abrtd - how to remove

[Tom Horsley]

> > Which file sizes are you talking about?
> 
> Depends what crashes - but probably of the order of 1,000 times less than
> a big set of debug packages

I can think of a couple of ways to improve things:

1. Separate out the debuginfo rpms into a zillion independent
rpms with one and only one library per rpm. That way instead
of downloading the entire glibc-debuginfo, abrt could download
just the debuginfo for libraries that appear in the stack
walkback (wouldn't even need all the libraries which were in the
process, but only the ones required to dig up a backtrace).
Bonus points for separating the source files from the object
files as well (don't need the source files to generate a
backtrace).

2. Make the retrace server actually be a remote debugger. Instead
of downloading the entire core file, it would just peek at the
bits of it needed to generate the backtrace.