[LLVMdev] Develop on trunk.

[Andrew Trick]

On Jun 13, 2012, at 7:15 AM, Sergei Larin <slarin at codeaurora.org> wrote:

> Again, this is branch code, but if I can see something wrong in platform
> independent portion, I'll track it on the trunk.


I hate to be the one to start this thread, but I have to say it at least once...

General advice. Working off trunk will be less painful than cherry-picking. When it comes to merging, one revision at a time is the only sane approach. If you run sanity tests after merging a set of new trunk revisions, before checking them into your development branch, you can avoid dealing with the rare, temporary breakage, or "instability" as people like to say. If the breakage lasts more than a day, then it's specific to your branch and you want to find out about it ASAP. It will be much harder to fix close to the next release.

For example, I just checked in changes to AA scheduling. If you run into a problem with it, I want to know about it now, as in this week. It will be a much bigger burden for me change it again next week, or any time after that.

I know for some reason people like their own releases to be "based on llvm release X". That could work if your release cycle is longer than llvm's (after branching your own development you could continue pulling llvm changes until next llvm release). For frequent releases, it's nonsense. Either way, your development branch should still follow trunk.

-Andy
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.llvm.org/pipermail/llvm-dev/attachments/20120613/41ba84c4/attachment.html>