[cvsnt] Re: vendor branch and import problem

Oliver Giesen ogware at gmx.net
Fri Feb 27 15:09:54 GMT 2004


Tony Hoyle wrote:

>>It also still allows me to import with a different branch name than what 
>>I used before without any error or warning, again messing up the 
>>repository (well, now that I know what's going on thanks to John Hoey 
>>it's actually rather easy to remedy but still...).
> 
> It's actually a lot harder to check for that.  As I said there's no
> code to generate the branch ID that I can find, so they all get number
> 1.1.1... thus messing up the repository.  

Couldn't you just check the existing symbolic name on the given branch 
ID (implying 1.1.1 - or better yet: the default branch) and let CVS barf 
on it if it doesn't match the given vendor branch name? AFAICT this info 
is right at the top of the RCS file, e.g:

head	1.1;
branch	1.1.1;    <<<< imply this if no explicit branch ID given
access;
symbols
	arelease1:1.1.1.1
	testvendor:1.1.1;    <<<< lookup name here
locks; strict;
comment	@# @;


I.e.:

1) receive "cvs import mymodule theothervendor_branch release1"
2) get default branch id, e.g. 1.1.1
3) lookup symbolic name for determined branch id
     (there should be only one), e.g. thevendor_branch
4) complain

or:

1) receive "cvs import -b1.1.3 mymodule theothervendor_branch release2"
2) lookup symbolic name for given branch id 1.1.3,
     e.g. theothervendor_branch
3) proceed


>>This doesn't feel quite finished... were you maybe talking about the 
>>upcoming 2.0.33 instead?
> 
> Umm... yea... that was it... :)

Thought so. ;)

Cheers,

-- 
Oliver
----	------------------
JID:	ogiesen at jabber.org
ICQ:	18777742	(http://wwp.icq.com/18777742)



More information about the cvsnt mailing list