[cvsnt] Re: Crash on socket connect
timo.kauppinen at lut.fi
Tue Sep 24 18:29:05 BST 2002
At the moment there's no other option than re-installing the whole Windows.
Before that I'd like to try anything else.. It shouldn't be too difficult to
track down the problem since it's happening so soon in the execution. No
repositories are read at that point, I guess.
In fact the service has been crashes once or twice earlier (long long time
ago), and for that reason I'd like to see also a debug version of the
"Tony Hoyle" <tmh at nodomain.org> wrote in message
news:3d909483.104356609 at news.cvsnt.org...
> On Tue, 24 Sep 2002 19:34:59 +0300, "Timo Kauppinen"
> <timo.kauppinen at lut.fi> wrote:
> >I've had cvsnt running in my system for 6 months now and suddenly it
> >working. I tried to debug it with -test switch but as soon a socket is
> >connected, cvsservice crashes. Just a plain telnet was used to connect
> >"The instruction at "0x10008690" referenced memory at "0x00000000"....".
> >and here's the output from command prompt:
> >CVS 184.108.40.206 (Build 57i) (Sep 13 2002) starting in test mode.
> >Repository prefix set to D:/***
> >Adding repository D:/***/CVS
> >TEMP/TMP currently set to ***
> >Initialising socket...ok
> >*WARNING* Impersonation is disabled - all file access will be done as
> >Starting auth server on port 2401/tcp...
> >Starting named pipe server...
> >CVS initialised successfully
> ><and then nothing --- crashes silently>
> >So, anyone got an idea how to fix the cvs? No sw installations or
> >like that has been made recently (that I'm aware of).I updated to the
> >version of CVSNT but that apparently didn't help too much.
> Something much have changed on the box... I'd be tempted to
> re-install from backups just in case there's something stupid going on
> like someone has installed an older MSVCRT or something.
More information about the cvsnt