Re: Bug-cvs Digest, Vol 52, Issue 2

[Mark D. Baushke]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

David O'Brien <obrien@FreeBSD.org> writes:

> On Fri, May 02, 2008 at 12:01:23PM -0400, bug-cvs-request@nongnu.org wrote:
> > It does the same thing even without the rm.  Updating to a particular
> > revision (whether numeric or symbolic) *always* sets a sticky tag to
> > that revision.  What makes you think BASE should be special?
> 
> 1. Its rather useless if that's proper behavior.
>    'cvs up -r BASE' should give you back the same file you had checked
>    out - for instance if you hacked the file to pieces and just want
>    the original file back.
> 2. mdb said so.

Hmmm... I thought I said

   cvs update -rBASE -p foo.c > foo.c

would make you lose all of your foo.c changes and get you back the
version you want. Both of the commands:

  cvs update -rHEAD foo.c
  cvs update -rBASE foo.c

will make a sticky revision that is fixed at the current top-of-branch
version (-rHEAD) of foo.c or a frozen version of the current revision
(-rBASE) of the branch of foo.c ...

I think there may have been some talk of the proposed special symbolic
tag modifiers:

  .head .base .commitid .prev .trunk .origin .root .next

that might have the impact of sometimes not making static sticky tags.
I do not recall of the details at present.

        -- Mark

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.6 (FreeBSD)

iD8DBQFIHjbnCg7APGsDnFERAmIPAJ9nshJ8t7YRypOp0GqlpEE3X2iUqACg2YxQ
nIYQ5afn2FGm7oHVI7cHo4o=
=v0y3
-----END PGP SIGNATURE-----