Re: Possible Transition from SVN to Git?
... and please don't Cc: me offlist if I don't request it - especially
if you read my mail through the list and should thus be aware that I'm
subscribed to it, thanks.
* Andres Mejia <email@example.com> [2008-01-01 00:36:51 CET]:
> On Dec 31, 2007 4:14 AM, Gerfried Fuchs <firstname.lastname@example.org> wrote:
> > Even if this propably will be another mail from me that you simply
> > ignore and don't feel the need to respond to, like with the problems of
> > your BTS mails, I still think I should add some comments.
> Well, I only recall what you wrote for bug #451669. If this is the
> only one (I can't find other emails you directed toward me) I just
> simply forgot.
It wasn't just this one, I mentioned a more general problem (that's why
I did put #xxx into the subject to give you the hint that it is *not*
specific to that bugreport). It is a more general problem with your BTS
handling: you don't mail to -submitter when you want feedback from them
(note that mails to xxx@bugs only gets to the BTS, not the submitters),
you don't mention anything that could help people reading your mails
identify what package it is about (you didn't do that in your reply to
#451669 neither - totally ignored that part of my mail there), and again
also in this mail you had an empty subject which is another of your
usual problems with BTS mails.
> > * Andres Mejia <email@example.com> [2007-12-31 03:06:13 CET]:
> > > I have used git-buildpackage and maintain a package that's maintained
> > > at git.debian.org. I've seen the many advantages that git has over
> > > SVN.
> > Like? If you claim that there are many advantages that git has over SVN
> > it would be more than interesting to state them, otherwise such a claim
> > goes to /dev/null.
> OK fine. The advantages I pointed out and the statement above _WAS_ a
> personal opinion.
The thing is, you didn't really point out any advantages - in the
original mail. If you claim that it has /many/ advantages it would be
pretty wise to note them. I still don't see /many/ in your replies,
neither. Of course I took it as a personal opinion, but even then you
should try to find a reasoning for it.
> > How regularly do you rename files in the debian subdirectory anyway?
> Not very many, still why can't I just rename a file with whatever tool
> instead on relying on "svn mv", and not have to force a move, and not
> be able to move it again because of a typo? I find this a annoyance.
A rename in fact /is/ a move. And doing it with whatever tool you like
you are going to loose history. Even in git. git isn't much different
here, it doesn't have magic to recognize that you haven't deleted the
old file and added the new - and if it doesn't you lose history
informations on that file through that.
> Me personally I don't use different branches for any packages.
> However, I have thought about creating seperate branches for
> maintaining experimental packages for stuff in upstream between
> releases, upstream stuff like vegastrike which has been in development
> but has not seen any new releases in five years. In git it's just
> "git-checkout <some_branch>". Is it easier in SVN?
There's no problem with svn co <some_branch>, no. I can't follow you.
> Well, I'm not saying we should all now switch to git. I just want to
> know the thoughts of others here in regards to this.
And thus I gave mine. ;)