Re: [Jython-dev] Planning a Release Candidate for Jython 2.7.0

[Jim B. <jim...@py...>]

All urgent bugs have been fixed, along with some bugs that should have been
marked as such. I would like to highlight the last fix, by Santoso Wijaya,
in https://hg.python.org/jython/rev/b597f478d5ab:

Introduce and use an overloaded non-buffered readinto() method
> for BufferedReader when #read(int size) is given an explicit size to read
> from its underlying stream.


This changeset fixed a problem with relative seeks of files, and how that
interacted with buffering; see http://bugs.jython.org/issue1793

We are now ready to build the 2.7.0 release candidate. Expect this to be
available very shortly!

- Jim

On Tue, Mar 10, 2015 at 3:46 PM, Jim Baker <jim...@py...> wrote:

> There is in fact light at the end of the tunnel! ;) Just in time for PyCon
> next month, I do believe we are on the verge of a release candidate. The
> release candidate is our first attempt at what we consider to be a
> production, stable release for 2.7.0.
>
> Here's my current list of urgent bugs, that is bugs that will block our
> first attempt at a release candidate for Jython 2.7.0:
>
>    - http://bugs.jython.org/issue2120 - Creating directories with UNC
>    paths on Windows - in general, we should attempt to convert all usages of
>    the Java Native Runtime where possible to Java 7's NIO2, or at least try
>    NIO2 first, because JNR simply is flaky on Windows.
>    - http://bugs.jython.org/issue2282 - No compilation step to produce
>    $py.class files for stdlib for nondev targets - this impacts standalone
>    usage, due to how it interacts with the class loader. Also for non
>    standalone usage, not precompiling the standard library unnecessarily slows
>    down start up time, possibly every single time depending on permissions.
>    - http://bugs.jython.org/issue2283 - site.py should not fail if
>    python.home is not set
>
> Compared to what has been done so far, these should be easy bugs,
> especially #2283. I anticipate completion in the next few days.
>
> There are also a number of urgent bugs that are fixed but are currently
> pending. Of these, probably support for Werkzeug (
> http://bugs.jython.org/issue2272) and fixing a deadlock seen at shutdown
> between the shutdown closer thread and the finalizer thread (
> http://bugs.jython.org/issue2280) were the most important that have been
> recently fixed.
>
> The most important thing we can do now is make sure we have triaged the
> bugs properly. Take a look at http://bugs.jython.org/ - if the bug is not
> urgent, it does not block this release candidate, no matter how worthy the
> requested fix or feature is. But if it should, please, please add your
> comments to the bug so this can be potentially reprioritized.
>
> As we have time available, we may look at pulling in a few more bug fixes,
> and possibly some work on the installer to do such setup tasks as running
> ensurepip. But this must be balanced against wrapping up the release
> candidate.
>
> Naturally remaining bugs are on our backlog for 2.7.1, which we will be
> sprinting on at PyCon in Montreal, during the sprint days April 13-16.
>
> - Jim
>



-- 
- Jim

jim.baker@{colorado.edu|python.org|rackspace.com|zyasoft.com}
twitter.com/jimbaker
github.com/jimbaker
bitbucket.com/jimbaker