A sub named ∃

diotalevi has asked for the wisdom of the Perl Monks concerning the following question:

I recently figured out how to type unicode in Emacs so I thought it'd be fun to write some functions with unicode names. Nope. It didn't work. I recalled [5.8.0 Note] A Sub Named Sigma but the same magic doesn't work here. Am I doing this wrong or is this a bug in my perl? I tried both 5.8.8 and bleadperl.

use utf8;
use strict;
use warnings;
use Test::More tests => 2;

sub ∃ (&\@) { my $cb = shift @_; $cb->() and return 1 for @_; 0 }
sub ∄ (&\@) { not ∃( @_ ) }

my @numbers = 1 .. 10;
ok( ∃ { /\d/ } @numbers, '∃' );
ok( ∄ { /[[:alpha:]]/ } @numbers, '∄' );

1..2
Illegal declaration of anonymous subroutine at /home/josh/src/any.pl l
+ine 6.
# Looks like your test died before it could output anything.
[download]

⠤⠤ ⠙⠊⠕⠞⠁⠇⠑⠧⠊

Comment on A sub named ∃ Download Code

Replies are listed 'Best First'.
Re: A sub named ∃ (\w) by tye (Sage) on Nov 29, 2006 at 06:17 UTC
Subroutine names have to start with a letter. Sigma is a letter, just in a different language. "not a member of" is not a letter. Well, that's my guess, anyway. - tye	[reply]
Re^2: A sub named ∃ (\w) by diotalevi (Canon) on Nov 29, 2006 at 17:52 UTC
Yeah, you're right. ∃ isn't an identifier character per unicore/gc_sc/IdStart.pl. `ok( "\x{2203}" !~ /\p{IdStart}/ )` ⠤⠤ ⠙⠊⠕⠞⠁⠇⠑⠧⠊	[reply] [d/l]
Re: A sub named ∃ by monarch (Priest) on Nov 29, 2006 at 08:28 UTC
I tried this on NT Emacs (although with a different language) using a utf8 source file: use utf8; use Test::Simple tests=>1; use strict; sub трщь { return 1; } ok( трщь(), "Unicode Sub" ); The result was: `C:\temp>perl -w russian.pl 1..1 ok 1 - Unicode Sub` [download] on Windows 2000.	[reply] [d/l]
Re: A sub named ∃ by Steve_p (Priest) on Nov 30, 2006 at 03:41 UTC
I believe that this bug is to blame. There are plans to fix it, but, as of yet, I don't know that there has been any progress towards resolving it.	[reply]
Re^2: A sub named ∃ by vkon (Curate) on Nov 30, 2006 at 11:11 UTC
I suspect you know the situation better than me, but why in this particular case the explanation that ∃!~/\w/ will not suffice? Perl allows utf8-names from letters of non-English alphabets, although documentation strictly warns against this.	[reply]
Re^3: A sub named ∃ by diotalevi (Canon) on Nov 30, 2006 at 16:48 UTC
/w isn't the pattern for matching the first character in identifiers. I never consider \w when matching that first character. In unicode, the pattern to follow is likely /\A\p{IdFirst}\w\z/. In ASCII, the pattern is `/\A[[:alpha:]_]\w\z/`. The pattern `/\^\w+\z/` is always wrong when matching identifiers. ⠤⠤ ⠙⠊⠕⠞⠁⠇⠑⠧⠊	[reply] [d/l] [select]
Re^4: A sub named ∃ by vkon (Curate) on Nov 30, 2006 at 21:27 UTC
Re^5: A sub named ∃ by diotalevi (Canon) on Nov 30, 2006 at 21:39 UTC
Some notes below your chosen depth have not been shown here
Re^2: A sub named ∃ by ikegami (Patriarch) on Nov 30, 2006 at 16:36 UTC
Note: In the following source code, replace "`∃`" with the UTF-8 representation of "∃", and replace "`трщь`" with the UTF-8 representation of "трщь". `use strict; use warnings; no warnings 'redefine'; for ( [ "Symbol, use utf8: " => "use utf8; sub ∃ {}" ], [ "Symbol, no utf8: " => "no utf8; sub ∃ {}" ], [ "Russian, use utf8: " => "use utf8; sub трщ&#11 +00; {}" ], [ "Russian, no utf8: " => "no utf8; sub трщ&#11 +00; {}" ], ) { print $_->[0], eval "$_->[1]; 1" ? "Success\n" : $@; }` [download] outputs `Symbol, use utf8: Illegal declaration of anonymous subroutine at (eva +l 1) line 1. Symbol, no utf8: Illegal declaration of anonymous subroutine at (eva +l 2) line 1. Russian, use utf8: Success Russian, no utf8: Illegal declaration of anonymous subroutine at (eva +l 4) line 1.` [download] If the problem is what you say it is, wouldn't `use utf8` have no effect on the error?	[reply] [d/l] [select]