Nothing new on the frontend, but the backend for GPNL has been almost completely replaced now with my updated changes. The difference in speed alone is pretty nice. Where it took about 30 to 45 minutes to import the entire tree, now it only takes around 5. Plus, I’m importing more stuff this time. That means a lot more flexibility on my end to play around with the code and do some debugging if I have to.
There’s still some minor sorting issues to worry about, but I’ll have that worked out by today. I’m also working on a really cool proof-of-concept using the postgres database as a backend, that will hopefully be ready soon within a few days.
The next stop is to get GPNL to do some more reporting on QA checks. I need some ideas of what to look for, but already there’s a few things I can do just by running system-wide queries against the entire tree. One issue that has come up recently is the splitting of DEPEND and RDEPEND correctly for building binary packages. Given a list of packages to look out for, I can easily report that. Another one that phreak came up with was to look at redundant data in the metadata.xml files. I can also duplicate that with a simple query, so that’ll be easy to track it’s progress in getting cleaned up.
Anyway, I could use some more ideas. It’s one thing to have all this data gathered, archived and indexed, but it’s another to come up with creative ideas on how to use it. Looking through the QA bugs usually helps to give me an idea.