Re: [Re: Bug reporting [Was: Promoting greater integration between testers!:)]]

[Luis Villa <louie ximian com>]

On Tue, 2003-07-01 at 11:01, Owen Taylor wrote:
> On Tue, 2003-07-01 at 02:51, Murray Cumming wrote:
> > Owen Taylor <otaylor redhat com> wrote:
> > > Beyond that, given a few hours a week, what I'd probably try to do is
> > > publish an easy-fix list for GTK+. (If you query for PATCH_NEEDED,
> > > that's a roughly the right result.)
> > 
> > This should be able to do that for you, assuming that you use the easy-fix
> > keyword:
> > http://bugzilla.gnome.org/gnome-23-report.html
> 
> I don't really like using the easy-fix keyword for GTK+, because
> people typically self-apply it (*), to bugs that variously:
> 
>  A) Would take an experienced GTK+ hacker several months of
>     concentrated work to fix. 
> 
>  B) Would be easy to fix, if we didn't mind breaking every
>     single existing GTK+ program out there.
> 
>  C) Include a 1 line patch that papers over some problem in 
>     entirely the wrong place.
> 
>  D) Really are easy to fix, but for that reason, aren't really
>     that useful for volunteers to work on:
> 
>      reporter> "separate is mispelled in a comment at line 45 of foo.c"
>      volunteer> "Here's a patch to fix it"
>      me> Looks good, fine to commit
>      volunteer> I don't have CVS access
>      me> OK, committed
> 
> At best, easy-fix ends up meaning "easy for the maintainers of
> the package to fix". 

Then really you need to remove it from those bugs; obviously you're
reading them and seeing that they are bogus.

> Browsing the easy-fix bugs from other modules, I see very little
> that would be appealing for an incoming hacker to work on. The
> whole easy-fix concept, I think, is a bit busted.

It's not well-maintained; like anything that's not well maintained it
falls apart. 

Luis