Terje Bless <[EMAIL PROTECTED]> writes:
> In a CGI app I'm using HTML::Parser to rewrite web pages before further
> processing. It appears that a document containing an ASCII NUL byte (in
> this case, as the last byte in the file) has the NUL byte stripped after
> processing with HTML::Parser.
It appears that you forgot to call the 'eof' method to flush remaining
buffered text. I don't think there should be any problems with NUL.
> The offending code is below, and the version
> of HTML::Parser is:
>
> % perl -MHTML::Parser -e 'print $HTML::Parser::VERSION'
> 3.26
>
>
> code:
>
>
> sub supress_doctype {
> no strict 'vars';
> my $file = shift; # $file = \@file
> local $HTML = '';
>
> HTML::Parser->new(
> default_h => [sub {$HTML .= shift}, 'text'],
> declaration_h => [sub {$HTML .= '<!-- ' . $_[0] . ' -->'}, 'text']
> )->parse(join "\n", @{$file});
^
insert '->eof' here.
> return [split /\n/, $HTML];
> }
Regards,
Gisle
