On Thu, 2008-12-04 at 14:21 -0500, Luis Villa wrote: > You don't need full symbols for crash information to be useful. It > certainly helps, but we shouldn't let the perfect be the enemy of the > good there.
Right, but I don't think there's an intermediate solution between resolving the crash trace symbols locally with a debugger and saving a minidump and let an external process gather the symbols afterwards. In the former case, you can either get incomplete information (if you lack debug some symbols locally) or a good trace, while in the latter (please correct me if I'm wrong) you either get a perfect trace or the minidump is completely useless. > > To be honest, I like more the way Ubuntu handles this, i.e. has its own > > crash server and pushes upstream only the good/unique traces. > > The crash information is (well, should be) the most important > bug-related information GNOME has. Relying on the judgment of others > to handle them should only be a last resort used if we're absolutely > incapable of handling it ourselves. I suppose distributors themselves are interested in fixing crashers (I'd go as far as saying the casual user would rather blame the distribution instead of upstream if something crashes!) and, at least the bigger ones, usually have people paid to take care to some degree of them, so I wouldn't worry too much about this. P.S: sorry if this is off-topic. Cheers, Cosimo _______________________________________________ desktop-devel-list mailing list desktop-devel-list@gnome.org http://mail.gnome.org/mailman/listinfo/desktop-devel-list
