5 messages in com.perforce.perforce-userHow to extract changes between labels...
FromSent OnAttachments
Raje...@fore.com22 Apr 1998 07:10 
Fred...@mydata.seFredric.Fredriksson23 Apr 1998 03:06 
Raje...@fore.com23 Apr 1998 08:19 
Matt...@ftlsol.comMatthew.Rice23 Apr 1998 10:09 
Nave...@vinayak.engr.sgi.com23 Apr 1998 12:50 
Subject:How to extract changes between labels or dates?
From:Raje...@fore.com (Raje@fore.com)
Date:04/23/1998 08:19:51 AM
List:com.perforce.perforce-user

Thanks for a very detailed answer.

But I still think that this could be just a simple linear diff of (maxchg # at latest label) - (maxchg # at older label), and listing of changes between these two changes (since that is already supported).

For example, this simple shell script does it for me now.

#!/bin/sh

if [ $# != 2 ]; then echo "Invalid usage" echo "Usage: $0 @label1 @label2" exit 1 fi

ochg=`p4 changes //depot/us/tag/4.2/...$1 | head -1 | awk '{print $2}'`

nchg=`p4 changes //depot/us/tag/4.2/...$2 | head -1 | awk '{print $2}'`

if [ $ochg -gt $nchg ]; then p4 changes //depot/us/tag/4.2/...@$nchg,@$ochg else p4 changes //depot/us/tag/4.2/...@$ochg,@$nchg fi

So, I am not sure why it might be difficult for perforce to support this usage.

Just a thought.

rv

-- using MH template repl.format -- In a previous message, Fredric.Fredriksson at mydata.se writes:

p4 changes //depot/us/tag/4.2/... at L_tag52_official_release, at L_tag52_stable_windows_nt Has anyone ever faced this before? I would think that something like the above construct would be most useful to extract fixes, and prepare release notes, etc. Any thoughts?

my 0.02$:

I solved this with a perl script that lists changes for both labels and filter out the ones that differ.

When I played around with this I realised that unless the labels does not contain the exact same set of files you can not get a simple list of changes that "differ". What You get is three lists:

* changes common to both labels * changes in the first label * changes in the second label

And this is not the whole truth: Lets say I submit a change affecting files A and B. If A is included only in the first label and B in the second I get this change listed as affecting both labels. (In reality this is not a common problem, though)

My conclusion is: this is an extremely useful function I use it a lot _but_ there is more to it than what you see at a first glance.

/Fredric Fredricson