Hello friends, how are you all?

I am trying to figure out the mouse click event for every button in my program,
the mouse events were working fine until I wrote code for ImageButton. Now my program is crashing whenever I hover over or mouse out of the 6th or 2nd button.

I tried debugging but even after lots of tries I couldn't find the problem.

Here's my code for Buttons:

I have attached the complete code.
I hope you are all safe and sound.

Good to hear from you Doctor!

Have a look at this, main.c, line 87 (approx)

        // close the display
        else if (event.keyboard.keycode == ALLEGRO_KEY_ESCAPE || event.type == ALLEGRO_EVENT_DISPLAY_CLOSE)
            break;

Checking for keycode being 'escape' but you didn't check first if that event was a key event. The event structure is a union and all the different possibilities are laid over each other and you need to check the event.type to figure out which one is valid.
Here, a mouse event with x=59, when interpreted (incorrectly) as a key event, just happens to have a key code of ALLEGRO_KEY_ESCAPE.
Here it is in LLDB

(ALLEGRO_EVENT) $0 = {
  type = 20
  any = {
    type = 20
    source = 0x0000000100199890
    timestamp = 2.275083
  }

...skip a bit...

keyboard = {
    type = 20
    source = 0x0000000100199890
    timestamp = 2.275083
    display = 0x0000000100675dd0
    keycode = 59
    unichar = 165
    modifiers = 0
    repeat = false
  }
  mouse = {
    type = 20
    source = 0x0000000100199890
    timestamp = 2.275083
    display = 0x0000000100675dd0
    x = 59
    y = 165
    z = 0
    w = 0
    dx = -1
    dy = 0
    dz = 0
    dw = 0
    button = 0
    pressure = 0
  }

So, when you move the mouse to x=59, the app will try to exit.

Now, in the last part of main:

    al_destroy_font(font);
    al_destroy_display(disp);
    al_destroy_timer(timer);
    al_destroy_event_queue(queue);
    al_destroy_bitmap(icon);
    al_destroy_font(font);
    return 0;

You've got al_destroy_font twice and that is what causes the crash (segfault)

That's as far as I got, there may be more.

What I did:
0. Ran in the debugger
1. Noted the segfault, found out where by doing a backtrace.
2. fixed that
3. Noted the app was now exiting 'for no reason'
4. There's only one line which could possibly cause an exit from the loop
5. Looked at that line and saw the logic error.

Hope that helps!
Peter

[edit] forgot to say, in your code you'd commented out al_install_mouse() and al_register_event_source(queue, al_get_mouse_event_source()) so I uncommented those.

Hi Peter, thanks for the detailed solution.
I can't thank you enough, it solved my problem and gave me a new perspective on ALLEGRO EVENTS. Now I look at it I can think of chaining events instead of checking a single event to do everything.

One more thing, can you tell me if using a single event queue is enough for everything or I should use different queues for different things such as drawing UI and App logic?

@Arthur Kalliokoski, I see you linked -lallegro -lallegro_primitives -lallegro_font -lallegro_image -lallegro_ttf in the compilation phase, I just link the monolith, does it make any difference by linking all the shared lib files separately?

It's a side-effect of cargo cult programming that I developed a habit of at my previous workplace, I'm trying to leave this bad habit but it still affects me one way or other. I'm trying to improve by making templates off of my previous projects so that I do not write Frankenstein code anymore.

I have attached the png file now.
It's from asset jesus's miracle pot. (Kenny Assets)

Edit:
I was wondering why my buttons weren't "clicking". So I changed the
controller_click_up(&buttons[i].M, &state);
to
controller_click_up(&buttons[i].M, &event.mouse);

IDK why it fixed it or what was causing the problem. If anybody can tell me, I would appreciate that.

I don't believe I've ever used more that one queue. Not sure what that would be useful for, maybe if you were doing multi-threaded apps?

In controller_click_up you've got

    if (state != NULL)
    if (state->buttons & 1)

I think state->buttons is always zero for the button you've just released in a mouse up event, therefore the following code never runs.

One more thing, it's normal to put the bulk of the code in .c files and just have the declarations in .h files (in C++, exceptions for very short functions which can be inlined)

Event queues are immensely useful in multi-threaded applications. Each thread gets its own synchronized access to the queue handled by a mutex. That's how my library does it anyway.

Rant:
I'm tired of seeing people struggle with things like keyboard and mouse events and state. My library has a full input handler that can handle anything you want. It is both event based and a state engine. Each event is like a tick of the clock for input. You can get presses, holds, releases, opens, up state down state, for each button joysticks and mice included. And then there are actual Input objects that can store a reference to a button or key and be checked automatically using a boolean conversion. I will support touch as soon as I have a device that has touch. ;P

What's more, the InputGroup class can hold combinations of keys and button states. It makes configuring your program immensely easy because you have a map of keys to actions you can change anytime.

You should have said something to me, I would have helped you immediately.

I can make some binaries for you, if you like, or help you get it to compile. Either is fine. Are you still on Discord? I am # 2 6 0 1 ,
M o d u l o 5 k .

I will send you a link to my eagle discord channel, where we can discuss your problems. The cmake version is probably too high for normal use. I always use the latest cmake, so it never bothers me.

Let me put together a clean release. I want to brush up the docs and make sure they're ready before I release anything.

As an aside, I have no idea what version to call it. I've been working on it so long, it's got to be close to version 1. Or maybe I should adopt the YY-MM dating that is popular these days.