More about ksh_glob and completion

More about ksh_glob and completion

[Bart Schaefer]

In zsh-users/3836 I wrote:
} 
} Unfortunately, zsh has options like kshglob that change the syntax of
} glob patterns.  When the completion system functions are autoloaded
} with kshglob in effect, the meanings of some of those patterns is
} altered -- and is compiled into the function, so even though kshglob is
} later turned off by the completion system, the [[ ]] and `case' patterns
} retain their ksh interpretations and fail to work.

Fortunately (and somewhat embarrassingly) I'm completely (ahem) wrong about
the above.  Pre-compilation of glob patterns is not broken by kshglob; I
had tried it outside the completion system and thought I'd verified it, but
I must have done something wrong (perhaps I typed `unset' for `unsetopt')
because today I can't reproduce it.

And further, the completion system didn't turn off kshglob.  It should.  See
if this doesn't fix you up, Andre.

Index: Completion/compinit
===================================================================
--- Completion/compinit	2001/04/09 20:14:08	1.1.1.1
+++ Completion/compinit	2001/04/29 18:14:05
@@ -134,6 +134,7 @@
     NO_globsubst
     NO_shwordsplit
     NO_shglob
+    NO_kshglob
     NO_ksharrays
     NO_cshnullglob
     NO_allexport

-- 
Bart Schaefer                                 Brass Lantern Enterprises
http://www.well.com/user/barts              http://www.brasslantern.com

Zsh: http://www.zsh.org | PHPerl Project: http://phperl.sourceforge.net   

Re: More about ksh_glob and completion

[Sven Wischnowsky]

Bart Schaefer wrote:

> Fortunately (and somewhat embarrassingly) I'm completely (ahem) wrong about
> the above.  Pre-compilation of glob patterns is not broken by kshglob; I
> had tried it outside the completion system and thought I'd verified it, but
> I must have done something wrong (perhaps I typed `unset' for `unsetopt')
> because today I can't reproduce it.

Yes, patterns are not pre-compiled, they are cached (under certain
circumstances, e.g. they must not contain parameter expansions).

Actually, I once thought about pre-compiling them, then had a look at
the pattern code, was reminded of the options controlling pattern
matching and...

> And further, the completion system didn't turn off kshglob.  It should.  See
> if this doesn't fix you up, Andre.

We should have remembered it when we put NO_shglob in the array...

Bye
  Sven


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

Re: More about ksh_glob and completion

[Andre Pang]

On Sun, Apr 29, 2001 at 06:30:22 +0000, Bart Schaefer wrote:

> And further, the completion system didn't turn off kshglob.  It should.  See
> if this doesn't fix you up, Andre.
> 
> Index: Completion/compinit
> ===================================================================
> --- Completion/compinit	2001/04/09 20:14:08	1.1.1.1
> +++ Completion/compinit	2001/04/29 18:14:05
> @@ -134,6 +134,7 @@
>      NO_globsubst
>      NO_shwordsplit
>      NO_shglob
> +    NO_kshglob

that worked very nicely.  (although i just unsetopt'ed kshglob
in my .zshrc before this official patch came out, and that
worked fine too.)

never ceases to amaze me how quickly you guys fix stuff / get
zsh to make coffee and cure the common cold ... thanks!


-- 
#ozone/algorithm <ozone@algorithm.com.au>          - trust.in.love.to.save