[p4] Algorithm for Generating a Build Report
Jeff A. Bowles
jab at pobox.com
Fri Nov 7 23:34:55 PST 2008
If the two labels are against the same codeline, certainly you could use "p4
files @label1" and "p4 files @label2" to create the list of all revisions
that exist between the two labels, and use the likes of "p4 files
specific-file#4" to find exact change numbers. It's hardly efficient,
however, running the risk of machine-gunning a server with lots of
itsy-bitsy requests. (Think of the Lilliputians tying down Gulliver, and
you have the picture.) If it's a one-time report, have at it, but you might
consider running it outside of development checkin-times if you are
comparing large trees over many revisions. (It depends on your data and your
Is there a reason why you believe that the changes -- from other branches --
are uninteresting? It sounds like they *did* contribute to the current line,
so it might be worth asking more specifics on that part of the topic.
On Fri, Nov 7, 2008 at 4:58 PM, Scott Marshall <Scott.Marshall at citrix.com>
> Hi there,
> I was wondering what others do to produce a build report using Perforce.
We have a script that takes the output from a "p4 changes -i" command run
against two labels and diffs the output to determing the changelists that
are part of the current build. Where this breaks down is the "-i" argument
is causing Changelists from other branches to be included in the report
since there can be a shared file history between the branches. The report in
this case misrepresents what exactly ended up in the build.
> Is there another algorithm that can/should be followed to ensure the
report is more accurate?
> I appreciate any help/insights that you can provide.
> Scott Marshall | Build Engineer
> Citrix Online Division Phone: 805.690.5832
> Citrix Systems, Inc. Fax: 805. 690.6471
> 6500 Hollister Avenue Email: scott.marshall at citrix.com
> Goleta, CA 93117 USA Web: www.citrix.com
> perforce-user mailing list - perforce-user at perforce.com
Jeff Bowles - jeff.a.bowles at gmail.com
More information about the perforce-user