This is the mail archive of the
gcc@gcc.gnu.org
mailing list for the GCC project.
Re: CVS and its problems
Phil Edwards wrote:
On Thu, Jun 20, 2002 at 12:25:08AM -0400, Nathanael Nerode wrote:
I'd suggest not putting too much effort into fixing problems caused by
CVS in the short term; quick hacks are probably good enough. Why?
Because subversion (subversion.tigris.org) actually looks like it's on
schedule to come out with a stable version 1.0 before the end of 2002.
And there was already agreement that it would be good to switch over
once it was released.
Yeah, I spent some time looking at this yesterday, prompted by the CVs
discussions.
The only two things that bother me are:
1) Subversion 1.0 will not have the equivalent of "cvs annotate"[*]
due to schedule slippage (something all of us are all too familiar
with), and
Yes. :-( Any help with that would be appreciated, of course. We're still
three weeks to Alpha and feature freeze. :-)
2) It uses HTTP for its underlying transport, unless I've seriously
misread their docs. Which means that people behind firewalls now face
a whole new set of problems, since simple tunneling through SSH is out.
Work is being done on a "pipe" access method that could be tunnelled
directly through SSH, without having to set up Apache on the server.
Sadly, that's not progressing very well, either ...
In the meantime, a SSH tunnel should work.
While I'll continue to encourage GCC to replace CVS with Subversion, if
for no other reason than most of the stupid O(n) operations become O(1),
I personally wouldn't be able to do any more work on the repository,
because our HTTP proxy is truly facist.
Phil
[*] amusingly called "svn blame" in Subversion.
Isn't that what you use "cvs annotate" for? :-)
--
Brane Čibej <brane@xbc.nu> http://www.xbc.nu/brane/