Received: from martin.ravenbrook.com (martin.ravenbrook.com [193.112.141.241]) by raven.ravenbrook.com (8.9.3/8.9.3) with ESMTP id VAA12326 for ; Tue, 5 Dec 2000 21:25:05 GMT Received: from [193.237.68.82] (grouse.ravenbrook.com [193.112.141.254]) by martin.ravenbrook.com (8.8.8/8.8.7) with ESMTP id VAA26503 for ; Tue, 5 Dec 2000 21:17:53 GMT (envelope-from gdr@ravenbrook.com) Mime-Version: 1.0 X-Sender: gdr@pop3 Message-Id: Date: Tue, 5 Dec 2000 21:24:34 +0000 To: p4dti-staff@ravenbrook.com From: Gareth Rees Subject: Test report for release 0.4.0 Content-Type: text/plain; charset="us-ascii" ; format="flowed" This is a report on testing release 0.4.0 on 2000-12-05. 1. In the AG, section 5.2, Bugzilla configuration parameters should be in a typewriter font. 2. In the AG, section 5.2, the replicator_address configuration parameter should come after the replicated_fields paramater (the parameters should be in alphabetical order except for rid and sid). 3. I wonder if "rid" should be "replicator_id" and "sid" should be "p4_server_id". There's no need to be gratuitously cryptic. 4. In section 5.4.1 we recommend people create a "VC Update Group" registry key. I think it would make sense to remove this, because we don't support editing, adding or deleting fixes or filespecs from TeamTrack anyway (and John McGinley says he's going to remove this feature from build 4404). 5. The screen shot for the "Add User: General tab" shows the P4DTI user without a user name or e-mail address. Should we specify these? 6. A bug in replicator.py prevents it sending e-mail when something goes wrong. Nick's fixed this in the master sources in change 5459. There needs to be a general method "send this e-mail to all the people associated with this job and issue". 7. The manual uses an em-dash surrounded by spaces for punctuation. Is that right? Should you use an en-dash surrounded by spaces, or else an em-dash not surrounded by spaces? 8. The replicator was unable to replicate a changelist because of a bug in replicator.py (p4 module not imported so the p4.error on line 337 was not recognized). 9. I created a pending changelist in Perforce. I selected "Add Job Fix" to associate the changelist with a job. The changelist and fix were replicated to TeamTrack. I then realised that I was in the wrong client, so I reverted the change, selected "Unfix Job", and deleted the changelist. I then made an illegal change to the job in Perforce (that is, one that couldn't be replicated to TeamTrack), so the replicator tried to replicate back to Perforce. But it could not replicate the fix back to Perforce because the changelist had been deleted (the replicator got an error from Perforce of the form "Change 84 unknown"). The replicator needs to notice this case when replicating fixes from the defect tracker to Perforce and at least it should avoid falling over by catching the error. What's the right thing to do here? Maybe we shouldn't replicate fixes to pending changelists? We should at least make it clear in the DT interface that the changelist is pending, otherwise it looks like it's been done. 10. The replicator started replicating from the beginning of time in TeamTrack, not from the time it was installed (see job000047). Fixed in version 0.4. 11. I was using P4Win 2000.1 17732 with P4D/NTX86/2000.2beta/18911. I tried to submit a pending changelist with an attached fix. I got the following error message: "Error in change specification. Wrong number of words for field 'Jobs'." The same thing happened with 2000.2 18934.