From 33a4d5705aa96ffc8ad6ef732af8de53c1a2e024 Mon Sep 17 00:00:00 2001
From: acli <acli>
Date: Thu, 12 Feb 2004 17:38:42 +0000
Subject: [PATCH] It now knows what << Prev is, but it may still be confused by
 other kinds of tag lookalikes

---
 misc/translator/text-extract2.pl | 6 ++++--
 1 file changed, 4 insertions(+), 2 deletions(-)

diff --git a/misc/translator/text-extract2.pl b/misc/translator/text-extract2.pl
index 8186fdf063..274f21f5b9 100755
--- a/misc/translator/text-extract2.pl
+++ b/misc/translator/text-extract2.pl
@@ -7,7 +7,8 @@
 
 # This script is meant to be a drop-in replacement of text-extract.pl
 
-# FIXME: Strings like "<< Prev" confuses *this* filter
+# FIXME: Strings like "<< Prev" or "Next >>" may confuse *this* filter
+# TODO: Need to detect unclosed tags, empty tags, and other such stuff.
 
 use Getopt::Long;
 use strict;
@@ -93,7 +94,8 @@ sub next_token_internal (*) {
 	;
     } elsif ($readahead =~ /^\s+/s) {	# whitespace
 	($kind, $it, $readahead) = (KIND_TEXT, $&, $');
-    } elsif ($readahead =~ /^[^<]+/s) {	# non-whitespace normal text
+    # FIXME the following (the [<\s] part) is an unreliable HACK :-(
+    } elsif ($readahead =~ /^(?:[^<]|<[<\s])+/s) {	# non-space normal text
 	($kind, $it, $readahead) = (KIND_TEXT, $&, $');
     } else {				# tag/declaration/processing instruction
 	my $ok_p = 0;
-- 
2.39.5