Hi All, hi Gerald,
I wrote some code for a forum, that displays user entered text,
turning all URLs found into real links. Here it is:
[- @t = split(/\x0d\x0a/, $message); $r = '';
{
local $/ = "\x0d\x0a"; # for chomp below
for($i=0;$i < scalar @t;$i++) {
$out = ''; $t = $t[$i];
HTML::Embperl::Execute({'escmode' => 1,
'input' => \$t,
'output' => \$out,
'mtime' => undef});
chomp $out;
$out =~ s@((?:ftp|http|news)://[^ ]*[^ .,:;!?<>()])@<A
HREF="$1">$1</A>@gi;
$r .= "<BR>\n" if $r;
$r .= $out;
}
} -]
[+ local $escmode=0; $r +]
It may not be optimal, but it works well. I tried to put that code in
a Perl module, since it looks quite reusable:
package XXX;
use strict;
use lib qw(..);
use Exporter ();
use HTML::Embperl ();
use vars qw(@ISA @EXPORT_OK);
@ISA = qw(Exporter);
@EXPORT_OK = qw(UserText);
sub UserText($) {
my($Text) = @_;
{
local $/ = "\x0d\x0a"; # for chomp below
return join("<BR>\n",
map { my $out;
HTML::Embperl::Execute({'escmode' => 1,
'input' => \$_,
'output' => \$out,
'mtime' => undef});
chomp $out;
$out =~ s@((?:ftp|http|news)://[^ ]*[^ .,:;!?<>()])@<A
HREF="$1">$1</A>@gi;
$out }
split(/\x0d\x0a/, $Text)
);
}
}
1;
When called from command-line using a Perl script, all seems fine.
>From a HTML page, code becomes :
[- use XXX qw(UserText); -]
...
[+ local $escmode=0; UserText($message) +]
...
But now, if $message contains several lines, UserText($message)
contains the conversion of the first line of $message, several times.
It looks like HTML::Embperl::Execute() only modifies $out during the
first loop.
What am I missing ?
--
Linux blade.workgroup 2.4.0-test10 #1 Thu Dec 7 12:47:22 CET 2000 i686 unknown
10:20am up 1 day, 14:50, 8 users, load average: 1.34, 1.73, 1.55
--
Linux blade.workgroup 2.4.0-test10 #1 Thu Dec 7 12:47:22 CET 2000 i686 unknown
4:09pm up 1 day, 20:40, 7 users, load average: 1.16, 1.15, 1.18
