Nucleus 2.0.0 about nucleus HOT 16 CLOSED

interesting how you removed OperatorSemaphore 👍

from nucleus.

Going to try this out, looks very promising!

from nucleus.

@konmik, without getView(), how can I call the view without using the restartable methods?
I was doing something like this:
Observable.just(null).compose(deliver()).subscribe(o -> { getView().doStuff(); });

Thnks =)

from nucleus.

Hi, you can do the same but now after deliver you will get the View and data bundled into Delivery object.

from nucleus.

OK, since deliver() was removed, is the deliverFirst() that have the same behavior?
Ins`t that too verbose? Because after subscription I need to get the Delivery and create two more Action2 to call the view object.

from nucleus.

By example: This is the after and the before.

If I change the deliver() to deliverFirst() can produce NullPointer

from nucleus.

This is how it looks like now:

    add(db
        .updatable(
            Label.TABLE,
            () -> db.query(Label.SELECT_ALL_SQL, Label.fromCursor))
        .compose(this.<SolidList<Label>>deliverLatestCache())
        .subscribe(this.<SolidList<Label>>split(InputView::onLabels)));

from nucleus.

GOT IT!

Thanks again @konmik

from nucleus.

Everything has been going great, and I honestly love how much better things feel in the 2.0 beta release. I have run into one problematic situation though, and was wondering if anyone else had a problem like mine.

Currently, I have an app with a NavigationView widget, and a TabLayout with child fragments inside of the container fragment. The child fragments all have their own presenter, along with the parent fragment having it's own presenter.

My problem is, whenever I have a configuration change (read: rotation) when that fragment with tabs and child fragments is in view (or have been shown/loaded once), the app will crash due to some error thrown by Nucleus.

The lines it hits are here:

public void onRestoreInstanceState(Bundle presenterState) {
    if (presenter != null)
        throw new IllegalArgumentException("onRestoreInstanceState() should be called before onResume()");
    this.bundle = presenterState;
}

Which is found in the PresenterLifeCycleDelegate class (82-86). I'm not quite sure why it's hitting this, from what I've tested (it's been hard to narrow down specifically what is wrong, but I think it has to do with Child Fragments). If more information is needed, I'll gladly test anything you request and let you know my findings.

from nucleus.

Hi, @cnnranderson!

Are you re-attaching the same fragment a second time or you're re-creating them? Or you just create them once and let the fragment manager to manage them?

I've just tried Nucleus with a child fragment and it works fine!
(I never tried child fragments before, but they work! :)

If you need a help, please create a simplest possible example that can reproduce your problem. I normally help during one-two days.

from nucleus.

BTW, I'm going to release Nucleus 2.0 as soon as your issue will be resolved. It (Nucleus 2.0) is pretty stable now.

from nucleus.

Oh, thanks for the quick response! I will try and throw together a basic app/repo demonstrating the problem (hopefully I can reproduce it along the way -- or better yet, not reproduce it and figure out where I went wrong!).

I am using a fragment pager adapter to create new instances of the fragments. Here is a snippet of the parent fragment class and the pager adapter, if it's relevant at all:

Parent: https://gist.github.com/cnnranderson/5770141012f80407468f
Adapter: https://gist.github.com/cnnranderson/c865e1da4ad380e4f7d2

Also just out of curiosity, does the parent Activity need to be of type NucleusAppCompatActivity<>?

from nucleus.

You inherit from Nucleus view classes if you need a presenter.

from nucleus.

Apologies for the delay. Got a little pre-occupied the past two weeks with other projects.

I've discovered the reason for the configuration problems lied in when I was injecting my fragments/presenters. Didn't realize I had been injecting in the onAttached method, which it should have been in onCreate.

I would like to ask, if you ever get a chance, to review for leaks using leakcanary. I'm not sure if it's on my end, if I'm not using restartable's correctly, or if it's the way the presenters work. I keep getting fragment instance leaks from their respective presenters.

from nucleus.

Hi, I tested for leaks thoroughly. I normally use MAT memory analyzer during development a lot.

I'm not sure if LeakCanary will work properly here - we're doing some background jobs that it can consider to be leaks.

from nucleus.

Ah okay, that must be it then. Thanks for the heads up.

from nucleus.