[cvsnt] Major regression in CVSNT 2.5.05.3649
arthur.barrett at march-hare.com
Sat Feb 6 13:03:25 GMT 2010
> > Can you send me privately your actual examples - un-obfuscated?
> > particularly showing with -d failing and without -d working?#
> I will do that later.
In light of your comments below - please also include the server version.
If possible please also sent the results of rlsacl:
"C:\Program Files\CVSNT\cvs.exe" -q -d :sspi:host:/cvs/rep rlsacl module/impl
"C:\Program Files\CVSNT\cvs.exe" -q -d :sspi:host:/cvs/rep rlsacl module
"C:\Program Files\CVSNT\cvs.exe" -q -d :sspi:host:/cvs/rep rlsacl .
If possible please include a trace of the checkout with server tracing enabled - if this will take additional time or is difficult - just skip it - with the onther info (server version, rlsacl) hopefully I can reproduce it here.
> > I think this is a server error - so will not affect TortoiseCVS's use
> > of the client.
> I doubt it, as this was against the server where I work.
> The old client worked, the new one didn't.
OK - this must be a completely different error to what I've seen then - and all I can deduce so far is that it's limited to particular server versions (since the command works OK against sourceforge and also against 2.5.05 itself).
I suppose there is a very very small chance that it is something to do with your local install. So running test commands against sourceforge on the same PC would be very very interesting (please do not use ext - stick to inbuilt protocols - if you need ssh to get through your firewall please use :ssh:host:/repo).
More information about the cvsnt