Karma's Nucleus DevBlog

After some digging, I've found some zipfiles for old releases (pre GPL = pre v2.0). They can now be found in the Nucleus Museum.

The pre-v0.9 directory contains versions that were never officially released. The Nucleus0.01.zip file even contains code that uses XML files to store data instead of XML.

<STYLE>@;/* (info)

For your reference (and for my own), here's an overview of the branches and tags applied to the CVS repository:

• Upgraded Gallery software (Nucleus screenshots) to v1.4.3-pl2
• Upgraded phpMyAdmin to v2.5.7-pl1, removed the phpMyAdmin software on the plugin.nucleuscms.org subdomain.
• Used the almost-released Nucleus v3.1 builds to upgrade almost all nucleus installs on the server (www, dev, plugins, karma, demuynck.org). As for japan.nucleuscms.org (still using v2.5beta), made sure it is protected against the security issues
• Updated the screenshots, languages and convert pages to list v3.1 as current version. the download/upgrade pages will get updated when the actual release is done. (update: released & updated)
• Updated the documentation/ directory
• Added Atom feed for nucleuscms.org
• On nucleuscms.org, made aliases for the RSS/Atom files: rss.xml, atom.xml, feed.atom, index.rdf by adding the following code to .htaccess:

  RewriteEngine On
  RewriteRule ^rss.xml xml-rss2.php
  RewriteRule ^index.rdf xml-rss2.php
  RewriteRule ^atom.xml atom.php
  RewriteRule ^feed.atom atom.php

• Added an overview of the latest Inside Nucleus posts on the right side of this blog.
• Added Google AdSense ads on the forum.

In this post, I'll try to explain how URLs (links to items/archives/...) are built in Nucleus. It turned out to be quite a lengthy article. I'll start in a fancy URL free world and will end up explaining how fancy URLs work and highlighting some problems.

I'm planning to release a Nucleus v3.1 version in the next few days. One of the reasons is another vulerability has been discovered (SQL insertion this time). After looking at the changes between the CVS version and 3.0, I feel like it's safe to base the v3.1 release on the most recent CVS code.

This time, I've created a special Nucleus-3-1 branch, which will be used for the 3.1 release and any critical fixes that might follow (v3.11 etc). I had some more luck creating the branch this time (read about my previous attempt on a Nucleus-3-0-1 branch). For help on selecting the 3.1 branch, see the WinCvs Daily Use Guide (provided you use the WinCvs software). A nice way to see branches & labels visualized is the 'graph selection' option in WinCVS.

Over the weekend, I've done some site-management on nucleuscms.org and responded to some of the waiting e-mails.

• Upgraded the forum to phpBB v2.0.10 (security fixes).

• Went through the list of pending approvals for the site list. Also removed a site which now points to a pornographic website (the owner must have abandoned the URL). The link also was on the donators page and welcome post, btw.

  It would be great if we had an automated script (NP_SiteList extended) that could run as a cron job and checks all sites once in a while (5 checks every hour or so). Automated checking could be done as follows:

  1. fetch the page, check if Generator: Nucleus vx is in the page headers
  2. if it's not, check the page for links like ?itemid=xxx, ?archivelist, ...
  3. URLs for which no automated checking could be done end up back on the manual approval pile

  Also, I'ld keep track of the time of last manual and automatic verification and require even the sites that can be validated automatically to be manually verified once every X months.

• Submitted a Nucleus CMS project to Freshmeat. Unfortunately, the Freshmeat staff decided that they already list too much similar projects and denied the request.

• Started investigating Google AdSense. I'll probably add them at the bottom of forum pages and see what the effect is.

1. I have a small pile of unanswered e-mail. If you sent me e-mail recently, don't worry: your mail will be answered (sooner or later :))
2. Yesterday, while preparing v3.01, I slightly messed up the CVS repository: part of the files have a Nucleus-3-0-1 branch (branchpoint on the Nucleus-3-0 tag). Since I couldn't manage to get all files branched, I gave up in the end. The branch is thus useless and does _not_ contain the actual v3.01 release files. If you happen to come across the branch, it can be safely ignored. If you're used to taking latest versions, there's no need to worry.
3. I ended up retrieving all files tagged as Nucleus-3-0 and applied the fixes starting from there. However, I forgot that before releasing v3.0 a quick fix was made to backup.php (rev 1.11: "sometimes fulltext keys are exported as KEY instead of FULLTEXT KEY, resulting in import problems"). That fix was committed to CVS after tagging v1.10 as Nucleus-3-0, so the updated backup.php is not in the v3.01 release :( (hcgtv pointed this out)

Hey! Nucleus v3.01 has been released!