[PATCH] multiple inlining scopes for a single source location

[Eric Christopher]

Hi Adrian,

Thinking about this I wonder if we might want two EmitLocation routines
instead, with one to force the column. Another thought is just to always
emit it and ignore it in the backend. Can you take a look at that and see
whether it would be better? Also please update the comment to not reference
internal apple clang versions and instead reference svn revisions (or,
instead and better, just describe the changes).

-eric


On Fri, Mar 1, 2013 at 5:53 PM, Adrian Prantl <aprantl at apple.com> wrote:

> On Wed Feb 27 12:45:37 Manman Ren wrote:
> > Given an example with two function calls in a single line, without
> -gcolumn-info, both function calls will be at the same source location.
> > And Clang will generate a single coalesced inlining scope since we
> unique scopes by the source location.
> >
> > Even without -gcolumn-info, we should generate two inlining scopes. This
> means we should differentiate the different call sites with the same source
> location.
> > One option is to use the column info in SourceLoc to store the call site
> id when -gcolumn-info is not on.
> >
> > Any problem with this? Any other better suggestions?
>
> Here’s a patch for review that implements this. I’m using the actual
> column info rather than the (arbitrary?) call site id.
>
> thanks,
> Adrian
>
> ---------
>
> Force column info to be generated for call expressions so we
> can differentiate multiple call sites on the same line in the
> debug info.
>
> Fixes rdar://problem/13036237
>
>
> _______________________________________________
> cfe-commits mailing list
> cfe-commits at cs.uiuc.edu
> http://lists.cs.uiuc.edu/mailman/listinfo/cfe-commits
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.llvm.org/pipermail/cfe-commits/attachments/20130305/ec13e2dd/attachment.html>