[xen-tools-discuss] Bug Tracking System / Commit access for others / Git or Hg?
abe at deuxchevaux.org
Sat Apr 10 15:14:17 CEST 2010
On Sat, Apr 10, 2010 at 02:33:18PM +0200, Jonathan Aquilina wrote:
> axel why not just use SF for git and use something like launchpad for the
> bug tracker?
Basically because I want to choose hosting and bugtracker based on
needs and preferences not based on number of hosted projects,
publicity or age of the hoster.
I'm no big fan of Launchpad either, but IMHO it's far better than SF's
bug tracker. I personally consider Debian's Bug Tracking System as a
bug tracker as it should be, but that's no option for the project.
Ideas for a bug tracker rumouring in my head:
* flyspray (http://flyspray.org/)
* Roundup (http://www.roundup-tracker.org/)
Distributed within the VCS/SCM:
* bugs-everywhere (http://bugseverywhere.org/)
* ditrack (http://www.ditrack.org/)
* Bugtracker of the VCS/SCM hoster whatever they offer
On Sat, Apr 10, 2010 at 02:35:14PM +0100, Steve Kemp wrote:
> I have no preference, since I'll almost certainly not be
(Hey, that's more than I expected! :-)
> but I find SF painful to navigate, slow, and clunky for issuing new
Yeah, I forgot "slow".
> I'd suggest hg + hgweb on a personal server + mailing list is
I must admit that the main reason for tending to hosted VCS/SCM
hosting is that I have not much expererience with hosting repositories
where arbitrary people have commit access.
And those experiences I made (with Subversion) were always covered
with Unix permission hassles. (And I do not like to create local Unix
accounts on my server for every committer anyway.)
So one of the reputable (i.e. not SF ;-) VCS/SCM hosters seems an easy
and good solution for me there.
> but I guess github is the new sexy collaborative way to do things
> so that's probably as good and requires less dedicated personal
Well, for me it's less the sexy thing than more that I don't have to
care much about administrating the hosting as well as that I in the
meanwhile prefer git over hg.
On Sat, Apr 10, 2010 at 02:41:34PM +0100, Steve Kemp wrote:
> In practice I guess we'll end up in a situation with:
> * 1-5 people making commits, regularly.
> * 1-5 people declaring "new release".
> * 10-100 people downloading a release tarball.
> * 100-2000 people getting the package via Debian, Ubuntu, repos,
> On that basis I think whatever solution is picked should be one that
> requires the least work for the 1-5 people doing the
Yeah. If we'd have a developer ML I probably would have asked there
instead of on the general discussion ML.
> Anyway I have no real stake so I'll bow out now, and leave either
> a consensus to develop or a firm decision to be made by Axel :)
Thanks for the nice summaries. :-)
Axel Beckert - abe at deuxchevaux.org, abe at noone.org - http://noone.org/abe/
More information about the xen-tools-discuss