I maintain the Old School RPG Planet. The list of feeds it manages is saved on a wiki page. I wanted to write a little script that will allow me to quickly add feeds to that list. And I did! There’s now a way to submit new feeds to the feed instead of editing the wiki page.
The problem? The thing tries to parse web pages, trying to discover feed addresses. And that works well for sites that validate. But the two first Blogspot sites I tried each had over two hundred errors! Once the markup is borked, parsing doesn’t work, and thus feed discovery doesn’t work.
Now, if I need to work around broken markup, I start wondering why tried to standardize HTML… What a glorious waste of time! In the end, we just treat it as tag soup anyway.
If you’re still interested in the source code, no problem. Lately all my CGI-scripts are able to spew forth their source code.
Unfortunately it is not complete, yet. It doesn’t update the wiki page. I didn’t bother once I realized that the entire parsing idea was not going to work.
Update: Wohoo, replaced HTML and XML parsing with regular expression matching, wrote what I needed, and finished the script!