I thought I'd post an update on this work so I don't leave any people hanging who are wondering about this.

I haven't had the time in the last week or so to work on this as I have been very busy recently with other responsibilities. With my day job in particular, I will be in Berlin most of next week hosting a conference with my company. Because it's a small world, by pure coincidence Sociomantic are actually sponsoring this conference. (I doubt the marketing guys talk to the D programmers that much, though.)

http://www.performancemarketinginsights.com/14/europe/sponsor/sociomantic/

However, after I get back I will defintely work on this redesign some more. Unless somebody pins me down, I'll keep going with it. I only just realised that I made the terrible mistake of not pushing my most recent changes, so you can find them now on GitHub.

I started working on discovering Markdown files at runtime for filling in a sections of pages, like center part of changelog pages, and generating the table of contents from the HTML automatically. Because the prospect of parsing HTML myself scares me, I just copied in Adam D Ruppe's DOM library for that. I hope that I have properly attributed the authors for that code, please advise me if I should add any additional attribution in there otherwise.

The early result of that I think has been pretty good. I can edit a file and see a change pretty quickly. I lost the functions I was using to generating the Bugzilla links, but I'm going to replacing them with links to /bug/<id> instead, and making that page redirect to the right place, just so I can make my Markdown files a little smaller. The amount of memory consumed while compiling is better, going from something like 2GB to 1.2GB, because there aren't tons of Diet templates to build in anymore. The Markdown parsing is just the library functionality that comes with vibe.d.

I totally have not converted all of the pages from Diet to Markdown yet. In fact, if you run the site as is now, you can't see any of the changelog pages except 2.000 anymore. Plus, you could perhaps do something involving caching files instead of loading them from the drive all the time, but I just want something that works for now.

So there's my update for now.

Reply via email to