hello, Urban~

er, Urban!

Hey Mark.

how can I help you today?

Can I proceed with a question?

go right ahead!

Before I do, I would just like to thank you for amazing books! Really inspiring

thanks for the kind words!

SO, my first question is regarding View loading. So basically at configuration change, you can get nullpointer if the view that you are reffering to has not been loaded

One of the soultion for it that I have been using is PostDelayed and you delay it for some time

however, this is clearly not the optimal solution as for some it might be 50ms and for some 500ms

So, my question is if there is any option to run a function AFTER the view has been loaded?

you are the one who is causing the Views to be created: setContentView(), onCreateView(), etc.

so, you know when they are ready, in general

there are odd cases, such as Google Maps, but usually you know when your views are ready

So basically, if I call this function after I create a view it should not throw NullPointer?

generally, yes

and if it does, you need to ask yourself why that view is not ready when you think that it shoudl be

er, should be

thanks. But are there still any functions to delay the run of a funciton anyway (such as PostDelayed) but would acctually wait for completion of something?

it depends on the "something" :-)

for example, in the case of Google Maps, there are things that you cannot do until the map is fully laid out

such as move the camera

for that, you can use things like addOnLayoutChangeListener()

So if I understand you correctly, there is really not point of delaying stuff if you "know" what you are doing nad what are the processes?

and*

yes

of course, this can get tricky with third-party libraries that do not behave well, such as the Google Maps case

so I reserve the post()/postDelayed() stuff for cases where a library is not well-behaved and I need *something* to get the code to work

I also sometimes use it as a proof-of-concept, to confirm that my issue is "a timing thing", then try to find some cleaner fix to the real problem

but what time should one give in postDelayed if you decide for it? Seems pretty much a gamble to me

oh, absolutely

frequently, I find that post() is sufficient, as you just need to get your work onto the end of the event queue

if postDelayed() is needed... it's largely guesswork, based in part on your test devices and how you think they compare speed-wise to your users' devices

I have also stumbled upon "Idle Handler" class instead of "postDelayed()". Does it has any value?

I have not used it personally

Well, first of all thanks for your answer to my first question(s).

you are welcome!

Now to the second one if that is okay :)

sure -- nobody else is here, so go right ahead!

is it usually that empty btw?

it seems that you are very popular guy to get with contact

in contact with*

frequently, there is only one person here

there used to be more, back when I had more subscribers

Oh, okay

SO, the second thing that I found problematic in my app was the popupWindow. I wanted to prevent its dismissal with .setFocusable(false) which worked up to a some point. However, no matter how I was abusing the .setFocusable(), it still wouldnt work

The idea I had for fixing this was to use TouchInterceptor and catching touch events outisde

However, I would still like to keep my menu buttons on Action Mode clickable

But after trying several MotionEvent. combinations, I did not manage to find an click interceptor for a menu

why are you using a PopupWindow, rather than just views in your activity/fragment?

What do you mean with views? I have a popupwindow populated with ListView that has an Adapter, ...

I just do not see many people using PopupWindow versus either views in the activity/fragment or some form of Dialog, that's all

I found it preatty neat, that is all ;P

Dont really have a specific reason

OK

Fits into the app nicely

And it also works fine BUT the issue with dismissal

outside of a few samples (e.g., pre-O mouse-based context menus), I have not used PopupWindow or its variants very much

After long click on one of the items, it display actionMode whereas I would like buttons to be clickable at the start and not the click dissmiss the popupwindow first

oh okay, no worries

was just wondering if you perhaps knew how should I fix it

nothing comes to mind, sorry

No problem :)

Can I also ask you some of more general questions on Android developing while we are alone here?

go right ahead!

So basically, I am writing an app that would use internet online base (of my own) to give some data back to the user. For this, I was thinking I should use mySQL with PHP interceptor with Retrofit call. However, I have recently found that there is also a Firebase DB and several others. Is my thinking of doing this okay or is it slightly outdated principle of doing this?

it is a question of control

personally, I will not use Firebase services, as I would rather either retain control (using my servers) or give business to smaller competitors

so, in terms of running your own server, if you feel comfortable in doing that, there is nothing wrong with that approach

Would you reccomend anything better or does this seem sufficient enough?

well, I'm not a PHP guy, so I wouldn't be using that :-)

but PHP is fairly popular, so you should have lots of support for it

So, if that is not a secret, can you outline how would YOU achieve the same functionality that I have described?

well, for the new Warescription site, I used Kotlin and SQLite

SQLite because I don't have a particularly large database

Kotlin because I wanted to explore Kotlin in a non-Android setting

the previous Warescription site was Ruby and SQLite

But up to my knowledge, SQLite is a single-device based database whereas I would require an online one. Can it be extended to online access?

unless you have a server farm or database cluster, your MySQL is a single-device based database

you are wrapping it in a Web service (I think) implemented in PHP

in my case, I wrapped SQLite in a Web service written in Ruby, then in Kotlin

(though actually mine is a plain Web site, not a Web service, except for a few things for payment processors)

PHP is typically used with MySQL, and so that is probably the simplest database to use if you are going to go with PHP for the Web stuff

in principle, you could use PHP with whatever other database that PHP supports

Oh, so there is not that big of a difference between them... I thought I would gather all data on mySQL and than perform quiries and save data on personal SQLite

SQLite can be used on clients or servers

MySQL is not really designed for use on clients

your approach is perfectly fine, but there is no strict requirement to use MySQL on a server

for example, lots of businesses use Oracle

I happen to use SQLite

Thank you for that insight!

And (hopefully) last thing: I have gone through your book on Data Binding/JavaRx/MVVM...

And they seem very nice principles to master. However, they are sort-of-a intimidating for a "beginner" Android developer to use

I have tried using them in my app but it does go rather slowly

That being said, would you suggest me trying to master those principles or would you say that they are not worth losing time on=

what is your objective for Android development? for example, is the app that you are building just for personal enjoyment, or are you looking for it to be something that you show to prospective employers?

Both. I am planning to spend few months on it and than publish it

(If I manage to come that far)

I have written lets say half of it already using those principles as much as possible

But they do take a long time. SO my question would be if they are worth losing time on

for the benefit of employers, I would recommend that you demonstrate some experience with current development techniques, notably reactive programming and some structured GUI architecture

whether the reactive programming is RxJava, LiveData, RxJava+LiveData, Kotlin coroutines, or something else, is really up to you

similarly, whether your GUI architecture is MVC, MVP, MVVM, MVI, or something else, is really up to you

and if you were not going to be using the code as "proof of ability", you would be welcome to skip all of that, if you feel that you can deliver a quality app to your users in other ways

So, you are low-key urging me to go with these principles, correct? :-)

given your objectives, yes

but, that is why I asked about your objectives -- if the app was for personal education and nothing more, you're welcome to do what you want

employers are going to be less interested in the specifics of what your app does as much as how quickly you are likely to "come up to speed" on how *they* build apps

and the current trends are strongly towards reactive programming, formal GUI architectures, and Kotlin

the more you can demonstrate those things, the more useful your app will be in terms of employment

That is true. I should probably focus more on mastering these and less time complaining about the time I spend doing so ;P

And one veeeeeery last thing: I intended to use Firebase Cloud Messaging as a part of detecting mySQL changes to confer notifications + data changes to users. Would you say that using FCM is okay for that purpose or you have a better sugestion?

if you need real time asynchronous delivery of updates, FCM is "the only game in town", for better or worse

if you do not need it to be real time, and you want to use JobScheduler or WorkManager to periodically poll your server (e.g., every few hours), that's fine too

I like your "only game in town" comparison

I wish that we had more competition in the real-time push messaging space, but, then again, I wish that we had more competition with lots of things in Android

Are there any Android aspects (apart from ones we discussed) that you suggest for a "beginner" Android developer to learn?

fragments and configuration changes are important

Perhaps slightly less beginner ;P

"Intermediate" is perhaps a better put

well, after that and architectural concerns, it starts to either get into specifics based on app type (media players versus productivity apps versus games) or into "professional polish" (e.g., crash logging, performance analysis)

so once we get past the really beginner things, it is difficult to give blanket recommendations

Oh, okay.

Is it worth learning the Testing with Unit etc.=

yes

you may not need to get too fancy with it (e.g., Robolectric)

Which is the one testing method/program you would recommend me to learn?

If you can give one that is

having unit tests for things that do not depend on Android, and having instrumented tests for the things that *do* depend on Android, would be a good "proof of ability" for employers

in terms of particular technologies, Espresso and Mockito would be useful

And are these of much value in Android debugging processes?

I am not completely certain what you mean by "Android debugging processes"

tests are useful for detecting where you make mistakes with changes to code in the future

I was asking if these tests are more for complex scenarios or also "more basic" ones

But you have already replied on that one with the above

I view tests mostly as making sure that "future you" does not screw up what "current you" did

Haha well said :-)

All in all, I am very grateful that you had this discussion with me

I have learned A LOT

I am happy to help!

And hopefully this help me to become a better Android developer

Have a nice day :)

BTW, the transcript for this will be posted to https://commonsware.com/office-hours/ shortly after the chat ends

okay, thank you

Enjoy the rest of the day!

you too!