[IPython-dev] IPython 5.0.0 beta 1 now available !

Matthias Bussonnier bussonniermatthias at gmail.com
Thu Jun 9 13:57:03 EDT 2016

Hi Matthew and Aaron,

In IPython you can use the following:

(rootenv)~/dev/ipython master$ ipython

In [1]: import a<tab>
                 abc                alabaster          appnope
   argparse           ast                asynchat           atexit
                 aem                anaconda_navigator appscript
   array              astroid            asyncio            audioop
                 aifc               antigravity        argcomplete
   arrow              astropy            asyncore           autoreload

Not that the completion is still technically in a dropdown anchored
close to the cursor, so does not take the real full width of the
terminal especially if the current line is long.
We’ll try to improve that.

The Ctrl-L shortcut seem to be a prompt_toolkit default as I seem to
be able to use it in IPython as well, though I need to focus the
completer by pressing tab first.

Hope that helps.


On Thu, Jun 9, 2016 at 10:33 AM, Aaron Meurer <asmeurer at gmail.com> wrote:
> The way xonsh (which also uses prompt_toolkit) handles this is much
> better. The dropbox is full width, with many columns, and even fills
> up the whole terminal if you ctrl-L.
> Aaron Meurer
> On Thu, Jun 9, 2016 at 1:27 PM, Matthew Brett <matthew.brett at gmail.com> wrote:
>> Hi,
>> On Mon, Jun 6, 2016 at 10:08 AM, Matthias Bussonnier
>> <bussonniermatthias at gmail.com> wrote:
>>> On Sun, Jun 5, 2016 at 2:12 PM, klo uo <klonuo at gmail.com> wrote:
>>>> - When I invoke completion, dropbox shows modestly with 5 items. Is it
>>>> possible to set number of items shown to preferred value?
>>> yes:
>>> c.TerminalInteractiveShell.space_for_menu=<int>
>> I'm finding that the dropbox is less useful to me than the the
>> previous IPython behavior that gives me a pageful of options.  For
>> example, I often find myself searching through the methods and
>> attributes of an object for something that looks useful, and there are
>> often 20 or so methods / attributes on an object, so it is useful to
>> do:
>> import numpy as np
>> np.<tab>
>> and get the whole page of possible completions, with multiple columns.
>>    On the other hand a dropbox of that length is harder to grok.   Is
>> it possible to get the multiple column completion list instead of the
>> dropbox?
>> See you,
>> Matthew
>> _______________________________________________
>> IPython-dev mailing list
>> IPython-dev at scipy.org
>> https://mail.scipy.org/mailman/listinfo/ipython-dev
> _______________________________________________
> IPython-dev mailing list
> IPython-dev at scipy.org
> https://mail.scipy.org/mailman/listinfo/ipython-dev

More information about the IPython-dev mailing list