Re: distributed version control systems
On Fri, May 19, 2006 at 07:49:20AM +0200, Frans Pop wrote:
> On Friday 19 May 2006 06:10, Sven Luther wrote:
> > Also, there is another thing to consider, i will need to do some
> > nobootlaoder changes in the near future, before we can release the new
> > version of the pegasos firmware, so that debian will not break on it
> > (and since we ship those boxes with debian on it, ...), and i will
> > possibly not be able to wait while all of you are taking post-debconf
> > vacations in mexico :)
> So, submit a bug report with a patch and I'm sure that Colin or Joey will
> apply it promptly. If not, feel free to remind them on IRC.
Yeah, i will probably do that, still it is inconvenient, and like said, there
is no real justification for not re-establishing the commit access.
Forcing Colin or Joey to do extra work like this is hardly an efficient way to
> nobootloader is such a simple udeb that I can hardly imagine that it will
> be a major patch.
Well, the firmware counterpart changes are a major piece of work, so it is
possible that the nobootloader changes will be modified before it is finally
uploadable. I hope not though, but you never know with such things.
> Testing changes is entirely possible without committing them. I routinely
> do so before I commit anything.
Well, there are two philosophies here, either the 'commit often' one, or the
test everything and commit only finalized work. The first is more logical for
centralized repos like d-i.