Received: from [192.168.0.73] (root@raven.ravenbrook.com [193.82.131.18]) by raven.ravenbrook.com (8.9.3/8.9.3) with ESMTP id CAA10333 for ; Sat, 28 Oct 2000 02:13:09 +0100 (BST) Mime-Version: 1.0 X-Sender: gdr@pop3 Message-Id: Date: Fri, 27 Oct 2000 18:13:10 -0700 To: p4dti-staff@ravenbrook.com From: Gareth Rees Subject: Alpha test report: Perforce, 2000-10-27 Content-Type: text/plain; charset="us-ascii" ; format="flowed" This is a report on alpha testing P4DTI release 0.3.2 (with some patches) at Perforce. We showed the integration to Karen, Albert, Wally, Michael, Chris, Nigel, Laura, Kathy, Leigh. These are the defects that came up and the comments of the testers: 1. The replicator can't cope if a job appears in the output of "p4 logger" but doesn't exist. (It fails an assertion.) It should do something more sensible. 2. Many assertions aren't really assertions, but runtime errors. Make it so. 3. It might be sensible for the replicator to check that it can ping the smtp server (say). But then we won't be able to demonstrate it offline? (But we can, by setting 'smtp-server' to None so that it doesn't send e-mail at all.) 4. In TeamTrack, if the transition buttons fill up the issue pane then the actions drop down and the help button get lost (in IE 5). 5. Why doesn't the VC History section show up? Because the user doesn't have it set in its preferences. 6. When we added a fix to a job, the replicator doesn't actually manage to replicate the fix until or unless you change a field in the job. Why is this? Is the fix being added to the database but with the wrong RECID or something like that? I have no idea. 7. When no transition is found, we shouldn't replicate but should throw an error anyway. 8. When the replicator sets the issue/job action to "wait", it's also making changes that it shouldn't be allowed to do. 9. Reports. 10. There was a mysterious error in the TeamTrack event log. Probably a bug in TeamTrack: the offending query isn't from the replicator. See the event log for details. 11. If you do a "generic copy", it copies the P4DTI_ fields. This is wrong. They should go back to their default values. You can do this in the TeamTrack administrator, but it would be better for the replicator to do it automatically (for all copy transitions?) 12. Dates aren't replicated properly when fixes are replicated. 13. If you press Control-C while the replicator is busy replicating, then it doesn't stop the replicator. This is because the replicator has a try... except wrapper around the replication and it catches all errors. We shouldn't do that: we should catch only errors that we expect to happen. KeyboardErrors (or whatever) should be propagated through to the top level. 14. When data is about to be thrown away (by overwriting after a conflict is resolved) we shouldn't just e-mail it to the person concerned, we should log it as well. This will make sure no data is lost if e-mail is down, or the address is wrong, or the recipient trashes the message. It also makes it possible for the administrator to find it. 15. What we're doing in the integration is pushing the defect tracker's security and enforcement into Perforce jobs. This is a product feature that we should document, both as a policy decision and in the FAQ. See Paul Goffin's e-mail. 16. TeamTrack doesn't work with Netscape 4.5 on Unix or Windows NT. You can't go Find -> General -> All (or any other link or button in the Find window). 17. We need to agree on the support protocol with Perforce and TeamShare support. Who gets which calls? When does it come to Ravenbrook? Should there be an agreement in place for Ravenbrook to support the software for the medium term? Can Perforce support send us e-mail for example? Needs sorting out. Any need to know who the integration expert is on the TeamShare? Get back to Perforce support on that. 18. Perforce support would like a beta release early so they can test it and get used to it. 19. What's the right thing to do with P4DTI-filespecs? Currently it's "default" but there's no default. This is wrong (but p4 jobspec doesn't diagnose the problem). I could make it optional, but then it wouldn't show up in the output of p4 -G job -o, but this could be got around with suitable changes to the replicator. I should supply a default of the empty string in quotes: "".