4 messages in com.perforce.perforce-user[p4] Integration question| From | Sent On | Attachments |
|---|---|---|
| Williams, Kenneth (Ken) (TLR Corp) | 28 Sep 2004 15:10 | |
| jab | 28 Sep 2004 15:43 | |
| Robert Cowham | 29 Sep 2004 03:19 | |
| Williams, Kenneth (Ken) (TLR Corp) | 29 Sep 2004 09:01 |
| Subject: | [p4] Integration question |
|---|---|
| From: | Robert Cowham (rob...@vaccaperna.co.uk) |
| Date: | 09/29/2004 03:19:54 AM |
| List: | com.perforce.perforce-user |
You're on the right track. Of course your life would be easier if as part of your rename "process" you made a point of updating the branch spec as you go (or not if you want to propagate the rename) - isn't hindsight wonderful...
Something that could provide you with a starting point is Tony Smith's integrename.rb script. See:
http://public.perforce.com/guest/tony_smith/perforce/P4Rubylib/index.html
Please note that the new ancestor detection of r04.2 does a great job with this scenario. If you update your branch spec to not propagate the rename, then any edits get propagated to old files without any requirement to use the -i flag. This is very nice and removes some painful steps!
I am interested if anyone has tested this new branching feature on a repository of a decent size with lots of branching history (just to get an idea of performance comparison).
Robert
Wondering about a good way to do a tricky integration. I've got a codeline at //mainline/foo/ that I created a release branch for, at //release/foo/2.2/ . After doing that, the code in //mainline/foo/ got heavily rearranged for version 3.0 (it's code from an external group), which I painstakingly implemented as branch/delete operations using P4Win's "rename" command. Now I've made a few changes in //mainline/foo/ that I'd like to propagate to //release/foo/2.2/ *without* propagating the renames.
Is there some way to do this automatically? Will I need to set up a branch spec that includes all [relevant] mappings between old locations and new locations, or is there some way to avoid this? The mapping is pretty complicated, and I've got no record of it except the actual branch/delete operations in the perforce history.
Any suggestions or alternate ways of thinking about this are appreciated.