[p4] [OT] recommendations for a code-review system

Grills, Jeff jgrills at soe.sony.com
Mon Mar 21 11:39:07 PST 2005


That's certainly a workable option.  You don't have to have the client
sync'd to head unless you want to be able to build it - it could remain
empty most of the time, and only have the files sync'd as they are being
reviewed.  If you allow developers to review at a time convenient for
them, then they could do reviews in their existing clients immediately
after submitting whatever pending changes they had.

Another thing you could do (that I just thought of) is have
per-developer branches, and review the submitted changelist.  Once the
changelist was reviewed and approved, it could get integrated to the
mainline.  That certainly could cause integration problems, though.

j

-----Original Message-----
From: Jay Glanville [mailto:Jay.Glanville at naturalconvergence.com] 
Sent: Monday, March 21, 2005 1:01 PM
To: Grills, Jeff; Perforce Users Mailing List
Subject: RE: [p4] [OT] recommendations for a code-review system

I'm assuming that the best way to use this solution is for everyone to
have a spare workspace with no files open for edit, and sync'ed to the
head.

This is the only way that I can conceive of preventing collisions.  For
example, if Bob sends me a tar'ed changelist for review, and also happen
to have a file open for edit that is in his changelist, then a collision
occurs between his version of the file and my changes.


JDG

> -----Original Message-----
> From: Grills, Jeff [mailto:jgrills at soe.sony.com]
> Sent: Monday, March 21, 2005 12:42 PM
> To: Jay Glanville; Perforce Users Mailing List
> Subject: RE: [p4] [OT] recommendations for a code-review system
> 
> We do code reviews here in a pretty casual way, so I can't comment on 
> the tools you mention below.
> 
> Here, if a reviewer can't sit down directly with the change author, 
> the author will use the p4tar tool to wrap their changelist up into a 
> single file and send it to the reviewer in email.  The reviewer will 
> then extract the change into a workspace, and it will look to the 
> reviewer exactly as if they had made the change.  They can use the 
> standard p4 diff tools to review the deltas, and provides feedback via

> email.
> 
> You can get the p4tar tools here:
> 
> http://public.perforce.com:8080/@md=d&cd=//guest/jeff_grills/&
> ra=s&c=TfC
> @//guest/jeff_grills/p4tar/?ac=83
> 
> j
> 
> -----Original Message-----
> From: perforce-user-bounces at perforce.com
> [mailto:perforce-user-bounces at perforce.com] On Behalf Of Jay Glanville
> Sent: Friday, March 11, 2005 10:40 AM
> To: Perforce Users Mailing List
> Subject: [p4] [OT] recommendations for a code-review system
> 
> Hello all.
>  
> Sorry for being off-topic.
>  
> We currently have a manual code-inspection process, and we're finding 
> the administration work due to this process is rather time consuming 
> (print off diffs, invite people to meeting, review the code at a pace 
> that everyone is comfortable with, etc, etc, etc).
> 
> Thus, we'd like to be able to replace the manual/meeting style code 
> review with some form of remote-review system.  Basically, what we're 
> envisioning is a system whereby a developer submits their uncommitted 
> changelist to a server, and then invites feedback from others.  These 
> reviewers review the deltas on their own time and provide appropriate 
> feedback.
> 
> My question is this: does anyone have any recommendations or feedback 
> on existing systems?
> 
> Some systems that we're currently investigating are things like 
> CodeStriker (http://codestriker.sf.net) and CodeReviewer 
> (http://www.codehistorian.com/codereviewer-overview.php).  Has anyone 
> used these systems?  Experiences?  How well could they work with P4?
> 
> 
> Thanks, and again, sorry for the off-topic request.
> 
> JDG
> 
> 
>  
> ---
> Jay Glanville
> Application Software Designer
> Natural Convergence
> 




More information about the perforce-user mailing list