Cygwin path completion

Cygwin path completion

[Andrej Borsenkow]

Cygwin has magic prefix ``/cygdrive'' that can be used to access any
drive under NT (obviating need to explicitly mount them).
``/cygdrive/d/foo'' is the same as ``D:\foo''. Currently Zsh cannot
complete path beginning with /cygdrive because it tries globbing and
this does not work (hmm ... may be, it may be considered a bug in
Cygwin):

mw1g017@MW1G17C:/cygdrive/h/src% ls /cygdrive
ls: /cygdrive: No such file or directory
mw1g017@MW1G17C:/cygdrive/h/src% ls /cygdrive/c
AUTOEXEC.BAT                SYSBUILD
Aladdin                     System Volume Information
......
mw1g017@MW1G17C:/cygdrive/h/src% ls /
PalmDev                 lib                     uninst.bat
bin                     login.README            usr
etc                     prc-tools               var
inetutils-1.3.2.README  tmp

(note current dir as /cygdrive/h/... - this is $HOME on server). Even if
we cannot find out all drives (is it possible?) _path_files still has to
treat /cygdrive/?/ specially, and not try to glob it. It can also always
offer ``cygdrive'' for the first component (of course, if it matches
current prefix/suffix).

-andrej

Have a nice DOS!
B >>

Re: Cygwin path completion

[Peter Stephenson]

Andrej wrote:
> Even if
> we cannot find out all drives (is it possible?) _path_files still has to
> treat /cygdrive/?/ specially, and not try to glob it. It can also always
> offer ``cygdrive'' for the first component (of course, if it matches
> current prefix/suffix).

Sven will have to answer for the feasibility of all of this, but...

One thing that might fit our needs here and elsewhere (e.g. speedups of
path completion when you know you don't want initial path components
re-jigged) is a style along the lines of fix-path-prefix, only maybe with a
better name.  A number would fix that many components (`2', here), while
something else e.g. `all' would turn off completion of earlier bits of the
path altogether.  A more generic `fix-prefix' might possibly be useful in
certain other completions.

-- 
Peter Stephenson <pws@cambridgesiliconradio.com>
Cambridge Silicon Radio, Unit 300, Science Park, Milton Road,
Cambridge, CB4 0XL, UK                          Tel: +44 (0)1223 392070

RE: Cygwin path completion

[Andrej Borsenkow]

> Andrej wrote:
> > Even if
> > we cannot find out all drives (is it possible?) _path_files
> still has to
> > treat /cygdrive/?/ specially, and not try to glob it. It
> can also always
> > offer ``cygdrive'' for the first component (of course, if it matches
> > current prefix/suffix).
>
> Sven will have to answer for the feasibility of all of this, but...
>
> One thing that might fit our needs here and elsewhere (e.g.
> speedups of
> path completion when you know you don't want initial path components
> re-jigged) is a style along the lines of fix-path-prefix,
> only maybe with a
> better name.  A number would fix that many components (`2',
> here), while
> something else e.g. `all' would turn off completion of
> earlier bits of the
> path altogether.  A more generic `fix-prefix' might possibly
> be useful in
> certain other completions.
>

While this may be useful, I meant something different. I'd prefer if
/c/d/t would still complete to /cygdrive/d/temp. I meant, that
_path_files (BTW I agree that it already deserves to be converted to
shell code. It may even give old compctl the ability to complete paths)
should simply stuff cygdrive as first component. In the above case the
possible completions would consist of

/c*/d*/t* AND
/cygdrive/d/t*

-andrej

Re: Cygwin path completion

[Peter Stephenson]

> I meant something different. I'd prefer if
> /c/d/t would still complete to /cygdrive/d/temp. I meant, that
> _path_files (BTW I agree that it already deserves to be converted to
> shell code. It may even give old compctl the ability to complete paths)
> should simply stuff cygdrive as first component. In the above case the
> possible completions would consist of

This is far too specific to be hard-wired into _path_files, but maybe
there's some way of specifying transformations.  I still don't understand
how that helps you since I don't see when you want cygdrive to be the first
component.  It seems to me you should be using a named directory or
something, e.g.

c=/cygdrive
: ~c
echo ~c/d/...

You can hardly get fewer characters than that.

-- 
Peter Stephenson <pws@cambridgesiliconradio.com>
Cambridge Silicon Radio, Unit 300, Science Park, Milton Road,
Cambridge, CB4 0XL, UK                          Tel: +44 (0)1223 392070

Re: Cygwin path completion

[Sven Wischnowsky]

Peter Stephenson wrote:

> Andrej wrote:
> > Even if
> > we cannot find out all drives (is it possible?) _path_files still has to
> > treat /cygdrive/?/ specially, and not try to glob it. It can also always
> > offer ``cygdrive'' for the first component (of course, if it matches
> > current prefix/suffix).
> 
> Sven will have to answer for the feasibility of all of this, but...
> 
> One thing that might fit our needs here and elsewhere (e.g. speedups of
> path completion when you know you don't want initial path components
> re-jigged) is a style along the lines of fix-path-prefix, only maybe with a
> better name.  A number would fix that many components (`2', here), while
> something else e.g. `all' would turn off completion of earlier bits of the
> path altogether.  A more generic `fix-prefix' might possibly be useful in
> certain other completions.

[Btw. I forgot to remove the optimation in the C-code again that makes 
initial existing directories be accepted immediately. We can't do
that, because, if you have `foo/x' and `foobar/y' and try `./foo/y<TAB>'
it won't complete to `foobar/y'. (It will work with `foo/y' for
another reason...) ]

Is the number really enough? But only if statting the thing shows that 
there is such a file, right? I think I'd still prefer patterns -- they 
are more flexible and can still made to match a certain number of
components.

Maybe we'll just have to try.


Andrej Borsenkow answered:

> While this may be useful, I meant something different. I'd prefer if
> /c/d/t would still complete to /cygdrive/d/temp. I meant, that
> _path_files (BTW I agree that it already deserves to be converted to
> shell code. It may even give old compctl the ability to complete paths)

I guess you mean C. And we started already. And _path_files (with some 
more optimisations I sent this morning) isn't really that slow anymore.

I, at least, will *not* do anything to help compctl. Never again.

> should simply stuff cygdrive as first component. In the above case the
> possible completions would consist of
> 
> /c*/d*/t* AND
> /cygdrive/d/t*

That's of course a completely different thing from the accept-thingy.

I know nothing about cygwin, but maybe a style for _path_files that
gives pairs of patterns and (sets of) strings. If the word on the line 
matches a pattern, the strings will be taken to be possible
completions. So you could do:

  zstyle '...:paths' fake-paths '/[^/]#:cygdrive' '/cygdrive/:a b c d'

Maybe even with ways to say `only that' and `this plus whatever you
find yourself'.

Bye
 Sven


--
Sven Wischnowsky                         wischnow@informatik.hu-berlin.de

Re: Cygwin path completion

[Sven Wischnowsky]

Peter Stephenson wrote:

> > I meant something different. I'd prefer if
> > /c/d/t would still complete to /cygdrive/d/temp. I meant, that
> > _path_files (BTW I agree that it already deserves to be converted to
> > shell code. It may even give old compctl the ability to complete paths)
> > should simply stuff cygdrive as first component. In the above case the
> > possible completions would consist of
> 
> This is far too specific to be hard-wired into _path_files, but maybe
> there's some way of specifying transformations.  I still don't understand
> how that helps you since I don't see when you want cygdrive to be the first
> component.  It seems to me you should be using a named directory or
> something, e.g.
> 
> c=/cygdrive
> : ~c
> echo ~c/d/...
> 
> You can hardly get fewer characters than that.

Right.


I forgot to say... the fake-path thing I just suggested could also
help us to do something I've been missing some time ago (and then
forgot again, of course, sigh): completion of auto-mounted directories.

Bye
 Sven


--
Sven Wischnowsky                         wischnow@informatik.hu-berlin.de

RE: Cygwin path completion

[Andrej Borsenkow]

> >
> > This is far too specific to be hard-wired into _path_files,
> but maybe
> > there's some way of specifying transformations.  I still
> don't understand
> > how that helps you since I don't see when you want cygdrive
> to be the first
> > component.  It seems to me you should be using a named directory or
> > something, e.g.
> >
> > c=/cygdrive
> > : ~c
> > echo ~c/d/...
> >
> > You can hardly get fewer characters than that.
>
> Right.
>
>

The point is not how many characters I type (but wiht Zsh I tend to use
less :-) The point is: I expect zsh to complete /c/d/t to a valid path
(all possibly valid paths) that match this string. And /cygdrive/d/temp
is valid path in my case. It is the same, as if Zsh simply always
skipped one directory.

Small explanation to Sven:

To access other drives under Cygwin I need either to

- mount each drive explicitly somewhere under root
or
- simply use /cygdrive that gives you "single filesystem tree" image

For this reason I really wish Zsh could assist here.

-andrej