I don't think there is a specific "release process" document yet. The
approach has been to triage issues (not necessarily just bugs; also
documentation, etc.) with a milestone target, and release when by consensus
the important issues (including features) for a specific target have been
resolved.
Issues are discussed and prioritized on the Github issue tracker:
https://github.com/julialang/julia/issues
There is a fairly extensive test suite, targeting both language
functionality and numerical accuracy, and including regression tests for a
number of non-trival bugs. The tests are run via Travis on every commit:
https://travis-ci.org/JuliaLang/julia/builds
(if your question is really "when will 0.3 be released", the unscientific
answer is "very soon")
On Mon, Apr 28, 2014 at 9:48 AM, Ryan Gardner <[email protected]> wrote:
> Can anyone point me to something that describes or briefly describe the
> process for determining/ensuring that a release is stable. A few sentences
> is fine.
>
> Is there essentially a large set of test cases that are run on the code
> before the release it made, while those test cases aren't run on the
> nightly builds? How are those test cases chosen/written/obtained? How are
> unresolved issues determined to be acceptable, or is there an effort to
> resolve all bug-related issues? Thanks.
>
