[cvsnt] Re: cvs checkout -d broken in 2.0.0?

Oliver Koltermann okoltermann at gmx.de
Fri Apr 11 09:24:50 BST 2003


Community technical support mailing list was retired 2010 and replaced with a professional technical support team. For assistance please contact: Pre-sales Technical support via email to sales@march-hare.com.


Tony Hoyle <tmh at nodomain.org> writes:

> Oliver Koltermann wrote:
> 
> > I forgot to mention that this worked for me with all older versions of
> > cvsnt. I can't say which version was the latest, but I have no problem
> > with CVSNT 1.11.1.3 b72. So I'm sorry that I have to disagree on
> > this. We used this functionality all the time - there must have been
> > some recent changes which broke it.
> 
> If it worked it was by accident... it couldn't possibly have been correct...
> even the manual documents that it gets the paths wrong - the client has to
> deal with that crap itself (which admittedly it seems to manage as it's
> fairly robust).
> 
> Making a 'clean' solution is going to take some time & testing, as it
> involves some quite hairy changes.

Hello Tony,

in the ChangeLog you write

2003-04-01 15:45  tmh

	* cvsnt.vcproj, src/annotate.c, src/checkout.c, src/client.c,
	src/cvs.h, src/log.c, src/ls.c, src/modules.c, src/patch.c,
	src/server.c, src/tag.c, cvsnt.vcproj, src/checkout.c,
	cvsnt.vcproj, src/checkout.c, src/server.c: Rewrite co -d handling.
	 I don't like doing this so close to a release but it's becoming a
	FAQ and I'm fed up with it.

That was release candidate 5 (3 days before the release). I don't know
much about how cvs works in detail. Especially I don't understand the
ampersand modules handling. I try to browse through the code and
gather information, but I'm not familiar enough with it to be of help.

I understand that it worked by some kind of magic in the old
releases. I don't know if it was intentionally or by chance. I'm sure
that the new handling of checkout -d is better than the former, but in
my eyes it's incomplete. It doesn't provide an essential functionality
that was available with the old code and that users may rely on.

So when there are further hairy changes needed to get back to the
stable state, I think this should go on in the development version. I
was really happy about the new version numbering scheme just
startet. It brought back the kind of safe feeling about a release. I
lived through the up's and down's of the 1.11.1.3 builds and always
hated the decision "Is it better to update or does this introduce new
problems?".

Please don't misunderstand me, but this change disappointed my new
hope to always know which release is best to distribute to the
developers. You're doing a really good job and indeed improving cvsnt
over the original unix version. And you made cvs more usable,
integrating and interconnecting in a windows environment. I think
changes are good, but as cvsnt evolutes in big steps it is also
important to have stable settle points. For me 2.0.0rc4 is the stable
point I use now.

Sorry for the long explanation. You invest so many time in the project
that my effort to select the right version fades. Thank you for all
the work! I hope it will get even more distributed as soon as the new
version numbers have stabilized.

Best wishes,
  OK.


More information about the cvsnt mailing list
Download the latest CVSNT, TortosieCVS, WinCVS etc. for Windows 8 etc.
@CVSNT on Twitter   CVSNT on Facebook