[Circle-dev] Tab-complete in frontends

[Kiyoshi Aman]

On Wed, Mar 25, 2009 at 9:15 AM, Alastair Douglas
<alastair.douglas at gmail.com> wrote:
> 2009/3/18 Paul LeoNerd Evans <leonerd at leonerd.org.uk>:
>>
>> That was roughly where I was aiming, possibly via some laziness. Perhaps
>> roundtrip is "good enough" to lazily populate the options to a command
>> the first time the command's name gets completed?
>>
>
> This conversation is circling the word 'cache', which makes the thin,
> dumb client no longer a) thin or b) dumb.
>
> How much thin and dumbness are we prepared to sacrifice?
> Tab-completion is such a commonly-used function of thicker clients
> that it would be folly to omit it, but as you say, we can't poll the
> server to answer the request every time you press tab.
>
> Perhaps we could add another tab press to it, similar to the
> double-tab that shows you a list of options. First tab = I want to
> tab-complete; Second tab = I mean it; Third tab = show me the options.

I feel that this is suboptimal; when I hit tab, I expect either an
immediate response (irssi, bash) or a beep indicating multiple
responses are possible (bash). It would be convenient (and, I feel,
more optimal) if the client (which already knows all the objects)
could be intelligent enough to assemble tabcompletion trees as data
comes in for other things and as the user tabcompletes.

...And I'm confused as to why I'm the only person to raise this
suggestion when that's what 'thick' clients do.