I seem to have sorted out the main code for making blosxom at least pretend to be RESTy, now I am working on the archives system to clean that up. It is largely based on the main code, but needs to take into account the “forced” /archives/ directory when working out relative paths and stuff like that.
Once this is complete, I will then start working on creating some templates for the actual site (coverting my Visio wireframes into real XHTML!).
I have added a simple calendar which will be displayed in the place of a @ CALENDAR@ tag in your RESTxom templates (head or foot preferably!)
The calendar will include links to any days which contain posts, and those links will go directly to the archives, where that day will be shown, and you can locate whichever post interests you. I’d like to add support for allowing the browsing of months, but the handling of the 403 errors doesn’t seem to allow that…?
When I thought I had “finished” the mainline code for RESTxom, I forgot that I hadn’t thoroughly tested the RSS code (allowing you to request rss.xml from any topic and get a correct feed from that point ‘down’). Turns out it doesn’t work properly… double-slashing problems and whatnot, shouldn’t be too hard to fix hopefully, but darned annoying in the meantime.
So off I go to fix the RSS code, then have to fix it in the archives as well (which are working now!)
I have made it so that the /archives/ directory, from the root of the blog, houses all of the archives for the entire blog, organised by date. It simply parses the dates and then requests the details from blosxom. To sort out the issues related to relative vs. absolute references, I have added in # # #URL# # # (actually no spaces between any of the #’s) to the story template, which is replaced by my scripts with the root of the blog-tree when displaying it.
The only problem with this system at the moment is that I don’t think it’s actually passing the internal page anchor (after the #) when it displays the page, so it might not link down on the page to the specific post… I will have to experiment with this some more.
UPDATED: (about 3 seconds later) — it does load the correct place in the page, so it’s all good!
There appears to be something dodgy going on with my paths. They aren’t resolving properly, which is causing internal links to be one or more levels (directory-speaking) out of whack.
This is really annoying… until I fix this I can’t really work on the commenting system, and I’m not launching the new Dented Reality website without the ability for people to comment on posts… no way!
Am working on a commenting system for RESTxom which will store a parallel file for each entry called “<entry>.comments” and will act like a small flatfile database to store comments for each post, which will also, of course, have REST-friendly URLs.
The plan with these helper scripts that I am writing is that blosxom will be able to operate in a “REST-like” environment. This means that directing a browser to a URL like /blogging/blogger.com/ should load the blog entries in that directory (if it is within a defined blog-tree).
It’s working so far, and I have added in the option of linking to ‘rss.xml’ within any directory in the blog-tree to get an RSS feed of that level/category/sub-blog. Sweeeet. Next up is to eliminate the permalinks needing to use the blosxom.pl script in the URL (i.e. so that archives don’t have to link to blosxom.pl at all).
Changes are afoot
I am re-writing a large proportion of the code which runs RESTxom, to make it more portable and reliable. Also trying to make sure that *everything* works properly, rather than getting any nasty surprises down the line!
Once this is done properly, then RESTments should be pretty simple(?) to add in.