Re: track bugs in VCS, not the other way around
Jelmer Vernooij <firstname.lastname@example.org> writes:
> Have you seen bugs everywhere (http://www.bugseverywhere.org) ? It's a
> distributed bug tracker that stores bug information as metadata in VCS
I'm currently working on <URL:http://bugs.debian.org/477125> a Debian
package for Bugs Everywhere
> I don't think it's exactly what you're talking about but it may be a
> source of inspiration.
Bugs Everywhere is a simple idea that is a good way to have a BTS for
people whose main mode of collaborating is via a DVCS, and won't or
can't set up a central bug tracker. This fits enough projects that I'm
on that I was motivated to package it in Debian.
However, I think that for people who have 'debbugs' to work with (i.e.
the Debian project), Bugs Everywhere might actually be a step
*backward* from 'debbugs'.
Currently, a Bugs Everywhere database is visible only to those who use
the VCS to get a copy of the branch, and use the Bugs Everywhere
command-line tool to query the database. The 'debbugs' available to
Debian developers has a web interface that over the last few years
provides comprehensive order out of the chaos of our bug database.
Currently, a Bugs Everywhere database can be manipulated only be
people who will use the VCS to get a branch, use the command-line tool
to manipulate bugs, and use the VCS again to propagate their changes
to others. Using 'debbugs' requires only that one know how to send
These interface issues alone are enough to disqualify something like
Bugs Everywhere from being proposed as a serious alternative to
Debian's 'debbugs' instance. Likely a quick look at the actual
features of Bugs Everywhere would easily come up with other
shortcomings compared to 'debbugs', but I won't.
\ “I think there is a world market for maybe five computers.” |
`\ —Thomas Watson, chairman of IBM, 1943 |