I mean... SQLite is famous for having almost a thousand lines of test code per line of application code. So yeah, I think they view the functions not just in terms of implementing, but also in terms of a commitment to robust maintenance.
That's a large part of it, yes. The sqlite project has only 2 full-time core developers and, AFAIK, only 3 current part-timers who tend to work on more "outer rim" features like the shell app, the .Net bindings, and the new WASM/JS parts.
Richard (sqlite architect/lead) has an explicit goal of supporting sqlite through at least 2050, and sqlite has a long history of strong backwards compatibility guarantees. Every feature (e.g. new SQL-facing function) added to it may well have to be maintained until 2050 by that small group of people. Thus Richard is necessarily extremely selective about what goes in.
I’m wondering if they have a probation (—edit— prohibition even, stupid spellcheck) on new utility functions (so they would only have to be written once and everyone would benefit) hidden behind feature flags or if nobody has bothered to send in a patch?
Also wondering if there’s a benefit to having the database call the functions over just doing it yourself in whatever language you are using? Not a db person so don’t know.
> if nobody has bothered to send in a patch
Quoting from SQLite:
> SQLite is open-source, meaning that you can make as many copies of it as you want and do whatever you want with those copies, without limitation. But SQLite is not open-contribution.
They pretty tightly control who can contribute to the code base for a few reasons .. they don't accept patches from the outside.