Code Search


I'm currently working on a poor man's code search using the library I made and Universe as the backend. The first problem is I've never actually written an inverted index before so that was fun. It's a simple idea but the hard part is the actual searching and weighting. For now I just list out the index, but the plan is to add weights for modified time and who wrote the program originally.

The other thing I've run into is that I have bugs in my library that I need to fix. The biggest one being record IDs are left as utf8 which means they are all sorts of screwed up because universe works with iso8859. I'll need to convert record ids properly. I skipped it consciously because I thought blowing up would be better as IDs shouldn't have delimiters in them but now I'm thinking my library should just handle it rather than than erroring out.

I'm also using vue and one big issue that I've run into is that by having a deeply nested object and a ton of data, the reactiveness of vue is causing a slow down. I want it to be reactive on loading things but after that it should stop I think. In this case it's basically acting like ejs which makes using a frontend framework pointless. Good too see. Unless we need reactiveness throughout, it might make sense to focus on a server side page and then add vue on top of it.

Overall, doing something new and using my new stuff is quite the learning experience. It's not exactly fun but it is worth doing.