utf-8 support seems to be working now

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");

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
 

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

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)

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();