Digging a little deeper, it looks like things here haven’t exactly converted over exactly as I wanted them.
Also, in the future we are going to need a means of converting WordPress posts to Ludum Dare data, and this is a good place to start. i.e. Use Markdownify.
YouTube videos are missing
Interestingly, YouTube support on GitHub is handled by <iframe> tags, which are allowed.
I managed to fix this problem myself, by introducing a small hack to Markdownify. Here’s a patch:
https://github.com/benbalter/wordpress-to-jekyll-exporter/pull/115
So technically, this one is fixed.
ALSO: This fixed an HTML5 graphic demo embed. https://toonormal.com/2014/08/21/phat-js-demo/
ALSO: A Vine Video embed is broken, since vine no longer exists. http://localhost:4000/2014/01/06/quadruped-animation-test/
Change the embed to this:
https://junk.mikekasprzak.com/vine/videos/1031923650816806912.mp4
Ocassional direct URL
I’ve prepared a shell script to fix my files.
#!/bin/sh
find _posts/ -type f -exec sed -i 's/http:\/\/blog.toonormal.com//g' {} \;
Where http://blog.toonormal.com begins some URLs, and that needs to stop.
ALSO: An HTML5 graphic demo embed needed fixing. https://toonormal.com/2014/08/21/phat-js-demo/
The original pointed to an http:// URL. It was changed to //, because HTTPS is allowed.
Source Code incorrectly being converted
I didn’t notice at first, but it turns out ALL the source code I merged over is incorrect.
I rely on a WordPress plugin called Crayon. Crayon unfortunately made some mistakes in its implementation of preformatted source code blocks. They used <pre> tags, but according to the HTML spec you’re supposed to use <pre><code> tags together. Oops!
On the plus side, I never used <pre> tags anywhere I didn’t want source code. When I added <pre><code> tag pairs by hand, it didn’t break the output of Crayon, so it might be possible to do something scary to fix this (regex). Something like replace <pre with <pre><code, noting that the Crayon <pre> tags are often longer (i.e. <pre class="lang:default decode:true " >).
Tracking down the issue I’ve been testing with this post:
Here’s a post I’ve been working to convert: /2012/10/18/http-get-and-posts-from-scratch-with-sockets/
Here’s the live version: http://blog.toonormal.com/2012/10/18/http-get-and-posts-from-scratch-with-sockets/
And the original: http://blog.toonormal.com/wp-admin/revision.php?revision=5546
At the time of this writing, this is still an issue.
Missing HTML tags in code blocks
This same post is throwing away some HTML tags.
http://blog.toonormal.com/2012/10/18/http-get-and-posts-from-scratch-with-sockets/
<script>, <area>, <map>, and others. Technically these tags should be preserved if they are found within a <code> block (or <pre> in case of Crayon, but it should really be <pre><code>).
At the time of this writing this is still an issue.