Upgrades
The observant out there may have noticed that this site has just undergone some rather dramatic changes. Mainly the following:
-
Non-crufty URIs. This means no more
?start=45&count=1
junk in the weblog permalinks etc. This has been by far the most complicated change, involving use of mod_rewrite and my own URI translation script that should hopefully ensure that links to pages with the old URI scheme end up in the right place. - You can navigate the weblog by date, by category or just by viewing one big list (see the archive).
- My Digest Authentication implementation is now used for site security instead of sessions/cookies.
- The HTML and XHTML1.1 versions of the site are now generated by performing an XML parse of the source (which is originally in an XHTML2-like format) instead of the extremely hacky (and easily broken if you weren't very careful) regular expressions used before.
-
Now using full and proper
Acceptheader analysis to decide which format out of the available formats (in most cases XHTML1.1 or HTML4) to send to the client. No more evil user agent detection used. - XML validation of new weblog posts. This will hopefully ensure that I won't accidentally invalidate the front page by adding a new weblog post that isn't valid.
- Plus it has a lovely new stylesheet. Alternate stylesheets may be reintroduced at some point.
- Media-specific style rules (try print preview and see what it does to the links… - you'll need a web browser that doesn't suck though).
This has been several months in the making, but I hope it'll be worth it. There are still a few things to be fixed. The news
module of Omoicore (which is what the weblog uses) has been completely rewritten, and the Pingback client and server have yet to be implemented (I got impatient and decided to release early). Hopefully they'll be up and running better than ever in a few days.
Posted: 2003-11-30 23:38:18 UTC by Xiven | Cross-references (1) | Comments (6)
Comments
Xiven (Registered) (2003-12-01 21:36:42 UTC)
Oops, looks like I forgot to redirect the old RSS/ESF/CDF feeds URIs to their new location.
mpal1 (2003-12-02 12:16:13 UTC)
Have you got a logout ability with HTTP Authentication? If so, how do you stop the browser from caching the authentication stuff?
Xiven (Registered) (2003-12-02 12:31:10 UTC)
Simple answer: no. There is no proper way of logging out with HTTP Authentication. I would consider this to be its single biggest flaw. There are (I believe) certain hacks that can force the browser to log out, but I think these are rather browser-dependant.
mpal1 (2003-12-04 02:32:02 UTC)
That's what I thought. Perhaps, if you control the http Authentication in PHP, you could, in theory, generate a unique authentication realm, and when the user clicks "logout" you could invalidate that realm, and to check if someone's logged in, make sure that the realm is valid as well as the userid and password.
If you have multiple realms, you can append the cookie to the existing realm.
(It's either that or my understanding of HTTP Authentication is flawed... :)
Aquarion (2003-12-11 12:22:39 UTC)
Your RSS & ESF feeds are broken :-)
Xiven (Registered) (2003-12-11 22:59:47 UTC)
Thank you very much. For some reason it was displaying the page twice, which it didn't like...