On 23/10/16 06:58, Thomas Adam wrote: >> Is there going to be some sort of naming and usage convention around >> > branches, e.g. like Git flow¹? Or is 'master' where the action happens? > Did you read DEVELOPERS.md? Don't misunderstand me, it's in that document, > and if it's not clear, I again need to know so I can improve it. But what I > don't want to do is have to keep repeating the same bits of information when I > can point people towards that file.
No problems, no, I hadn't read it, as I'm mostly an end user, not a developer. I was mostly asking on the basis of figuring out how I obtain the latest stable release via git. On Git flow projects, master branch is always the latest stable release. One can set up their system to always pull in that branch, and they'll always get a stable release. Tornado do something similar: they just call theirs 'stable', and do development on 'master'. (Theirs isn't "git flow", but it works just fine.) We do this at work with projects that use this convention when we want to build Debian packages with them, our CI server just tracks their "stable release" branch (whether it be called "master", "stable" or whatever) and pushes new packages built off that, meaning we're always testing with the latest stable release of a package. For me with FVWM, it'd mean I'd set up Gentoo on my machines to pull that branch in an ebuild, and so "emerge fvwm" would pull in the latest stable release via Git, rather than the bleeding edge release. Tagged master is okay, but it does mean that whatever pulls the repository has to know which tag is the latest release, if that is what one intends. -- Stuart Longland (aka Redhatter, VK4MSL) I haven't lost my mind... ...it's backed up on a tape somewhere.
signature.asc
Description: OpenPGP digital signature
