Re: Sticky tags

[Eric Siegerman]

On Thu, Jul 12, 2001 at 08:26:24AM +1200, Chris Cameron wrote:
> We use non branch sticky tags for preserving 'contours' through our code
> (e.g. release 1.0, integration build 2, etc.).  This is very usefull for
> determining changes from one 'release' to another and also for ensuring that
> we can always deliver the same 'release'.

That's a non-branch tag (aka release tag); I agree that they're
extremely useful for this among other things.  It's *not* sticky,
though -- more accurately, you likely don't use the tag in a
sticky fashion.

"Stickiness" isn't an attribute of the tag itself; it refers to
the fact that when you do "cvs update -rTAG", CVS keeps track of
that in your sandbox and changes its behaviour accordingly.  The
point of the thread is that this behaviour change is useful when
TAG happens to be a branch tag, but is (usually) just annoying
when TAG happens to be a release tag.

When I referred to "sticky non-branch tags", that was shorthand
for "the sticky behaviour of 'cvs update -rNON-BRANCH-TAG'".
That's a common idiom on this list (and perhaps in the CVS
documentation), but it's misleading nonetheless.  My apologies.

--

|  | /\
|-_|/  >   Eric Siegerman, Toronto, Ont.        address@hidden
|  |  /
With sufficient thrust, pigs fly just fine. However, this is not
necessarily a good idea.
        - RFC 1925 (quoting an unnamed source)