[cvsnt] vendor branch vs. HEAD

Bo Berglund Bo.Berglund at system3r.se
Fri May 14 12:56:43 BST 2004


Don't import to the vendor branch then...
There are command switches for cvs import for just that use (only available with CVSNT).
See here: http://www.cvsnt.org/wiki/CvsCommand#head-62fdfbd55d19b2a4671102ad7bca17d875f8207a

Then the module you import will not have the vendor branch and initially all files are revision 1.1.

/Bo


-----Original Message-----
From: cvsnt-bounces at cvsnt.org [mailto:cvsnt-bounces at cvsnt.org]On Behalf Of Dianne Britton
Sent: den 14 maj 2004 07:11
To: cvsnt at cvsnt.org
Subject: [cvsnt] vendor branch vs. HEAD


When I create a module via "import", I'm expecting to have a HEAD branch (trunk) rooted at 1.1 for every file in the module, but I don't actually get one unless I modify the file.   This has repercussions for branching, as follows.

Say I have just two files in my newly created module, foo.txt and baz.txt.  Let's say I modify baz.txt and commit it, so it is at revision 1.2.  Now I create a new branch with tag LOCAL_BRANCH, and move my local copy to the branch via "update -r LOCAL_BRANCH".  Let's say I now modify foo.txt and commit it.  Now here's the counter-intuitive part.  The revision number for foo.txt is now 1.1.1.1.2.1.  Doing a "status" shows that LOCAL_BRANCH is 1.2.2 for baz.txt and 1.1.1.1.2 for foo.txt.  Thus LOCAL_BRANCH is coming off the trunk in one case and off the vendor branch in the other case!  This is counter-intuitive, to say the least.

I've only used CVSNT, so I don't know if it works this way in CVS.  Regardless, I think this might be a problem, though perhaps practically-speaking it doesn't make any difference, despite its counter-intuitiveness.  Any opinions?

? Dianne  



More information about the cvsnt mailing list