[cfe-commits] r154749 - in /cfe/trunk: lib/AST/Decl.cpp test/CodeGenCXX/visibility.cpp

[John McCall]

On Apr 16, 2012, at 12:32 PM, Rafael Espíndola wrote:

>>> If at all possible I would like to propose the visibility changes to
>>> 3.1. The particular case in PR12552 was extracted from a mozilla test
>>> that failed with clang.
>> 
>> This is really, really late.  That said, we definitely don't want to break mozilla.
>> Is there a less aggressive fix?
> 
> I am not sure we broke it. Everything used to pass with an old clang
> and a old firefox. This is very likely just a new test in firefox that
> found the problem.

If it's not a regression, I think I'd rather not screw around with visibility
computation until tomorrow.  Visibility right now is just a minefield of
subtle expectations and inadequate testing.  For example, apparently
libc++ didn't pass tests (at least on Darwin) with this patch in — the
dylib stopped exporting some symbol.  That might be a bug in libc++'s
annotations, or it might be a subtle error in your patch (something
getting overridden that shouldn't), or it might be evidence of a
deeper flaw in the visibility logic.

If you're interested in investigating this, I believe all you need to do
is run ./testit in libc++.

John.