media-playback: Fix possible crash on startup

[jeiea]

Description

Newly added frame width/height check crash if the frame is not given.

We can move the width/height check after other precondition checks.

Motivation and Context

I tested the master branch on my machine, and encountered the following.

How Has This Been Tested?

My local windows 11.

Types of changes

• Bug fix (non-breaking change which fixes an issue)

Checklist:

• My code has been run through clang-format.
• I have read the contributing document.
• My code is not on the master branch.
• The code has been tested.
• All commit messages are properly formatted and commits squashed where appropriate.
• I have included updates to all appropriate documentation.

[RytoEX]

It seems that on load/play (or end?) of playing an audio file, mp_media_next_video is called and d->frame is NULL.

	obs-ffmpeg.dll!mp_media_next_video(mp_media * m, bool preload) Line 383	C
	obs-ffmpeg.dll!mp_media_reset(mp_media * m) Line 593	C
	obs-ffmpeg.dll!mp_media_thread(mp_media * m) Line 748	C
	obs-ffmpeg.dll!mp_media_thread_start(void * opaque) Line 840	C

An audio file will end up in the early return at Line 400:

obs-studio/shared/media-playback/media-playback/media.c

Lines 391 to 401 in 0fc69fb

	if (!preload) {
	if (!mp_media_can_play_frame(m, d))
	return;
	d->frame_ready = false;
	if (!m->v_cb)
	return;
	} else if (!d->frame_ready) {
	return;
	}

Thus, moving this frame width/height checks after that resolves that issue, and avoids adding a nullptr check for f. I'm not sure if we should add the nullptr check anyway, but that can be discussed separately. We could also move this block further down to right before the if (preload) { section at the end of the function. @Lain-B ?

[Lain-B]

Either way is fine in my opinion. This is probably fine too.

[Lain-B]

For additional clarification, the check just needs to be after the:

	} else if (!d->frame_ready) {

check. That will filter out any invalid pointer usage.

[RytoEX]

For additional clarification, the check just needs to be after the:

	} else if (!d->frame_ready) {

check. That will filter out any invalid pointer usage.

Yeah, I'd figured out as much while testing this PR. We could move the check later to just before we assign width and height, but this is fine.