Re: bash convert: new completion system skeleton?

Re: bash convert: new completion system skeleton?

[Clemens Fischer]

John Buttery <john@io.com>:

>   I second that; I realize that zsh's new zstyle stuff is more powerful
> inherently than bash's system, but still I see almost-daily updates to
> the "bash-completion" project on freshmeat with new stuff and it would
> be great to be able to use it.

sorry, i've been googling and even searching freshmeat (returns more
then 100 bash-related projects!) for this "bash-completion" project.

could someone please post the URL?  maybe i can send them something
they don't already have.

i do almost all repetitive tasks with completers, because my memory
serves me not, and my typing is adequate (to the memory).  i start
with aliases, sometimes make them into scripts, and finally i make
the completer.  couldn't live without it, and once you get the hang
of it, it's quite easy.

btw, has somebody pointers to the "Reactive Keyboard"?  i have a
severly hacked version that doesn't want to run on my freebsd, and
the version i could make to run was slow and the code was ugly.  but
its prediction (by looking at the history of what a user types)
worked astonishingly well.

  clemens

Re: bash convert: new completion system skeleton?

[Felix Rosencrantz]

--- Clemens Fischer <ino-qc@spotteswoode.de.eu.org> wrote:
> sorry, i've been googling and even searching freshmeat (returns more
> then 100 bash-related projects!) for this "bash-completion" project.
> 
> could someone please post the URL?  maybe i can send them something
> they don't already have.
Not sure why you didn't find it, maybe the dash:
Home page: 
http://www.caliban.org/bash/index.shtml#completion
Freshmeat project page:  (Contains comments, including some by Oliver and other
Zsh-workers)
http://freshmeat.net/projects/bashcompletion/?topic_id=253%2C884%2C861



> btw, has somebody pointers to the "Reactive Keyboard"?  i have a
> severly hacked version that doesn't want to run on my freebsd, and
> the version i could make to run was slow and the code was ugly.  but
> its prediction (by looking at the history of what a user types)
> worked astonishingly well.
The most active version seems to be: ( I didn't look too hard.)  The latest
development on this version is from March 2000.
http://www.csoft.net/~dummy/robert/software/rk.shtml

It is originally from University of Calgary by John J. Darragh and Ian Witten.
They have some papers and a book as part of a PhD.  Though there hasn't been
much development by the original developers, since the early 1990s.
Some links:
http://pharos.cpsc.ucalgary.ca/Dienst/UI/2.0/Describe/ncstrl.ucalgary_cs/1989-343-05
http://www.cs.waikato.ac.nz/~ihw/publications.html

It would be great if someone wanted to make the prediction code a zsh module so
prediction could be native to zsh.

-FR.

-FR.


__________________________________________________
Do you Yahoo!?
Yahoo! Mail Plus - Powerful. Affordable. Sign up now.
http://mailplus.yahoo.com

Re: bash convert: new completion system skeleton?

[clemens fischer]

Felix Rosencrantz <f_rosencrantz@yahoo.com>:

> Not sure why you didn't find it, maybe the dash:
> Home page: 
> http://www.caliban.org/bash/index.shtml#completion

i used the "search OSDN" function on freshmeat, looking for "bash
completion".  this had more than 100 hits, i gave up too early,
obviously.

> http://freshmeat.net/projects/bashcompletion/?topic_id=253%2C884%2C861

... and i remember such a comment.  then i got distracted with the
problem of FM not recognizing my source-forge UID  :)

> [about reactive keyboard]

> It would be great if someone wanted to make the prediction code a
> zsh module so prediction could be native to zsh.

that would be a very good idea!  of course the one taking a stab at
this should first try it out.  of somebody finds a _portable_
version, i'd be glad to at least report on which predictions return
the most value for the users.  actually, the version i once had was
way too slow, but i think markov-chains are much better suited for
todays XXXX-GHz CPUs otherwise idling away ...

  clemens

Re: bash convert: new completion system skeleton?

[Clemens Fischer]

Felix Rosencrantz <f_rosencrantz@yahoo.com>:

> http://www.csoft.net/~dummy/robert/software/rk.shtml
> http://pharos.cpsc.ucalgary.ca/Dienst/UI/2.0/Describe/ncstrl.ucalgary_cs/1989-343-05
> http://www.cs.waikato.ac.nz/~ihw/publications.html
>
> It would be great if someone wanted to make the prediction code a zsh module so
> prediction could be native to zsh.

why didn't anybody mention `predict-on'?  i'm having oodles of fun
with it, and it's _the_ improvement i've been looking for.

laziness sometimes make me searching so hard for this sort of thing
that it begins to resemble real work (oouuch).

  clemens