[RFC] commit --fixes xxx: please, store actual bug id in addition to full URL

[Robert Collins]

On Thu, 2009-02-26 at 11:28 +0100, Nicholas Allen wrote:

> That makes more sense to me. It would be better if only the bug id was
> stored and there was another property on the branch as a whole (not on
> a
> revision) that defined the URL for bugs in general.

The bzr bug handling code should handle turning a bug url back into
other metadata; if thats not implemented yet, patches appreciated :).

> Then you just need --fixes=bug_number instead of --fixes=lp:bug_number
> etc. It's unlikely that a single branch will use more than one bug
> tracker.

Its pretty common actually: Nearly *every* piece of packaged software is
present in multiple bug trackers: the upstream, and N distros. 

>  This is much better because the URL for the bug tracker may
> change over time

Thats true; OTOH one could always provide a mapping later.

> , or the project might migrate from one bug tracker to
> another. If just the bug IDs are stored then this is no problem - you
> just update the branch's bug tracker location.

I think most bug tracker migrations lead to new bug ID's. Particularly
if you look at migrations from sf|savannah|lp|other_aggregate_trackers
to-or-from bugzilla|trac|roundup|other-single-trackers.

I don't think it makes sense to duplicate the data that is stored; It
may make sense to split it out - but that gets into enforcing a
particular model of how bugtrackers work, whereas the URL code is really
very nice and general. (Imagine a bug tracker using human descriptions
rather than numbers).

-Rob
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 197 bytes
Desc: This is a digitally signed message part
Url : https://lists.ubuntu.com/archives/bazaar/attachments/20090226/b04adc4e/attachment.pgp