Re: preventing disturbance of branch tags.

[Stephen Cameron]

Hi folks,

I've updated my patch to "cvs tag/rtag" with sanity.sh test cases, ChangeLog
entries, cvs.texinfo changes...  I think it's done, but more eyes looking it
over couldn't hurt.

It's here:
http://www.geocities.com/dotslashstar/branch_patch.html#tagpatch
Or the patch itself:
http://www.geocities.com/dotslashstar/tagpatch.06.06.2001.txt


--- "Derek R. Price" <dprice@collab.net> wrote:
> "Derek R. Price" wrote:
> 
> > Stephen Cameron wrote:
> >
> > > "cvs tag -F -b -B" may convert non-branch tags to branch tags
> > > and vice versa.
> 
> I hope and expect that you meant '"cvs tag -F -b -B" may convert non-branch
> tags
> to branch tags and may move branch tags to new branches' here, but I'm still
> reading.

Yes.  I have put a big table into cvs.texinfo which describes all the
permutations of previous tag state (branch/non-branch/non-existent), and -F,
-d, -b, -B (r)tag options and what they do.

Let me know if you see any problems...



BTW, 

Some section about how to go about properly moving/renaming/recovering a branch
tag via "cvs (r)tag -b..." + "cvs admin -n" might be useful.  I didn't see
anything like that in cvs.texinfo, but there is an example of recovering a
branch tag I saw in sanity.sh in the tagf tests (which is where I added my
tests).

I'm not sure I understand it all well enough to write such a section though.
My -B was mainly intended to allow warnings to be printed in its absense. 
Presense of "-B" just allows access to the previous behavior which I wasn't
sure I was comfortable removing.

I was also thinking it might be cool to have a program (or extension to cvs?)
that could help with recovery of lost branch tags (or check integrity of
existing ones).  You might say:

cvs branch_tag_check branch_tag_foo [...some other options go here...]

with the "some other options" telling the program/cvs extension certain known
facts about the branch_tag_foo, such as non-branch tags that should be on this
branch, branch tags that pre-date it/post-date it, etc.  Then it could try to
figure out what revision(s) could possibly be the ones from which
branch_tag_foo could conceivably sprout, and report those revision(s) if
branch_tag_foo were missing, possibly repair it, if only one revision were
possible and directed to do so, or report inconsistency if the branch tag
existed but not on a revision consistent with the stated facts.  Just an idea,
possibly infeasible.

Well, supposing I were to attempt that, I'd probably be able to write the
section in cvs.texinfo describing how to go about recovering a branch tag...
 
[...]

-- steve

__________________________________________________
Do You Yahoo!?
Get personalized email addresses from Yahoo! Mail - only $35 
a year!  http://personal.mail.yahoo.com/