No, I got that. My point was that while the initial error may have been bogus,
after that I can see a way the error might have continued. It was actually
telling me the file was not synced AND it thought it would need to resolve a
conflict (possibly because of the earlier error the integration record was not
up to date).
The file is fine, and everything is acting normally. It’s a curiosity at this point.
I had done a nearly identical branch operation two hours earlier using the
same source branch and even more files without any issues. (The main
difference was the earlier branch was new and this one did contain some
files in the target from earlier partial integrations – not this particular file,
though.)
-Wm
From: Gabor Maghera [mailto:***@gmail.com]
Sent: Thursday, July 15, 2010 4:10 PM
To: Ivey, William
Cc: perforce-***@perforce.com
Subject: Re: [p4] Has anyone else seen this bogus integration error
I think I failed to emphasize that the use case I described ran subsequent integrations (within the same user+client configuration), without submitting between them (and the first integration used the -v flag). It was an attempt at refreshing the pending changelist with modifications to the source codeline since the first integration was ran.
Sounds like you are experiencing something different. I'd take a look at what p4 verify reports for this one file, but I am sure you've tried that already.
--
Gabor
On Thu, Jul 15, 2010 at 9:41 AM, Ivey, William <***@bmc.com<mailto:***@bmc.com>> wrote:
The error originally occurred during step 1, and the subsequent
attempts were all without the -v, but from what you say below
that doesn't matter, it was the initial branching with -v that
set up the continuing error.
The curious part is that while it said it couldn't integrate the
file, it actually did. The submit of the first integration added
the file to the target branch as expected.
-Wm
-----Original Message-----
From: Gabor Maghera [mailto:***@gmail.com<mailto:***@gmail.com>]
Sent: Thursday, July 15, 2010 11:23 AM
To: Ivey, William
Cc: perforce-***@perforce.com<mailto:perforce-***@perforce.com>
Subject: Re: [p4] Has anyone else seen this bogus integration error
I have seen this in the following situation:
1. Integrate using -v
2. Let some time go by, and do not submit pending change from step 1
3. Integrate again
If any file was branched by step 1, but the source of the integration
has changed during step 2, P4V will try to update the file during step
3, but the file is not available locally due to the -v step flag used
in step 1.
I contacted Perforce support about this and they advised not to use -v
in cases where the integration might have to be repeated before
submitting.
Not sure if this is what you had experienced, but here it is. In my
case the behavior was consistent between P4 and P4V.
--
Gabor
Sent from my iPhone.
Post by Ivey, WilliamI was doing a mass integration (>180,000 files) and in one directory of
maybe 20 files all of them integrated correctly except one which
"can't integrate file branched with integrate -v"
So I tried it by itself without the -v (the P4V equivalent) with the same
result. Thinking it was one of those cases where P4V had lost track of
reality, I tried it from the command line, same result.
Went to a different system and tried it. Worked OK but when I tried to
submit it, I was told it needed to be resolved. Sure enough, the original
integration had worked, even though it appeared to have produced the
error above. That explains subsequent errors, but not the error from
the original successful integration.
The only thing I could think of was a conflict between a directory and
link stored somewhere in the path (I've seen a few of those, our
users are a bit sloppy with links) but that didn't seem to be the case here.
-Wm
_______________________________________________
http://maillist.perforce.com/mailman/listinfo/perforce-user