Many are wondering why development is so slow lately. The reasons are mainly two:
- the new server required a bit of work for configuration and data migration
- the activity in the forum has increased a lot, thus requiring more time than before.
So, what is under development?
Basically, it’s a giant debug phase. Version 2.0.0 suffered a lot of bugs when released, and I want v3 to be rock solid. Many bugs have been fixed since version 3.0 RC0, but there is still one important problem: the search engine index gets corrupted quite often.
The reason is that words are identified with a 64-bit integer, generated by hashing the word. This turned out to be a problem because the algorithm I decided to use (named SDBM) seems to produce many collisions with real-world words. I did an extensive testing with randomly-generated strings and never hit a collision after millions of different words. Well, there is nothing worse that unrealistic test data, I guess.
I’m performing a change in the search engine so that it will use a sequential 32-bit integer for identifying words (and documents). It’s a bit more complex but it’s required.
Also, the WYSIWYG editor turned out to bee a bug mine, so I’m not sure about its future in the next versions (I mean 3.1+).