[p4] Partial branch integration causes problems

Jeff A. Bowles jab at pobox.com
Wed Mar 11 11:10:28 PDT 2009


You are, effectively, trying to do two things in parallel, with a tool that
will do one, or the other, but not both at the same time.

If you keep changelists simple: one bug fix or one feature modification per
checkin, you can play the game you're doing. It sounds like you want to pull
over part-of-the-checkin now, and part-of-it-later, for the same checkin of
the same file.

Changelists will actively fight this particular approach, and rightly
so.It's a case where you could either change the way you develop the
original
work (meaning, check in at the granularity you hope to eventually use as
integration-sources) or you change the integration workflow (perhaps, doing
a integrate-resolve-test-revert followed by the larger integration for the
final work / checking).

   -Jeff Bowles




On Wed, Mar 11, 2009 at 6:13 AM, Finn Normann Pedersen
<merenpha at gmail.com>wrote:

> Hi,
>
> Thank you all for the nice answers.
>
> To Dan Wilder: That may indeed be the problem - I will take that up
> with people, but unfortunately it doesnt solve our problem.
>
> To Chuck: That is the problem. A changelist may contain multiple files
> and each file may contain multiple changes, BUT only some of the
> changes within the files is wanted at first - the remainder of the
> changes is wanted integrated later on.
>
> In the example I recently had it is an interface definition which
> multiple functional definitions. The first test only focus on some
> some of the functions so the remaining functions are left out to not
> disturb the first test.
>
> Cheers,
>   Finn Normann Pedersen
>
>
>
> On Tue, Mar 10, 2009 at 8:50 PM, Dan Wilder
> <Dan.Wilder at eng.watchguard.com> wrote:
> >
> > If you find yourself having to integrate part of a changelist in order
> > to have something QA-able, think about whether your changelists are
> > well-targeted on a single topic, or whether they're big omnibus
> > wrap-up-everthing-one-developer-did-in-a-week things.
> >
> > It's possible you've got a mismatch between your QA requirements and how
> > your developers are working.
> >
> > Personally I favor narrowly-targeted changelists.  I'll fix one bug or
> > add one well-defined and free-standing portion of one feature, then
> > submit.   If it takes several days to fix one bug or add a feature I'll
> > submit more than once, provided the submits don't break the build or the
> > smoke test.  That is, each change must be one complete and consistent
> > unit of work.
> >
> > If I'm working on something else at the same time, I'm either very
> > selective about files to be submitted, or else I'll work in two or three
> > different work areas, one per bug or feature.
> >
> > The developers here all follow this, at least roughly.  Frequent
> > automatic smoketest builds with results emailed to everybody who has
> > submitted since last successful build, and also to a few key managers,
> > help enforce a discipline of don't-break-the-build.
> >
> > --
> > Dan Wilder
> >
> > -----Original Message-----
> > From: perforce-user-bounces at perforce.com
> > [mailto:perforce-user-bounces at perforce.com] On Behalf Of steve at vance.com
> > Sent: Tuesday, March 10, 2009 11:51 AM
> > To: merenpha at gmail.com; perforce-user at perforce.com
> > Subject: Re: [p4] Partial branch integration causes problems
> >
> > When you tell Perforce to force an integration, you're not only telling
> > it to do it despite having done it before. You're also telling it to
> > ignore any previous integration history and go all the way back to the
> > beginning in evaluating the diffs.
> >
> > Steve
> >
> > Original Message:
> > -----------------
> > From: Finn Normann Pedersen merenpha at gmail.com
> > Date:   Tue, 10 Mar 2009 14:12:12 +0100
> > To: perforce-user at perforce.com
> > Subject: [p4] Partial branch integration causes problems
> >
> >
> > Hi,
> >
> > I have a problem with branching that I hope some of you might know a
> > solution on.
> >
> > 1. I branch some files from branch A to B.
> > 2. Once work is done in B and I do a filespec integration from B to A,
> > but only some of the files are branched back into A to make testing A
> > more simple.
> > 3. Once testing is done on the partial changes I want to integrate the
> > rest of the changes in B back into A ... but now perforce tells me that
> > the integration is "done" and there is nothing to integrate. I then
> > force integration of file already integrated (check under options), but
> > now the perforce merge application tells me that the files are totally
> > different, even though only minor changes exist.
> >
> > Is this a perforce problem? A merge application problem? ... or me ;o)
> > ??
> >
> > Hope you can help !
> >
> > Cheers,
> >  Finn Normann Pedersen
> > _______________________________________________
> > perforce-user mailing list  -  perforce-user at perforce.com
> > http://maillist.perforce.com/mailman/listinfo/perforce-user
> >
> >
> > --------------------------------------------------------------------
> > mail2web.com - Microsoft(r) Exchange solutions from a leading provider -
> > http://link.mail2web.com/Business/Exchange
>  >
> >
> > _______________________________________________
> > perforce-user mailing list  -  perforce-user at perforce.com
> > http://maillist.perforce.com/mailman/listinfo/perforce-user
> >
>
> _______________________________________________
> perforce-user mailing list  -  perforce-user at perforce.com
> http://maillist.perforce.com/mailman/listinfo/perforce-user
>



-- 
---
Jeff Bowles - jeff.a.bowles at gmail.com



More information about the perforce-user mailing list