Steve Youngs <youngs(a)xemacs.org> writes:
|--==> "JF" == Jarl Friis <jarl(a)diku.dk> writes:
JF> Gerd Boerrigter <gerd.boerrigter(a)i2c-systems.com> writes:
[...]
>>| Message: Parser Error: 'CVS server: Updating .'
[...]
JF> I have experienced this too some times (with 21.1 and 21.4 ) using
JF> 'CVS_RSH=ssh', but only over "slower" connections (128kb/s)
and
JF> hitting cvs-status a couple of times made it go away. I have never
JF> experienced this when using local (same computer or LAN) cvs
JF> repository.
JF> It could be nice if it was possible to set a cvs-debug variable to let
JF> the cvs output go to a debug buffer to have a closer look at what
JF> pcl_cvs receives.
Look in 'pcl-cvs-parse.el'. I'd say that's where you might be able to
fix this.
Yep, Adrian allready pointed out the line, so I guess we're just
waiting for a patch from Stefan. But maybe he/we wait to isolate the
problem more, cause I don't think this is the root of the problem. I
have seen other non-parseable situations, that eventually went away
after a couple of cvs-status.
JF> I have not yet experienced it with the xemacs cvs repository yet, what
JF> cvs version runs on that repository?
I've seen this from
cvs.xemacs.org (it happens a lot for me). The CVS
server is a "Martin Buchholz hybrid", but I'm not sure what version it
is based on.
Could you send us (or me) the content of " *cvs-tmp*" buffer (note the
initial space in the buffer name).
Does "it happens a lot for me" means that it also sometimes parses
without problems?
What version do you use locally (as client)?
Jarl