MRB does not initialise GitFlow
mrb gitCheckout --fork larcoreobj end up in a
larcoreobj local repository that is not initialised to GitFlow.
LArSoft has abandoned the GitFlow model for a pull request model, so it is not clear to me whether the one in this report is an intended change.
For sure, it makes it more complicate for an experiment user who needs to follow GitFlow command patterns on some repositories but not on others.
Especially since the experiment repositories checked out with MRB also do not get GitFlow initialisation.
Could you clarify MRB and LArSoft stance on the matter?
In principle, enabling GitFlow has no adverse effect.
If LArSoft or MRB decides or has decided that this is the intended behaviour, could you advise on how experiments/users should support GitFlow on their side?
#1 Updated by Lynn Garren about 1 month ago
The only reason to enable gitflow in a clone of a github repository would be to make it easy to create a feature branch. LArSoft has not taken an official stance, but we have not been requiring that pull requests be feature branches. The documentation, however, still advises feature branches.
Thank you for pointing this out. We will clarify current best practices.