Re: Are we (nearly) ready for 3.4 yet?

[Søren Hauberg]

ons, 15 12 2010 kl. 05:06 -0500, skrev John W. Eaton:
> Looking back in my mail archive, I see we discussed releasing 3.4
> about a year ago, then again in July, then yet again in September.  So
> now you must be thinking that if I start to talk about a release, it
> is probably just talk.  But really, I would like to make a release
> relatively soon.  I think it is way overdue.  We are seeing a lot of
> reports for bugs that have already been fixed.  We have a lot of new
> features that people would probably like to be using.

I agree.

> So what
> important things need to happen before we can create the 3.4.0
> tarball?  My unordered list is
> 
>   * Take one last shot at editing and updating the NEWS file.
> 
>   * Update copyright years.

You actually need to do that? I did not know that.

>   * Add @DOCSTRING entries (and more descriptive info if possible) for
>     all functions that are currently missing from the manual.

I seem to remember Rik creating a script that could detect if a function
did not have a corresponding @DOCSTRING. Is this true or did I invent
that?

> Some things that would be nice to have but not essential if they can't
> or don't happen:
> 
>   * Other documentation edits/improvements.
> 
>   * Reduce the number of open bug reports.  I'm not concerned with
>     trying to close all or even most of the current set of reports
>     before the release, but we should try to fix critical problems,
>     regressions, or problems that many people will likely notice.
>     Browsing the list of open reports, it looks like many are graphics
>     related, and then there are a few obscure problems or feature
>     requests.  Are there any absolutely critical bugs that must be
>     fixed before a release happens?  Maybe the problem with indexed
>     assignment and empty arrays (bug #31287).  Are there any critical
>     plotting problems?
> 
>   * Merge John Swensen's imread/imwrite changes.

This would be nice indeed.

>   * Make FLTK plotting work well enough to use by default.  I doubt
>     this is a reasonable goal for a time horizon of a few weeks, so I
>     don't expect it.  But I think this should be a high priority for
>     the next stable release after 3.4.

I too doubt that this is a reasonable goal for 3.4.

> Anything else?

I would like to see the www.octave.org/missing.html page added to the
web server before the release.

Is it somehow possible to see the list of open bug reports where a patch
has been attached. I think it would be nice if we could apply as many of
such patches as possible.

> I could also list a lot of features that I would like to have, but
> none are likely to happen in a few weeks, so it would probably be best
> to not delay a release much longer.  We could do that forever (as it
> seems we already have).

I think we should more or less go into feature-freeze, i.e. don't accept
new features until after the release.

Soren