Hi Jim,

Sending to cfe-commits. I'm preparing a test case and will commit if
there is no objections from cfe guys.

--renato

On 24 July 2012 22:50, Jim Grosbach <[email protected]> wrote:
> LGTM. Though especially since we've apparently been getting this wrong for 
> quite a while now, test cases seem a good idea.
>
> -Jim
>
> On Jul 24, 2012, at 2:45 PM, Renato Golin <[email protected]> wrote:
>
>> As discussed in the main list, triples like "thumbv7m-none-eabi" still
>> produce ARM code. This is a flaw in the Driver that is fixed by this
>> patch.
>>
>> To test, compile any C code like:
>>
>> $ clang -ccc-host-triple thumbv7-none-eabi -S foo.c
>>
>> It'll produce ARM code. With this patch, it should produce Thumb code.
>> In Darwin, all v7 defaults to Thumb, so test this on Linux.
>>
>> Where should I add a test to match the code generated to Thumb in Clang's 
>> suite?
>>
>> --
>> cheers,
>> --renato
>>
>> http://systemcall.org/
>> <clang-driver.patch>
>



-- 
cheers,
--renato

http://systemcall.org/

Attachment: clang-driver.patch
Description: Binary data

_______________________________________________
cfe-commits mailing list
[email protected]
http://lists.cs.uiuc.edu/mailman/listinfo/cfe-commits

Reply via email to