utf-8 support seems to be working now

5 files changed, 29 insertions, 8 deletions

diff --git a/IkiWiki/Render.pm b/IkiWiki/Render.pm
index f33d6e66e..08f5e7e95 100644
--- a/IkiWiki/Render.pm
+++ b/IkiWiki/Render.pm
@@ -28,10 +28,13 @@ sub htmlize ($$) { #{{{
 		$blosxom::version="is a proper perl module too much to ask?";
 		use warnings 'all';
 		do "/usr/bin/markdown";
+		require Encode;
 	}
 	
 	if ($type eq '.mdwn') {
-		$content=Markdown::Markdown($content);
+		# Markdown does character based stuff that does not work
+		# well with utf-8 strings.
+		$content=Encode::decode_utf8(Markdown::Markdown(Encode::encode_utf8($content)));
 	}
 	else {
 		error("htmlization of $type not supported");
diff --git a/debian/changelog b/debian/changelog
index 12dc06178..5be298136 100644
--- a/debian/changelog
+++ b/debian/changelog
@@ -14,11 +14,17 @@ ikiwiki (1.4) UNRELEASED; urgency=low
   * Rebuilding on upgrade to this version is recommended.
   * Add a html validity check to the test suite, using the wdg-html-validator,
     if available.
-  * Make the html valid when there is nothing in the actions list by adding an
-    empty <li> to the end of it.
+  * Make the html valid when there is nothing in the actions list.
   * Reordered some function call parameters for consistency.
+  * Enable full utf-8 support for page input and output.
+  * Add a workaround for markdown, which does not work well with utf-8
+    strings.
+  * --getctime had bitrotted (well I only ever used it the once so far..), 
+    fix and make it a bit more flexible
+  * rcs_getctime is changed, now rather than needing to loop over all pages,
+    it should just use the rcs to get the ctime of the passed file.
 
- -- Joey Hess <joeyh@debian.org>  Fri, 26 May 2006 04:11:57 -0400
+ -- Joey Hess <joeyh@debian.org>  Fri, 26 May 2006 04:49:49 -0400
 
 ikiwiki (1.3) unstable; urgency=low
 
diff --git a/debian/postinst b/debian/postinst
index 76d826357..9135af754 100755
--- a/debian/postinst
+++ b/debian/postinst
@@ -4,7 +4,7 @@ set -e
 
 # Change this when some incompatible change is made that requires
 # rebuilding all wikis.
-firstcompat=1.1
+firstcompat=1.4
 
 if [ "$1" = configure ] && \
    dpkg --compare-versions "$2" lt "$firstcompat"; then
diff --git a/doc/todo/utf8.mdwn b/doc/todo/utf8.mdwn
index 68195b729..b905e4633 100644
--- a/doc/todo/utf8.mdwn
+++ b/doc/todo/utf8.mdwn
@@ -25,4 +25,17 @@ The following problems have been observed when running ikiwiki this way:
 	Malformed UTF-8 character (fatal) at /usr/bin/markdown line 1317.
 
   In this example, a literal 0x97 character had gotten into a markdown
-  file. 
+  file.
+
+  Running this before markdown can avoid it:
+
+  $content = Encode::encode_utf8($content);
+
+  I'm not sure how, or what should be done after markdown to get the string
+  back into a form that perl can treat as utf-8.
+
+* Apache "AddDefaultCharset on" settings will not play well with utf-8
+  pages.
+
+* CGI::FormBuilder needs to be told to set `charset => "utf-8"` so that
+  utf-8 is used in the edit form. (done)
diff --git a/ikiwiki b/ikiwiki
index be7f86a45..3ea6b7aa0 100755
--- a/ikiwiki
+++ b/ikiwiki
@@ -1,4 +1,4 @@
-#!/usr/bin/perl -T
+#!/usr/bin/perl -T -CSD
 $ENV{PATH}="/usr/local/bin:/usr/bin:/bin";
 
 package IkiWiki;
@@ -101,7 +101,6 @@ sub main () { #{{{
 		loadindex();
 		require IkiWiki::Render;
 		rcs_update();
-		rcs_getctime() if $config{getctime};
 		refresh();
 		rcs_notify() if $config{notify};
 		saveindex();