On Sat, 19 Nov 2016, Jo-Philipp Wich wrote:
Hi all,
I am currently on working on automating the required steps to cut a
proper release - in order to make this process as standardized and
repeatable as possible, I intend to script any required steps to avoid
the need for manual setup tasks as much as possible.
Ideally I want to reach a point where one runs a "make_release.sh" with
just two arguments: a release number and a code name.
In order to achieve that goal, we must ensure that any related resources
like download URLs, Git repositories etc. are using predictable names
which can be constructed from the version number (or the nickname) alone.
Below is a list of things I'd propose for automating release cutting,
please let me know if you agree or think otherwise.
REPOSITORY PREPARATION
1) Branch "source.git" and name it "lede-$version"
2) In the used package feeds, branch "lede-$version" and use it
does it really make sense to make a branch instead of just tagging commits? Is
there really an expectation that there will be different changes in a release
branch vs the mainline (especially for feeds, I can see an argument for the
lede source in that you may backport some fixes into a release branch)
Other than this, it looks reasonable.
David Lang
_______________________________________________
Lede-dev mailing list
Lede-dev@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/lede-dev
