Re: I miss tags

On Fri, 24 Sep 2004 15:34:31 -0500, Sean Laurent <sean@neuronfarm.com> wrote:
> A typical scenario we encountered follows:
>
> 1) Install team generates Build #42 of product XYZ based on revision 1000 and
> produces a label, associating XYZ_BUILD_42 with revision 1000.

Install team generates a build by making a copy into
"project/builds/42" and building that source

> 2) Developers continue to make changes to the branch.

Developers are off on trunk, or their own branches.

> 3) QA team reports a nasty, difficult bug in Build #42 that gets assigned to
> me.
> 4) At this point, the repository is up to revision 1100. To simply my ability
> to reproduce and track down the bug, I checkout XYZ_BUILD_42.

svn co the stuff at project/build/42 ?

> Yes, this ~can~ be accomplished using branches, but it's not exactly the same
> approach. I don't need the repository to store a _copy_ of the source tree
> at revision 1000. I just need to be able to easily get there, WITHOUT having
> to track down the revision number.

I don't see any revision numbers involved above...

> Otherwise, we have to have documentation
> showing the mapping between build numbers and revision numbers.

No, just a reasonably organized repository setup.

(I'm sure this has been thought of before, but...) Thinking of a tag
in subversion terms is really an association of a particular {URL,
revision #} tuple with a name, isn't it? While this could be adopted
pretty easily (I suspect) by svn itself, you could version control
this with scripts and get a pretty reasonable first cut at it.

Personally, we're finding here at work that the copy model means we
can move patches around for specific regressions a lot easier and that
whole tag thing isn't being missed much at all.

.chris

---------------------------------------------------------------------
To unsubscribe, e-mail: users-unsubscribe@subversion.tigris.org
For additional commands, e-mail: users-help@subversion.tigris.org
Received on Fri Sep 24 22:49:18 2004