vunderba 2 days ago

I really like Pocketbase as a simpler and locally hosted Firebase/Supabase alternative (yes I know Supabase can be self-hosted but its pretty heavyweight) except for one major disadvantage - the lack of a NULL primitive type.

There's no ability to set null as a value of an item despite null being a perfectly valid data type in SQLite (which is what Pocketbase uses).

Having to arbitrarily assign a MAGIC CONSTANT value (like -1 for example) to represent a "lack of information" in a column is kludgy at best, and impossible at worst - e.g. if all floating point values have meaning in a column, you're out of luck. The only workaround is to create an unnecessary additional boolean column isNumFilled to indicate if the data in the numerical column is valid.

  • hahn-kev 2 days ago

    Yeah that's strange, it's like they heard that null was a mistake and removed it without realizing that it provides a useful function

jasfi 2 days ago

For a project with 42.2K stars I thought there would be sponsors by now.

demarq 2 days ago

The answer from the maintainer is PERFECT.

This is the situation for a lot of open source projects.

ianschmitz 2 days ago

(2023)

  • d1sxeyes 2 days ago

    Yep, but the situation hasn’t changed.

    In fact, ~~development velocity has slowed recently~~ PocketBase has been in a feature freeze recently.

    But like with any project like this, you need to weigh up the pros and cons of whatever tech stack you choose.

    PocketBase at least is MIT licensed, so you can fork and fix/add new features if you want. If you’re running a successful business with PocketBase as a backend, it might be sensible to do this anyway, and potentially hire a Go dev or consultant to do whatever work you need doing, rather than hoping Gani has it on his priority list.

    • joseda-hg 2 days ago

      Wasn't the feature freeze in preparation to v1? Not saying is hasn't slowed down, but it seems reasonable

      • d1sxeyes 2 days ago

        Not sure he explicitly said that, but he’s said that there’s a bunch of technical debt he wants to fix through refactoring etc, which will be v0.23, I’m not sure path goes to V1 next (at least directly) but it may. He also stopped accepting donations in Jan saying that he felt like donations came with expectations. I’m not saying it’s not reasonable, just that the whole ethos behind PocketBase is “I’m building this, you folks are all welcome to use it, but I don’t owe you anything” (completely fairly).

        Asking “will it survive” is the wrong question. It’s already there, works, and is used by thousands of people. Whether it gets updates or not depends entirely on whether Gani wants to and is able to keep updating it. No one can answer that except Gani, and he’s saying “no promises”.

    • demarq 2 days ago

      There’s already a maintainer, why not give him a contract?

      • d1sxeyes 2 days ago

        Not sure he wants it. PocketBase is just a backend he developed for his main project, Presentator, and everything I’ve read from his conversations are “Yep, I’m publishing this because someone else might get some use out of it but I’m not developing by consensus here”, and I don’t think he wants it to be his full time job (he’s always super polite and helpful, not trying to imply he’s rude or dismissive or anything).

        Obviously everyone has a price, so maybe I’m wrong, but even then you’ve not solved the “bus factor” problem.