From c55f71887d620f13e99014f5a25350dc79890e92 Mon Sep 17 00:00:00 2001
From: Peter Karlsson <peterk>
Date: Sat, 18 Jan 2003 18:34:13 +0000
Subject: Since we are forced to use entities at some places in the English
 source files, when copying the page first decode any entities for the ISO
 8859-1 area. The entities are then re-introduced if required by the output
 encoding.

CVS version numbers

copypage.pl: 1.22 -> 1.23
---
 copypage.pl | 32 ++++++++++++++++++++++++++++----
 1 file changed, 28 insertions(+), 4 deletions(-)

(limited to 'copypage.pl')

diff --git a/copypage.pl b/copypage.pl
index 723a1c354d1..de66d21bb10 100755
--- a/copypage.pl
+++ b/copypage.pl
@@ -7,7 +7,7 @@
 # Makefiles are not supported anymore for they bear too much space for errors.
 
 # Originally written 2000-02-26 by Peter Karlsson <peterk@debian.org>
-# © Copyright 2000-2002 Software in the public interest, Inc.
+# © Copyright 2000-2003 Software in the public interest, Inc.
 # This program is released under the GNU General Public License, v2.
 
 # $Id$
@@ -245,15 +245,27 @@ sub copy
 		}
 		else
 		{
+			# Transform the string into a string that is fit for the encoding
+			# of the output language. We do that by first converting any
+			# SGML entities in the input stream into 8-bit ISO 8859-1
+			# encoding, and then convert extended characters (back) into
+			# entities if necessary for the target encoding.
+
+			# Decode
+			s/(&[^#;]+;)/&decodeentity($1)/ge;
+			s/&#(1[6-9][0-9]|2[0-4][0-9]|25[0-5]);/chr($1)/ge;
+
+			# Encode
 			if (defined $compattable)
 			{
-				# Recode non-ASCII characters that are not identical to
-				# ISO 8859-1 into entitites
+				# Output encoding is in part compatible with ISO 8859-1, only
+				# convert incompatible characters into entities.
 				s/([\xA0-\xFF])/$$compattable[ord($1)-160]?$1:$entities[ord($1)-160]/ge;
 			}
 			elsif ($recodelatin1)
 			{
-				# Recode any non-ASCII characters into entities
+				# Output encoding is incompatible with ISO 8859-1, convert all
+				# 8-bit characters into entities.
 				s/([\xA0-\xFF])/$entities[ord($1)-160]/ge;
 			}
 
@@ -274,3 +286,15 @@ sub copy
 	print "and to remove $dsttitle when finished\n"
 		if defined $dsttitle;
 }
+
+# Return the ISO-8859-1 character that corresponds to the given entity
+sub decodeentity
+{
+	my $ent = shift;
+	# Start at one to avoid decoding &nbsp;
+	for (my $i = 1; $i < $#entities; ++ $i)
+	{
+		return chr($i + 160) if $entities[$i] eq $ent;
+	}
+	return $ent;
+}
-- 
cgit v1.2.3