Re: Patch ping (stage1-ish patches)

[Jeff Law <law at redhat dot com>]

On 11/28/13 00:17, Jakub Jelinek wrote:
> On Wed, Nov 27, 2013 at 01:11:59PM -0700, Jeff Law wrote:
> > On 11/27/13 00:36, Jakub Jelinek wrote:
> > > Use libbacktrace for libsanitizer's symbolization (will need tweaking,
> > > depending on next libsanitizer merge, whether the corresponding
> > > sanitizer_common changes are upstreamed or not, and perhaps to compile
> > > libbacktrace sources again with renamed function names and other tweaks
> > > - different allocator, only subset of files, etc.; but, there is a P1
> > > bug for this anyway):
> > > http://gcc.gnu.org/ml/gcc-patches/2013-11/msg02055.html
> > Isn't libsanitizer maintained outside GCC?  In which case making
> > significant changes of this nature ought to be avoided.
>
> libsanitizer contains some files imported from upstream (pretty much all of
> *.cc and *.h) and the rest (configury/Makefiles etc.) is owned by GCC, as
> the LLVM buildsystem is very different.
OK.  I actually kindof came to the same conclusion while looking at 
other sanitizer library patches.
> The changes to the *.cc/*.h files actally have been committed to upstream,
> so a next merge from upstream will bring those changes automatically and
> we'll just need the build system etc. changes.  When that happens (I think
> Kostya said he'll work on that), I'll update the patch accordingly.
OK.  Go ahead and check it in then.

Thanks for clarifying things,
jeff