Prevent segfault due to division by zero in music.get_pos

[oddbookworm]

@pmp-p reported an issue in discord where this code can crash due to a division by zero:

import pygame
pygame.init()
pygame.mixer.init()
pygame.mixer.music.get_pos()

I tried to write a regression test for this fix, but I couldn't reliably reproduce it in a test environment. Because the variables involved are static to the music.c file, I believe this is only possible before any music has ever been loaded, and because other tests run before the created test, those static variables retain their last value, even if pygame.mixer.quit() or pygame.quit() has been called and mixer has been reinitialized

@bilhox reported that they didn't get the floating point exception on windows 11 (pmp-p and I are both using linux), but I suspect that is windows doing windows things and not actually showing the parachute info.

For reference, here is the parachute dump I get:

pygame-ce 2.4.0.dev1 (SDL 2.28.2, Python 3.11.3)
Fatal Python error: pygame_parachute: (pygame parachute) Floating Point Exception
Python runtime state: initialized

Current thread 0x00007f55234a3740 (most recent call first):
  File "/home/andrew/Projects/pygame-ce/test.py", line 4 in <module>

Extension modules: pygame.base, pygame.constants, pygame.rect, pygame.rwobject, pygame.surflock, pygame.bufferproxy, pygame.math, pygame.surface, pygame.display, pygame.draw, pygame.joystick, pygame.event, pygame.imageext, pygame.image, pygame.key, pygame.mouse, pygame.time, pygame.mask, pygame.pixelcopy, pygame.transform, pygame.font, pygame.mixer_music, pygame.mixer, pygame.scrap, pygame.system, pygame._freetype (total: 26)
zsh: IOT instruction (core dumped)  /home/andrew/Projects/pygame-ce/venv/bin/python 

And the gdb result on receipt of the exception

Thread 1 "python" received signal SIGFPE, Arithmetic exception.
0x00007ffff626da71 in music_get_pos (self=<optimized out>, _null=<optimized out>) at src_c/music.c:258
258         ticks = (long)(1000 * music_pos /

[pmp-p]

what about returning -1 instead of raising

[ankith26]

I also think returning a -1 is more consistent with existing code that already seems to return a -1

[oddbookworm]

Updated

[pmp-p]

works fine for me

[bilhox]

By returning -1, does it raise an error now ? Because I still have the same result as before (nothing printed).

[ankith26]

It doesn't raise an error. The code in the original comment prints nothing now because that is what is expected. Modifying the reproducer a bit to

import pygame
pygame.init()
pygame.mixer.init()
print(pygame.mixer.music.get_pos())

Should print a -1 after this PR. Before this PR, on systems where the segfault message isn't explicitly put into stderr for whatever reason, the program would still abruptly exit before printing anything so this PR is still a fix.

[ankith26]

The code and doc changes LGTM, thanks for the PR 🎉

I also just remembered, a simple unittest addition would be nice, but 'm approving this PR nonetheless

[bilhox]

The program still crashes in my side. I'm on Windows 11, perhaps I setup wrongly pygame-ce (I'm using oddbookworm's version).

[oddbookworm]

The program still crashes in my side. I'm on Windows 11, perhaps I setup wrongly pygame-ce (I'm using oddbookworm's version).

You should be able to grab a wheel from the CI runs and install the wheel to check it

[oddbookworm]

I also just remembered, a simple unittest addition would be nice

I agree, but as said in my first message, I couldn’t get it to work. If you have a suggestion to resolve the issue I was seeing with unit testing, I’ll happily write a test