[pandaman]

Int3 is a one byte instruction, how do you imagine this code could possibly compile into one byte? This will compile into several moves/stores. Not to mention the code is not x86 specific at all.

It's a forced memory access violation and the 3 is used to distinguish it from a natural one: when you hit this you will see an immediate 3 load nearby and that's a constant that is rarely seen in code normally.

[leetNightshade]

"Not to mention the code is not x86 specific at all."

Technically no, however it's only used in x86. It's only ever called if PLATFORM_EXCEPTIONS_DISABLED is 0. If you look up where PLATFORM_EXCEPTIONS_DISABLED is defined in Platform.h, you'll see it's only defined when not using PLATFORM_DESKTOP. PLATFORM_DESKTOP is disabled for Android, HTML5, iOS, WinRT, and WinRTArm. PLATFORM_DESKTOP is enabled for Linux, Mac, and Windows. So that instruction is only intended to be used in x86 desktop environments.

Considering this, I don't think RaiseException is well defined. If it's only meant to be used for x86, it should be coded that way instead of relying on other code to enforce the behavior.

[pandaman]

>It's only ever called if PLATFORM_EXCEPTIONS_DISABLED is 0

Quite the opposite. The ! operator means a logical negation.

[leetNightshade]

if PLATFORM_EXCEPTIONS_DISABLED means if it is not 0. if !PLATFORM_EXCEPTIONS_DISABLED means if it is 0, if it is not disabled. I said nothing incorrect, I'm well aware of what it means.

[pandaman]

You said the code is being called only if PLATFORM_EXCEPTIONS_DISABLED is 0. This is incorrect. Just compile that function and set the breakpoints if you cannot figure out the #if/#else logic.

[leetNightshade]

It's getting hit because you're running on Windows/Mac and Windows/Mac specific methods are calling it. In platform generic code like OutputDevices.cpp you see this:

    #if PLATFORM_EXCEPTIONS_DISABLED
    		FPlatformMisc::DebugBreak();
    #else
		FPlatformMisc::RaiseException( 1 );
    #endif

[pandaman]

I see. You are saying that because a library/framework is not calling one of its own functions in a certain configuration nobody ever going to call this function in that configuration, right? I am not sure I can join you in this belief, after all there are not that many fully recursive libraries so some functions will never be called in almost every library.