Poggit came first. Then me, @Edwardthedog2 , @Thunder33345 , and @archie426 made Plug-On, a new plugin list. @Primus joined in our efforts later. We then merged with @SOFe , combining our efforts on Poggit (Poggit is supported by the PMMP team) . At some point, @Primus left, and decided to make Sharrit. The PMMP team probably doesn't want tons of Plugin lists running around, so that people making plugins don't get spread thin across multiple platforms.
I agree with all you just said, but Sharrit from Poggit and Plug-On differs because it was released, although short time after I took it down for public. I'll watch how things go with Poggit and if nothing changes, I will try to publish/announce it here if Staff (@Intyre, @99leonchang, @SOFe, @dktapps) does not mind.
Personally, I support using Composer for plugin installation. Poggit is trying to do too much, and I feel as if the end result will be poor. On the other hand, Composer is a well-tested code base that is under much more active use. Oh, that's right. I'm a strange person who cares about PHP standards (however few and flawed they are).
Composer is not meant for PocketMine plugin fetching. 80% Minecraft servers are hosted by kids who doesn't know how to use Composer and has seen CLI only in tutorial videos.
I wonder why you don't like it. Sure, it's not a perfect solution, but Poggit isn't either. What's the advantage in using Poggit over Composer? The only serious argument against Composer would probably be the lack of support for Composer on shared hosting (but that is a different can of worms).
Among lots of other things, I'm currently doing a lot of work to bring Composer to PocketMine. We shall see where that goes.
And if staff minds... You are already announcing it here. I can see the advantage of using Composer within PocketMine, but I don't see the advantage using Composer for plugins. How plugins work is simply different from how normal PHP projects do. I'm not sure about how plugin release and installation will work in Composer, but I am sure that Poggit can do everything designed for PocketMine. Call me NIH if you want, but things locally designed after all suits locals better. Also, Poggit 2.0 is going to offer the feature of Virion, which can allow plugins to use libraries more widely without easy duplication or even collision.
We can critique the API design of PocketMine at a later time, but Composer already offers what "Virion" promises to do. This is one of the most clear-cut cases of NIH I've seen. You've promised continuous integration, dependency management, release management and localization management in a single tool. In my opinion, you're either going to do one thing well or you're going to do everything poorly - there's no continuum in between. Composer tackles dependency management very well, and it is extendable to managing plugin setups. It's not a perfect solution, and perhaps Poggit might end up as the clear front-runner. But at the moment, I don't see the benefit of this 4-in-1 tool.
@Tux of course I am going to do one thing by one thing. But they are interconnected. Dependency management and localization management are going at lower priority. Continuous integration is already completed. Release management is being worked on. There is no rush, and everything is going progressively.
I fail to see how your reach that conclusion; every tool (and person) I've ever come across lies somewhere on that continuum. However, I can help you with the latter problem: it's because you haven't actually used poggit yet SOFe is doing an amazing job building a tailor-made github based tool for plugin distribution that will make life easier for server owners, devs and the review team - with a well designed dev workflow, pocketmine-specific LINT and much more... why not just keep a positive open mind and wait until you've seen what it can do for yourself?
My first impression still stands. My mind is open but highly skeptical. Perhaps this is because my expectations are set too high in proportion to what's expected.
In my opinion one should always reuse things. That way you can focus on the bits that are important/unique rather than things that have been done elsewhere. I am not much of a fun of "Composer", but it is well used and a lot of people know of it. The good thing about using it, specially for the kids doing PocketMine is that if they learn how to use it, when they go to the real world, they would already have the experience, so they would be more qualified for jobs. I myself use for my plugins Travis-CI which provides me with CI and Release management. Shoghi used to use Jenkins. So, it is in a way for me a little bit of a surprise that we are moving to a plugin specific CI tool instead of re-using something that is already there. On the subject of Poggit itself, I am wondering how is it going to support automated testing. (Which is one of the big features of CI).