I'd like to hear what techniques you've found successful for fixing bugs in a previous release while simultaneously
a) Working on the next release b) Propagating the bug fix to the next release c) Not propagating new code with bug fixes for the previous release
I'm currently labeling each build (thanks to Anthill), but I'm wondering if I need to make use of branching (one for the previous release, one for next release) as well. The way we did this at the my last job feels to manual, and error prone:
1) Fix the bug in the main/only branch 2) Extract all files with revision label (or date) of previous release 3) Extract the current revisions for the file(s) containing the bug fix, overwriting the versions obtained in step 2. 4) Run the build script.
The big problem I see is that the collection of versions that I've just assembled has no label/tag, or easy way to get back to it if another bug is found. It seems like it would be cleaner to have a maintenance branch for the previous release, where all changes were immediately merged to the main branch. The point of the branch would be to not pick-up changes for the next release, while maintaining the latest bug fix versions of the previous release. I'm hoping it would then be a one-step/automated process to 'build maintenance branch', by supplying a label or something else to denote what revisions to build.
Does this sound right, or can I achieve what I'm aiming for with the current labels/tags I've been assigning without branching?
Previous Topic
