Don't ask why i was trying to do this...

Don't ask why i was trying to do this...

[Mikael Magnusson]

I'm not sure if this is supposed to work or not,
print $(eval "echo -n '\\U'{1..200}")
outputs the first 200 codepoints on stdout more or less correctly,
print -z $(eval "echo -n '\\U'{1..200}")
only pushes up until 3 characters after 'y' to the editing stack.

-- 
Mikael Magnusson

Re: Don't ask why i was trying to do this...

[Bart Schaefer]

On Nov 4,  4:23pm, Mikael Magnusson wrote:
}
} print -z $(eval "echo -n '\\U'{1..200}")
} only pushes up until 3 characters after 'y' to the editing stack.

Hmm, with the latest cvs update I get

schaefer<502> print -z $(eval "echo -n '\\U'{1..200}")
zsh: character not in range
schaefer<503> ^A ^B ^C ^D ^E ^F ^G ^H ^P ^Q ^R ^S ^T ^U ^V ^W ^X ^Y ! " # $ % &
' ( ) 0 1 2 3 4 5 6 7 8 9 @ A B C D E F G H I P Q R S T U V W X Y ` a b c d e f
g h i p q r s t u v w x y ^À ^Á ^Â ^@^Ä ^Å ^Æ ^Ç ^È ^É ^Ð ^Ñ ^Ò ^Ó ^Ô ^Õ ^Ö ^×
^Ø ^Ù

I suspect that NUL byte (displayed as ^@ above) has something to do with
your problem.

Re: Don't ask why i was trying to do this...

[Mikael Magnusson]

On 05/11/06, Bart Schaefer <schaefer@brasslantern.com> wrote:
> On Nov 4,  4:23pm, Mikael Magnusson wrote:
> }
> } print -z $(eval "echo -n '\\U'{1..200}")
> } only pushes up until 3 characters after 'y' to the editing stack.
>
> Hmm, with the latest cvs update I get
>
> schaefer<502> print -z $(eval "echo -n '\\U'{1..200}")
> zsh: character not in range
> schaefer<503> ^A ^B ^C ^D ^E ^F ^G ^H ^P ^Q ^R ^S ^T ^U ^V ^W ^X ^Y ! " # $ % &
> ' ( ) 0 1 2 3 4 5 6 7 8 9 @ A B C D E F G H I P Q R S T U V W X Y ` a b c d e f
> g h i p q r s t u v w x y ^À ^Á ^Â ^@^Ä ^Å ^Æ ^Ç ^È ^É ^Ð ^Ñ ^Ò ^Ó ^Ô ^Õ ^Ö ^×
> ^Ø ^Ù
>
> I suspect that NUL byte (displayed as ^@ above) has something to do with
> your problem.

(locale is sv_SE)
% print -z $(eval "echo -n '\\U'{1..200}")
% ^A ^B ^C ^D ^E ^F ^G ^H ^P ^Q ^R ^S ^T ^U ^V ^W ^X ^Y ! " # $ % & '
( ) 0 1 2 3 4 5 6 7 8 9 @ A B C D E F G H I P Q R S T U V W X Y ` a b
c d e f g h i p q r s t u v w x y ^À ^Á ^Â
% export LC_ALL=C
% print -z $(eval "echo -n '\\U'{1..200}")
zsh: character not in range
% ^A ^B ^C ^D ^E ^F ^G ^H ^P ^Q ^R ^S ^T ^U ^V ^W ^X ^Y ! " # $ % & '
( ) 0 1 2 3 4 5 6 7 8 9 @ A B C D E F G H I P Q R S T U V W X Y ` a b
c d e f g h i p q r s t u v w x y
% export LC_ALL=en_US
% print -z $(eval "echo -n '\\U'{1..200}")
zsh: character not in range
% ^A ^B ^C ^D ^E ^F ^G ^H ^P ^Q ^R ^S ^T ^U ^V ^W ^X ^Y ! " # $ % & '
( ) 0 1 2 3 4 5 6 7 8 9 @ A B C D E F G H I P Q R S T U V W X Y ` a b
c d e f g h i p q r s t u v w x y ^À ^Á ^Â ^@^Ä ^Å ^Æ ^Ç ^È ^É ^Ð ^Ñ
^Ò ^Ó ^Ô ^Õ ^Ö ^× ^Ø ^Ù

-- 
Mikael Magnusson

Re: Don't ask why i was trying to do this...

[Peter Stephenson]

On Sun, 5 Nov 2006 01:23:37 +0100
"Mikael Magnusson" <mikachu@gmail.com> wrote:
> I'm not sure if this is supposed to work or not,
> print $(eval "echo -n '\\U'{1..200}")
> outputs the first 200 codepoints on stdout more or less correctly,
> print -z $(eval "echo -n '\\U'{1..200}")
> only pushes up until 3 characters after 'y' to the editing stack.

Right, normal print handles unmetafied output correctly, but options -z
as well as -s don't.  \U83 contains a character which, when unmetafied,
looks just like the Meta character.  It's therefore very useful for
testing.  (As you probably already know, \U assumes the argument is
hexadecimal, so the '\\U'{1..200}, which gives it decimal numbers, is a
bit illogical, but that's not relevant to the bug.)

This bug isn't new---someone decided it would be fantastic for
getkeystring(), which does interpretation of character escapes for
print, to return unmetafied output, since it would be easy to fix up
bin_print() to handle that, wouldn't it?  It isn't even documented.

(In case you think I sound bitter, the motto of the CSR firmware group
used to be "non sumus amari", which is Latin for "we're not bitter".)

By the way, a rewrite of getkeystring() is part of my improvement for
unquoting in completion, if that ever sees the light of day.  Sorry,
doing it again.

Index: Src/builtin.c
===================================================================
RCS file: /cvsroot/zsh/zsh/Src/builtin.c,v
retrieving revision 1.164
diff -u -r1.164 builtin.c
--- Src/builtin.c	5 Oct 2006 21:53:26 -0000	1.164
+++ Src/builtin.c	5 Nov 2006 21:38:40 -0000
@@ -3454,6 +3454,14 @@
     else \
 	count += fprintf(fout, spec, width, VAL);
 
+/*
+ * Because of the use of getkeystring() to interpret the arguments,
+ * the elements of args spend a large part of the function unmetafied
+ * with the lengths in len.  This may have seemed a good idea once.
+ * As we are stuck with this for now, we need to be very careful
+ * deciding what state args is in.
+ */
+
 /**/
 int
 bin_print(char *name, char **args, Options ops, int func)
@@ -3727,6 +3735,14 @@
     
     /* normal output */
     if (!fmt) {
+	if (OPT_ISSET(ops, 'z') || OPT_ISSET(ops, 's')) {
+	    /*
+	     * We don't want the arguments unmetafied after all.
+	     */
+	    for (n = 0; n < argc; n++)
+		metafy(args[n], len[n], META_NOALLOC);
+	}
+
 	/* -z option -- push the arguments onto the editing buffer stack */
 	if (OPT_ISSET(ops,'z')) {
 	    queue_signals();

-- 
Peter Stephenson <p.w.stephenson@ntlworld.com>
Web page now at http://homepage.ntlworld.com/p.w.stephenson/