llvm dev - Apr 2011 - [LLVMdev] A problem from XCode 4 - help sought tracking it down

Sorry if this isn't the right place; but I'm facing an XCode 4 (LLVM 2.0
to Apple, I heard it's LLVM 2.9 under the hood) issue that may involve
LLVM, and maybe some people on this list could help me short-circuit the
source.

I have a code pattern that, when used in XCode 4 in Objective C++ files,
causes the whole IDE to go crash.  It is a stripped sample to highlight
the issue.

In the call stack, I see libclang.dylib, calling through
clang_visitChildren, up through various CursorVisitor::Visit... and
CursorVisitor::VisitChildren... calls.  I have LLVM compiling, and
imagine I could replace XCode's libclang.dylib in order to add debugging
aids/logging.

Am I on the right track?  Should I look near the point of failure, or
should I inspect the integrity of the data sent to clang_visitChildren
first?

If I'm not on the right track, I appreciate any help in getting me moving.


If anyone happens to be able to test-rig XCode 4 (in iPhone mode with
LLVM2 as compiler) vs the code below, it reproduces the problem:

template <typename I0,typename I1>
void (^ cc(void (^f)(I0,I1), I0 i0))(I1)
{
    ^ void (I1 i1) { };
}


int main(int argc, char *argv[]) {
    NSAutoreleasePool * pool = [[NSAutoreleasePool alloc] init];
   
       cc(^ void (NSString * a, id b) { NSLog(a, b); }, @"%@");
   
    [pool drain];
}


Thanks,
Matt

Matt,

This is definitely an issue with Clang's integration in Xcode.  This
isn't the right list for such discussions (cfe-dev is more appropriate).

Whether you are on the right track or not depends on the *full* stack trace (and
where the failure is occurring).  clang_visitChildren() represents an AST walk
by the IDE, but the specific problem is usually in the last couple of stack
frames.

If you could submit a project via bugreporter.apple.com that reproduces the
issue that would be awesome.  That's the appropriate place since this
concerns iOS code (although it doesn't really look specific to iOS, and
probably has more to do with templates and blocks).

Ted

On Apr 21, 2011, at 6:18 AM, Matt Lamari wrote:
> 
> Sorry if this isn't the right place; but I'm facing an XCode 4
(LLVM 2.0
> to Apple, I heard it's LLVM 2.9 under the hood) issue that may involve
> LLVM, and maybe some people on this list could help me short-circuit the
> source.
> 
> I have a code pattern that, when used in XCode 4 in Objective C++ files,
> causes the whole IDE to go crash.  It is a stripped sample to highlight
> the issue.
> 
> In the call stack, I see libclang.dylib, calling through
> clang_visitChildren, up through various CursorVisitor::Visit... and
> CursorVisitor::VisitChildren... calls.  I have LLVM compiling, and
> imagine I could replace XCode's libclang.dylib in order to add
debugging
> aids/logging.
> 
> Am I on the right track?  Should I look near the point of failure, or
> should I inspect the integrity of the data sent to clang_visitChildren
> first?
> 
> If I'm not on the right track, I appreciate any help in getting me
moving.
> 
> 
> If anyone happens to be able to test-rig XCode 4 (in iPhone mode with
> LLVM2 as compiler) vs the code below, it reproduces the problem:
> 
> template <typename I0,typename I1>
> void (^ cc(void (^f)(I0,I1), I0 i0))(I1)
> {
>    ^ void (I1 i1) { };
> }
> 
> 
> int main(int argc, char *argv[]) {
>    NSAutoreleasePool * pool = [[NSAutoreleasePool alloc] init];
> 
>       cc(^ void (NSString * a, id b) { NSLog(a, b); }, @"%@");
> 
>    [pool drain];
> }
> 
> 
> Thanks,
> Matt
> _______________________________________________
> LLVM Developers mailing list
> LLVMdev at cs.uiuc.edu         http://llvm.cs.uiuc.edu
> http://lists.cs.uiuc.edu/mailman/listinfo/llvmdev