Re: Question about version numbers.

[stremler]

Jon Miner <address@hidden> wrote:
[snip]

> No, you do just like Greg said:

Did you bother to /look/ at the effect of that sequence of commands? No?

>     cvs co -r bar_branch -d QUX-Bar QUX
>     cvs co -r baz_branch -d QUX-Baz QUX
>     cvs co QUX

And this gives me three directories instead of one (well, one and a link,
but the directory contains a lot more), cluttering up my filespace. And
there are other unwanted effects as well (f'instance, since the module
name is well-known, there are scripts and other utility programs that
can be run in the checked out tree that can determine where they are
by walking up the path until they encounter the module name. If they
couldn't do this, I'd have to hard-code it to a particular configuration,
which sucks, or rely on a slew of environment variables, which would 
suck, or do it all by hand, which would -- you guessed it -- suck!).

> Although, I have no idea why you would want to.  In general, people work
> on one branch at a time.  If they have changes that they need to make on
> all the branches, they make them and merge the changes (or something
> like that.)

I don't have the good fortune of working in 'just one branch at a time'.

> Maybe I'm missing your point, but perhaps you're misusing branches.  In
> general, branches are used for prod/dev revisions, and or releases and
> bugfixes..

Perhaps.  I'm using branches to keep developers from interfering with
one another as they work on related pieces, without having them go off
for a week or two without making any commits. 

My customer has never needed me to go back and fix a bug or otherwise
enhance an "old" release. 

-- 
 -----------------------------------------------------------------------------
  I AM NOT A MAN! I AM A FREE NUMBER!          |    address@hidden
                                               |       Stewart Stremler
  Wait a minute....                            |