Am Mittwoch, den 02.01.2008, 14:49 -0500 schrieb Behdad Esfahbod: > On Wed, 2008-01-02 at 14:39 -0500, Diego Escalante Urrelo wrote: > > On 1/2/08, Bruno Boaventura <brunoboaventura gmail com> wrote: > > > On Jan 1, 2008 6:14 PM, Behdad Esfahbod <behdad behdad org> wrote: > > > > I would like to ask no bugsquad work done on my modules: > > > > > > > > gnome-terminal > > > > vte > > > > pango > > > > gucharmap > > > > > > > > Specifically, I don't want NEEDINFO bugs automatically closed INCOMPLETE > > > > after a month, or two, or six FWIW. I have asked this before on IRC, > > > > but every once in a while a new bugsquader shows up and closes tens of > > > > bugs in gnome-terminal. > > > > > > Sorry, Behdad. I think I did it sometimes too, when I was starting. > > > Maybe can we have something more specifically about this on > > > http://live.gnome.org/Bugsquad/TriageGuide? BTW, I recommend new > > > triagers do not close any bug as INCOMPLETE when it have a stacktrace > > > and (comments > 1). > > > > > > Bugsquaders: Ideas? > > > > > > > Usually I close as INCOMPLETE when the reporter is a random email > > address with 0 points and comment-less. I also check if the trace can > > be read or if it's just a ton of ?? that doesn't even let you guess > > what's the real crash. > > Sure, a bunch of ??'s are definitely an INCOMPLETE right away. My main > issue is with automatic closing of NEEDINFO to INCOMPLETE. We all know > that in 99% of the cases the reporter is not going to provide that > requested info, either because they don't get the mail at all (I've > reported a few crashes myself and never got any followup), or because > they cannot reproduce the crash, or because they have no idea whatsoever > what's happening. They just clicked on a button to shut the dialog > up... > > If we keep closing those not-so-useful-but-not-useless traces as > INCOMPLETE, what we've essentially done is pumping up the total number > of bugs reported, so we can brag about having half million bugs while > KDE has only 150000, with all those extra bugs adding nothing to the > quality of our software. > > What I want to see is: > > - No bug with dups closed INCOMPLETE > - No bug with comments closed INCOMPLETE > - When a Simple dup search shows a trace is a dup of an INCOMPLETE > bug, the INCOMPLETE is REOPENED and the dup added > > That way, I can feel confident that a real but infrequent crasher in my > software doesn't go unnoticed just because all the reports were closed > INCOMPLETE. > > Thanks all. Thanks. though i also definitely would love to see behdad wants to see, realism shows that we just don't have the manpower and time to search for the dups and close them correctly as a dup instead of an incomplete, especially for bigger projects. closing them correctly would also make it easier to identify our worst crashers and problems. i think we had a discussion about the policy for nautilus and gtk+ to not quickly close needinfo bugs as incomplete, see http://live.gnome.org/Bugsquad/DirectiveDiscussion . we should pick that discussion up again. we also have http://live.gnome.org/Bugsquad/TriageGuide/ProductSpecificGuidelines , but i am afraid that not so many triagers know about that page. and it's partially outdated. i will add "update the triage guide" to my todo-list so i finally read it. :-P andre -- mailto:ak-47 gmx net | failed http://www.iomc.de/ | http://blogs.gnome.org/aklapper
Attachment:
signature.asc
Description: Dies ist ein digital signierter Nachrichtenteil