This would be a great addition. Rather than hacking the script we could use an API and you could have a plugins repository.
Yeah. We definitely don't have a full API...although there are some possibilities moving forward.
Right now the usersc folder offers a lot of options for different scripts. You can replace existing pages with your own and you can provide your own auto included custom functions as well as provide scripts for pre and post logout.
I'd love to see an example of where an API & Plugins would be helpful.
Right now I see UserSpice as a plug-in itself into my app(s) rather than the other way around.
What type of things would people want to write plug-ins for?
I sort of write plugins for different things...
I wrote one that takes advantage of the "account owner" feature that ties sub accounts to "master" accounts and enforces a limit on how many subs a master can have.
I write little plugins in the usersc folder that only certain access levels can access.
My thought on plugins is more for "sharabiltiy" that functionality. Anyone can write anything they want now, so plugins are more convenience for others, but I'm sure they would break with upgrades pretty frequently.
I totally agree with this.
If UserSpice starts taking all the best hacks, tweaks, etc... and can add a way for that feature to be used as a plug-in.
It would add more features, lessen user errors, reduce forum posts and support issues, etc...
Need some type of re-vert/back out mechanism in case plug-in fails and breaks like before the writes to files, etc... it copies the original files in a backup folder or something similar.
I think it would increase use of userspice for less experienced users.
I've gone back and forth on this. Ideally we would pull some things like MQTT and maybe some of our current features out and make them plugins. The downside is that I think it would be easy for these things to get left behind and make them more vulnerable things like that. People always want to delete files and db tables etc, which I get the logic, but it just doesn't translate into a performance gain.
That said, yes, I definitely want plugins and I've always thought that the MQTT (which controls Internet of Things devices was a good test case.
I am in agreement with a plugin/api/repository system, but with that, as mudmin suggested, could easily get left behind. This requires a strong level of community-development and support, regular updates and users following updates on an-as released basis, which just doesn't always happen. We're fortunate for those who are extremely active in the community, but we run into vulnerabilities, instability, and compatibility issues when plugins get left behind.
