Upgrade to Laravel 5 about framework

[mikedugan]

I've worked on 2 applications built on L5, as with anything it's a rather tedious process of fixing everything when it breaks from a composer update. As of right now, Laravel 5 isn't even alpha, and package loading is currently being restructured, so I would strongly recommend you stick it out with L4 through January (there will be an accessible upgrade path afaik)

[cryptiklemur]

Honestly, the whole upgrade path of laravel makes me surprised you guys chose it as the base.

[tobyzerner]

@mikedugan Alright, thanks, we'll postpone the upgrade till it's stable.

@aequasi From what I gather, the upgrade from L4 to L5 won't actually be that painful at all. It's just at this point, L5 is changing so much that upgrading early causes a lot of pain.

[cryptiklemur]

@tobscure i meant the upgrade path in general. Laravel. BC breaks, undocumented changes, lack of community additions..... All kind of scary for a project like this

[lancepioch]

No, don't upgrade, it's not ready yet.

[mikedugan]

Suggest closing this thread to prevent it from turning into L5 thread :)

[yulianovdey]

Since L5 is finally out I'll start working on this.

[tobyzerner]

Thanks @yulianovdey. I have a few thoughts:

• We can ditch the laracasts/commander command bus and domain events for the native L5 one (which you've already mentioned :)
• I wonder if we should refactor the whole Api/Actions thing to be more "native"? Make use of controllers and middlewares and stuff? Haven't thought about this in detail, and I do like the current setup (it's a nice layer of decoupling from Laravel), but I'm open to suggestions.
• Will also need to work on the flarum/flarum repo and installation instructions. Specifically, workbench is gone.

Let me know if you have any questions.

[yulianovdey]

@tobscure I like the current setup, it helps enforce simple classes. I don't think there is any advantage to using controllers if you like the way it is now, unless making things more 'native' is worth the extra bit of familiarity new contributors will have.

As far as middleware, it may be a nice way to move your attemptLogin filter to its own class. I gotta look around more to see what else would make sense for it. Either way I'm assuming you want to keep as little as possible in flarum/flarum so we'll need a clean way to register any middleware out of the core. At least the default stuff doesn't even look like it's going to be needed since you're not using (cookie) sessions at all :).

[thangngoc89]

@yulianovdey It looks like we have to remove everything inside app/Kernel.php

[tobyzerner]

@yulianovdey OK, let's keep the current Api/Actions stuff :)

Don't worry too much about all of the auth stuff at the moment. I'm going to implement OAuth2 soon so it's gonna get a makeover. Will probably also make use of cookies somewhat (maybe).

[tobyzerner]

Hey @yulianovdey turns out I have been working quite heavily on the Laravel app — mainly auth stuff, and I'm refactoring API Actions so that it's easier to call them from within other routes. Will be pushing soon. Let me know how far along you are (if you've made a start) and I'm happy to help merge.

[yulianovdey]

@tobscure I haven't started yet, still working on something else. Looking forward to seeing your changes :).

[tobyzerner]

@yulianovdey No worries. The refactoring has become a bit of a mess and the next logical step is really to do the L5 upgrade. So I'll be making a start tomorrow!

[solhuebner]

Great work with the upgrade!