New VCS Choice; SCM SIG

[Kristian Høgsberg]

Ralf Corsepius wrote:
> On Tue, 2007-01-23 at 08:25 -0500, Jesse Keating wrote:
>> On Tuesday 23 January 2007 01:44, Ralf Corsepius wrote:
>>> What I find arguable/questionable in FE's CVS-repos, is the way
>>> "FE-branches" have been implemented into it. They are implemented as
>>> separate directories instead of CVS-branches. If real CVS-branches had
>>> been used several details would have been much easier.
>>>
>>> In the early days of FE, I had been told the reason for this design
>>> decision had been AVCs, because CVS storing branches in files would
>>> prevent AVCs to be applicable.
>> Way back when, the dist-cvs method did use actual cvs branches, however the 
>> workflow was not easy to apply changes across all branches.  And to some 
>> extent having actual directories made life a bit easier for many things.  
>> Now, directories COULD be implemented at the same time as cvs branches, just 
>> a little extra work on the tool side when bringing them down.  But here we 
>> are, and most of us feel that it would be better to move to a new SCM with 
>> new possibilities rather than put significant effort into CVS.
> Well, from a user perspective things look a bit different:
> 
> CVS has a long history, probably everybody has his own record of
> experiences with it, probably everybody knows about the weaknesses of
> CVS => Everybody is bashing on it.
> 
> IMO, svn is heading along the same lane. Initially it had not been much
> more but a "weak promise", then it had been pushed/hyped, now people are
> about to move away from it. To me personally, gcc having moved from CVS
> to svn, gave me sufficient reasons to hate it and not to consider it
> anymore for my own work :(
> 
> Now, dist-hg, git, bazaar seem to be on high on the list of personal
> preferences for some "leading RH people", for reasons only partially
> known to me.

There is no conspiracy here.  It's not a secret club of RH people pushing an 
agenda.  What it is, probably, is that some people have actually tried out 
these new tools, and some use them on a daily basis.  Once you've done that, 
it's pretty obvious that these tools give you new powerful ways of working 
with your source code.  And at the same time they allow pretty much the same 
centralized model as centralized tools such as cvs or svn.

> Please don't forget that these still qualify as exotic and rarely used
> tools, which have had comparable little exposure, and therefore probably
> suffer from a lot of bugs and unknowns. Also, don't forget that FE is
> supposed to be easily usable by many people. IMO, it would qualify as a
> fault to switch to using something "exotic".

I sure beg to differ.  Open your eyes - all of X.org, mesa, the kernel, cairo, 
hal is developed in git.  Don't call that 'little exposure'.  git is different 
from cvs in many ways, but that's typically in a good way, and certainly 
doesn't warrent the label 'exotic'.

thanks,
Kristian