[IPython-dev] 0.11 update

[MinRK]

Hello,

Another update on currently outstanding issues marked for 0.11, so we
can see how we are doing. I think we are quite close, and
documentation is really where the focus needs to be this week.  I'm
sure there are several places where docs and examples refer to stale
0.10 APIs or patterns.

critical issues:
---------

#281: pyreadline doesn't appear to play nice with the matplotlib qt
backend on Windows.  I don't know anything about how to address this
issue.
    Jörgen said that it does not affect the qtconsole (logical,
because the qtconsole doesn't use readline).

    This is a pretty big deal, but it doesn't *technically* appear to
be an IPython issue, though IPython is where interactive matplotlib
and pyreadline meet each other.
    It does *only* affect qt+terminal IPython with pyreadline, though.

#66: What's New.  It appears to be fairly up to date
(https://github.com/ipython/ipython/blob/master/docs/source/whatsnew/development.txt),
but anything devs notice as missing should be added. We can leave this
open as a reminder until the last minute.

#8: mayavi+qt - This appears to be fixed by the recently released
ETS4, but Evan said he will test to be sure

#269: mainloop() API - this code seems to be ready, so the issue now
just involves a doc update, and the embedding IPython docs and
examples don't appear to have been touched since 0.10, so they
obviously need to be revisited.  I've started working on this, and
I'll push updates probably this evening.

high:
-----

#440: qtconsole crash on %run completion on Linux+EPD when run from a
launcher without a terminal window - since this is solved by requiring
the qtconsole to have a terminal window, I'm not too concerned with
letting this stand in 0.11, though we should definitely figure it out.

#378: Windows path issues - Windows path completion appears to have
never been particularly good when using absolute paths and/or
backslash escapes.  I thought this was a bad regression, but when I
tried testing with 0.10, it doesn't appear to be any worse as far as I
can tell, though I am not well experienced with Windows.  Since fixing
this would be a major improvement over 0.10, rather than fixing a
regression, I'm okay with slipping to 0.12.  Any Windows developers
who can explore this would be much appreciated.


At this point, I think all medium and low priority issues will not
make the cut for 0.11, unless someone feels the need to take care of a
few of them.  Anyone who thinks we should block on one of these,
should suggest it be promoted to high/critical.  I don't think letting
these slip is a huge deal, since we plan to have a relatively quick
succession of releases as bugs are inevitably revealed once we
actually get 0.11 out the door.

I think we should be able to cut 0.11 (or possibly 0.11 rc1) by this
Friday, 07/01.

-MinRK