[cfe-commits] r172601 - in /cfe/trunk: lib/Format/Format.cpp unittests/Format/FormatTest.cpp
djasper at google.com
Wed Jan 16 02:54:44 PST 2013
Yes, we need an option that describes what clang-format should output and
that might be different from what it reads. And, as Chandler said, we need
to take the different warnings that users might have enabled into account
so that we don't produce code that generates warnings. So this might need
some more thinking.
On Wed, Jan 16, 2013 at 11:36 AM, Kim Gräsman <kim.grasman at gmail.com> wrote:
> On Wed, Jan 16, 2013 at 10:42 AM, Chandler Carruth <chandlerc at google.com>
> > On Tue, Jan 15, 2013 at 11:30 PM, David Blaikie <dblaikie at gmail.com>
> >> On Jan 15, 2013 11:21 PM, "Daniel Jasper" <djasper at google.com> wrote:
> >> >
> >> > Author: djasper
> >> > Date: Wed Jan 16 01:19:28 2013
> >> > New Revision: 172601
> >> >
> >> > URL: http://llvm.org/viewvc/llvm-project?rev=172601&view=rev
> >> > Log:
> >> > Never merge < and ::, as it produces different tokens.
> >> Shouldn't this be std-specific, since this works in C++11? Speaking of
> >> which - what does clang-format do with the >> in nested template types?
> > Agreed on both fronts -- this needs to be conditioned on the standard
> > version, and >> in nested templates should be handled similarly....
> Just to be clear -- do you think this should be conditioned on -std or
> on a separate compatibility switch?
> I don't think conditioning it on -std works because I may be
> formatting portable source code with clang-format that must be
> buildable with an older (C++98) or incompatible (e.g. MSVC) compiler.
> I think this is what your follow-up paragraph said, but I wanted to
> - Kim
> cfe-commits mailing list
> cfe-commits at cs.uiuc.edu
-------------- next part --------------
An HTML attachment was scrubbed...
More information about the cfe-commits