3
votes

Situation: I have seen a practice where SVN branches are created from branches and keep on going. For example: Branch1 was created from TRUNK; Branch2 from Branch1; etc. See below to get an idea:

....goes on and on.......Branch4--from--> Branch3--from--> Branch2--from--> Branch1--from-->Trunk.

So each new branch is created from the latest branch; instead of creating it from Trunk or Main.

Questions:

  1. Is this kind of practice faulty?
  2. Will this cause issues?
  3. Is this kind of practice followed anywhere?

Update: First and third questions have satisfactory answers. Thanks for the help. Are there any issues to be wary of in particular while following this method?

1
SVN is a tool that can support very different use cases, and there might be use cases for which this is a good solution. That means: First describe your use cases. Then it's easy to explain, how you can support your use cases with SVN.nosid
I see. I just wanted to make sure that the practice is not wrong per se. I get that depending on the requirements this can be right (or not). Thanks.sivv

1 Answers

0
votes

Is this kind of practice faulty?

No.

Will this cause issues?

No new types of issues for "here and back" merges, compared to "trunk-mediator" case, but (possible) *N old types (where N is amount of synchronized branches)

Is this kind of practice followed anywhere?

Probably yes - it's rather common case and doesn't have anything exotic and irrational


Note: I think, using SVN 1.8 automatic merges give you a noticeable amount of benefits in the amount of needed operations in final reintegration, comparing to pre-1.8 merge --reitegrate