From a9e13f000adaeb9eb8c9c956f7ecaaf5f58a0512 Mon Sep 17 00:00:00 2001
From: Tatsuya Kinoshita <tats@debian.org>
Date: Fri, 2 Aug 2013 06:58:09 +0900
Subject: Correct underline processing and more UTF-8 support for
 w3mman2html.cgi

Patch from <https://bugs.launchpad.net/ubuntu/+source/w3m/+bug/680202>,
provided by Piotr P. Karwasz.
---
 scripts/w3mman/w3mman2html.cgi.in | 8 +++++---
 1 file changed, 5 insertions(+), 3 deletions(-)

(limited to 'scripts')

diff --git a/scripts/w3mman/w3mman2html.cgi.in b/scripts/w3mman/w3mman2html.cgi.in
index f430307..2e3576c 100644
--- a/scripts/w3mman/w3mman2html.cgi.in
+++ b/scripts/w3mman/w3mman2html.cgi.in
@@ -126,12 +126,14 @@ while(<F>) {
   s/\&/\&amp;/g;
   s/\</\&lt;/g;
   s/\>/\&gt;/g;
+  # non ASCII UTF-8 codepoint
+  my $utf8="[\300-\337][\200-\277]|[\340-\357][\200-\277]{2}|[\360-\367][\200-\277]{3}|[\370-\373][\200-\277]{4}|[\374\375][\200-\277]{5}";
 
-  s@([\200-\377].)(\010{1,2}\1)+@<b>$1</b>@g;
+  s@($utf8)(\010\1)+@<b>$1</b>@g;
   s@(\&\w+;|.)(\010\1)+@<b>$1</b>@g;
-  s@__\010{1,2}((\<b\>)?[\200-\377].(\</b\>)?)@<u>$1</u>@g;
+  s@_\010((\<b\>)?($utf8)(\</b\>)?)@<u>$1</u>@g;
   s@_\010((\<b\>)?(\&\w+\;|.)(\</b\>)?)@<u>$1</u>@g;
-  s@((\<b\>)?[\200-\377].(\</b\>)?)\010{1,2}__@<u>$1</u>@g;
+  s@((\<b\>)?($utf8)(\</b\>)?)\010_@<u>$1</u>@g;
   s@((\<b\>)?(\&\w+\;|.)(\</b\>)?)\010_@<u>$1</u>@g;
   s@.\010(.)@$1@g;
 
-- 
cgit v1.2.3