Taking on every NYI we hit while parsing a large code base #71
Conversation
…ended the AST to typesymbol mapper with the mapping id of the declared alias template
|
This needs some testing. @linuswagner could you give it as spin? and perhaps supply a small code example that exercises this new feature? |
… bridge between Rascal and CDT
…non type-value arguments as template parameters
src/lang/cpp/AST.rsc
Outdated
| @@ -554,3 +554,59 @@ java rel[loc,loc] parseForMacros(loc file, str charset=DEFAULT_CHARSET, bool inf | |||
| @synopsis{All functions in this module that have a charset parameter use this as default.} | |||
| public str DEFAULT_CHARSET = "UTF-8"; | |||
|
|
|||
| public map[str, str] linusMacros = ( | |||
There was a problem hiding this comment.
strange.. I must have git commit -amed again.. it's in my muscle memory.
There was a problem hiding this comment.
What's the status now? are we ready to merge?
src/lang/cpp/TypeSymbol.rsc
Outdated
|
|
||
| | \functionSetType(loc decl, list[TypeSymbol] templateArguments) | ||
| | \functionSetTypePointer(loc decl, list[TypeSymbol] templateArguments) | ||
|
|
||
| | \unaryTypeTransformation(str operator, TypeSymbol operand) |
There was a problem hiding this comment.
I'm not a big fan of encoding the operator as a string parameter. Can we specialize this per operator? Looking at the code, there's (currently) only one value in the operator enum anyway
There was a problem hiding this comment.
I don't like that either. I'll have a look. If it's a small finite set we can have one transformation constructor per operator. Otherwise, we introduce an TypeTransformationOperator sort.
| @@ -376,7 +376,7 @@ private ISourceLocation resolveIEnumeration(IEnumeration binding, ISourceLocatio | |||
| throw new RuntimeException("NYI" + binding.getClass().getSimpleName() + ": " + binding + " @ " + origin); | |||
| } | |||
|
|
|||
| private ISourceLocation resolveICPPBinding(ICPPBinding binding, ISourceLocation origin) throws URISyntaxException { | |||
There was a problem hiding this comment.
Was this change intentional?
There was a problem hiding this comment.
Yes this is dead code.
| @@ -840,6 +840,16 @@ public ISourceLocation resolveBinding(IASTNameOwner node, ISourceLocation origin | |||
| return failedBinding("unknown"); | |||
| } | |||
|
|
|||
| public ISourceLocation resolveBinding(ICPPBinding binding, ISourceLocation origin) { | |||
There was a problem hiding this comment.
Is there a particular reason that the preexisting interface method did not suffice?
There was a problem hiding this comment.
Yes, due to multiple inheritance of interfaces, the code for reaching ICPPBinding was not always reachable, in particular when the object did not accidentally also implement IASTNameOwner.
| @@ -1459,29 +1457,6 @@ private int visit(ICPPASTDesignatedInitializer initializer) { | |||
| return PROCESS_ABORT; | |||
| } | |||
|
|
|||
| // InitializerClauses | |||
There was a problem hiding this comment.
Was this unreachable code?
There was a problem hiding this comment.
yes. the compiler warned about that
| @@ -1557,12 +1532,6 @@ private int visit(IASTArrayDeclarator declarator) { | |||
| return PROCESS_ABORT; | |||
| } | |||
|
|
|||
| private int visit(IASTFieldDeclarator declarator) { | |||
There was a problem hiding this comment.
Was this unreachable code?
So true. We are under a lot of time pressure here so we decided to take a sprint and solve every remaining issue as it hits us. The code base we are trying to parse and map is incredibly large and decades old, so it seems we are bound for hitting every |
jurgenvinju
changed the title
|
Just added GNU's computed goto's. Will try and simplify the TypeTransformationOperator design now. |
This is to fix this stracktrace: