[Top][All Lists]

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

Re: CVS export

From: John Minnihan
Subject: Re: CVS export
Date: Fri, 19 Oct 2001 06:01:59 -0700
User-agent: Mozilla/5.0 (Windows; U; Win98; en-US; rv:0.9.2) Gecko/20010628

address@hidden wrote:

In message <address@hidden>
          address@hidden (John Minnihan) wrote:

In all the years I've used CVS, I've institutionalized use of 'cvs export' only briefly at one client. And that was only to satisfy some nimrod (who didn't understand cvs) who screamed - repeatedly - "get rid of all these CVS files!". Presence - or absence - of the CVS admin structure hasn't had the slightest impact (positively or negatively on its own merit)on the success of failure of any of the thousands of releases I've produced. Period. 'cvs co' with a known tag is the way to go.

I agree that I prefer 'cvs co'.  We run automated builds of several products
and prefer to have incremental updates to the source tree (thus requiring
checkout not export).

Plus I seem to recall that in CVS 1.10.8, you cannot export directories that
have default log messages specified in rcsinfo (you get an error about not
being able to create CVS/Entries files). We would use 1.11.1p1, but we've had
to revert to 1.10.8 as CVS 1.11.1p1 dies with protocol errors all over the
place during the nightly checkouts.

I had the 'protocol error:uncounted data' checkout failures too. The frequency of the errors increased as the repository grew. I took a gamble that it was less of a CVS-caused error and more of a CVS reaction to the system. This was after reading the source, and groking it only a little bit. After adding more RAM (very inexpensive these days) to my repository server, the problem hasn't reappeared.

John Minnihan

reply via email to

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