "Name has to refer to a struct, enum or contract."); + typeError(_typeName.location(), "Name has to refer to a struct, enum or contract."); } void ReferencesResolver::endVisit(FunctionTypeName const& _typeName) @@ -112,17 +121,25 @@ void ReferencesResolver::endVisit(FunctionTypeName const& _typeName) case VariableDeclaration::Visibility::External: break; default: - fatalTypeError(_typeName.location(), "Invalid visibility, can only be \"external\" or \"internal\"."); + typeError(_typeName.location(), "Invalid visibility, can only be \"external\" or \"internal\"."); + return; } if (_typeName.isPayable() && _typeName.visibility() != VariableDeclaration::Visibility::External) - fatalTypeError(_typeName.location(), "Only external function types can be payable."); + { + typeError(_typeName.location(), "Only external function types can be payable."); + return; + } + if (_typeName.visibility() == VariableDeclaration::Visibility::External) for (auto const& t: _typeName.parameterTypes() + _typeName.returnParameterTypes()) { solAssert(t->annotation().type, "Type not set for parameter."); if (!t->annotation().type->canBeUsedExternally(false)) - fatalTypeError(t->location(), "Internal type cannot be used for external function type."); + { + typeError(t->location(), "Internal type cannot be used for external function type."); + return; + } } _typeName.annotation().type = make_shared(_typeName); @@ -142,15 +159,21 @@ void ReferencesResolver::endVisit(Mapping const& _typeName) void ReferencesResolver::endVisit(ArrayTypeName const& _typeName) { TypePointer baseType = _typeName.baseType().annotation().type; + if (!baseType) + { + solAssert(!m_errorReporter.errors().empty(), ""); + return; + } if (baseType->storageBytes() == 0) fatalTypeError(_typeName.baseType().location(), "Illegal base type of storage size zero for array."); if (Expression const* length = _typeName.length()) { - if (!length->annotation().type) - ConstantEvaluator e(*length, m_errorReporter); - auto const* lengthType = dynamic_cast(length->annotation().type.get()); + TypePointer lengthTypeGeneric = length->annotation().type; + if (!lengthTypeGeneric) + lengthTypeGeneric = ConstantEvaluator(m_errorReporter).evaluate(*length); + RationalNumberType const* lengthType = dynamic_cast(lengthTypeGeneric.get()); if (!lengthType || !lengthType->mobileType()) - fatalTypeError(length->location(), "Invalid array length, expected integer literal."); + fatalTypeError(length->location(), "Invalid array length, expected integer literal or constant expression."); else if (lengthType->isFractional()) fatalTypeError(length->location(), "Array with fractional length specified."); else if (lengthType->isNegative()) @@ -206,7 +229,7 @@ bool ReferencesResolver::visit(InlineAssembly const& _inlineAssembly) // Will be re-generated later with correct information assembly::AsmAnalysisInfo analysisInfo; - assembly::AsmAnalyzer(analysisInfo, errorsIgnored, false, resolver).analyze(_inlineAssembly.operations()); + assembly::AsmAnalyzer(analysisInfo, errorsIgnored, assembly::AsmFlavour::Loose, resolver).analyze(_inlineAssembly.operations()); return false; } @@ -321,17 +344,13 @@ void ReferencesResolver::endVisit(VariableDeclaration const& _variable) type = ref->copyForLocation(typeLoc, isPointer); } else if (varLoc != Location::Default && !ref) - fatalTypeError(_variable.location(), "Storage location can only be given for array or struct types."); - - if (!type) - fatalTypeError(_variable.location(), "Invalid type name."); + typeError(_variable.location(), "Storage location can only be given for array or struct types."); + _variable.annotation().type = type; } else if (!_variable.canHaveAutoType()) - fatalTypeError(_variable.location(), "Explicit type needed."); + typeError(_variable.location(), "Explicit type needed."); // otherwise we have a "var"-declaration whose type is resolved by the first assignment - - _variable.annotation().type = type; } void ReferencesResolver::typeError(SourceLocation const& _location, string const& _description) diff --git a/libsolidity/analysis/SyntaxChecker.cpp b/libsolidity/analysis/SyntaxChecker.cpp index b6cc04dac648..5a3745b0ad99 100644 --- a/libsolidity/analysis/SyntaxChecker.cpp +++ b/libsolidity/analysis/SyntaxChecker.cpp @@ -224,3 +224,17 @@ bool SyntaxChecker::visit(FunctionTypeName const& _node) return true; } + +bool SyntaxChecker::visit(VariableDeclaration const& _declaration) +{ + bool const v050 = m_sourceUnit->annotation().experimentalFeatures.count(ExperimentalFeature::V050); + + if (!_declaration.typeName()) + { + if (v050) + m_errorReporter.syntaxError(_declaration.location(), "Use of the \"var\" keyword is deprecated."); + else + m_errorReporter.warning(_declaration.location(), "Use of the \"var\" keyword is deprecated."); + } + return true; +} diff --git a/libsolidity/analysis/SyntaxChecker.h b/libsolidity/analysis/SyntaxChecker.h index d5d72f14366b..871bf0a98437 100644 --- a/libsolidity/analysis/SyntaxChecker.h +++ b/libsolidity/analysis/SyntaxChecker.h @@ -69,6 +69,8 @@ class SyntaxChecker: private ASTConstVisitor virtual bool visit(FunctionDefinition const& _function) override; virtual bool visit(FunctionTypeName const& _node) override; + virtual bool visit(VariableDeclaration const& _declaration) override; + ErrorReporter& m_errorReporter; /// Flag that indicates whether a function modifier actually contains '_'. diff --git a/libsolidity/analysis/TypeChecker.cpp b/libsolidity/analysis/TypeChecker.cpp index 73047e769454..d67142e40290 100644 --- a/libsolidity/analysis/TypeChecker.cpp +++ b/libsolidity/analysis/TypeChecker.cpp @@ -171,13 +171,7 @@ void TypeChecker::checkContractDuplicateFunctions(ContractDefinition const& _con ssl.append("Another declaration is here:", (*it)->location()); string msg = "More than one constructor defined."; - size_t occurrences = ssl.infos.size(); - if (occurrences > 32) - { - ssl.infos.resize(32); - msg += " Truncated from " + boost::lexical_cast(occurrences) + " to the first 32 occurrences."; - } - + ssl.limitSize(msg); m_errorReporter.declarationError( functions[_contract.name()].front()->location(), ssl, @@ -219,12 +213,7 @@ void TypeChecker::findDuplicateDefinitions(map> const& _defini if (ssl.infos.size() > 0) { - size_t occurrences = ssl.infos.size(); - if (occurrences > 32) - { - ssl.infos.resize(32); - _message += " Truncated from " + boost::lexical_cast(occurrences) + " to the first 32 occurrences."; - } + ssl.limitSize(_message); m_errorReporter.declarationError( overloads[i]->location(), @@ -570,6 +559,17 @@ bool TypeChecker::visit(FunctionDefinition const& _function) m_errorReporter.typeError(var->location(), "Type is required to live outside storage."); if (_function.visibility() >= FunctionDefinition::Visibility::Public && !(type(*var)->interfaceType(isLibraryFunction))) m_errorReporter.fatalTypeError(var->location(), "Internal or recursive type is not allowed for public or external functions."); + if ( + _function.visibility() > FunctionDefinition::Visibility::Internal && + type(*var)->category() == Type::Category::Struct && + !type(*var)->dataStoredIn(DataLocation::Storage) && + !_function.sourceUnit().annotation().experimentalFeatures.count(ExperimentalFeature::ABIEncoderV2) + ) + m_errorReporter.typeError( + var->location(), + "Structs are only supported in the new experimental ABI encoder. " + "Use \"pragma experimental ABIEncoderV2;\" to enable the feature." + ); var->accept(*this); } @@ -604,6 +604,8 @@ bool TypeChecker::visit(FunctionDefinition const& _function) { if (_function.visibility() < FunctionDefinition::Visibility::Public) m_errorReporter.typeError(_function.location(), "Functions in interfaces cannot be internal or private."); + else if (_function.visibility() != FunctionDefinition::Visibility::External) + m_errorReporter.warning(_function.location(), "Functions in interfaces should be declared external."); } if (_function.isConstructor()) m_errorReporter.typeError(_function.location(), "Constructor cannot be defined in interfaces."); @@ -873,7 +875,7 @@ bool TypeChecker::visit(InlineAssembly const& _inlineAssembly) assembly::AsmAnalyzer analyzer( *_inlineAssembly.annotation().analysisInfo, m_errorReporter, - false, + assembly::AsmFlavour::Loose, identifierAccess ); if (!analyzer.analyze(_inlineAssembly.operations())) @@ -1060,7 +1062,7 @@ bool TypeChecker::visit(VariableDeclarationStatement const& _statement) _statement.initialValue()->location(), "Invalid rational " + valueComponentType->toString() + - " (absolute value too large or divison by zero)." + " (absolute value too large or division by zero)." ); else solAssert(false, ""); @@ -1551,8 +1553,12 @@ bool TypeChecker::visit(FunctionCall const& _functionCall) if (!functionType->takesArbitraryParameters() && parameterTypes.size() != arguments.size()) { + bool isStructConstructorCall = _functionCall.annotation().kind == FunctionCallKind::StructConstructorCall; + string msg = - "Wrong argument count for function call: " + + "Wrong argument count for " + + string(isStructConstructorCall ? "struct constructor" : "function call") + + ": " + toString(arguments.size()) + " arguments given but expected " + toString(parameterTypes.size()) + @@ -1668,10 +1674,12 @@ void TypeChecker::endVisit(NewExpression const& _newExpression) SecondarySourceLocation ssl; for (auto function: contract->annotation().unimplementedFunctions) ssl.append("Missing implementation:", function->location()); + string msg = "Trying to create an instance of an abstract contract."; + ssl.limitSize(msg); m_errorReporter.typeError( _newExpression.location(), ssl, - "Trying to create an instance of an abstract contract." + msg ); } if (!contract->constructorIsPublic()) diff --git a/libsolidity/analysis/ViewPureChecker.cpp b/libsolidity/analysis/ViewPureChecker.cpp index 7e41fc16cce1..13c3ab68eee1 100644 --- a/libsolidity/analysis/ViewPureChecker.cpp +++ b/libsolidity/analysis/ViewPureChecker.cpp @@ -40,19 +40,20 @@ class AssemblyViewPureChecker: public boost::static_visitor void operator()(assembly::Label const&) { } void operator()(assembly::Instruction const& _instruction) { - if (eth::SemanticInformation::invalidInViewFunctions(_instruction.instruction)) - m_reportMutability(StateMutability::NonPayable, _instruction.location); - else if (eth::SemanticInformation::invalidInPureFunctions(_instruction.instruction)) - m_reportMutability(StateMutability::View, _instruction.location); + checkInstruction(_instruction.location, _instruction.instruction); } void operator()(assembly::Literal const&) {} void operator()(assembly::Identifier const&) {} void operator()(assembly::FunctionalInstruction const& _instr) { - (*this)(_instr.instruction); + checkInstruction(_instr.location, _instr.instruction); for (auto const& arg: _instr.arguments) boost::apply_visitor(*this, arg); } + void operator()(assembly::ExpressionStatement const& _expr) + { + boost::apply_visitor(*this, _expr.expression); + } void operator()(assembly::StackAssignment const&) {} void operator()(assembly::Assignment const& _assignment) { @@ -102,6 +103,13 @@ class AssemblyViewPureChecker: public boost::static_visitor private: std::function m_reportMutability; + void checkInstruction(SourceLocation _location, solidity::Instruction _instruction) + { + if (eth::SemanticInformation::invalidInViewFunctions(_instruction)) + m_reportMutability(StateMutability::NonPayable, _location); + else if (eth::SemanticInformation::invalidInPureFunctions(_instruction)) + m_reportMutability(StateMutability::View, _location); + } }; } @@ -267,6 +275,22 @@ void ViewPureChecker::endVisit(FunctionCall const& _functionCall) reportMutability(mut, _functionCall.location()); } +bool ViewPureChecker::visit(MemberAccess const& _memberAccess) +{ + // Catch the special case of `this.f.selector` which is a pure expression. + ASTString const& member = _memberAccess.memberName(); + if ( + _memberAccess.expression().annotation().type->category() == Type::Category::Function && + member == "selector" + ) + if (auto const* expr = dynamic_cast(&_memberAccess.expression())) + if (auto const* exprInt = dynamic_cast(&expr->expression())) + if (exprInt->name() == "this") + // Do not continue visiting. + return false; + return true; +} + void ViewPureChecker::endVisit(MemberAccess const& _memberAccess) { StateMutability mutability = StateMutability::Pure; diff --git a/libsolidity/analysis/ViewPureChecker.h b/libsolidity/analysis/ViewPureChecker.h index fec060b6a31a..0b882cd82fe7 100644 --- a/libsolidity/analysis/ViewPureChecker.h +++ b/libsolidity/analysis/ViewPureChecker.h @@ -56,6 +56,7 @@ class ViewPureChecker: private ASTConstVisitor virtual bool visit(ModifierDefinition const& _modifierDef) override; virtual void endVisit(ModifierDefinition const& _modifierDef) override; virtual void endVisit(Identifier const& _identifier) override; + virtual bool visit(MemberAccess const& _memberAccess) override; virtual void endVisit(MemberAccess const& _memberAccess) override; virtual void endVisit(IndexAccess const& _indexAccess) override; virtual void endVisit(ModifierInvocation const& _modifier) override; diff --git a/libsolidity/ast/Types.cpp b/libsolidity/ast/Types.cpp index 21daac2cd9fe..e4b7e4fd8ecf 100644 --- a/libsolidity/ast/Types.cpp +++ b/libsolidity/ast/Types.cpp @@ -949,11 +949,25 @@ bool RationalNumberType::operator==(Type const& _other) const return m_value == other.m_value; } +string RationalNumberType::bigintToReadableString(dev::bigint const& _num) +{ + string str = _num.str(); + if (str.size() > 32) + { + int omitted = str.size() - 8; + str = str.substr(0, 4) + "...(" + to_string(omitted) + " digits omitted)..." + str.substr(str.size() - 4, 4); + } + return str; +} + string RationalNumberType::toString(bool) const { if (!isFractional()) - return "int_const " + m_value.numerator().str(); - return "rational_const " + m_value.numerator().str() + '/' + m_value.denominator().str(); + return "int_const " + bigintToReadableString(m_value.numerator()); + + string numerator = bigintToReadableString(m_value.numerator()); + string denominator = bigintToReadableString(m_value.denominator()); + return "rational_const " + numerator + " / " + denominator; } u256 RationalNumberType::literalValue(Literal const*) const diff --git a/libsolidity/ast/Types.h b/libsolidity/ast/Types.h index 635279abe898..2e7d05ba9c8c 100644 --- a/libsolidity/ast/Types.h +++ b/libsolidity/ast/Types.h @@ -257,7 +257,7 @@ class Type: private boost::noncopyable, public std::enable_shared_from_this parseRational(std::string const& _value); + + /// @returns a truncated readable representation of the bigint keeping only + /// up to 4 leading and 4 trailing digits. + static std::string bigintToReadableString(dev::bigint const& num); }; /** diff --git a/libsolidity/codegen/ABIFunctions.cpp b/libsolidity/codegen/ABIFunctions.cpp index 6648be06631e..00f59065301a 100644 --- a/libsolidity/codegen/ABIFunctions.cpp +++ b/libsolidity/codegen/ABIFunctions.cpp @@ -120,7 +120,7 @@ string ABIFunctions::tupleDecoder(TypePointers const& _types, bool _fromMemory) Whiskers templ(R"( function (headStart, dataEnd) -> { - switch slt(sub(dataEnd, headStart), ) case 1 { revert(0, 0) } + if slt(sub(dataEnd, headStart), ) { revert(0, 0) } } )"); @@ -151,7 +151,7 @@ string ABIFunctions::tupleDecoder(TypePointers const& _types, bool _fromMemory) R"( { let offset := (add(headStart, )) - switch gt(offset, 0xffffffffffffffff) case 1 { revert(0, 0) } + if gt(offset, 0xffffffffffffffff) { revert(0, 0) } := (add(headStart, offset), dataEnd) } )" : @@ -1134,7 +1134,7 @@ string ABIFunctions::abiDecodingFunctionArray(ArrayType const& _type, bool _from R"( // function (offset, end) -> array { - switch slt(add(offset, 0x1f), end) case 0 { revert(0, 0) } + if iszero(slt(add(offset, 0x1f), end)) { revert(0, 0) } let length := array := ((length)) let dst := array @@ -1169,7 +1169,7 @@ string ABIFunctions::abiDecodingFunctionArray(ArrayType const& _type, bool _from else { string baseEncodedSize = toCompactHexWithPrefix(_type.baseType()->calldataEncodedSize()); - templ("staticBoundsCheck", "switch gt(add(src, mul(length, " + baseEncodedSize + ")), end) case 1 { revert(0, 0) }"); + templ("staticBoundsCheck", "if gt(add(src, mul(length, " + baseEncodedSize + ")), end) { revert(0, 0) }"); templ("retrieveElementPos", "src"); templ("baseEncodedSize", baseEncodedSize); } @@ -1197,11 +1197,11 @@ string ABIFunctions::abiDecodingFunctionCalldataArray(ArrayType const& _type) templ = R"( // function (offset, end) -> arrayPos, length { - switch slt(add(offset, 0x1f), end) case 0 { revert(0, 0) } + if iszero(slt(add(offset, 0x1f), end)) { revert(0, 0) } length := calldataload(offset) - switch gt(length, 0xffffffffffffffff) case 1 { revert(0, 0) } + if gt(length, 0xffffffffffffffff) { revert(0, 0) } arrayPos := add(offset, 0x20) - switch gt(add(arrayPos, mul(, )), end) case 1 { revert(0, 0) } + if gt(add(arrayPos, mul(, )), end) { revert(0, 0) } } )"; else @@ -1209,7 +1209,7 @@ string ABIFunctions::abiDecodingFunctionCalldataArray(ArrayType const& _type) // function (offset, end) -> arrayPos { arrayPos := offset - switch gt(add(arrayPos, mul(, )), end) case 1 { revert(0, 0) } + if gt(add(arrayPos, mul(, )), end) { revert(0, 0) } } )"; Whiskers w{templ}; @@ -1235,13 +1235,13 @@ string ABIFunctions::abiDecodingFunctionByteArray(ArrayType const& _type, bool _ Whiskers templ( R"( function (offset, end) -> array { - switch slt(add(offset, 0x1f), end) case 0 { revert(0, 0) } + if iszero(slt(add(offset, 0x1f), end)) { revert(0, 0) } let length := (offset) array := ((length)) mstore(array, length) let src := add(offset, 0x20) let dst := add(array, 0x20) - switch gt(add(src, length), end) case 1 { revert(0, 0) } + if gt(add(src, length), end) { revert(0, 0) } (src, dst, length) } )" @@ -1268,7 +1268,7 @@ string ABIFunctions::abiDecodingFunctionStruct(StructType const& _type, bool _fr Whiskers templ(R"( // function (headStart, end) -> value { - switch slt(sub(end, headStart), ) case 1 { revert(0, 0) } + if slt(sub(end, headStart), ) { revert(0, 0) } value := () <#members> { @@ -1296,7 +1296,7 @@ string ABIFunctions::abiDecodingFunctionStruct(StructType const& _type, bool _fr dynamic ? R"( let offset := (add(headStart, )) - switch gt(offset, 0xffffffffffffffff) case 1 { revert(0, 0) } + if gt(offset, 0xffffffffffffffff) { revert(0, 0) } mstore(add(value, ), (add(headStart, offset), end)) )" : R"( @@ -1501,7 +1501,7 @@ string ABIFunctions::arrayAllocationSizeFunction(ArrayType const& _type) Whiskers w(R"( function (length) -> size { // Make sure we can allocate memory without overflow - switch gt(length, 0xffffffffffffffff) case 1 { revert(0, 0) } + if gt(length, 0xffffffffffffffff) { revert(0, 0) } size := } @@ -1620,7 +1620,7 @@ string ABIFunctions::allocationFunction() memPtr := mload() let newFreePtr := add(memPtr, size) // protect against overflow - switch or(gt(newFreePtr, 0xffffffffffffffff), lt(newFreePtr, memPtr)) case 1 { revert(0, 0) } + if or(gt(newFreePtr, 0xffffffffffffffff), lt(newFreePtr, memPtr)) { revert(0, 0) } mstore(, newFreePtr) } )") diff --git a/libsolidity/codegen/Compiler.cpp b/libsolidity/codegen/Compiler.cpp index 44264a07a181..d3afada5549e 100644 --- a/libsolidity/codegen/Compiler.cpp +++ b/libsolidity/codegen/Compiler.cpp @@ -51,6 +51,7 @@ void Compiler::compileClone( map const& _contracts ) { + solAssert(!_contract.isLibrary(), ""); ContractCompiler runtimeCompiler(nullptr, m_runtimeContext, m_optimize); ContractCompiler cloneCompiler(&runtimeCompiler, m_context, m_optimize); m_runtimeSub = cloneCompiler.compileClone(_contract, _contracts); diff --git a/libsolidity/codegen/CompilerContext.cpp b/libsolidity/codegen/CompilerContext.cpp index ce9c3b7f69e4..7a88475a51e9 100644 --- a/libsolidity/codegen/CompilerContext.cpp +++ b/libsolidity/codegen/CompilerContext.cpp @@ -319,14 +319,19 @@ void CompilerContext::appendInlineAssembly( ErrorList errors; ErrorReporter errorReporter(errors); auto scanner = make_shared(CharStream(_assembly), "--CODEGEN--"); - auto parserResult = assembly::Parser(errorReporter).parse(scanner); + auto parserResult = assembly::Parser(errorReporter, assembly::AsmFlavour::Strict).parse(scanner); #ifdef SOL_OUTPUT_ASM cout << assembly::AsmPrinter()(*parserResult) << endl; #endif assembly::AsmAnalysisInfo analysisInfo; bool analyzerResult = false; if (parserResult) - analyzerResult = assembly::AsmAnalyzer(analysisInfo, errorReporter, false, identifierAccess.resolve).analyze(*parserResult); + analyzerResult = assembly::AsmAnalyzer( + analysisInfo, + errorReporter, + assembly::AsmFlavour::Strict, + identifierAccess.resolve + ).analyze(*parserResult); if (!parserResult || !errorReporter.errors().empty() || !analyzerResult) { string message = @@ -347,6 +352,9 @@ void CompilerContext::appendInlineAssembly( solAssert(errorReporter.errors().empty(), "Failed to analyze inline assembly block."); assembly::CodeGenerator::assemble(*parserResult, analysisInfo, *m_asm, identifierAccess, _system); + + // Reset the source location to the one of the node (instead of the CODEGEN source location) + updateSourceLocation(); } FunctionDefinition const& CompilerContext::resolveVirtualFunction( diff --git a/libsolidity/codegen/CompilerContext.h b/libsolidity/codegen/CompilerContext.h index 7743fd3f017b..a155a3a5194c 100644 --- a/libsolidity/codegen/CompilerContext.h +++ b/libsolidity/codegen/CompilerContext.h @@ -174,6 +174,9 @@ class CompilerContext eth::AssemblyItem appendData(bytes const& _data) { return m_asm->append(_data); } /// Appends the address (virtual, will be filled in by linker) of a library. void appendLibraryAddress(std::string const& _identifier) { m_asm->appendLibraryAddress(_identifier); } + /// Appends a zero-address that can be replaced by something else at deploy time (if the + /// position in bytecode is known). + void appendDeployTimeAddress() { m_asm->append(eth::PushDeployTimeAddress); } /// Resets the stack of visited nodes with a new stack having only @c _node void resetVisitedNodes(ASTNode const* _node); /// Pops the stack of visited nodes @@ -187,8 +190,8 @@ class CompilerContext CompilerContext& operator<<(u256 const& _value) { m_asm->append(_value); return *this; } CompilerContext& operator<<(bytes const& _data) { m_asm->append(_data); return *this; } - /// Appends inline assembly. @a _replacements are string-matching replacements that are performed - /// prior to parsing the inline assembly. + /// Appends inline assembly (strict mode). + /// @a _replacements are string-matching replacements that are performed prior to parsing the inline assembly. /// @param _localVariables assigns stack positions to variables with the last one being the stack top /// @param _system if true, this is a "system-level" assembly where all functions use named labels. void appendInlineAssembly( diff --git a/libsolidity/codegen/ContractCompiler.cpp b/libsolidity/codegen/ContractCompiler.cpp index a81ba518ce9f..f463db9478db 100644 --- a/libsolidity/codegen/ContractCompiler.cpp +++ b/libsolidity/codegen/ContractCompiler.cpp @@ -64,6 +64,12 @@ void ContractCompiler::compileContract( ) { CompilerContext::LocationSetter locationSetter(m_context, _contract); + + if (_contract.isLibrary()) + // Check whether this is a call (true) or a delegatecall (false). + // This has to be the first code in the contract. + appendDelegatecallCheck(); + initializeContext(_contract, _contracts); appendFunctionSelector(_contract); appendMissingFunctions(); @@ -75,8 +81,13 @@ size_t ContractCompiler::compileConstructor( ) { CompilerContext::LocationSetter locationSetter(m_context, _contract); - initializeContext(_contract, _contracts); - return packIntoContractCreator(_contract); + if (_contract.isLibrary()) + return deployLibrary(_contract); + else + { + initializeContext(_contract, _contracts); + return packIntoContractCreator(_contract); + } } size_t ContractCompiler::compileClone( @@ -122,6 +133,7 @@ void ContractCompiler::appendCallValueCheck() void ContractCompiler::appendInitAndConstructorCode(ContractDefinition const& _contract) { + solAssert(!_contract.isLibrary(), "Tried to initialize library."); CompilerContext::LocationSetter locationSetter(m_context, _contract); // Determine the arguments that are used for the base constructors. std::vector const& bases = _contract.annotation().linearizedBaseContracts; @@ -163,6 +175,7 @@ void ContractCompiler::appendInitAndConstructorCode(ContractDefinition const& _c size_t ContractCompiler::packIntoContractCreator(ContractDefinition const& _contract) { solAssert(!!m_runtimeCompiler, ""); + solAssert(!_contract.isLibrary(), "Tried to use contract creator or library."); appendInitAndConstructorCode(_contract); @@ -188,6 +201,34 @@ size_t ContractCompiler::packIntoContractCreator(ContractDefinition const& _cont return m_context.runtimeSub(); } +size_t ContractCompiler::deployLibrary(ContractDefinition const& _contract) +{ + solAssert(!!m_runtimeCompiler, ""); + solAssert(_contract.isLibrary(), "Tried to deploy contract as library."); + + CompilerContext::LocationSetter locationSetter(m_context, _contract); + + solAssert(m_context.runtimeSub() != size_t(-1), "Runtime sub not registered"); + m_context.pushSubroutineSize(m_context.runtimeSub()); + m_context.pushSubroutineOffset(m_context.runtimeSub()); + m_context.appendInlineAssembly(R"( + { + // If code starts at 11, an mstore(0) writes to the full PUSH20 plus data + // without the need for a shift. + let codepos := 11 + codecopy(codepos, subOffset, subSize) + // Check that the first opcode is a PUSH20 + switch eq(0x73, byte(0, mload(codepos))) + case 0 { invalid() } + mstore(0, address()) + mstore8(codepos, 0x73) + return(codepos, subSize) + } + )", {"subSize", "subOffset"}); + + return m_context.runtimeSub(); +} + void ContractCompiler::appendBaseConstructor(FunctionDefinition const& _constructor) { CompilerContext::LocationSetter locationSetter(m_context, _constructor); @@ -244,11 +285,26 @@ void ContractCompiler::appendConstructor(FunctionDefinition const& _constructor) _constructor.accept(*this); } +void ContractCompiler::appendDelegatecallCheck() +{ + // Special constant that will be replaced by the address at deploy time. + // At compilation time, this is just "PUSH20 00...000". + m_context.appendDeployTimeAddress(); + m_context << Instruction::ADDRESS << Instruction::EQ; + // The result on the stack is + // "We have not been called via DELEGATECALL". +} + void ContractCompiler::appendFunctionSelector(ContractDefinition const& _contract) { map, FunctionTypePointer> interfaceFunctions = _contract.interfaceFunctions(); map, const eth::AssemblyItem> callDataUnpackerEntryPoints; + if (_contract.isLibrary()) + { + solAssert(m_context.stackHeight() == 1, "CALL / DELEGATECALL flag expected."); + } + FunctionDefinition const* fallback = _contract.fallbackFunction(); eth::AssemblyItem notFound = m_context.newTag(); // directly jump to fallback if the data is too short to contain a function selector @@ -260,7 +316,7 @@ void ContractCompiler::appendFunctionSelector(ContractDefinition const& _contrac if (!interfaceFunctions.empty()) CompilerUtils(m_context).loadFromMemory(0, IntegerType(CompilerUtils::dataStartOffset * 8), true); - // stack now is: 1 0 + // stack now is: ? for (auto const& it: interfaceFunctions) { callDataUnpackerEntryPoints.insert(std::make_pair(it.first, m_context.newTag())); @@ -272,6 +328,7 @@ void ContractCompiler::appendFunctionSelector(ContractDefinition const& _contrac m_context << notFound; if (fallback) { + solAssert(!_contract.isLibrary(), ""); if (!fallback->isPayable()) appendCallValueCheck(); @@ -291,6 +348,13 @@ void ContractCompiler::appendFunctionSelector(ContractDefinition const& _contrac CompilerContext::LocationSetter locationSetter(m_context, functionType->declaration()); m_context << callDataUnpackerEntryPoints.at(it.first); + if (_contract.isLibrary() && functionType->stateMutability() > StateMutability::View) + { + // If the function is not a view function and is called without DELEGATECALL, + // we revert. + m_context << dupInstruction(2); + m_context.appendConditionalRevert(); + } m_context.setStackOffset(0); // We have to allow this for libraries, because value of the previous // call is still visible in the delegatecall. @@ -441,6 +505,7 @@ void ContractCompiler::registerStateVariables(ContractDefinition const& _contrac void ContractCompiler::initializeStateVariables(ContractDefinition const& _contract) { + solAssert(!_contract.isLibrary(), "Tried to initialize state variables of library."); for (VariableDeclaration const* variable: _contract.stateVariables()) if (variable->value() && !variable->isConstant()) ExpressionCompiler(m_context, m_optimise).appendStateVariableInitialization(*variable); diff --git a/libsolidity/codegen/ContractCompiler.h b/libsolidity/codegen/ContractCompiler.h index 7c5ee59f7a0c..1fd80d05f3cd 100644 --- a/libsolidity/codegen/ContractCompiler.h +++ b/libsolidity/codegen/ContractCompiler.h @@ -75,10 +75,19 @@ class ContractCompiler: private ASTConstVisitor /// with a new and initialized context. Adds the constructor code. /// @returns the identifier of the runtime sub assembly size_t packIntoContractCreator(ContractDefinition const& _contract); + /// Appends code that deploys the given contract as a library. + /// Will also add code that modifies the contract in memory by injecting the current address + /// for the call protector. + size_t deployLibrary(ContractDefinition const& _contract); /// Appends state variable initialisation and constructor code. void appendInitAndConstructorCode(ContractDefinition const& _contract); void appendBaseConstructor(FunctionDefinition const& _constructor); void appendConstructor(FunctionDefinition const& _constructor); + /// Appends code that returns a boolean flag on the stack that tells whether + /// the contract has been called via delegatecall (false) or regular call (true). + /// This is done by inserting a specific push constant as the first instruction + /// whose data will be modified in memory at deploy time. + void appendDelegatecallCheck(); void appendFunctionSelector(ContractDefinition const& _contract); void appendCallValueCheck(); /// Creates code that unpacks the arguments for the given function represented by a vector of TypePointers. diff --git a/libsolidity/codegen/ExpressionCompiler.cpp b/libsolidity/codegen/ExpressionCompiler.cpp index bb8c4a94039a..8e1cf0194813 100644 --- a/libsolidity/codegen/ExpressionCompiler.cpp +++ b/libsolidity/codegen/ExpressionCompiler.cpp @@ -1014,6 +1014,30 @@ bool ExpressionCompiler::visit(MemberAccess const& _memberAccess) _memberAccess.expression().accept(*this); return false; } + // Another special case for `this.f.selector` which does not need the address. + // There are other uses of `.selector` which do need the address, but we want this + // specific use to be a pure expression. + if ( + _memberAccess.expression().annotation().type->category() == Type::Category::Function && + member == "selector" + ) + if (auto const* expr = dynamic_cast(&_memberAccess.expression())) + if (auto const* exprInt = dynamic_cast(&expr->expression())) + if (exprInt->name() == "this") + if (Declaration const* declaration = expr->annotation().referencedDeclaration) + { + u256 identifier; + if (auto const* variable = dynamic_cast(declaration)) + identifier = FunctionType(*variable).externalIdentifier(); + else if (auto const* function = dynamic_cast(declaration)) + identifier = FunctionType(*function).externalIdentifier(); + else + solAssert(false, "Contract member is neither variable nor function."); + m_context << identifier; + /// need to store store it as bytes4 + utils().leftShiftNumberOnStack(224); + return false; + } _memberAccess.expression().accept(*this); switch (_memberAccess.expression().annotation().type->category()) diff --git a/libsolidity/formal/SMTChecker.cpp b/libsolidity/formal/SMTChecker.cpp index 1050621e24d6..a64024b368f3 100644 --- a/libsolidity/formal/SMTChecker.cpp +++ b/libsolidity/formal/SMTChecker.cpp @@ -55,7 +55,7 @@ void SMTChecker::analyze(SourceUnit const& _source) void SMTChecker::endVisit(VariableDeclaration const& _varDecl) { if (_varDecl.isLocalVariable() && _varDecl.type()->isValueType() &&_varDecl.value()) - assignment(_varDecl, *_varDecl.value()); + assignment(_varDecl, *_varDecl.value(), _varDecl.location()); } bool SMTChecker::visit(FunctionDefinition const& _function) @@ -71,6 +71,7 @@ bool SMTChecker::visit(FunctionDefinition const& _function) m_interface->reset(); m_currentSequenceCounter.clear(); m_nextFreeSequenceCounter.clear(); + m_pathConditions.clear(); m_conditionalExecutionHappened = false; initializeLocalVariables(_function); return true; @@ -90,15 +91,16 @@ bool SMTChecker::visit(IfStatement const& _node) checkBooleanNotConstant(_node.condition(), "Condition is always $VALUE."); - visitBranch(_node.trueStatement(), expr(_node.condition())); + auto countersEndFalse = m_currentSequenceCounter; + auto countersEndTrue = visitBranch(_node.trueStatement(), expr(_node.condition())); vector touchedVariables = m_variableUsage->touchedVariables(_node.trueStatement()); if (_node.falseStatement()) { - visitBranch(*_node.falseStatement(), !expr(_node.condition())); + countersEndFalse = visitBranch(*_node.falseStatement(), !expr(_node.condition())); touchedVariables += m_variableUsage->touchedVariables(*_node.falseStatement()); } - resetVariables(touchedVariables); + mergeVariables(touchedVariables, expr(_node.condition()), countersEndTrue, countersEndFalse); return false; } @@ -178,8 +180,7 @@ void SMTChecker::endVisit(VariableDeclarationStatement const& _varDecl) else if (knownVariable(*_varDecl.declarations()[0])) { if (_varDecl.initialValue()) - // TODO more checks? - assignment(*_varDecl.declarations()[0], *_varDecl.initialValue()); + assignment(*_varDecl.declarations()[0], *_varDecl.initialValue(), _varDecl.location()); } else m_errorReporter.warning( @@ -208,7 +209,10 @@ void SMTChecker::endVisit(Assignment const& _assignment) { Declaration const* decl = identifier->annotation().referencedDeclaration; if (knownVariable(*decl)) - assignment(*decl, _assignment.rightHandSide()); + { + assignment(*decl, _assignment.rightHandSide(), _assignment.location()); + defineExpr(_assignment, expr(_assignment.rightHandSide())); + } else m_errorReporter.warning( _assignment.location(), @@ -230,7 +234,81 @@ void SMTChecker::endVisit(TupleExpression const& _tuple) "Assertion checker does not yet implement tules and inline arrays." ); else - m_interface->addAssertion(expr(_tuple) == expr(*_tuple.components()[0])); + defineExpr(_tuple, expr(*_tuple.components()[0])); +} + +void SMTChecker::checkUnderOverflow(smt::Expression _value, IntegerType const& _type, SourceLocation const& _location) +{ + checkCondition( + _value < minValue(_type), + _location, + "Underflow (resulting value less than " + formatNumber(_type.minValue()) + ")", + "value", + &_value + ); + checkCondition( + _value > maxValue(_type), + _location, + "Overflow (resulting value larger than " + formatNumber(_type.maxValue()) + ")", + "value", + &_value + ); +} + +void SMTChecker::endVisit(UnaryOperation const& _op) +{ + switch (_op.getOperator()) + { + case Token::Not: // ! + { + solAssert(_op.annotation().type->category() == Type::Category::Bool, ""); + defineExpr(_op, !expr(_op.subExpression())); + break; + } + case Token::Inc: // ++ (pre- or postfix) + case Token::Dec: // -- (pre- or postfix) + { + solAssert(_op.annotation().type->category() == Type::Category::Integer, ""); + solAssert(_op.subExpression().annotation().lValueRequested, ""); + if (Identifier const* identifier = dynamic_cast(&_op.subExpression())) + { + Declaration const* decl = identifier->annotation().referencedDeclaration; + if (knownVariable(*decl)) + { + auto innerValue = currentValue(*decl); + auto newValue = _op.getOperator() == Token::Inc ? innerValue + 1 : innerValue - 1; + assignment(*decl, newValue, _op.location()); + defineExpr(_op, _op.isPrefixOperation() ? newValue : innerValue); + } + else + m_errorReporter.warning( + _op.location(), + "Assertion checker does not yet implement such assignments." + ); + } + else + m_errorReporter.warning( + _op.location(), + "Assertion checker does not yet implement such increments / decrements." + ); + break; + } + case Token::Add: // + + defineExpr(_op, expr(_op.subExpression())); + break; + case Token::Sub: // - + { + defineExpr(_op, 0 - expr(_op.subExpression())); + if (auto intType = dynamic_cast(_op.annotation().type.get())) + checkUnderOverflow(expr(_op), *intType, _op.location()); + break; + } + default: + m_errorReporter.warning( + _op.location(), + "Assertion checker does not yet implement this operator." + ); + } } void SMTChecker::endVisit(BinaryOperation const& _op) @@ -268,16 +346,14 @@ void SMTChecker::endVisit(FunctionCall const& _funCall) solAssert(args.size() == 1, ""); solAssert(args[0]->annotation().type->category() == Type::Category::Bool, ""); checkCondition(!(expr(*args[0])), _funCall.location(), "Assertion violation"); - m_interface->addAssertion(expr(*args[0])); + addPathImpliedExpression(expr(*args[0])); } else if (funType.kind() == FunctionType::Kind::Require) { solAssert(args.size() == 1, ""); solAssert(args[0]->annotation().type->category() == Type::Category::Bool, ""); - m_interface->addAssertion(expr(*args[0])); - checkCondition(!(expr(*args[0])), _funCall.location(), "Unreachable code"); - // TODO is there something meaningful we can check here? - // We can check whether the condition is always fulfilled or never fulfilled. + checkBooleanNotConstant(*args[0], "Condition is always $VALUE."); + addPathImpliedExpression(expr(*args[0])); } } @@ -290,7 +366,7 @@ void SMTChecker::endVisit(Identifier const& _identifier) // Will be translated as part of the node that requested the lvalue. } else if (dynamic_cast(_identifier.annotation().type.get())) - m_interface->addAssertion(expr(_identifier) == currentValue(*decl)); + defineExpr(_identifier, currentValue(*decl)); else if (FunctionType const* fun = dynamic_cast(_identifier.annotation().type.get())) { if (fun->kind() == FunctionType::Kind::Assert || fun->kind() == FunctionType::Kind::Require) @@ -306,10 +382,10 @@ void SMTChecker::endVisit(Literal const& _literal) if (RationalNumberType const* rational = dynamic_cast(&type)) solAssert(!rational->isFractional(), ""); - m_interface->addAssertion(expr(_literal) == smt::Expression(type.literalValue(&_literal))); + defineExpr(_literal, smt::Expression(type.literalValue(&_literal))); } else if (type.category() == Type::Category::Bool) - m_interface->addAssertion(expr(_literal) == smt::Expression(_literal.token() == Token::TrueLiteral ? true : false)); + defineExpr(_literal, smt::Expression(_literal.token() == Token::TrueLiteral ? true : false)); else m_errorReporter.warning( _literal.location(), @@ -326,36 +402,30 @@ void SMTChecker::arithmeticOperation(BinaryOperation const& _op) case Token::Add: case Token::Sub: case Token::Mul: + case Token::Div: { solAssert(_op.annotation().commonType, ""); solAssert(_op.annotation().commonType->category() == Type::Category::Integer, ""); + auto const& intType = dynamic_cast(*_op.annotation().commonType); smt::Expression left(expr(_op.leftExpression())); smt::Expression right(expr(_op.rightExpression())); Token::Value op = _op.getOperator(); smt::Expression value( op == Token::Add ? left + right : op == Token::Sub ? left - right : + op == Token::Div ? division(left, right, intType) : /*op == Token::Mul*/ left * right ); - // Overflow check - auto const& intType = dynamic_cast(*_op.annotation().commonType); - checkCondition( - value < minValue(intType), - _op.location(), - "Underflow (resulting value less than " + formatNumber(intType.minValue()) + ")", - "value", - &value - ); - checkCondition( - value > maxValue(intType), - _op.location(), - "Overflow (resulting value larger than " + formatNumber(intType.maxValue()) + ")", - "value", - &value - ); + if (_op.getOperator() == Token::Div) + { + checkCondition(right == 0, _op.location(), "Division by zero", "value", &right); + m_interface->addAssertion(right != 0); + } + + checkUnderOverflow(value, intType, _op.location()); - m_interface->addAssertion(expr(_op) == value); + defineExpr(_op, value); break; } default: @@ -383,7 +453,7 @@ void SMTChecker::compareOperation(BinaryOperation const& _op) /*op == Token::GreaterThanOrEqual*/ (left >= right) ); // TODO: check that other values for op are not possible. - m_interface->addAssertion(expr(_op) == value); + defineExpr(_op, value); } else m_errorReporter.warning( @@ -400,9 +470,9 @@ void SMTChecker::booleanOperation(BinaryOperation const& _op) { // @TODO check that both of them are not constant if (_op.getOperator() == Token::And) - m_interface->addAssertion(expr(_op) == expr(_op.leftExpression()) && expr(_op.rightExpression())); + defineExpr(_op, expr(_op.leftExpression()) && expr(_op.rightExpression())); else - m_interface->addAssertion(expr(_op) == expr(_op.leftExpression()) || expr(_op.rightExpression())); + defineExpr(_op, expr(_op.leftExpression()) || expr(_op.rightExpression())); } else m_errorReporter.warning( @@ -411,30 +481,50 @@ void SMTChecker::booleanOperation(BinaryOperation const& _op) ); } -void SMTChecker::assignment(Declaration const& _variable, Expression const& _value) +smt::Expression SMTChecker::division(smt::Expression _left, smt::Expression _right, IntegerType const& _type) { - // TODO more checks? - // TODO add restrictions about type (might be assignment from smaller type) - m_interface->addAssertion(newValue(_variable) == expr(_value)); + // Signed division in SMTLIB2 rounds differently for negative division. + if (_type.isSigned()) + return (smt::Expression::ite( + _left >= 0, + smt::Expression::ite(_right >= 0, _left / _right, 0 - (_left / (0 - _right))), + smt::Expression::ite(_right >= 0, 0 - ((0 - _left) / _right), (0 - _left) / (0 - _right)) + )); + else + return _left / _right; +} + +void SMTChecker::assignment(Declaration const& _variable, Expression const& _value, SourceLocation const& _location) +{ + assignment(_variable, expr(_value), _location); } -void SMTChecker::visitBranch(Statement const& _statement, smt::Expression _condition) +void SMTChecker::assignment(Declaration const& _variable, smt::Expression const& _value, SourceLocation const& _location) { - visitBranch(_statement, &_condition); + TypePointer type = _variable.type(); + if (auto const* intType = dynamic_cast(type.get())) + checkUnderOverflow(_value, *intType, _location); + m_interface->addAssertion(newValue(_variable) == _value); } -void SMTChecker::visitBranch(Statement const& _statement, smt::Expression const* _condition) +SMTChecker::VariableSequenceCounters SMTChecker::visitBranch(Statement const& _statement, smt::Expression _condition) +{ + return visitBranch(_statement, &_condition); +} + +SMTChecker::VariableSequenceCounters SMTChecker::visitBranch(Statement const& _statement, smt::Expression const* _condition) { VariableSequenceCounters sequenceCountersStart = m_currentSequenceCounter; - m_interface->push(); if (_condition) - m_interface->addAssertion(*_condition); + pushPathCondition(*_condition); _statement.accept(*this); - m_interface->pop(); + if (_condition) + popPathCondition(); m_conditionalExecutionHappened = true; - m_currentSequenceCounter = sequenceCountersStart; + std::swap(sequenceCountersStart, m_currentSequenceCounter); + return sequenceCountersStart; } void SMTChecker::checkCondition( @@ -446,7 +536,7 @@ void SMTChecker::checkCondition( ) { m_interface->push(); - m_interface->addAssertion(_condition); + addPathConjoinedExpression(_condition); vector expressionsToEvaluate; vector expressionNames; @@ -472,7 +562,7 @@ void SMTChecker::checkCondition( } smt::CheckResult result; vector values; - tie(result, values) = checkSatisifableAndGenerateModel(expressionsToEvaluate); + tie(result, values) = checkSatisfiableAndGenerateModel(expressionsToEvaluate); string conditionalComment; if (m_conditionalExecutionHappened) @@ -518,13 +608,13 @@ void SMTChecker::checkBooleanNotConstant(Expression const& _condition, string co return; m_interface->push(); - m_interface->addAssertion(expr(_condition)); - auto positiveResult = checkSatisifable(); + addPathConjoinedExpression(expr(_condition)); + auto positiveResult = checkSatisfiable(); m_interface->pop(); m_interface->push(); - m_interface->addAssertion(!expr(_condition)); - auto negatedResult = checkSatisifable(); + addPathConjoinedExpression(!expr(_condition)); + auto negatedResult = checkSatisfiable(); m_interface->pop(); if (positiveResult == smt::CheckResult::ERROR || negatedResult == smt::CheckResult::ERROR) @@ -554,7 +644,7 @@ void SMTChecker::checkBooleanNotConstant(Expression const& _condition, string co } pair> -SMTChecker::checkSatisifableAndGenerateModel(vector const& _expressionsToEvaluate) +SMTChecker::checkSatisfiableAndGenerateModel(vector const& _expressionsToEvaluate) { smt::CheckResult result; vector values; @@ -584,9 +674,9 @@ SMTChecker::checkSatisifableAndGenerateModel(vector const& _exp return make_pair(result, values); } -smt::CheckResult SMTChecker::checkSatisifable() +smt::CheckResult SMTChecker::checkSatisfiable() { - return checkSatisifableAndGenerateModel({}).first; + return checkSatisfiableAndGenerateModel({}).first; } void SMTChecker::initializeLocalVariables(FunctionDefinition const& _function) @@ -614,6 +704,22 @@ void SMTChecker::resetVariables(vector _variables) } } +void SMTChecker::mergeVariables(vector const& _variables, smt::Expression const& _condition, VariableSequenceCounters const& _countersEndTrue, VariableSequenceCounters const& _countersEndFalse) +{ + set uniqueVars(_variables.begin(), _variables.end()); + for (auto const* decl: uniqueVars) + { + int trueCounter = _countersEndTrue.at(decl); + int falseCounter = _countersEndFalse.at(decl); + solAssert(trueCounter != falseCounter, ""); + m_interface->addAssertion(newValue(*decl) == smt::Expression::ite( + _condition, + valueAtSequence(*decl, trueCounter), + valueAtSequence(*decl, falseCounter)) + ); + } +} + bool SMTChecker::createVariable(VariableDeclaration const& _varDecl) { if (dynamic_cast(_varDecl.type().get())) @@ -695,6 +801,18 @@ smt::Expression SMTChecker::maxValue(IntegerType const& _t) smt::Expression SMTChecker::expr(Expression const& _e) { if (!m_expressions.count(&_e)) + { + m_errorReporter.warning(_e.location(), "Internal error: Expression undefined for SMT solver." ); + createExpr(_e); + } + return m_expressions.at(&_e); +} + +void SMTChecker::createExpr(Expression const& _e) +{ + if (m_expressions.count(&_e)) + m_errorReporter.warning(_e.location(), "Internal error: Expression created twice in SMT solver." ); + else { solAssert(_e.annotation().type, ""); switch (_e.annotation().type->category()) @@ -716,7 +834,12 @@ smt::Expression SMTChecker::expr(Expression const& _e) solAssert(false, "Type not implemented."); } } - return m_expressions.at(&_e); +} + +void SMTChecker::defineExpr(Expression const& _e, smt::Expression _value) +{ + createExpr(_e); + m_interface->addAssertion(expr(_e) == _value); } smt::Expression SMTChecker::var(Declaration const& _decl) @@ -724,3 +847,31 @@ smt::Expression SMTChecker::var(Declaration const& _decl) solAssert(m_variables.count(&_decl), ""); return m_variables.at(&_decl); } + +void SMTChecker::popPathCondition() +{ + solAssert(m_pathConditions.size() > 0, "Cannot pop path condition, empty."); + m_pathConditions.pop_back(); +} + +void SMTChecker::pushPathCondition(smt::Expression const& _e) +{ + m_pathConditions.push_back(currentPathConditions() && _e); +} + +smt::Expression SMTChecker::currentPathConditions() +{ + if (m_pathConditions.size() == 0) + return smt::Expression(true); + return m_pathConditions.back(); +} + +void SMTChecker::addPathConjoinedExpression(smt::Expression const& _e) +{ + m_interface->addAssertion(currentPathConditions() && _e); +} + +void SMTChecker::addPathImpliedExpression(smt::Expression const& _e) +{ + m_interface->addAssertion(smt::Expression::implies(currentPathConditions(), _e)); +} diff --git a/libsolidity/formal/SMTChecker.h b/libsolidity/formal/SMTChecker.h index 8e07d74da172..b57f0f96baec 100644 --- a/libsolidity/formal/SMTChecker.h +++ b/libsolidity/formal/SMTChecker.h @@ -26,6 +26,7 @@ #include #include +#include namespace dev { @@ -57,6 +58,7 @@ class SMTChecker: private ASTConstVisitor virtual void endVisit(ExpressionStatement const& _node) override; virtual void endVisit(Assignment const& _node) override; virtual void endVisit(TupleExpression const& _node) override; + virtual void endVisit(UnaryOperation const& _node) override; virtual void endVisit(BinaryOperation const& _node) override; virtual void endVisit(FunctionCall const& _node) override; virtual void endVisit(Identifier const& _node) override; @@ -66,12 +68,21 @@ class SMTChecker: private ASTConstVisitor void compareOperation(BinaryOperation const& _op); void booleanOperation(BinaryOperation const& _op); - void assignment(Declaration const& _variable, Expression const& _value); + /// Division expression in the given type. Requires special treatment because + /// of rounding for signed division. + smt::Expression division(smt::Expression _left, smt::Expression _right, IntegerType const& _type); - // Visits the branch given by the statement, pushes and pops the SMT checker. - // @param _condition if present, asserts that this condition is true within the branch. - void visitBranch(Statement const& _statement, smt::Expression const* _condition = nullptr); - void visitBranch(Statement const& _statement, smt::Expression _condition); + void assignment(Declaration const& _variable, Expression const& _value, SourceLocation const& _location); + void assignment(Declaration const& _variable, smt::Expression const& _value, SourceLocation const& _location); + + /// Maps a variable to an SSA index. + using VariableSequenceCounters = std::map; + + /// Visits the branch given by the statement, pushes and pops the current path conditions. + /// @param _condition if present, asserts that this condition is true within the branch. + /// @returns the variable sequence counter after visiting the branch. + VariableSequenceCounters visitBranch(Statement const& _statement, smt::Expression const* _condition = nullptr); + VariableSequenceCounters visitBranch(Statement const& _statement, smt::Expression _condition); /// Check that a condition can be satisfied. void checkCondition( @@ -88,14 +99,21 @@ class SMTChecker: private ASTConstVisitor Expression const& _condition, std::string const& _description ); + /// Checks that the value is in the range given by the type. + void checkUnderOverflow(smt::Expression _value, IntegerType const& _Type, SourceLocation const& _location); + std::pair> - checkSatisifableAndGenerateModel(std::vector const& _expressionsToEvaluate); + checkSatisfiableAndGenerateModel(std::vector const& _expressionsToEvaluate); - smt::CheckResult checkSatisifable(); + smt::CheckResult checkSatisfiable(); void initializeLocalVariables(FunctionDefinition const& _function); void resetVariables(std::vector _variables); + /// Given two different branches and the touched variables, + /// merge the touched variables into after-branch ite variables + /// using the branch condition as guard. + void mergeVariables(std::vector const& _variables, smt::Expression const& _condition, VariableSequenceCounters const& _countersEndTrue, VariableSequenceCounters const& _countersEndFalse); /// Tries to create an uninitialized variable and returns true on success. /// This fails if the type is not supported. bool createVariable(VariableDeclaration const& _varDecl); @@ -124,15 +142,27 @@ class SMTChecker: private ASTConstVisitor static smt::Expression minValue(IntegerType const& _t); static smt::Expression maxValue(IntegerType const& _t); - using VariableSequenceCounters = std::map; - - /// Returns the expression corresponding to the AST node. Creates a new expression - /// if it does not exist yet. + /// Returns the expression corresponding to the AST node. Throws if the expression does not exist. smt::Expression expr(Expression const& _e); + /// Creates the expression (value can be arbitrary) + void createExpr(Expression const& _e); + /// Creates the expression and sets its value. + void defineExpr(Expression const& _e, smt::Expression _value); /// Returns the function declaration corresponding to the given variable. /// The function takes one argument which is the "sequence number". smt::Expression var(Declaration const& _decl); + /// Adds a new path condition + void pushPathCondition(smt::Expression const& _e); + /// Remove the last path condition + void popPathCondition(); + /// Returns the conjunction of all path conditions or True if empty + smt::Expression currentPathConditions(); + /// Conjoin the current path conditions with the given parameter and add to the solver + void addPathConjoinedExpression(smt::Expression const& _e); + /// Add to the solver: the given expression implied by the current path conditions + void addPathImpliedExpression(smt::Expression const& _e); + std::shared_ptr m_interface; std::shared_ptr m_variableUsage; bool m_conditionalExecutionHappened = false; @@ -140,6 +170,7 @@ class SMTChecker: private ASTConstVisitor std::map m_nextFreeSequenceCounter; std::map m_expressions; std::map m_variables; + std::vector m_pathConditions; ErrorReporter& m_errorReporter; FunctionDefinition const* m_currentFunction = nullptr; diff --git a/libsolidity/formal/SolverInterface.h b/libsolidity/formal/SolverInterface.h index c9adf863e458..884873100ea5 100644 --- a/libsolidity/formal/SolverInterface.h +++ b/libsolidity/formal/SolverInterface.h @@ -72,6 +72,11 @@ class Expression }, _trueValue.sort); } + static Expression implies(Expression _a, Expression _b) + { + return !std::move(_a) || std::move(_b); + } + friend Expression operator!(Expression _a) { return Expression("not", std::move(_a), Sort::Bool); @@ -120,6 +125,10 @@ class Expression { return Expression("*", std::move(_a), std::move(_b), Sort::Int); } + friend Expression operator/(Expression _a, Expression _b) + { + return Expression("/", std::move(_a), std::move(_b), Sort::Int); + } Expression operator()(Expression _a) const { solAssert( diff --git a/libsolidity/formal/Z3Interface.cpp b/libsolidity/formal/Z3Interface.cpp index e5c1aef4434f..769e6edb1f33 100644 --- a/libsolidity/formal/Z3Interface.cpp +++ b/libsolidity/formal/Z3Interface.cpp @@ -127,7 +127,8 @@ z3::expr Z3Interface::toZ3Expr(Expression const& _expr) {">=", 2}, {"+", 2}, {"-", 2}, - {"*", 2} + {"*", 2}, + {"/", 2} }; string const& n = _expr.name; if (m_functions.count(n)) @@ -173,6 +174,8 @@ z3::expr Z3Interface::toZ3Expr(Expression const& _expr) return arguments[0] - arguments[1]; else if (n == "*") return arguments[0] * arguments[1]; + else if (n == "/") + return arguments[0] / arguments[1]; // Cannot reach here. solAssert(false, ""); return arguments[0]; diff --git a/libsolidity/inlineasm/AsmAnalysis.cpp b/libsolidity/inlineasm/AsmAnalysis.cpp index 2804ddfcbaed..2d6e58deaae6 100644 --- a/libsolidity/inlineasm/AsmAnalysis.cpp +++ b/libsolidity/inlineasm/AsmAnalysis.cpp @@ -54,14 +54,15 @@ bool AsmAnalyzer::analyze(Block const& _block) bool AsmAnalyzer::operator()(Label const& _label) { - solAssert(!m_julia, ""); + solAssert(m_flavour == AsmFlavour::Loose, ""); m_info.stackHeightInfo[&_label] = m_stackHeight; + warnOnInstructions(solidity::Instruction::JUMPDEST, _label.location); return true; } bool AsmAnalyzer::operator()(assembly::Instruction const& _instruction) { - solAssert(!m_julia, ""); + solAssert(m_flavour == AsmFlavour::Loose, ""); auto const& info = instructionInfo(_instruction.instruction); m_stackHeight += info.ret - info.args; m_info.stackHeightInfo[&_instruction] = m_stackHeight; @@ -140,22 +141,34 @@ bool AsmAnalyzer::operator()(assembly::Identifier const& _identifier) bool AsmAnalyzer::operator()(FunctionalInstruction const& _instr) { - solAssert(!m_julia, ""); + solAssert(m_flavour != AsmFlavour::IULIA, ""); bool success = true; for (auto const& arg: _instr.arguments | boost::adaptors::reversed) if (!expectExpression(arg)) success = false; // Parser already checks that the number of arguments is correct. - solAssert(instructionInfo(_instr.instruction.instruction).args == int(_instr.arguments.size()), ""); - if (!(*this)(_instr.instruction)) - success = false; + auto const& info = instructionInfo(_instr.instruction); + solAssert(info.args == int(_instr.arguments.size()), ""); + m_stackHeight += info.ret - info.args; m_info.stackHeightInfo[&_instr] = m_stackHeight; + warnOnInstructions(_instr.instruction, _instr.location); + return success; +} + +bool AsmAnalyzer::operator()(assembly::ExpressionStatement const& _statement) +{ + size_t initialStackHeight = m_stackHeight; + bool success = boost::apply_visitor(*this, _statement.expression); + if (m_flavour != AsmFlavour::Loose) + if (!expectDeposit(0, initialStackHeight, _statement.location)) + success = false; + m_info.stackHeightInfo[&_statement] = m_stackHeight; return success; } bool AsmAnalyzer::operator()(assembly::StackAssignment const& _assignment) { - solAssert(!m_julia, ""); + solAssert(m_flavour == AsmFlavour::Loose, ""); bool success = checkAssignment(_assignment.variableName, size_t(-1)); m_info.stackHeightInfo[&_assignment] = m_stackHeight; return success; @@ -217,14 +230,14 @@ bool AsmAnalyzer::operator()(assembly::FunctionDefinition const& _funDef) Block const* virtualBlock = m_info.virtualBlocks.at(&_funDef).get(); solAssert(virtualBlock, ""); Scope& varScope = scope(virtualBlock); - for (auto const& var: _funDef.arguments + _funDef.returns) + for (auto const& var: _funDef.parameters + _funDef.returnVariables) { expectValidType(var.type, var.location); m_activeVariables.insert(&boost::get(varScope.identifiers.at(var.name))); } int const stackHeight = m_stackHeight; - m_stackHeight = _funDef.arguments.size() + _funDef.returns.size(); + m_stackHeight = _funDef.parameters.size() + _funDef.returnVariables.size(); bool success = (*this)(_funDef.body); @@ -405,13 +418,13 @@ bool AsmAnalyzer::operator()(Block const& _block) return success; } -bool AsmAnalyzer::expectExpression(Statement const& _statement) +bool AsmAnalyzer::expectExpression(Expression const& _expr) { bool success = true; int const initialHeight = m_stackHeight; - if (!boost::apply_visitor(*this, _statement)) + if (!boost::apply_visitor(*this, _expr)) success = false; - if (!expectDeposit(1, initialHeight, locationOf(_statement))) + if (!expectDeposit(1, initialHeight, locationOf(_expr))) success = false; return success; } @@ -495,7 +508,7 @@ Scope& AsmAnalyzer::scope(Block const* _block) } void AsmAnalyzer::expectValidType(string const& type, SourceLocation const& _location) { - if (!m_julia) + if (m_flavour != AsmFlavour::IULIA) return; if (!builtinTypes.count(type)) @@ -522,11 +535,11 @@ void AsmAnalyzer::warnOnInstructions(solidity::Instruction _instr, SourceLocatio "the Metropolis hard fork. Before that it acts as an invalid instruction." ); - if (_instr == solidity::Instruction::JUMP || _instr == solidity::Instruction::JUMPI) + if (_instr == solidity::Instruction::JUMP || _instr == solidity::Instruction::JUMPI || _instr == solidity::Instruction::JUMPDEST) m_errorReporter.warning( _location, - "Jump instructions are low-level EVM features that can lead to " + "Jump instructions and labels are low-level EVM features that can lead to " "incorrect stack access. Because of that they are discouraged. " - "Please consider using \"switch\" or \"for\" statements instead." + "Please consider using \"switch\", \"if\" or \"for\" statements instead." ); } diff --git a/libsolidity/inlineasm/AsmAnalysis.h b/libsolidity/inlineasm/AsmAnalysis.h index e484b876d0f6..7a81dbf8bd95 100644 --- a/libsolidity/inlineasm/AsmAnalysis.h +++ b/libsolidity/inlineasm/AsmAnalysis.h @@ -54,9 +54,9 @@ class AsmAnalyzer: public boost::static_visitor explicit AsmAnalyzer( AsmAnalysisInfo& _analysisInfo, ErrorReporter& _errorReporter, - bool _julia = false, + AsmFlavour _flavour = AsmFlavour::Loose, julia::ExternalIdentifierAccess::Resolver const& _resolver = julia::ExternalIdentifierAccess::Resolver() - ): m_resolver(_resolver), m_info(_analysisInfo), m_errorReporter(_errorReporter), m_julia(_julia) {} + ): m_resolver(_resolver), m_info(_analysisInfo), m_errorReporter(_errorReporter), m_flavour(_flavour) {} bool analyze(assembly::Block const& _block); @@ -65,6 +65,7 @@ class AsmAnalyzer: public boost::static_visitor bool operator()(assembly::Identifier const&); bool operator()(assembly::FunctionalInstruction const& _functionalInstruction); bool operator()(assembly::Label const& _label); + bool operator()(assembly::ExpressionStatement const&); bool operator()(assembly::StackAssignment const&); bool operator()(assembly::Assignment const& _assignment); bool operator()(assembly::VariableDeclaration const& _variableDeclaration); @@ -77,7 +78,7 @@ class AsmAnalyzer: public boost::static_visitor private: /// Visits the statement and expects it to deposit one item onto the stack. - bool expectExpression(Statement const& _statement); + bool expectExpression(Expression const& _expr); bool expectDeposit(int _deposit, int _oldHeight, SourceLocation const& _location); /// Verifies that a variable to be assigned to exists and has the same size @@ -96,7 +97,7 @@ class AsmAnalyzer: public boost::static_visitor std::set m_activeVariables; AsmAnalysisInfo& m_info; ErrorReporter& m_errorReporter; - bool m_julia = false; + AsmFlavour m_flavour = AsmFlavour::Loose; }; } diff --git a/libsolidity/inlineasm/AsmData.h b/libsolidity/inlineasm/AsmData.h index a792a1b8dbe0..2982d5e0254f 100644 --- a/libsolidity/inlineasm/AsmData.h +++ b/libsolidity/inlineasm/AsmData.h @@ -58,23 +58,25 @@ struct StackAssignment { SourceLocation location; Identifier variableName; }; /// Multiple assignment ("x, y := f()"), where the left hand side variables each occupy /// a single stack slot and expects a single expression on the right hand returning /// the same amount of items as the number of variables. -struct Assignment { SourceLocation location; std::vector variableNames; std::shared_ptr value; }; +struct Assignment { SourceLocation location; std::vector variableNames; std::shared_ptr value; }; /// Functional instruction, e.g. "mul(mload(20:u256), add(2:u256, x))" -struct FunctionalInstruction { SourceLocation location; Instruction instruction; std::vector arguments; }; -struct FunctionCall { SourceLocation location; Identifier functionName; std::vector arguments; }; +struct FunctionalInstruction { SourceLocation location; solidity::Instruction instruction; std::vector arguments; }; +struct FunctionCall { SourceLocation location; Identifier functionName; std::vector arguments; }; +/// Statement that contains only a single expression +struct ExpressionStatement { SourceLocation location; Expression expression; }; /// Block-scope variable declaration ("let x:u256 := mload(20:u256)"), non-hoisted -struct VariableDeclaration { SourceLocation location; TypedNameList variables; std::shared_ptr value; }; +struct VariableDeclaration { SourceLocation location; TypedNameList variables; std::shared_ptr value; }; /// Block that creates a scope (frees declared stack variables) struct Block { SourceLocation location; std::vector statements; }; /// Function definition ("function f(a, b) -> (d, e) { ... }") -struct FunctionDefinition { SourceLocation location; std::string name; TypedNameList arguments; TypedNameList returns; Block body; }; +struct FunctionDefinition { SourceLocation location; std::string name; TypedNameList parameters; TypedNameList returnVariables; Block body; }; /// Conditional execution without "else" part. -struct If { SourceLocation location; std::shared_ptr condition; Block body; }; +struct If { SourceLocation location; std::shared_ptr condition; Block body; }; /// Switch case or default case struct Case { SourceLocation location; std::shared_ptr value; Block body; }; /// Switch statement -struct Switch { SourceLocation location; std::shared_ptr expression; std::vector cases; }; -struct ForLoop { SourceLocation location; Block pre; std::shared_ptr condition; Block post; Block body; }; +struct Switch { SourceLocation location; std::shared_ptr expression; std::vector cases; }; +struct ForLoop { SourceLocation location; Block pre; std::shared_ptr condition; Block post; Block body; }; struct LocationExtractor: boost::static_visitor { diff --git a/libsolidity/inlineasm/AsmDataForward.h b/libsolidity/inlineasm/AsmDataForward.h index d627b41a36eb..3a9600fece1d 100644 --- a/libsolidity/inlineasm/AsmDataForward.h +++ b/libsolidity/inlineasm/AsmDataForward.h @@ -43,10 +43,22 @@ struct FunctionDefinition; struct FunctionCall; struct If; struct Switch; +struct Case; struct ForLoop; +struct ExpressionStatement; struct Block; -using Statement = boost::variant; +struct TypedName; + +using Expression = boost::variant; +using Statement = boost::variant; + +enum class AsmFlavour +{ + Loose, // no types, EVM instructions as function, jumps and direct stack manipulations + Strict, // no types, EVM instructions as functions, but no jumps and no direct stack manipulations + IULIA // same as Strict mode with types +}; } } diff --git a/libsolidity/inlineasm/AsmParser.cpp b/libsolidity/inlineasm/AsmParser.cpp index 8f171005fe60..306b07e64a9b 100644 --- a/libsolidity/inlineasm/AsmParser.cpp +++ b/libsolidity/inlineasm/AsmParser.cpp @@ -77,9 +77,7 @@ assembly::Statement Parser::parseStatement() { assembly::If _if = createWithLocation(); m_scanner->next(); - _if.condition = make_shared(parseExpression()); - if (_if.condition->type() == typeid(assembly::Instruction)) - fatalParserError("Instructions are not supported as conditions for if - try to append \"()\"."); + _if.condition = make_shared(parseExpression()); _if.body = parseBlock(); return _if; } @@ -87,9 +85,7 @@ assembly::Statement Parser::parseStatement() { assembly::Switch _switch = createWithLocation(); m_scanner->next(); - _switch.expression = make_shared(parseExpression()); - if (_switch.expression->type() == typeid(assembly::Instruction)) - fatalParserError("Instructions are not supported as expressions for switch - try to append \"()\"."); + _switch.expression = make_shared(parseExpression()); while (m_scanner->currentToken() == Token::Case) _switch.cases.emplace_back(parseCase()); if (m_scanner->currentToken() == Token::Default) @@ -107,14 +103,14 @@ assembly::Statement Parser::parseStatement() return parseForLoop(); case Token::Assign: { - if (m_julia) + if (m_flavour != AsmFlavour::Loose) break; assembly::StackAssignment assignment = createWithLocation(); advance(); expectToken(Token::Colon); assignment.variableName.location = location(); assignment.variableName.name = currentLiteral(); - if (!m_julia && instructions().count(assignment.variableName.name)) + if (instructions().count(assignment.variableName.name)) fatalParserError("Identifier expected, got instruction name."); assignment.location.end = endPosition(); expectToken(Token::Identifier); @@ -127,18 +123,21 @@ assembly::Statement Parser::parseStatement() // Simple instruction (might turn into functional), // literal, // identifier (might turn into label or functional assignment) - Statement statement(parseElementaryOperation(false)); + ElementaryOperation elementary(parseElementaryOperation()); switch (currentToken()) { case Token::LParen: - return parseCall(std::move(statement)); + { + Expression expr = parseCall(std::move(elementary)); + return ExpressionStatement{locationOf(expr), expr}; + } case Token::Comma: { // if a comma follows, a multiple assignment is assumed - if (statement.type() != typeid(assembly::Identifier)) + if (elementary.type() != typeid(assembly::Identifier)) fatalParserError("Label name / variable name must precede \",\" (multiple assignment)."); - assembly::Identifier const& identifier = boost::get(statement); + assembly::Identifier const& identifier = boost::get(elementary); Assignment assignment = createWithLocation(identifier.location); assignment.variableNames.emplace_back(identifier); @@ -146,43 +145,43 @@ assembly::Statement Parser::parseStatement() do { expectToken(Token::Comma); - statement = parseElementaryOperation(false); - if (statement.type() != typeid(assembly::Identifier)) + elementary = parseElementaryOperation(); + if (elementary.type() != typeid(assembly::Identifier)) fatalParserError("Variable name expected in multiple assignemnt."); - assignment.variableNames.emplace_back(boost::get(statement)); + assignment.variableNames.emplace_back(boost::get(elementary)); } while (currentToken() == Token::Comma); expectToken(Token::Colon); expectToken(Token::Assign); - assignment.value.reset(new Statement(parseExpression())); + assignment.value.reset(new Expression(parseExpression())); assignment.location.end = locationOf(*assignment.value).end; return assignment; } case Token::Colon: { - if (statement.type() != typeid(assembly::Identifier)) + if (elementary.type() != typeid(assembly::Identifier)) fatalParserError("Label name / variable name must precede \":\"."); - assembly::Identifier const& identifier = boost::get(statement); + assembly::Identifier const& identifier = boost::get(elementary); advance(); // identifier:=: should be parsed as identifier: =: (i.e. a label), // while identifier:= (being followed by a non-colon) as identifier := (assignment). if (currentToken() == Token::Assign && peekNextToken() != Token::Colon) { assembly::Assignment assignment = createWithLocation(identifier.location); - if (!m_julia && instructions().count(identifier.name)) + if (m_flavour != AsmFlavour::IULIA && instructions().count(identifier.name)) fatalParserError("Cannot use instruction names for identifier names."); advance(); assignment.variableNames.emplace_back(identifier); - assignment.value.reset(new Statement(parseExpression())); + assignment.value.reset(new Expression(parseExpression())); assignment.location.end = locationOf(*assignment.value).end; return assignment; } else { // label - if (m_julia) + if (m_flavour != AsmFlavour::Loose) fatalParserError("Labels are not supported."); Label label = createWithLocation