Bug: Warp_Flash broken on nightly builds past 2024.07.02

[skdKitsune]

Starting with nightly 2024.07.03, the warp flash effect does not show up on some pc configurations.

https://fsnebula.org/log/669ecad1c241ad89f80010bb - log for 2024.07.02 (the last nightly that works)

https://fsnebula.org/log/669eaca0d9fd840ff60010bd - log for 2024.07.19

[MjnMixael]

I tried to track this down but was unable to reproduce the issue. It's clearly tied to #6214 where, notably, Warp Flash used to require a game restart. I removed that requirement because my testing showed (and still shows) it shouldn't be necessary. It is possible I missed something, though, so I did additional testing on a fresh Release build to make sure no code was optimized out which has caused issues with In Game Options in the past. Here is what I said on Discord after doing the testing.

I built a fresh Release build (that takes upwards of 20 minutes on my shitty work laptop) and launched the MediaVPs, attaching debugger. I set a watch on Fireball_warp_flash, the boolean the option toggles. Break when the value changes. I also set a breakpoint where that variable is checked during the warpout rendering to decide if a warp flash should be drawn.

Changing the option always correctly set the boolean. Didn't matter if I did this during a mission or not. This ruled out the usual issue I've had in the past where Release builds had optimized out the code that actually set the variable.

I did in-mission testing of both the player warpout and NPC. I didn't think they used different code paths but I wanted to be sure. Each warpout showed the Fireball_warp_flash boolean still being set to what I expected it to and the warp flash drew when it was true.

Finally I verified one other wrinkle; that the warp flash bitmap was properly loaded. I theorized that perhaps if Fireball_warp_flash was false at game start or during mission load then the bitmap would not be loaded which would cause rendering to be skipped. However, this proved not to be the case as even if warp flash is disabled the bitmap is still loaded into bmpman and the ID correctly set in the fireball data.

This was all in addition to visual verification that the warp flash was, in fact, visible.

Outside of that, I'm not sure what else to test. That feels pretty thorough, but maybe I missed something?

Beyond this I'm not sure how else to try and track this down without being able to reproduce it.