Re: Bug tracking

[Tom Tromey <tromey at cygnus dot com>]

>>>>> "Nathan" == Nathan Sidwell <nathan@codesourcery.com> writes:

Nathan> The lack of a proper bugtracking system is hampering the GCC
Nathan> effort.

Nathan> All the existing bug trackers fell down in some way,

Nathan> * Most only accepted data via web pages, and wouldn't allow
Nathan> upload of (big) source. Or they only accepted via email. Never
Nathan> both.

Gnats does both.  Gnats doesn't have upload size limits that I know
of.

Nathan> * The formating and classification didn't really fit with our
Nathan> needs, and/or was hard to change.

I'm told it is possible to change the categories in the latest Gnats.
I don't know how hard or easy this might be.

For formatting you'd have to hack gnatsweb, I guess.

What are the gcc needs and how do they differ from the needs of all
the other projects hosted on sourceware?

Nathan> * several implemented their own data base, rather than using
Nathan> an existing one

Yes, Gnats does this.  I don't understand why this is an a priori
constraint.

Nathan> I could probably get that working, or deploy a student on it,
Nathan> but they'll need financial incentive, I guess.

I personally think it is a bad idea to write a new bug tracking
system.  If you are going to hire somebody to work on one, have them
enhance an existing one instead of starting from scratch.  For
instance, a student could just as easily change Gnats to use a
different database, if that is really important.

FWIW, I use Gnats basically every day; I work on autoconf, automake,
and libgcj, all of which are hosted on sourceware.  I like Gnats.  It
has come a long, long way in the past few years.  For instance, when I
first used it years ago there was no web interface.  Now I find I only
use the web interface, because it is so convenient.

Tom