Within an SVN repo, I have an old branch (lets call it branch1) which is now about 2 months out of sync from trunk, and I now need to create a new branch (branch2) to extend the functionality of branch1.
It's very likely there are going to be conflicts and merged files, and due to the amount of time/work passed, I can imagine that an开发者_运维技巧 automatic merge is not going to be flawless, or maybe I should just have more faith.
Either way, I was wondering if anybody could provide any useful pointers for how I might tackle this scenario, and what sort of things I should watch out for, or will this simply be a case of simply doing the entire merge manually?
Thanks for any advice!
Generally, I run all my automated tests to ensure all is good. I then make a copy of the branches and trunk (e.g. copy/paste an up to date local repo). Then attempt an auto merge. Afterwords I rerun all the unit, integration and functional tests that are automated. If any thing is off that looks bad, I revert to original version and go at it by hand. Usually, if you got a horrible auto merge it won't even compile, let alone pass tests. Normally I decide to revert on a case by case basis assessing which ever looks less time consuming at that point (i.e. manually re-merge a few busted files, or redo the whole kitten caboodle).
精彩评论