[fixed /closed]Scab-Clan Giant

[stassy]

Describe the Bug:
Scab-Clan Giant generate an error window when entering the bf and if opponent has more than one creature in bf

Which card did behave improperly ?
Scab-Clan Giant

Which update are you using?(date,name)Which type(Duel,Gauntlet,Sealed Deck)
OIv2

What exactly should be the correct behavior/interaction ?
Scab-Clan Giant does not generate an error window when entering the bf and if opponent has more than one creature in bf

Are any other cards possibly affected by this bug ?
-

This seems to no affect the game at all but corrupt savegame but this is an unusual error with trashed registers mention (I think I got on of those long long time ago)

[Aswan jaguar]

I got one couple of days ago,for the first time after 3 years.But I don't remember more(probably AI turn - beginning the game for sure)

[Korath]

Aswan Jaguar, your popup's unrelated. It's for an EVENT_ABILITIES (0x34) call to RukhEggFx (0x00457e10); that's the legacy card that got rewritten so it calls arbitrary code, so it'll be essentially impossible to debug unless you can find some way to reliably trigger the popup, much the same way the "Wrong parameter at address:" popups are when they only give the address of the C wrapper for get_card_instance.

As for Scab-Clan Giant, I'm not sure. That popup is generated when a card's function changes the ebp, ebx, esi, or edi registers; these are all callee-saved, so if gcc generates code that uses them, it'll save their previous values at the start of the function it's compiling and restore them at the end. And, indeed, that's what it does for card_scab_clan_giant() in the OtIv2 version of manalinkeh.dll - so even if it one of the other functions it calls trashed the registers, they should still have been properly restored.

I'm a little tempted to change call_cards_function(), which 99% of the calls to individual cards go through and is the only function that pops up this error, so that it saves and restores ebx/esi/edi itself instead of just checking to see if they weren't properly preserved. Problem is, if a card function genuinely doesn't clean up one of these registers, it'll cause random errors in the 1% of calls that don't go through call_cards_function().

[Korath]

Never mind, I see why - you can't do "int cr_array[] = { 0 };" - that just allocates an array of length 1, with cr_array[0] initialized to 0. When you then try to store a second creature at cr_array[1], it overwrites the previous data on the stack; that happens to be one of the saved registers, so it gets restored wrong. Try int cr_array[500] (and there's no need to initialize it, since cr_count keeps track of the last-used index).

Also, Scab-Clan Giant doesn't do anything if opponent controls exactly one creature; that's because trg is left at 0 instead of being assigned to cr_array[0].

Gargaroz, I'll install both these fixes when you send me the code. (edit:done)

(I haven't fixed the part where it's supposed to target the random creature, so can't pick shrouded/hexproof/protection-from-red creatures, nor scare a Skulking Ghost into the graveyard before the effect resolves.)