X-Git-Url: https://sipb.mit.edu/gitweb.cgi/ikiwiki.git/blobdiff_plain/afc45616c731601d2a00e88264d9f216e02e7a4f..eb5f8531fe3d9cc752a10343bb52abf8d0858196:/doc/patchqueue/more_class__61____34____34___for_css.mdwn

diff --git a/doc/patchqueue/more_class__61____34____34___for_css.mdwn b/doc/patchqueue/more_class__61____34____34___for_css.mdwn
index 52905abe6..49affd29b 100644
--- a/doc/patchqueue/more_class__61____34____34___for_css.mdwn
+++ b/doc/patchqueue/more_class__61____34____34___for_css.mdwn
@@ -2,28 +2,58 @@ I'm writing my own CSS for ikiwiki. During this effort I often found the need of
 
 In this patch I plan to collect changes in this direction.
 
-The first, one-liner, patch is to use a <div class="actions"> for inline page as is done with non-inlined page. This way the same CSS formatting can be applied to div.actions in the CSS, while at the moment it must be duplicated for a span.actions (which I believe is also incorrect, since it will contain an <ul>, not sure though). In case the markup should be differentiated it will still be possible relying on the fact that a div.actions is contained or not in a div.inlinepage.
+The first, one-liner, patch is to use a "div" element with a
+class="actions" attribute for inline page as is done with non-inlined page.
+This way the same CSS formatting can be applied to div.actions in the CSS,
+while at the moment it must be duplicated for a span.actions (which I
+believe is also incorrect, since it will contain a "ul" element, not sure
+though). In case the markup should be differentiated it will still be
+possible relying on the fact that a div.actions is contained or not in a
+div.inlinepage.
 
 Here's the one-liner:
 
-    --- /usr/share/ikiwiki/templates/inlinepage.tmpl        2006-12-28 16:24:06.000000000 +0100
-    +++ inlinepage.tmpl.new 2006-12-28 16:24:04.000000000 +0100
-    @@ -31,7 +31,7 @@
+> applied --[[Joey]]
+
+The following adds a div element with class="trailer" around the meta-information
+added after an inlined page (namely: the post date, the tags, and the actions):
+
+    --- inlinepage.tmpl.orig        2006-12-28 16:56:49.000000000 +0100
+    +++ inlinepage.tmpl     2006-12-28 17:02:06.000000000 +0100
+    @@ -17,6 +17,8 @@
      </span>
+     <TMPL_VAR CONTENT>
 
-     <TMPL_IF NAME="HAVE_ACTIONS">
-    -<span class="actions">
-    +<div class="actions">
-     <ul>
-     <TMPL_IF NAME="EDITURL">
-     <li><a href="<TMPL_VAR EDITURL>">Edit</a></li>
-    @@ -40,7 +40,7 @@
-     <li><TMPL_VAR DISCUSSIONLINK></li>
-     </TMPL_IF>
-     </ul>
-    -</span>
-    +</div>
+    +<div class="trailer">
+    +
+     <span class="pageinfo">
+     Posted <TMPL_VAR CTIME>
+     </span>
+    @@ -44,3 +46,5 @@
      </TMPL_IF>
 
      </div>
+    +
+    +</div>
+
+> Unfortunately, the inlinepage content passes through markdown, and markdown
+> gets confused by these nested div's and puts p's around one of them, generating
+> broken html. If you can come up with a way to put in the div that passes
+> the test suite, or a fix to markdown, I will accept it, but the above patch
+> fails the test suite. --[[Joey]] 
+
+>> Just a note...  This discrepancy doesn't exist in [pandoc](http://code.google.com/p/pandoc/) as
+>> demonstrated in the relevant [page](http://code.google.com/p/pandoc/wiki/PandocVsMarkdownPl).
+>> Pandoc is a _real parser_ for markdown (contrasting the regexp based implementation of 
+>> markdown.pl).  I've almost finished the Debian packaging.  John is working on a `--strict` mode
+>> which will hopefully make pandoc a drop-in replacement for markdown.  I'll upload pandoc after 
+>> his work has finished.  Whether it could be used in IkiWiki is an open question, but having
+>> alternatives is always a good thing and perhaps, the fact that pandoc can make markdown->LaTeX
+>> conversion may lead to new possibilities. --[[Roktas]]
 
+>>> I confirm that this ([[debbug 405058]]) has just been fixed in markdown
+>>> [`1.0.2b7`](http://packages.debian.org/experimental/web/markdown) (BTW, thanks to your bug
+>>> report Joey).  FYI, I've observed some performance drop with `1.0.2b7` compared to `1.0.1`,
+>>> especially noticable with big files.  This was also confirmed by someone else, for example,
+>>> see this [thread](http://six.pairlist.net/pipermail/markdown-discuss/2006-August/000152.html)
+>>> --[[Roktas]]
\ No newline at end of file