Rebuild for abseil_cpp20211102 #7

regro-cf-autotick-bot · 2022-02-15T15:45:30Z

This PR has been triggered in an effort to update abseil_cpp20211102.

Notes and instructions for merging this PR:

Please merge the PR only after the tests have passed.
Feel free to push to the bot's branch to update this PR if needed.

Please note that if you close this PR we presume that the feedstock has been rebuilt, so if you are going to perform the rebuild yourself don't close this PR until the your rebuild has been merged.

If this PR was opened in error or needs to be updated please add the bot-rerun label to this PR. The bot will close this PR and schedule another one. If you do not have permissions to add this label, you can use the phrase @conda-forge-admin, please rerun bot in a PR comment to have the conda-forge-admin add it for you.

_{This PR was created by the regro-cf-autotick-bot. The regro-cf-autotick-bot is a service to automatically track the dependency graph, migrate packages, and propose package version updates for conda-forge. Feel free to drop us a line if there are any issues! This PR was generated by https://github.com/regro/autotick-bot/actions/runs/1847610558, please use this URL for debugging.}

…nda-forge-pinning 2022.02.15.14.05.09

conda-forge-linter · 2022-02-15T15:45:34Z

Hi! This is the friendly automated conda-forge-linting service.

I just wanted to let you know that I linted all conda-recipes in your PR (recipe) and found it was in an excellent condition.

BastianZim · 2022-02-15T16:24:13Z

@conda-forge-admin, please rerender

BastianZim · 2022-02-15T16:26:25Z

recipe/meta.yaml

-  number: 0
-  skip: true  # [not linux]
+  number: 1
+  skip: true  # [win]


Ref: conda-forge/abseil-cpp-feedstock#24

github-actions · 2022-02-15T16:50:06Z

Hi! This is the friendly conda-forge automerge bot!

I considered the following status checks when analyzing this PR:

linter: passed
azure: failed

Thus the PR was not passing and not merged.

wolfv · 2022-02-17T09:49:59Z

Ah, cool, I just wanted to link to the recent conda-forge discussion but it was you who asked on gitter :)

BastianZim · 2022-02-17T09:51:36Z

😄 I'm just waiting on upstream and will otherwise just increment and test it.

BastianZim · 2022-02-17T19:29:15Z

@conda-forge-admin, please rerender

…nda-forge-pinning 2022.02.17.15.32.38

BastianZim · 2022-02-17T21:35:54Z

@wolfv Would you mind having a look again? It's now the correct SDK but it's hitting an Undefined symbols for architecture x86_64 error which I'm not sure how to solve.

BastianZim · 2022-02-17T21:36:20Z

Also, I'm going to bot-rerun this before merging, just to get a clean git history.

BastianZim · 2022-02-21T10:01:13Z

@wolfv Gentle ping here, if you have a sec. :)

BastianZim · 2022-03-08T13:54:32Z

@wolfv Quick ping here :)

wolfv · 2022-03-09T09:02:26Z

maybe @Mizux has an idea?

I think what changed is that we switched to C++ 17 on macOS? And another Abseil version?

The error is

[ 97%] Linking CXX shared library lib/libortools.dylib
ld: warning: -pie being ignored. It is only used when linking a main executable
Undefined symbols for architecture x86_64:
  "CoinMessageHandler::setLogLevel(int)", referenced from:
      operations_research::CLPInterface::Solve(operations_research::MPSolverParameters const&) in clp_interface.cc.o
  "CoinMessageHandler::setLogLevel(int, int)", referenced from:
      operations_research::CBCInterface::Solve(operations_research::MPSolverParameters const&) in cbc_interface.cc.o
      operations_research::CLPInterface::Solve(operations_research::MPSolverParameters const&) in clp_interface.cc.o
  "CoinMessageHandler::CoinMessageHandler()", referenced from:
      operations_research::CBCInterface::Solve(operations_research::MPSolverParameters const&) in cbc_interface.cc.o
      operations_research::CLPInterface::Solve(operations_research::MPSolverParameters const&) in clp_interface.cc.o
  "CoinMessageHandler::~CoinMessageHandler()", referenced from:
      operations_research::CBCInterface::Solve(operations_research::MPSolverParameters const&) in cbc_interface.cc.o
      operations_research::CLPInterface::Solve(operations_research::MPSolverParameters const&) in clp_interface.cc.o
  "CoinBuild::addRow(int, int const*, double const*, double, double)", referenced from:
      operations_research::CLPInterface::ExtractNewConstraints() in clp_interface.cc.o
  "CoinBuild::CoinBuild()", referenced from:
      operations_research::CLPInterface::ExtractNewConstraints() in clp_interface.cc.o
  "CoinBuild::~CoinBuild()", referenced from:
      operations_research::CLPInterface::ExtractNewConstraints() in clp_interface.cc.o
  "CoinModel::addRow(int, int const*, double const*, double, double, char const*)", referenced from:
      operations_research::CBCInterface::Solve(operations_research::MPSolverParameters const&) in cbc_interface.cc.o
  "CoinModel::addColumn(int, int const*, double const*, double, double, double, char const*, bool)", referenced from:
      operations_research::CBCInterface::Solve(operations_research::MPSolverParameters const&) in cbc_interface.cc.o
  "CoinModel::CoinModel()", referenced from:
      operations_research::CBCInterface::Solve(operations_research::MPSolverParameters const&) in cbc_interface.cc.o
  "CoinModel::~CoinModel()", referenced from:
      operations_research::CBCInterface::Solve(operations_research::MPSolverParameters const&) in cbc_interface.cc.o
ld: symbol(s) not found for architecture x86_64

Mizux · 2022-03-09T09:44:22Z

https://github.com/google/or-tools/blob/dd28bdf6e56f1d71dfc33a77657a99c69dc32efb/cmake/dependencies/CMakeLists.txt#L179 ?
this may explain the bad build.

@wolfv did you try an otool -l on ortools.dylib to check if we effectively depends on system install of CoinOR (Cbc, Clp, Cgl, Osi, CoinUtils) ?

BTW in or-tools we use my CoinOR forks which are CMake based build (when building coinor as static lib to ship a standalone ortools library)

devNote:

%grepc -RIin "CoinModel::addRow(int" *-src
coinutils-src/CoinUtils/src/CoinModel.cpp:717:void CoinModel::addRow(int numberInRow, const int *columns,

EDIT: Maybe found the error with my last commit in FindCoinUtils.cmake
https://github.com/google/or-tools/blame/1b1cbc4fa5de40be9c7bf8180d5e8a1ed1f6b9a0/cmake/FindCoinUtils.cmake#L40-L42
s/UTILS/Utils/g -_-

h-vetinari · 2022-03-10T09:31:16Z

The new abseil builds in conda-forge use C++17 on unix. Unfortunately, abseil is ABI-sensitive to the C++ standard version used to compile, and each consuming library needs to match that. See conda-forge/abseil-cpp-feedstock#29

BastianZim · 2022-03-10T09:35:16Z

Yeah this comes from @Mizux comment and the resulting discussion in: conda-forge/staged-recipes#16147 (comment)

h-vetinari · 2022-03-10T09:49:32Z

Have you tried setting -DCMAKE_CXX_STANDARD=17 in build-cpp.sh?

BastianZim · 2022-03-10T10:33:44Z

Have you tried setting -DCMAKE_CXX_STANDARD=17 in build-cpp.sh?

No, but I am slightly hesitant to hard-code these things. Let me try though and see.

h-vetinari · 2022-03-10T10:45:12Z

No, but I am slightly hesitant to hard-code these things.

Global C++ standard flags change very rarely. Or at the very least, the next jump to C++20 will probably still take a loooooong time, not least because clang is still far behind conformance (much less through the CWG/LWG issues) even on main (and all the compiler diffusion timelines that implies). GCC also hasn't buttoned up modules yet.

And in this case, you should IMO be explicit rather than relying on CMake/compiler defaults, because the abseil-ABI forces you to match.

…nda-forge-pinning 2022.03.10.08.54.06

Mizux · 2022-03-10T11:59:35Z

FYI: For or-tools we are moving to c++20 because we have few designated initializers in struct so compilation should works with a gcc 8.
note: for make based build we require gcc 10+ because we hard code -std=c++20 while it is call c++2a on gcc <= 9 (CMake know it and do it nicely)
ref: https://en.cppreference.com/w/cpp/language/aggregate_initialization

IIRC gcc/g++ with -std=c++17 still manage to compile or-tools (with this C++20 feature) while Visual Studio is more pedantic and fail...

h-vetinari · 2022-03-10T12:10:04Z

OK, evidently this didn't work. 😑

The symbol collisions still seem to come from else-where (coin-or-clp & coin-or-cbc); maybe we'd need to compile those in C++17 too...? 🤔

BastianZim · 2022-03-10T12:32:04Z

The symbol collisions still seem to come from else-where (coin-or-clp & coin-or-cbc); maybe we'd need to compile those in C++17 too...? 🤔

That could make sense. Are you working on them anyways? Otherwise, I'd just rerender them and add that (Not a member there though)

Edit: Just remembered that they're not CMake based so well need to adapt that.

BastianZim · 2022-03-10T12:33:44Z

IIRC gcc/g++ with -std=c++17 still manage to compile or-tools (with this C++20 feature) while Visual Studio is more pedantic and fail...

Well, it's working for Linux but I'm not sure if something is hardcoded somewhere in the conda-forge infrastructure. 😓

h-vetinari · 2022-03-10T21:24:30Z

but I'm not sure if something is hardcoded somewhere in the conda-forge infrastructure.

No it's not; there was some discussion in conda-forge/clang-compiler-activation-feedstock#17, but it's left to the feedstocks. Abseil is a bit special because the ABI changes based on the standard version, which is not usually what happens.

BastianZim · 2022-03-11T17:40:39Z

Ahh ok, thanks for the pointers. From what I can see you're doing some stuff in the other feedstocks. I'll wait then for that ok?

h-vetinari · 2022-03-13T23:31:03Z

FYI: For or-tools we are moving to c++20 because we have few designated initializers

Just FYI in turn: using abseil with C++20 mode is subject to as-yet undetermined breakage: abseil/abseil-cpp#1127

Mizux · 2022-03-14T08:24:36Z

FYI: For or-tools we are moving to c++20 because we have few designated initializers

Just FYI in turn: using abseil with C++20 mode is subject to as-yet undetermined breakage: abseil/abseil-cpp#1127

Well in turn out that bazel is not compatible is any project that don't use the same C++ dialect for all of its sub-components (aka a project in C++14 and an other in C++17 may also break when trying to merge both in your super project build...), for Homebrew I manage to move "every packages" to C++17, it seems that or-tools is compilable with c++17 when using gcc (11?) or clang (13?) since they are not picky when using "designated initializers" so would suggest to replace lines
https://github.com/google/or-tools/blob/525162feaadaeef640783b2eaea38cf4b623877f/CMakeLists.txt#L15 and
https://github.com/google/or-tools/blob/525162feaadaeef640783b2eaea38cf4b623877f/cmake/dependencies/CMakeLists.txt#L307 to C++17 and see if it works with your build system (i.e.compiler) stack....

BastianZim · 2022-03-14T13:46:28Z

Ok, thanks for investigating @Mizux. Just to be clear: Do you just want the two lines changed or something else as well?

BastianZim · 2022-03-17T15:57:48Z

As we now have the new release, I'm going to re-run this so that we can get the new release out first and then try to enable OSX later.

regro-cf-autotick-bot · 2022-03-17T16:22:44Z

Due to the bot-rerun label I'm closing this PR. I will make another one as appropriate. This was generated by https://github.com/regro/autotick-bot/actions/runs/1999485720

Mizux · 2022-03-17T17:11:01Z

Ok, thanks for investigating @Mizux. Just to be clear: Do you just want the two lines changed or something else as well?

Well, currently we revert to C++17 the CMAKE_CXX_STANDARD otherwise it would imply to also bump abseil-cpp and all related projects (aka big bang). According to our CI it seems to pass even if there are some c++20 features in use 🤷

note: only VS2022 spot it and complained so if CMake detect MSVC generator it will set CXX_STANDARD to 20 otherwise it will stick to 17 like before.
TLDR: nothing to done on your side, should works as in v9.2...

regro-cf-autotick-bot added 2 commits February 15, 2022 15:44

Rebuild for abseil_cpp20211102

6c02cf5

MNT: Re-rendered with conda-build 3.21.8, conda-smithy 3.16.2, and co…

db0056c

…nda-forge-pinning 2022.02.15.14.05.09

regro-cf-autotick-bot requested review from BastianZim and wolfv as code owners February 15, 2022 15:45

BastianZim reviewed Feb 15, 2022

View reviewed changes

recipe/meta.yaml

number: 0

skip: true # [not linux]

number: 1

skip: true # [win]

Copy link

Member

BastianZim Feb 15, 2022

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Ref: conda-forge/abseil-cpp-feedstock#24

BastianZim added the automerge Merge the PR when CI passes label Feb 15, 2022

BastianZim removed the automerge Merge the PR when CI passes label Feb 15, 2022

This comment was marked as outdated.

Sign in to view

Remove OSX skip

6ce5eaa

BastianZim force-pushed the rebuild-abseil_cpp20211102-0-1_h0e16de branch from 91ff606 to 6ce5eaa Compare February 17, 2022 19:50

BastianZim added 8 commits February 17, 2022 20:54

OSX min 10.12

9e2fb27

MNT: Re-rendered with conda-build 3.21.8, conda-smithy 3.16.2, and co…

079dac4

…nda-forge-pinning 2022.02.17.15.32.38

Increase to 13

8b4590e

MNT: Re-rendered with conda-build 3.21.8, conda-smithy 3.16.2, and co…

ce4c172

…nda-forge-pinning 2022.02.17.15.32.38

Increase 14

90ca541

MNT: Re-rendered with conda-build 3.21.8, conda-smithy 3.16.2, and co…

b7c4e1d

…nda-forge-pinning 2022.02.17.15.32.38

Increase 15

8547e13

MNT: Re-rendered with conda-build 3.21.8, conda-smithy 3.16.2, and co…

aaaf26d

…nda-forge-pinning 2022.02.17.15.32.38

BastianZim mentioned this pull request Feb 21, 2022

Additional solvers #2

Open

BastianZim mentioned this pull request Mar 10, 2022

ENH: support for osx-{64,arm64}, win-64 #9

Open

BastianZim added 2 commits March 10, 2022 12:14

Add DCMAKE_CXX_STANDARD

35fb097

MNT: Re-rendered with conda-build 3.21.8, conda-smithy 3.18.0, and co…

5eb17b0

…nda-forge-pinning 2022.03.10.08.54.06

BastianZim added the bot-rerun Instruct the bot to retry the PR label Mar 17, 2022

regro-cf-autotick-bot closed this Mar 17, 2022

regro-cf-autotick-bot deleted the rebuild-abseil_cpp20211102-0-1_h0e16de branch March 17, 2022 16:22

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Rebuild for abseil_cpp20211102 #7

Rebuild for abseil_cpp20211102 #7

regro-cf-autotick-bot commented Feb 15, 2022

conda-forge-linter commented Feb 15, 2022

BastianZim commented Feb 15, 2022

BastianZim Feb 15, 2022

github-actions bot commented Feb 15, 2022

This comment was marked as outdated.

wolfv commented Feb 17, 2022

BastianZim commented Feb 17, 2022

BastianZim commented Feb 17, 2022

BastianZim commented Feb 17, 2022

BastianZim commented Feb 17, 2022

BastianZim commented Feb 21, 2022

BastianZim commented Mar 8, 2022

wolfv commented Mar 9, 2022

Mizux commented Mar 9, 2022 •

edited

Loading

h-vetinari commented Mar 10, 2022

BastianZim commented Mar 10, 2022 •

edited

Loading

h-vetinari commented Mar 10, 2022

BastianZim commented Mar 10, 2022

h-vetinari commented Mar 10, 2022 •

edited

Loading

Mizux commented Mar 10, 2022 •

edited

Loading

h-vetinari commented Mar 10, 2022

BastianZim commented Mar 10, 2022 •

edited

Loading

BastianZim commented Mar 10, 2022

h-vetinari commented Mar 10, 2022

BastianZim commented Mar 11, 2022

h-vetinari commented Mar 13, 2022

Mizux commented Mar 14, 2022 •

edited

Loading

BastianZim commented Mar 14, 2022

BastianZim commented Mar 17, 2022

regro-cf-autotick-bot commented Mar 17, 2022

Mizux commented Mar 17, 2022 •

edited

Loading

Rebuild for abseil_cpp20211102 #7

Rebuild for abseil_cpp20211102 #7

Conversation

regro-cf-autotick-bot commented Feb 15, 2022

conda-forge-linter commented Feb 15, 2022

BastianZim commented Feb 15, 2022

BastianZim Feb 15, 2022

Choose a reason for hiding this comment

github-actions bot commented Feb 15, 2022

This comment was marked as outdated.

wolfv commented Feb 17, 2022

BastianZim commented Feb 17, 2022

BastianZim commented Feb 17, 2022

BastianZim commented Feb 17, 2022

BastianZim commented Feb 17, 2022

BastianZim commented Feb 21, 2022

BastianZim commented Mar 8, 2022

wolfv commented Mar 9, 2022

Mizux commented Mar 9, 2022 • edited Loading

h-vetinari commented Mar 10, 2022

BastianZim commented Mar 10, 2022 • edited Loading

h-vetinari commented Mar 10, 2022

BastianZim commented Mar 10, 2022

h-vetinari commented Mar 10, 2022 • edited Loading

Mizux commented Mar 10, 2022 • edited Loading

h-vetinari commented Mar 10, 2022

BastianZim commented Mar 10, 2022 • edited Loading

BastianZim commented Mar 10, 2022

h-vetinari commented Mar 10, 2022

BastianZim commented Mar 11, 2022

h-vetinari commented Mar 13, 2022

Mizux commented Mar 14, 2022 • edited Loading

BastianZim commented Mar 14, 2022

BastianZim commented Mar 17, 2022

regro-cf-autotick-bot commented Mar 17, 2022

Mizux commented Mar 17, 2022 • edited Loading

Mizux commented Mar 9, 2022 •

edited

Loading

BastianZim commented Mar 10, 2022 •

edited

Loading

h-vetinari commented Mar 10, 2022 •

edited

Loading

Mizux commented Mar 10, 2022 •

edited

Loading

BastianZim commented Mar 10, 2022 •

edited

Loading

Mizux commented Mar 14, 2022 •

edited

Loading

Mizux commented Mar 17, 2022 •

edited

Loading