Hi, I am moving this conversation into gnome-bugsquad mailing list to
get more people commenting on it.

On 8/1/06, rushabh doshi <[EMAIL PROTECTED]> wrote:
> Cool. I took a look at the blog and went over your proposed solution.
> It sounds interesting and I would like to contribute to this server
> infrastructure project.
>
> What would be a good starting point to begin with?


Well, the first step is writting a proposal/design and get some
feedback on it from people interested: GNOME bugsquad people, maybe
Ubunut, etc...

Main points here:

- Binaries Collector thing
  * Should watch some configurable URLs to get binary+dbg packages from distros
  * Should unpack binaries and store them somewhere on the server with
some md5 thing on a table identifying them
  * This should support at least .deb and .rpm based distributions,
and should be extensible

  I guess this can be done in a beautiful way using Python objects.

- Crash dump format
  * We should research on what is the best format for the client to
send the crash information: full binary core? text based backtrace
with mem addresses on an XML?

- Crash Enricher:
  * Get crahses on the agreed format and matched them against
binaries. Could use jailed chrooted dir + gdb to do this or maybe
custom code.
_______________________________________________
Gnome-bugsquad mailing list
[email protected]
http://mail.gnome.org/mailman/listinfo/gnome-bugsquad

Reply via email to