TF2_OnIsHolidayActive does not catch all cases

[Mikusch]

Many years ago, #42 changed the TF2_OnIsHolidayActive forward from a TF_IsHolidayActive detour to a CTFGameRules::IsHolidayActive virtual hook.

I'm not quite sure how these functions looked like back then, but the current behavior of the TF2_OnIsHolidayActive forward is most certainly wrong. TF2 has direct calls to TF_IsHolidayActive all over game code, meaning that the forward does not get called for these invocations.

However, CTFGameRules::IsHolidayActive directly calls TF_IsHolidayActive, meaning that this can be fixed without breaking anything in the process by changing the forward to the signature call (again).

[asherkin]

This was an intentional trade off, the reasoning is given in the linked PR.

We could do both, but that would result in differing behaviour between Windows and Linux which I think would be much worse.

[Mikusch]

the reasoning is given in the linked PR

If this is about it being inlined; it's not. At least not anymore. The function was expanded over the years.

Here are the current signatures: Linux: @_Z18TF_IsHolidayActivei Windows: \x55\x8B\xEC\xA1\x2A\x2A\x2A\x2A\x83\x78\x30\x00\x74\x2A\x32\xC0

[psychonic]

I agree, with that explanation, switching back seems reasonable.

[KyleSanderson]

does hooking multiple vtables here help? or is it really that direct on the call