[PATCH 4/5] Support backtracing through elfutils (#70746)

[Peter Hutterer]

On Thu, Nov 14, 2013 at 05:00:29PM +0900, Keith Packard wrote:
> Peter Hutterer <peter.hutterer at who-t.net> writes:
> 
> > On Tue, Oct 29, 2013 at 10:13:28PM -0400, Jasper St. Pierre wrote:
> >> I'm not a fan at all. This API is really bad compared to what's provided by
> >> libbacktrace, which is super simple. This also seem to require callbacks
> >> that have "Linux" in the name. Why are we trying to get rid of
> >> libbacktrace, and what's the schedule for this?
> >
> > we're getting rid of libunwind, which was merged after 1.14 and isn't
> > available on e.g. RHEL and apparently not required either. as for
> > libbacktrace - haven't dealt with it yet, so I can't comment much on it.
> >
> > schedule - asap, before the 1.15 release would be nice so we don't introduce
> > a dependency that we don't need.
> 
> Are you OK with the current conditional dependency on libunwind? We
> could easily add a '--with-libunwind' and disable it by default so that
> the build would be consistent independent of whether libunwind was
> accidentally installed.
> 
> Sorting this out before 1.15 seems like a good idea.

>From the discussion I gathered that libunwind is the best option for now
until someone implements and external backtracer through fork() + exec().

So please ignore this patchset.

Cheers,
   Peter