We have two branches: 1. HEAD - latest version (AKA Trunc) 2. PROD - released version
When you fix bugs in the released version, which one do you make:
- Correct it in PROD, then merge in HEAD
- Correct it in HEAD, then merge in PROD
The advantage (1) is that in this way you absolutely cannot harm the released version by accidentally delivering unverified code from HEAD (it is assumed that PROD is always more stable / verified than HEAD).
The advantage of (2) is that for a part of the code in HEAD there may be more usages than in PROD, therefore, if you correct only in PROD, you may not find all such usages and thus copy errors in HEAD.
I personally for (1). What do you think?
source
share