[Circle-dev] Tab-complete in frontends

Philip Kovac pkovac at cs.uml.edu
Wed Mar 18 16:55:03 GMT 2009


> Not so much combinations, just simply a tree of commands -> option names
> -> option value enumerations. Obviously only the enumerations are
> particularly completable; either fixed ones, or ones based on some
> running data list (channels/nicks/etc...)

The problem is where there is interaction between options in the general
case. Flags changing what are valid targets, etc. But yes, I suppose for a
lot of things that may work just fine.


> 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?
>
> The problem is that lazy population will make the UI feel "sluggish." If we
do this we should demand it all upfront and populate/update the information
as we would anything else, or turn it off entirely. The "middle case" where
bandwidth is scarce but roundtrips are cheap/fast doesn't seem terribly
common. If anything, you usually have the case where bandwidth is cheap and
plentiful but you have high latency (satellite broadband, "good" cellular
links.)

To see exactly what I mean about "sluggish," use modern bash intelligent
completion on an older machine. It's absolutely maddening.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.leonerd.org.uk/pipermail/circle-dev/attachments/20090318/a0f2a321/attachment.htm>


More information about the Circle-dev mailing list