At 08:40 PM 7/23/99 +0200, Lars Appel wrote:
>Hi gang,
>
>just out of curiosity... Is anybody using GNU cvs on Windows or Linux?
>
Using GNU CVS on HPUX , Solaris, MPE/iX and Windows NT.  Others UNICIES as
well.   All use (even on MPE) is to a remote CVS repository and access is
through the client/server model.

The native MPE CVS works well except for failures on checkout related to
filenames that have illiegal characters for MPE (ie.  filenames with spaces
in them such as "a long NT filename").  MPE CVS stops dead in it's tracks
when it encounters these.

>Did you ever try to use it on MPE/iX sources that reside on a Samba/iX
>share? I wonder if this would allow using the revision control as well
>as other tools (fancy editors etc) on the PC, and have a telnet session
>into the 3000 to perform make and/or program tests.
>

Tried it.  Exported a SAMBA share directory from MPE and used a Windows NT
box for doing the actual CVS checkout.
It was very very very slow and a checkout that would take about 2 hours on
MPE  took about 12 hours on the NT.  This was an attempt to get around the
failures on checkout related to  NT filenames.  I won't mention the
compilation time.

>If such a PC also has Internet access, the GNU cvs could even be used
>to perform checkout/update/commit activities to a cvs server like the
>one on Jazz, and thus participate in Open Source development for MPE.
>

As indicated the MPE/ix version of GNU CVS works well enough that the
NT<->MPE<->UNIX scenario isn't needed. Checkout/update/commit can be done
natively on the MPE box.  With the MPE GCC compiler, PERL, etc.  you never
have to leave the sanctity and safety of the MPE OS.

Just my .03 cents CDN (.02 cents US)

.Mike.

>Just curious...
>
>Lars.
>
Michael G. Roche
Legato Systems (Canada) Inc.
3390 South Service Road
Burlington, Ontario CANADA L7N 3J5
vmail: (905) 637-4701 fax: (905) 637-4730 email: [log in to unmask]