[PATCH] First set of patches for type uniquing

[David Blaikie]

On Wed, Aug 28, 2013 at 2:01 AM, David Majnemer
<david.majnemer at gmail.com> wrote:
> On Tue, Aug 27, 2013 at 5:42 PM, Manman Ren <manman.ren at gmail.com> wrote:
>>
>>
>>
>>
>> On Tue, Aug 27, 2013 at 2:34 PM, David Blaikie <dblaikie at gmail.com> wrote:
>>>
>>> On Tue, Aug 27, 2013 at 2:28 PM, Reid Kleckner <rnk at google.com> wrote:
>>> > Why are we using the Itanium mangler for uuidof exprs?  Can we not do
>>> > that?
>>>
>>> We have a test case (see
>>> clang/test/CodeGenCXX/debug-info-template.cpp) because we seem to
>>> support the language feature on itanium platforms. If we shouldn't
>>> allow that language extension in that case, we should disable it &
>>> move this test out to a separate case.
>>>
>>> (+David Majnemer who implemented this test/debug info support)
>>>
>>> > If not, feel free to grab the approach used in the MS mangler and
>>> > pretend to
>>> > mangle a global variable called _GUID_1234....
>>>
>>> I'm OK-ish with this, but deferring to you & David if you want to
>>> avoid these language extensions cropping up outside of win32
>>> platforms.
>>
>>
>> Hi David M,
>>
>> Any comment on whether we should move it out to a separate case?
>
>
> We have yet to define a vendor mangling for it so I wouldn't worry about
> testing it when targeting the itanium ABI.

Are there places this could lead to a crash in the existing code?
Should we disable the language feature on Itanium targets to ensure
this doesn't happen? (& modify the test case so the GUID stuff is only
tested on win32 ABI targets)