On Aug 30, 2010, at 11:46 AM, Michael Spencer wrote:
> On Mon, Aug 30, 2010 at 11:04 AM, Douglas Gregor <[email protected]> wrote:
>> Author: dgregor
>> Date: Mon Aug 30 10:04:51 2010
>> New Revision: 112482
>>
>> URL: http://llvm.org/viewvc/llvm-project?rev=112482&view=rev
>> Log:
>> Emulate (some of) Microsoft's looser semantic checking of exception
>> specifications, from Martin Vejnar!
>>
>>
>> Modified:
>> cfe/trunk/lib/Sema/SemaExceptionSpec.cpp
>> cfe/trunk/test/SemaCXX/MicrosoftExtensions.cpp
>> cfe/trunk/test/SemaCXX/exception-spec.cpp
>>
>> Modified: cfe/trunk/lib/Sema/SemaExceptionSpec.cpp
>> URL:
>> http://llvm.org/viewvc/llvm-project/cfe/trunk/lib/Sema/SemaExceptionSpec.cpp?rev=112482&r1=112481&r2=112482&view=diff
>> ==============================================================================
>> --- cfe/trunk/lib/Sema/SemaExceptionSpec.cpp (original)
>> +++ cfe/trunk/lib/Sema/SemaExceptionSpec.cpp Mon Aug 30 10:04:51 2010
>> @@ -261,6 +261,14 @@
>>
>> bool OldAny = !Old->hasExceptionSpec() || Old->hasAnyExceptionSpec();
>> bool NewAny = !New->hasExceptionSpec() || New->hasAnyExceptionSpec();
>> + if (getLangOptions().Microsoft) {
>> + // Treat throw(whatever) as throw(...) to be compatible with MS headers.
>> + if (New->hasExceptionSpec() && New->getNumExceptions() > 0)
>> + NewAny = true;
>> + if (Old->hasExceptionSpec() && Old->getNumExceptions() > 0)
>> + OldAny = true;
>> + }
>
> I really don't like accepting bugs without any warnings. Clang should
> complain if it ever sees this in anything but a Windows Platform
> header (or macro expanded from such).
Yes, that's a good point. This should ExtWarn when it's actually accepting
incompatible exception specifications.
- Doug
_______________________________________________
cfe-commits mailing list
[email protected]
http://lists.cs.uiuc.edu/mailman/listinfo/cfe-commits
