[cfe-commits] [Patch 2 of 4] -verify fixes (stage 2)

[Jordan Rose]

Hm. This doesn't feel right either, because at this point errRec is no longer the actual client hooked up to DiagnosticsEngine. I guess the nested begin/end is a better way to go.

I'll try to look over all the patches one more time tomorrow morning, but I don't think I have any outstanding objections.

Jordan


On Aug 8, 2012, at 9:40 AM, Andy Gibbs <andyg1001 at hotmail.co.uk> wrote:

> Sorry, reposted.  This time with non-uuencoded attachment!!! (hopefully)
> 
> --------------------------------------------------
> 
> Hi,
> 
> I've revised my patch for ARCMT.cpp here.
> 
> This patch now no longer uses nested calls to Begin/EndSourceFile but is
> otherwise the same as that previously posted.
> 
> This means that now it is only the ast-merge feature that nests such calls.
> My suggestion, therefore, would be to commit everything as it stands,
> including patch 1 as it stands, even though it is not ideal, and then I can
> work towards fixing ast-merge with Doug so that it doesn't require nesting.
> For my part, I think it is better to commit now since at least one person
> has found a regression which is fixed in these patches.
> 
> What say you?
> 
> 
> In keeping with convention, the complete set of patches can now be found
> here (in order):
> 
> Patch0:
> http://lists.cs.uiuc.edu/pipermail/cfe-commits/attachments/20120726/cc445d54/attachment.obj
> 
> Patch1:
> http://lists.cs.uiuc.edu/pipermail/cfe-commits/attachments/20120726/985686a2/attachment.obj
> 
> Patch1a:
> http://lists.cs.uiuc.edu/pipermail/cfe-commits/attachments/20120726/45e0a028/attachment.obj
> 
> Patch2:
> (attached to this post)
> 
> Patch3:
> http://lists.cs.uiuc.edu/pipermail/cfe-commits/attachments/20120717/8155a9d2/attachment.obj
> 
> Patch4:
> http://lists.cs.uiuc.edu/pipermail/cfe-commits/attachments/20120717/9cbd8128/attachment.obj
> 
> 
> Cheers
> Andy
> 
> <patch2-revised.diff>