[cfe-commits] r64416 - /cfe/trunk/tools/ccc/ccclib/Tools.py

[Daniel Dunbar]

My intent was that clang's own tests would always use the clang
compiler binary, not the driver. The idea is that the driver is the
end-user tool, and as such should go as fast as possible. On the other
hand, for development and testing we get the alloc/free behavior so we
shake out bugs like you mentioned.

Does this seem reasonable, or are you still sad? :)

 - Daniel

On Thu, Feb 12, 2009 at 4:31 PM, Douglas Gregor <dgregor at apple.com> wrote:
>
> On Feb 12, 2009, at 4:29 PM, Daniel Dunbar wrote:
>
>> Author: ddunbar
>> Date: Thu Feb 12 18:29:22 2009
>> New Revision: 64416
>>
>> URL: http://llvm.org/viewvc/llvm-project?rev=64416&view=rev
>> Log:
>> Always pass -disable-free to clang when compiling.
>
> This makes me sad. I found an icky ownership problem a few days ago that I
> would have missed completely if we always passed "-disable-free". Since
> we're moving toward using the driver for *everything*, it really seems like
> we want -disable-free in Release-Asserts builds but not in Debug builds.
>
>        - Doug
>
>> Modified:
>>   cfe/trunk/tools/ccc/ccclib/Tools.py
>>
>> Modified: cfe/trunk/tools/ccc/ccclib/Tools.py
>> URL:
>> http://llvm.org/viewvc/llvm-project/cfe/trunk/tools/ccc/ccclib/Tools.py?rev=64416&r1=64415&r2=64416&view=diff
>>
>>
>> ==============================================================================
>> --- cfe/trunk/tools/ccc/ccclib/Tools.py (original)
>> +++ cfe/trunk/tools/ccc/ccclib/Tools.py Thu Feb 12 18:29:22 2009
>> @@ -229,6 +229,12 @@
>>            for arg in arglist.getArgs(arglist.parser.XanalyzerOption):
>>                cmd_args.extend(arglist.getValues(arg))
>>        else:
>> +            # The make clang go fast button.
>> +            #
>> +            # FIXME: Always pass this, once analyzer is fixed to still
>> +            # write plist output with it.
>> +            cmd_args.append('-disable-free')
>> +
>>            # Perform argument translation for LLVM backend. This
>>            # takes some care in reconciling with llvm-gcc. The
>>            # issue is that llvm-gcc translates these options based on
>>
>>
>> _______________________________________________
>> cfe-commits mailing list
>> cfe-commits at cs.uiuc.edu
>> http://lists.cs.uiuc.edu/mailman/listinfo/cfe-commits
>
>