r187994 - Add clang_Location_isFromMainFile() to libclang.

Eli Friedman eli.friedman at gmail.com
Wed Aug 21 17:31:40 PDT 2013


On Wed, Aug 21, 2013 at 5:16 PM, Du Toit, Stefanus <
stefanus.du.toit at intel.com> wrote:

>  On 2013-08-21, at 8:10 PM, "Eli Friedman" <eli.friedman at gmail.com> wrote:
>
>   On Thu, Aug 8, 2013 at 10:48 AM, Stefanus Du Toit <
> stefanus.du.toit at intel.com> wrote:
>
>> Author: sdt
>> Date: Thu Aug  8 12:48:14 2013
>> New Revision: 187994
>>
>> URL: http://llvm.org/viewvc/llvm-project?rev=187994&view=rev
>> Log:
>> Add clang_Location_isFromMainFile() to libclang.
>>
>> Also bump the minor version number and update libclang.exports.
>>
>> Reviewed by: Dmitri Gribenko, Doug Gregor
>>
>>
>  I'm currently working on a patch to split isFromMainFile into two
> versions; one which respects line markers (after '# 1 "XXX.h" 1', we treat
> the following lines as not part of the main file), and one which has the
> current behavior.  I think you want the behavior which respects line
> markers, but I'm not completely sure how you expect this to be used.
>
>
>  For the use case that led me to add this, I want to know if it's
> physically in the main source file. I can see how the line marker behaviour
> would be helpful too (actually I may want that behaviour in addition
> myself).
>

Okay.  I left the behavior the same for now; feel free to change it.


>
>  I think this should probably be exposed through a parameter in the
> libclang function (PhysicalOnly or RespectLineMarkers or something along
> those lines) and expect it should still be fine to change the API since it
> was added so recently (but I don't know if there's a specific rule about
> that).
>

I'm not sure either.  We generally draw the line between dot releases, but
if you wanted to be on the safe side, you could rename it.

-Eli
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.llvm.org/pipermail/cfe-commits/attachments/20130821/78a3f372/attachment.html>


More information about the cfe-commits mailing list