HTML::Entities - Unmatched [ in regex

andreas1234567 has asked for the wisdom of the Perl Monks concerning the following question:

In the neverending quest for enlightenment, I came across this problem when converting an old log parser. The file two-lines-with-one-slash-each.txt was the smallest data sample I could reduce the original log to while preserving this behaviour.

Could someone please explain why this is happening? Is it a bug?

two-lines-with-one-slash-each.txt

/
/
[download]

html.entities.pl

use strict;
use warnings;
use HTML::Entities;
print encode_entities(`cat two-lines-with-one-slash-each.txt`);
__END__
[download]

Runtime:

$ perl html.entities.pl
Unmatched [ in regex; marked by <-- HERE in m/([ <-- HERE / at (eval 1
+) line 2.
 while trying to turn range: "/
"
 into code: sub {$_[0] =~ s/([/
])/$char2entity{$1} || num_entity($1)/ge; }
  at /usr/lib/perl5/vendor_perl/5.8.3/i386-linux-thread-multi/HTML/Ent
+ities.pm line 428.
[download]

Environment:

$ perl -v | head -2

This is perl, v5.8.5 built for i386-linux-thread-multi

$ uname -r
2.6.9-42.0.10.EL

cpan[2]> install HTML::Entities
HTML::Entities is up to date (1.35).
[download]

Andreas
--

Comment on HTML::Entities - Unmatched [ in regex Select or Download Code

Replies are listed 'Best First'.
Re: HTML::Entities - Unmatched [ in regex by tinita (Parson) on May 30, 2007 at 14:36 UTC
try print encode_entities(scalar `cat two-lines-with-one-slash-each.txt`); you're executing qx() in list context. the second argument to encode_entities is a character range, though.	[reply] [d/l]
Re^2: HTML::Entities - Unmatched [ in regex by blazar (Canon) on May 30, 2007 at 20:55 UTC
you're executing qx() in list context. the second argument to encode_entities is a character range, though. ++ because this just as obvious now as subtle enough to make me suspect that I wouldn't have spotted it easily in the first place. Indeed I don't like qx and try to avoid it if possible. Here, even if only working on a given OS, I don't see why one should rely on an external `cat` utility to do something that perl can handle perfectly fine itself: we would want a cheap idiom for slurping a file, and one can be provided by File::Slurp which is also known to be more efficient than the builtins. OTOH if not wanting to use an external module, I (would) do: `print encode_entities do { open my $fh, '<', 'two-lines-with-one-slash-each.txt' or die "D'Oh! $!\n"; local $/; <$fh>; };` [download] If I wanted to stay even cheaper, I'd use `@ARGV` magic: `print encode_entities do { local (@ARGV,$/)='two-lines-with-one-slash- +each.txt'; <> };` [download]	[reply] [d/l] [select]
Re^2: HTML::Entities - Unmatched [ in regex by andreas1234567 (Vicar) on May 31, 2007 at 05:30 UTC
you're executing qx() in list context. Fooled by context issues. Again.. ++tinita for pointing out. I totally agree with blazar that using cat to open files is a bad idea next to File::Slurp or a proper 3 argument open with error handling. I just could not figure out what was going on. Thanks. Andreas --	[reply]