PATCH: small parameter fix

PATCH: small parameter fix

[Sven Wischnowsky]

...although I'm not sure if it was not intentional. If it was, I'd
like to hear the reason. This:

  ${${...}[...]}

previously didn't really work: if the inner ${...} produced an array
with only one element, the outer ${...} treated it as a string and the 
subscript gave the n'th character.


Bye
 Sven

--- os/subst.c	Fri Mar  5 20:10:39 1999
+++ Src/subst.c	Sat Mar  6 13:28:41 1999
@@ -263,7 +263,7 @@
 	    *isarr = 0;
 	return 0;
     }
-    if ((l = countlinknodes(foo)) > 1) {
+    if ((l = countlinknodes(foo)) > 1 || a) {
 	p = r = ncalloc((l + 1) * sizeof(char*));
 	while (nonempty(foo))
 	    *p++ = (char *)ugetnode(foo);

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

Re: PATCH: small parameter fix

[Bart Schaefer]

Sven Wischnowsky writes:
 > 
 > ...although I'm not sure if it was not intentional. If it was, I'd
 > like to hear the reason. This:
 > 
 >   ${${...}[...]}
 > 
 > previously didn't really work: if the inner ${...} produced an array
 > with only one element, the outer ${...} treated it as a string and the 
 > subscript gave the n'th character.

AHA!  This must be the reason that "${${(@)...}...}" is required in so
many mystifying circumstances.  I, too, would like to hear whether the
previous behavior was intentional ... this change does potentially
break scripts, including I think 3.0.5 scripts, so we'd better tread
carefully.

Re: PATCH: small parameter fix

[Peter Stephenson]

>  > 
>  >   ${${...}[...]}
>  > 
>  > previously didn't really work: if the inner ${...} produced an array
>  > with only one element, the outer ${...} treated it as a string and the 
>  > subscript gave the n'th character.
> 
> AHA!  This must be the reason that "${${(@)...}...}" is required in so
> many mystifying circumstances.  I, too, would like to hear whether the
> previous behavior was intentional ... this change does potentially
> break scripts, including I think 3.0.5 scripts, so we'd better tread
> carefully.

in particular it has this effect:

% print "${$(uname -n)[1]}"
ibmth

where you might guess $(...) in quotes produced a scalar for subscripting
--- indeed it did, but the subscripting decided to treat it as an array
element.  In fact, since the nested substitution recognises the quotes
whereas the subscripting doesn't, you get,

% print "${${path}[1]}"
/home/user2/pws/bin /home/user2/pws/links /home/user2/pws/mh/bin
/home/user2/pws/scripts /usr/local/bin /usr/sue/bin /usr/bin /usr/bin/X11
/usr/ucb /usr/local/bin/X11 /cern/pro/bin /usr/local/cap /etc .

(wrapped a bit --- well, you don't get that, but I do) which is distinctly
on the strange side.  I would suggest that in quotes the old behaviour
character-subscripting behaviour is the right one.

-- 
Peter Stephenson <pws@ibmth.df.unipi.it>       Tel: +39 050 844536
WWW:  http://www.ifh.de/~pws/
Dipartimento di Fisica, Via Buonarroti 2, 56127 Pisa, Italy

RE: PATCH: small parameter fix

[Andrej Borsenkow]

  I would suggest that in quotes the old behaviour
> character-subscripting behaviour is the right one.
>

And consistent with manuals. By definition, "...${array}..." gives a scalar
*unless* ``@'' is used

/andrej

Re: PATCH: small parameter fix

[Sven Wischnowsky]

I wrote:

> The patch at least fixes this. But maybe we should collect opinions
> which expression should expand to what and then make the code produce
> that.

And here is the patch that makes `_long_options' and `_multi_parts'
use this...

Bye
 Sven

--- oc/User/_long_options	Tue Mar  9 09:05:06 1999
+++ Completion/User/_long_options	Tue Mar  9 11:03:08 1999
@@ -87,9 +87,9 @@
   # option up to the end. Finally all elements with option strings
   # that contain uppercase letters are removed.
 
-  opts=("--${^${${${(M)${(@ps:\n:j:\n:)${(@)${(M)${(@f)$($words[1] --help)}:#[ 	]#-*}//,/
+  opts=("--${(@)^${(@)${(@)${(@M)${(@ps:\n:j:\n:)${(@)${(@M)${(@f)$("$words[1]" --help)}:#[ 	]#-*}//,/
 }}:#*--*}#*--}%%[, ]*}:#(*-[A-Z]*|)}")
-
+echo $#opts
   # The interpretation of the options is completely table driven. We
   # use the positional parameters we were given and a few standard
   # ones. Then we loop through this table.
@@ -144,7 +144,7 @@
 	# an array.
 
         tmp=("${(@)tmp:#*\[\=*}")
-        tmpo=("${${(@)tmpo%%\=*}//[^a-z0-9-]}")
+        tmpo=("${(@)${(@)tmpo%%\=*}//[^a-z0-9-]}")
         _lo_cache_actions[anum]="_lo_cache_optarg_$anum $action"
         eval "_lo_cache_optarg_${anum}=(\"\$tmpo[@]\")"
 	(( anum++ ))
@@ -156,7 +156,7 @@
       tmpo=("${(@M)tmp:#*\=*}")
       if (( $#tmpo )); then
         tmp=("${(@)tmp:#*\=*}")
-        tmpo=("${${(@)tmpo%%\=*}//[^a-z0-9-]}")
+        tmpo=("${(@)${(@)tmpo%%\=*}//[^a-z0-9-]}")
         _lo_cache_actions[anum]="_lo_cache_arg_$anum $action"
         eval "_lo_cache_arg_${anum}=(\"\$tmpo[@]\")"
 	(( anum++ ))
@@ -169,7 +169,7 @@
     # Now filter out any option strings we don't like and stuff them
     # in an array, if there are still some.
 
-    tmp=("${${(@)tmp%%\=*}//[^a-z0-9-]}")
+    tmp=("${(@)${(@)tmp%%\=*}//[^a-z0-9-]}")
     if (( $#tmp )); then
       _lo_cache_actions[anum]="$name $action"
       eval "${name}=(\"\$tmp[@]\")"
@@ -215,7 +215,7 @@
   for i in "$_lo_cache_actions[@]"; do
     name="${i%% *}"
     action="${i#* }"
-    if (( ${${(@P)name}[(I)$pre]} )); then
+    if (( ${(@)${(@P)name}[(I)$pre]} )); then
       if [[ "$action[1]" = (\[|\() ]]; then
         compadd - ${=action[2,-2]}
       elif (( $#action )); then
diff -u oc/Core/_multi_parts Completion/Core/_multi_parts
--- oc/Core/_multi_parts	Mon Mar  8 11:05:13 1999
+++ Completion/Core/_multi_parts	Tue Mar  9 10:21:02 1999
@@ -85,7 +85,7 @@
 
   tmp1="${orig%%${sep}*}${sep}"
   pref="$pref$tmp1"
-  matches=("${${(@)${(@M)matches:#${tmp1}*}#$tmp1}:#}")
+  matches=("${(@)${(@)${(@M)matches:#${tmp1}*}#$tmp1}:#}")
   orig="${orig#*${sep}}"
   patstr="${patstr#*${sep}}"
 done

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

Re: PATCH: small parameter fix

[Sven Wischnowsky]

I wrote:

> With the patch below we get:
> 
>   % echo ${$(uname -n)[1]}
>   beta
>   % echo "${$(uname -n)[1]}"
>   b
>   % echo ${${path}[1]}
>   /users/wischnow/bin/alpha
>   % echo "${${path}[1]}"
>   /
> 
> ...which seems sensible to me.

Grrr, this broke "${(@)$(foo)#...}" where the manual says that the
thing inside will be treated as an array if the `(@)' flags is given.

The patch at least fixes this. But maybe we should collect opinions
which expression should expand to what and then make the code produce
that.


Bye
 Sven

--- os/subst.c	Tue Mar  9 11:07:25 1999
+++ Src/subst.c	Tue Mar  9 11:10:13 1999
@@ -976,7 +976,8 @@
 	skipparens(*s, *s == Inpar ? Outpar : Outbrace, &s);
 	sav = *s;
 	*s = 0;
-	if (multsub(&val, ((quoted || aspar) ? NULL : &aval), &isarr, NULL) &&
+	if (multsub(&val, (((quoted || aspar) && !nojoin) ? NULL : &aval),
+		    &isarr, NULL) &&
 	    quoted) {
 	    isarr = -1;
 	    aval = alloc(sizeof(char *));

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

Re: PATCH: small parameter fix

[Sven Wischnowsky]

Peter Stephenson wrote:

> in particular it has this effect:
> 
> % print "${$(uname -n)[1]}"
> ibmth
> 
> where you might guess $(...) in quotes produced a scalar for subscripting
> --- indeed it did, but the subscripting decided to treat it as an array
> element.  In fact, since the nested substitution recognises the quotes
> whereas the subscripting doesn't, you get,
> 
> % print "${${path}[1]}"
> /home/user2/pws/bin /home/user2/pws/links /home/user2/pws/mh/bin
> /home/user2/pws/scripts /usr/local/bin /usr/sue/bin /usr/bin /usr/bin/X11
> /usr/ucb /usr/local/bin/X11 /cern/pro/bin /usr/local/cap /etc .
> 
> (wrapped a bit --- well, you don't get that, but I do) which is distinctly
> on the strange side.  I would suggest that in quotes the old behaviour
> character-subscripting behaviour is the right one.

With the patch below we get:

  % echo ${$(uname -n)[1]}
  beta
  % echo "${$(uname -n)[1]}"
  b
  % echo ${${path}[1]}
  /users/wischnow/bin/alpha
  % echo "${${path}[1]}"
  /

...which seems sensible to me.

Bye
 Sven

--- os/subst.c	Mon Mar  8 09:51:27 1999
+++ Src/subst.c	Tue Mar  9 09:16:13 1999
@@ -976,7 +976,8 @@
 	skipparens(*s, *s == Inpar ? Outpar : Outbrace, &s);
 	sav = *s;
 	*s = 0;
-	if (multsub(&val, (aspar ? NULL : &aval), &isarr, NULL) && quoted) {
+	if (multsub(&val, ((quoted || aspar) ? NULL : &aval), &isarr, NULL) &&
+	    quoted) {
 	    isarr = -1;
 	    aval = alloc(sizeof(char *));
 	    aspar = 0;

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