[p4] fstat digest of text file (CALL#2327524)
Jeff A. Bowles
jab at pobox.com
Fri Mar 27 07:05:09 PDT 2009
If *I* needed to compare locally generated files against the server's copy,
I would make sure that I could account for RCS keywords and a number of
similar things such as end-of-line.
The checksum idea has merit: md5-checksum the local content, then surf
through the server's stored-checksums and compare. (I have written a small
script for other purposes, that compares an MD5 checksum against a tree,
hoping to find where the file was retrieved from.)
For this specific case, have you considered the "sync" option that updates
database tables but doesn't retrieve content? That, plus a "p4 diff"
afterwards, would give you the diff information you seek with less content
traveling back/forth, and would deal with binary/text/RCS-text/unicode
better than many other solutions.
It would be more useful if you described WHY you are comparing... against
the server. I expect that the reasons is, "see if these files I just
generated need to be checked in because they have changed."
On Fri, Mar 27, 2009 at 8:24 AM, <tmcd at panix.com> wrote:
> On Fri, 27 Mar 2009, Sergey Klibanov <sklibanov at day1studios.com> wrote:
> > Benjamin, you asked why... Well, I have an automated process where I
> > need to compare locally-generated files against their head revision
> > in the depot. The wrinkle is that for a variety of reasons the files
> > I need to compare aren't synced on the machine running this
> > operation, and may even reside some place other than where the
> > client spec thinks they should be. I looked at p4 help diff for a
> > bit, but found no way to get it to report diffs for un-synced and
> > badly-located files.
> Would your problem be solved by "p4 print -q ... > another_local_file"
> and calling a local diff program on another_local_file versus your
> locally-generated file?
> I wonder whether the Perforce maintainers would consider extending
> "p4 diff2" to allow a local file as an argument.
> says that both regular arguments have to be depot paths.
> Tim McDaniel; Reply-To: tmcd at panix.com
> perforce-user mailing list - perforce-user at perforce.com
Jeff Bowles - jeff.a.bowles at gmail.com
More information about the perforce-user