Re: [GNU ELPA] New package proposal: aggressive-completion.El

[Tassilo Horn <tsdh@gnu.org>]

Philip Kaludercic <philipk@posteo.net> writes:

>>> Just tried out the package, and it seems to work very well! I just
>>> have two annoyances:
>>>
>>> 1. Where there are too many options, lag is noticeable (eg. with C-h o).
>>
>> There's a knob for that: `aggressive-completion-max-shown-completions'.
>
> Sadly it still persists. All in all it is barley noticeable, and it only
> affects the first letter in my case.

What do you mean it affects only the first letter?  I mean, obviously
that's where it is slowest because the number of completions is the
largest you can get for that letter, no?

> Maybe you can replicate it:
>
>         C-h o
>         a
>         [wait]
>         
> on my system there is about half a second of lag between typing a and
> a appearing on the screen.  If I typo more letter, they are all
> delayed by half a second.

Yes, because aggressive-completion (completion or completion help) kick
in after `aggressive-completion-delay' seconds.  The default value is
0.3.  You might want to try a smaller delay for added snappiness but I
guess that will become inconvenient in other case, e.g., when you
mistype.

>>> 2. The completion buffer keeps appearing and disappearing, shrinking
>>>    and growing.
>>
>> The *Completions* buffer should appear as soon as there is at least
>> one and at most `aggressive-completion-max-shown-completions'
>> completions, and it'll disappear when completion selected a unique
>> match.  Do you see something different?
>
> No, but if I remove a word and the completion is not unique any more,
> the buffer re-appears.

Yes, hopefully, that's intended.

>> Wrt. the resizing: that's the standard completion help behavior you'd
>> also see when double-TAB-ing during completion.  Not sure if there's
>> a way to restrict that window's height.
>
> Yes, but my point was that the change in window height is in response
> to a manual completion request. This might just be me, but I try to
> avoid "undeterministic" behaviour, or at least effects that I cannot
> infer from my actions.

Well, aggressive-completion is pretty much such a bird with spiral
spring legs hacking its beak onto your TAB key every 0.3 seconds after
you've edited your input. ;-)

> I cannot say for sure, but I think that if you could force the
> completion buffer to stay at one place during the entire completion,
> this would be preferable.

I can understand that desire but I'm not sure if
aggressive-completion.el is the right place for such a feature.  IMHO,
the standard *Completions* buffer window's size should be customizable
directly in emacs if at all...

Oh, well, does `temp-buffer-resize-mode' with its limits
`temp-buffer-max-height' and `temp-buffer-max-width' already do what you
want?

Bye,
Tassilo