This is the mail archive of the gcc@gcc.gnu.org mailing list for the GCC project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]

Re: GNATS documentation


On Sat, 21 Apr 2001, Mark Mitchell wrote:

> A lot of people seem to be creating GNATS reports along the lines of:
> 
>   CVS snapshot X didn't bootstrap on Y
> 
> Our policy is that these kinds of reports shouldn't go in GNATS.

We ought to be tracking such problems on less-used platforms that can stay
broken for some length of time without being widely noticed.  (When did
anyone last successfully bootstrap the mainline and report testresults on
an m68k native platform?)  What should avoid GNATS are short-lived and
actively discussed problems on widely used platforms - but if a bug is
reported, not discussed or discussed without any resolution, and persists
in the tree for over a week, I think it ought to be tracked in GNATS (in a
separate "bootstrap" category; GNATS ought to regularly request submitters
of bootstrap bugs to report whether the problem is still present), along
with any followups on the issues involved in the bootstrap failure, rather
than lost among 10Mb/month of gcc-bugs messages.

-- 
Joseph S. Myers
jsm28@cam.ac.uk


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]