tag-order problem

tag-order problem

[Tanaka Akira]

Z(4):akr@serein% Src/zsh -f
serein% bindkey -e; autoload -U compinit; compinit -D; compdef _tst tst
serein% zstyle ':completion:*' group-name ''
serein% zstyle ':completion:*:messages' format %d
serein% zstyle ':completion:*:descriptions' format %d
serein% zstyle ':completion:*:*:foo:*' tag-order '*' '*:-case'
serein% zstyle ':completion:*-case' matcher 'm:{a-z}={A-Z}'
serein% foo c<TAB>
file
config.cache    config.h        config.log      config.sub*     configure.in 
config.guess*   config.h.in     config.status*  configure*      
file
CVS/            Completion/     config.guess*   config.log      configure*
ChangeLog       Config/         config.h        config.status*  configure.in 
ChangeLog.3.0   config.cache    config.h.in     config.sub*     

I think it shouldn't completes two file sets at once.
-- 
Tanaka Akira

tag-order problem.

[Tanaka Akira]

I found that tag-order prevents some of completions.

Z(3):akr@serein% Src/zsh -f
serein% bindkey -e;autoload -U compinit; compinit -D
serein% zstyle ':completion:*' tag-order '*:label' 
serein% which <TAB>

This completes nothing.

Note that

serein% lynx <TAB>
serein% gzip <TAB>
serein% compress <TAB>

etc. have same problem.
-- 
Tanaka Akira

Re: tag-order problem.

[Bart Schaefer]

On May 30, 11:22pm, Tanaka Akira wrote:
} Subject: tag-order problem.
}
} I found that tag-order prevents some of completions.

(Peter, I'm afraid you're going to have to re-do this release.  Good thing
you didn't "really" announce it yet.)

} Z(3):akr@serein% Src/zsh -f
} serein% bindkey -e;autoload -U compinit; compinit -D
} serein% zstyle ':completion:*' tag-order '*:label' 
} serein% which <TAB>
} 
} This completes nothing.

I used the example of completing after `lynx' because _lynx has not been
edited since before this broke, whereas _which was changed recently.

The problem is with the change to _arguments to use a counter instead of
a shift when going around this nested loop:

    while _tags; do
      anum=1
      while [[ anum -le  $#descrs ]]; do

	action="$actions[anum]"
	descr="$descrs[anum]"
	subc="$subcs[anum++]"

        # lots deleted
      done
      # lots more deleted
    done

In the previous implementation, `shift descrs' at the point where I've put
`# lots deleted' would mean that the inner while loop consumed all the
descriptions on the first time around the _tags loop, so they were not
tried again for the second and succeeding tags.  Now, all the descriptions
are tried for every tag.  I'm not sure if that's what Sven intended, but
it breaks the `all tags not explicitly selected are tried last' behavior
in some way.

As a fast workaround, stick `descrs=()' after the inner `done', i.e. where
I've got `#lots more deleted' in the snippet above.  That's probably not
the correct fix ....

-- 
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   

tag-order problem

[Borsenkow Andrej]

Just found problem in my _service/_services

_service calls _services to complete list of available services. Because
_services normally completes both init and xinetd, I presumed I could
limit matches by setting tag-order; in this case:

zstyle :completion::complete:service:argument-1: tag-order init -

(meaning - for the first argument of service somplete only tag `init').

It did restrict services set to init only - but it also stopped dead
options completion.

Setting tags to `options init -' does allow options completion - both in
first and second word as well; third and subsequent are completed
normally (after -TAB three times:

{pts/2}% service --debug --full-restart httpd-naat
Completing init services
httpd-naat  syslog-ng


Looking in traces, it seems that

_arguments in _service correctly setups requested tags (that are
argument-1 and options) and does not use the above tag-order style at
all

first call to _tags in tags loop is O.K.; next `_requested argument-1'
also returns true

next call to _tags is still true

but call to `_requested options' fails.

If no tag-order is defined it works correctly but then both sets are
listed (that I'd like to avoid). 

-andrej

Re: tag-order problem

[Sven Wischnowsky]

Borsenkow Andrej wrote:

> ...
> 
> If no tag-order is defined it works correctly but then both sets are
> listed (that I'd like to avoid). 

The `-' special value is probably not as thoroughly tested as it
should be, I'll have a look.


Bye
  Sven

-- 
Sven Wischnowsky                          wischnow@berkom.de

Re: tag-order problem

[Sven Wischnowsky]

Borsenkow Andrej wrote:

> Just found problem in my _service/_services
> 
> _service calls _services to complete list of available services. Because
> _services normally completes both init and xinetd, I presumed I could
> limit matches by setting tag-order; in this case:
> 
> zstyle :completion::complete:service:argument-1: tag-order init -
> 
> (meaning - for the first argument of service somplete only tag `init').
> 
> It did restrict services set to init only - but it also stopped dead
> options completion.
> 
> Setting tags to `options init -' does allow options completion - both in
> first and second word as well; [...]

First, there was a buglet in _arguments that made option completion
use a context name containing the `argument-1'. Then there was a
buglet in _services because the loop wasn't left when completions were
produced. That's the reason why it didn't do what it should without
the `-' in tag-order.

With the two fixes it now works nicely for me (without `-').


Bye
  Sven

Index: Completion/Base/Utility/_arguments
===================================================================
RCS file: /cvsroot/zsh/zsh/Completion/Base/Utility/_arguments,v
retrieving revision 1.11
diff -u -r1.11 _arguments
--- Completion/Base/Utility/_arguments	18 Feb 2002 14:27:38 -0000	1.11
+++ Completion/Base/Utility/_arguments	14 May 2002 08:03:32 -0000
@@ -335,9 +335,11 @@
          [[ -z "$hasopts" &&
             -z "$matched" &&
             ( -z "$aret" || "$PREFIX" = "$origpre" ) ]] &&
-          { ! zstyle -T ":completion:${curcontext}:options" prefix-needed ||
+          { ! zstyle -T ":completion:${oldcontext%:*}:options" prefix-needed ||
             [[ "$origpre" = [-+]* || -z "$aret$mesg$tried" ]] } ; then
-	local prevpre="$PREFIX" previpre="$IPREFIX"
+	local prevpre="$PREFIX" previpre="$IPREFIX" prevcontext="$curcontext"
+
+        curcontext="${oldcontext%:*}:options"
 
 	hasopts=yes
 
@@ -389,6 +391,7 @@
         fi
 	PREFIX="$prevpre"
 	IPREFIX="$previpre"
+        curcontext="$prevcontext"
       fi
       [[ -n "$tried" && "${${alwopt:+$origpre}:-$PREFIX}" != [-+]* ]] && break
     done
Index: Completion/Unix/Type/_services
===================================================================
RCS file: /cvsroot/zsh/zsh/Completion/Unix/Type/_services,v
retrieving revision 1.1
diff -u -r1.1 _services
--- Completion/Unix/Type/_services	8 May 2002 09:18:13 -0000	1.1
+++ Completion/Unix/Type/_services	14 May 2002 08:03:32 -0000
@@ -18,6 +18,7 @@
     compadd  -a inits && ret=0
   _requested xinetd expl 'xinetd services' \
     compadd  -a xinetds && ret=0
+  (( ret )) || break
 done
 
 return $ret

-- 
Sven Wischnowsky                          wischnow@berkom.de