Linux

[Bill Barry]

It seems solid to me.
As for quirks, I haven't noticed any (we moved to hg from svn and had 
the same case sensitivity issues before). I have only been using it for 
a couple of months now though (I wouldn't consider myself very 
experienced, there is much that I don't know or understand; for instance 
I've been following these osutil.c code changes for a while now and 
still have only the vaguest ideas what any of it means).

I wouldn't use CVS for any new work (it has no advantage whatsoever over 
subversion). Given the choice between hg and svn, I would go with hg 
simply for the better merge tracking, speed and local history.

My precedence table for version control systems (based on what I have 
used and seen in use):
hg git
bzr
svn
cvs
tfs
nothing
"hard copies printed out and stored for every revision"
vss

bzr is a special case in that under certain environments it could be 
better than hg or git (though I like the freedom more than the workflow 
bzr pushes)
(to read: systems on a single row can be substituted for each other and 
are on roughly equal ground but are all better then anything below them).
Clark Hwang wrote:
> Based on your experience with Mercurial, would you say it's a solid SCM system.    Are there any quirks?   Any problems with corrupt or inaccessible repositories?
> I was just about to deploy CVS just before looking into more about Mercurial.   Luckily I still have time to switch to something better.
>