15 messages in com.perforce.perforce-usertwo questions!| From | Sent On | Attachments |
|---|---|---|
| Dave...@vignette.com | 16 Oct 1998 09:00 | |
| Dave...@vignette.com | 16 Oct 1998 09:42 | |
| Long...@mindq.com | 16 Oct 1998 09:44 | |
| Dave...@vignette.com | 16 Oct 1998 09:47 | |
| Mark...@pml.com | 16 Oct 1998 10:34 | |
| Rich...@netapp.com | 16 Oct 1998 10:48 | |
| Eric...@Adobe.COM | 16 Oct 1998 10:54 | |
| Tony...@informix.com | 16 Oct 1998 11:02 | |
| Robe...@netapp.com | 16 Oct 1998 11:25 | |
| Step...@vance.com | 16 Oct 1998 20:26 | |
| Gary...@home.com | 18 Oct 1998 17:30 | |
| Dave...@vignette.com | 18 Oct 1998 22:56 | |
| Pete...@esavionics.se | 19 Oct 1998 00:07 | |
| Ping...@mit.edu | 19 Oct 1998 01:35 | |
| Scot...@seanet.com | 19 Oct 1998 09:04 |
| Subject: | two questions! |
|---|---|
| From: | Mark...@pml.com (Mark...@pml.com) |
| Date: | 10/16/1998 10:34:54 AM |
| List: | com.perforce.perforce-user |
We have this problem, too. Our common code is ported to at least 6 different embedded platforms; all in the same codeline. The common code group wants to be sure that a change is tested on each platform before submitting it. We separate the platforms into different Perforce clients on their machine.
Other than branching, which we do for really risky changes, I don't have a good solution for this problem with Perforce.
---------------------------------------------------------------------- Mark_McIntosh at pml.com R&D Software Development Support Power Measurement Ltd. Voice: 250-652-7100, ext. 7571 2195 Keating Cross Rd. FAX: 250-652-0411 Saanichton, BC, V8M 2A5, Canada http://www.pml.com/
------Original Message----- From: Dave Lewis [mailto:dlewis at vignette.com] Sent: Friday, October 16, 1998 9:43 AM To: ldoan at mindq.com Cc: perforce-user at perforce.com Subject: Re: two questions!
> From: "Long Doan" <ldoan at mindq.com> > >Also, and more speculative, consider the following situation: > >Program is being developed for two platforms, so users typically > >have two clients. Then when making changes on one of them, they > >would like to test the changes on the other, without, of course, > >checking the changes in. Do you guys have any suggestions or > >techniques that make this less error-prone or troublesome? > >I remember somebody suggesting branching, but this doesn't > >sound too palatable to me. > ... > > Besides, it is usually a bad idea to work on code for two or more different > targets at the same time. The common code should be separated from the > target-specific code in the first place, so that each target can have a > reasonable degree of independence from the other targets. Then you can > evolve the code line on just target at a time.
This is, of course, our situation. It is common code, so when people make changes they want to test the changes on both machines *before* checking them in.
dave