[LLVMbugs] [Bug 15512] New: Clang generates invalid code when lambdas call lambdas.

bugzilla-daemon at llvm.org bugzilla-daemon at llvm.org
Thu Mar 14 08:29:56 PDT 2013


http://llvm.org/bugs/show_bug.cgi?id=15512

            Bug ID: 15512
           Summary: Clang generates invalid code when lambdas call
                    lambdas.
           Product: clang
           Version: trunk
          Hardware: PC
                OS: Windows XP
            Status: NEW
          Severity: normal
          Priority: P
         Component: C++11
          Assignee: unassignedclangbugs at nondot.org
          Reporter: ian.stewart at autodesk.com
                CC: dgregor at apple.com, llvmbugs at cs.uiuc.edu
    Classification: Unclassified

Created attachment 10179
  --> http://llvm.org/bugs/attachment.cgi?id=10179&action=edit
template workaround to avoid nested lambda bags.

I have created a simple program that crashes when running the resulting code
generated by Clang.

    int main( int argc, char *argv[])
    {
        auto first = [&]() { return argc; };
        auto second = [&]() { return first(); };
        return second();
    }

I have investigated a little, and what I have found is that Clang mistakenly
thinks that 'first' and 'second' should have the same implementation, only
generates code for one of them, yet calls it for both.

If I change the resulting type of 'first' to differ from 'second', it works:

    int main( int argc, char *argv[])
    {
        auto first = [&]() -> unsigned int { return argc; };
        auto second = [&]() -> int { return first(); };
        return second();
    }

As an ugly workaround, forcing a different return type using a templated
wrapper works (see attachment)

LLVM 3.3 @ SVN 176884
clang++.exe -Xclang -cxx-abi -Xclang microsoft -S -emit-llvm lambdafail.cpp -o
lambdafail.cpp.bc
lli.exe lambdafail.cpp.bc x y

expected result: 3


Reply from Richard Smith <richard at metafoo.co.uk> on the mailing list:

The implementation of the Microsoft ABI is incomplete. It's quite likely that
it does not correctly mangle lambdas yet.

-- 
You are receiving this mail because:
You are on the CC list for the bug.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.llvm.org/pipermail/llvm-bugs/attachments/20130314/01676908/attachment.html>


More information about the llvm-bugs mailing list