Re: [p4] Missed Integration

Posted on behalf of forum user 'Sambwise'.



[https://forums.perforce.com/index.php?app=forums&module=forums&section=findpost&pid=22599]
DClyde, on 2017/11/01 01:03:04 UTC, said:
>    It is my understanding that perforce requires an integration to occur to a file regardless of if there was a change or not. If it was committed, and then integrated even if the file is identical the record of the integration is needed. That Branch 2 is up to date as of Branch 1 commit 123. Even if there are changes that you do not want, you do the integration and then "Accept Mine" there is no "not doing anything."
>     
If this were strictly true, you would never be able to get two branches up to
date with one another, since each time you integrate you create a new revision
(which would need to be integrated back in the other direction if the
integration engine weren't smart enough to take the origin of each revision
into account).

This *is* a true statement if you're talking strictly about plain
'edit' revisions -- the actual content of revisions is examined during
resolve, not during integrate, so if there are no integration records indicating
otherwise, a plain edit is assumed by integrate to contain some actual change.



Quote
>   All of that said, bigger issue here isn't how I am resolving my integrations. It is that perforce is being inconsistent in its behavior. Given that the changelist in question is the last change to the file, and that the target branch was up to date before said changelist it should give me the same results in relation to that file regardless of which I do (assuming the same integration settings.) Integrating head for that file, integrating head for that branch, or integrating that changelist that happens to be head should all result in that file being in my local changelist an integrate.
>     
For a plain edit I'm pretty sure that makes sense, but if the changelist in
question were an integrate from somewhere else (or maybe an undo?) I'd
imagine you could easily get different behavior.  There are subtleties
around certain edge cases where a credit range gets expanded on one integrate to
avoid something more confusing (or harmful) in a later integrate.  If
I saw the exact case you're looking at I could probably explain in more
detail, but when speaking in generalities it's easier to say "trust the
magic".   :)



--
Please click here to see the post in its original format:
  http://forums.perforce.com/index.php?/topic/5528-missed-integration
_______________________________________________
perforce-user mailing list  -  [email protected]
http://maillist.perforce.com/mailman/listinfo/perforce-user