Lines 1156-1162
sub get_decoded_stripped_body_text_array
Link Here
|
1156 |
$self->{html}{ratio} = ($raw - $self->{html}{non_uri_len}) / $raw; |
1156 |
$self->{html}{ratio} = ($raw - $self->{html}{non_uri_len}) / $raw; |
1157 |
} # if ($raw > 0) |
1157 |
} # if ($raw > 0) |
1158 |
delete $self->{html_last_tag}; |
1158 |
delete $self->{html_last_tag}; |
1159 |
} |
1159 |
|
|
|
1160 |
$text =~ s/<![^>]*>//g; |
1161 |
} # if HTML |
1160 |
|
1162 |
|
1161 |
# whitespace handling (warning: small changes have large effects!) |
1163 |
# whitespace handling (warning: small changes have large effects!) |
1162 |
$text =~ s/\n+\s*\n+/\f/gs; # double newlines => form feed |
1164 |
$text =~ s/\n+\s*\n+/\f/gs; # double newlines => form feed |