8 messages in com.perforce.perforce-user[p4] checking file contents in a trigger| From | Sent On | Attachments |
|---|---|---|
| debr...@peoplesoft.com | 03 Jan 2001 17:37 | |
| Jeff A. Bowles | 03 Jan 2001 17:56 | |
| Blum, Robert | 03 Jan 2001 20:28 | |
| Paul Goffin | 04 Jan 2001 00:51 | |
| Slootweg Paul | 04 Jan 2001 01:22 | |
| Blum, Robert | 04 Jan 2001 07:54 | |
| Blum, Robert | 04 Jan 2001 08:01 | |
| Diane Holt | 04 Jan 2001 09:40 |
| Subject: | [p4] checking file contents in a trigger![]() |
|---|---|
| From: | Blum, Robert (rbl...@midwaygames.com) |
| Date: | 01/04/2001 08:01:33 AM |
| List: | com.perforce.perforce-user |
We've implemented a review demon which parses SUBMITTED files for the sort of things indicated below.
Interesting thought, but if we do not enforce coding style, 'diff' will report differences just because somebody forgot to reformat to the agreed-upon format.
We're also considering implementing a second stage of the review which actually corrects the problem - where the "correction" can be automated - by checking out the file, doing an automatic update and submitting the change. We're still thinking this through, though...
That would mean lots of more or less meaningless change lists in the history.... I'd rather reject it right away.
Use of the review system has one advantage over the use of triggers - it can cope with the "I know I've not run the XYZ tool on this file, but we must check it in now as the customer needs it RIGHT NOW!" problem.
I am in the lucky position to be able to refuse 'RIGHT NOW' fixes. Every fix that goes out has to be tested...
I a trigger is used, then the code can't go in at all and this increases the risk that it will be shipped when no controlled version exists...
Only if you're not following a controlled process. And if you (like us) need a special build machine for the final product, you minimize that risk even more. It's really hard to just do a quick hack with that system. Sometimes bad, but mostly beneficial. (Been there, felt the pain :-)
(Of course, this could be managed by Perforce implementing some form of "trigger override" mechanism. - but that seems unlikely for now.)
And it's pointless - a trigger is a policy you have in place. If you can override policies just because, you might as well have none. (If you really need it, your trigger can always check for a semaphore - but don't tell anybody :-)
Bye, Robert
Disclaimer:
Please be aware that messages sent over the Internet may not be secure and should not be seen as forming a legally binding contract unless otherwise stated. Furthermore, I am not speaking for Midway Games in any way - this is a strictly personal message.




