16 messages in com.perforce.perforce-user[p4] Unwanted new timestamps after su...
FromSent OnAttachments
Max.Ritter08 Dec 2004 05:58 
Noel Yap08 Dec 2004 06:28 
Robert Cowham08 Dec 2004 07:36 
Grills, Jeff08 Dec 2004 08:15 
Noel Yap08 Dec 2004 08:19 
chri...@ni.com08 Dec 2004 08:34 
Arnt Gulbrandsen08 Dec 2004 08:48 
Ivaylo Beltchev08 Dec 2004 08:55 
chri...@ni.com08 Dec 2004 09:20 
Max.Ritter08 Dec 2004 09:32 
Kevin Wang08 Dec 2004 10:51 
Noel Yap08 Dec 2004 11:45 
Alen Ladavac09 Dec 2004 00:45 
Chuck Karish17 Dec 2004 08:16 
Max.Ritter17 Dec 2004 08:36 
Chuck Karish17 Dec 2004 21:38 
Subject:[p4] Unwanted new timestamps after submit
From:Ivaylo Beltchev (iv@adelphia.net)
Date:12/08/2004 08:55:22 AM
List:com.perforce.perforce-user

I think the real problem here is that the same option is used for two
independent things: 1) To use the submit time or the sync time when syncing. 2) To modify or not the file time when expanding keywords.

What makes sense for C++ files is 1) to use the sync time when syncing, 2) to
not modify the file time when expanding keywords. That's how Source Safe works
by default. It would be really helpful if Perforce can do the same thing.

Ivo

This will work if your file is ktext but you don't actually have any keywords in it.

Perforce should do a couple things to help this. First, it should check if the file has actually changed after the submit and only update it if it has. Second, it should attempt to fill in the keywords when you open a file for edit (and when you resolve conflicts). It could certainly get Author, revision, and file right at that time. This would give you the added benefit that you are building the same thing that actually gets submitted.

This doesn't work with all keywords.

-----Original Message----- From: perforce-user-admin at perforce.com [mailto:perforce-user-admin at perforce.com] On Behalf Of Robert Cowham Sent: Wednesday, December 08, 2004 9:37 AM To: 'Noel Yap'; 'Max.Ritter' Cc: perforce-user at perforce.com Subject: RE: [p4] Unwanted new timestamps after submit

Another alternative is to use a build system like scons (www.scons.org) which defaults to using MD5 digests of contents to detect file changes rather than timestamps...

It seems to me that every company using Perforce in connection with big C++ projects and RCS keywords would run into this problem. This makes me hope that somebody can recommend a solution :-)

Basically, what I am looking for is an option to update the file timestamp after a sync operation, but to not update it after a submit.

Thanks in advance, Max.