info-cvs
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: Edit archive to correct corruption?


From: Matthew Rich
Subject: Re: Edit archive to correct corruption?
Date: Thu, 15 Nov 2007 16:46:42 -0800 (PST)

Just to clarify, cvs log ends with an error messages and does not finish.  In 
fact if I'm running it recursivly is dones not complete the command after the 
error messages.  Check outs complete with no error menssages  and the files 
compile just fine.  The issue is with accessing the file history\log info.

----- Original Message ----
From: Arthur Barrett <address@hidden>
To: Larry Jones <address@hidden>
Cc: Todd Denniston <address@hidden>; address@hidden
Sent: Thursday, November 15, 2007 4:41:26 PM
Subject: RE: Edit archive to correct corruption?

Larry,


> > 
> > > or will the CVS version of validate_repo work with CVSNT's 
> > > format?
> > 
> > No idea - I suspect there would be a good chance it'd consider some
> > things as errors.
> 
> I don't think so.  It's a fairly standard perl script that uses normal
> CVS commands to examine the files.  It does need direct access to the

In that case it could run but not tell you what you want to know, see
below...

> It runs "cvs rlog" to get the revision history for each file and then
> runs "cvs co" on the "interesting" revisions (the root of the 
> trunk and
> the tip of every other branch); as long as the checkouts 
> work, the file
> is considered OK.  

Without going back to the messages, I thought Matthew (who reported the
problems) said that the log completes OK (just with the odd messages)
and that checking out the versions log complains about also works ok -
so my guess is that the script would also run ok (since it's not looking
for specific 'odd' messages - just return codes).

CVSNT (for better or worse) seems to often return success where CVS
returns fail - it confuses the heck out of people who are moving from
one to the other, but CVSNT old-timers rely on the CVSNT return codes
making for a lovelly catch-22...  I think the logic is that if any of
the checkout has succeeded then the checkout returns success (just some
bits have failed?).  From memory update does one thing and checkout the
other...

Regards,


Arthur




reply via email to

[Prev in Thread] Current Thread [Next in Thread]