[cfe-dev] Invalid mangled names emitted for local types declared in Apple Block expressions used in non-static data member initializers

[John McCall via cfe-dev]

> On Sep 6, 2017, at 5:50 PM, Richard Smith <richard at metafoo.co.uk> wrote:
> 
> On 6 September 2017 at 14:16, Tom Honermann via cfe-dev <cfe-dev at lists.llvm.org <mailto:cfe-dev at lists.llvm.org>> wrote:
> Clang appears to be generating ill-formed mangled names for local types
> declared in Apple Block expressions used as non-static data member
> initializers:
> 
> $ cat t.cpp
> template<typename T> void tf(T) {}
> struct S {
>    void(^bp)() = ^{
>      enum E {} e;
>      tf(e);
>    };
> };
> auto x = S{}.bp;
> 
> $ clang++ --version
> clang version 4.0.0 (tags/RELEASE_400/final)
> Target: x86_64-unknown-linux-gnu
> Thread model: posix
> ...
> 
> $ clang++ -c -std=c++11 -fblocks t.cpp
> $ nm t.o
> ...
> 0000000000000000 W _Z2tfIZ1S2bpMUb0_E1EEvT_
> ...
> 
> The mangled name generated for the function template specialization
> instantiated for the local enum type 'E' uses a <pointer-to-member-type>
> production [1] (as signified by 'M'),
> 
> I believe you're mistaken. This is not a context in which a <type> production can appear. That's a <data-member-prefix>:
> 
> http://itanium-cxx-abi.github.io/cxx-abi/abi.html#mangle.data-member-prefix <http://itanium-cxx-abi.github.io/cxx-abi/abi.html#mangle.data-member-prefix>
> 
> ... which specifies a context of the default member initializer of S::bp.
> 
> The following construct (the Ub0_) is a vendor extension for naming a block closure.
> 
> but only one of the two required
> dependent productions is present (<class type> via 'Ub0_', but not
> <member type>).
> 
> If someone can confirm that this is not intentional (it could be since
> Apple Blocks aren't governed by the Itanium ABI), I'll file a new BZ for
> this.
> 
> This behavior looks correct to me, except that I'm surprised that this block is numbered as Ub0_ instead of Ub_. Looks like we have an off-by-one error here, introduced in r214699 while fixing a different off-by-one error.
> 
> John: should we restore the pre-r214699 ABI, or would you prefer that we just accept the new mangling as our ABI now?

Nothing about blocks ever has identity across translation units, so there's no harm in fixing the bug and restoring the original ABI to start counting at b_.

John.

>  
> Additionally, for unnamed local types, discriminators are not added to
> disambiguate such types.  Bug 34511 [2] has been filed for that issue.
> 
> Tom.
> 
> [1]:
> http://itanium-cxx-abi.github.io/cxx-abi/abi.html#mangle.pointer-to-member-type <http://itanium-cxx-abi.github.io/cxx-abi/abi.html#mangle.pointer-to-member-type>
> [2]: https://bugs.llvm.org/show_bug.cgi?id=34511 <https://bugs.llvm.org/show_bug.cgi?id=34511>
> _______________________________________________
> cfe-dev mailing list
> cfe-dev at lists.llvm.org <mailto:cfe-dev at lists.llvm.org>
> http://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-dev <http://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-dev>
> 

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.llvm.org/pipermail/cfe-dev/attachments/20170906/5381e056/attachment.html>