Threaded Mode | Linear Mode

Archer30 · 04.12.2023 14:08

Whether it's a bug or not, I don't see any case UN:U returning the wrong number of towns would be favoured.

XEPOMAHT · 04.12.2023 17:26

(04.12.2023 14:08)Archer30 Wrote: Whether it's a bug or not, I don't see any case UN:U returning the wrong number of towns would be favoured.

Not bug. But UN:U not support objects with two yellow cells - it is worked only on one. If you want towns count on map, don't use UN:U.

wessonsm · 04.12.2023 21:49

Герой не должен стоять на красной клетке, по идее.
Я думаю, скорее всего это вина забагованного скрипта, который его туда поставил, надо править сам скрипт.
Остальное следствия этого.

Berserker · 04.12.2023 23:19

Согласен. Нельзя накладывать один объект поверх другого.

Archer30 · 05.12.2023 03:37

As I know there are two ways to put a hero in the red tile of a town.

- By bug of Flying spell. Sometimes the hero stays at the red tile after moving with Fly. I can't reproduce it now but it was reported from time to time.
- By scripts. We try to avoid this situation.

XEPOMAHT, thanks. Could you tell me what UN:C should be used to obtain the real town number? We don't have access to the Russian Erm help.

Berserker · 05.12.2023 04:16

Yep, getNumTowns() function would be a stable solution.

XEPOMAHT · 05.12.2023 08:39

(05.12.2023 03:37)Archer30 Wrote: thanks. Could you tell me what UN:C should be used to obtain the real town number? We don't have access to the Russian Erm help.

На Си - одна строчка кода:

Code:

GetTownsCount() {return ((_int_)((*(_dword_*)(((_ptr_)this) + 0x21618) - *(_dword_*)(((_ptr_)this) + 0x21614)) / 360));}

На вого-ассемблерной вставке так:

Code:

int GetCastleMapCount(void)

{

    STARTNA(__LINE__, 0)

    _CastleSetup_ *strt,*end;

    __asm{

        mov  eax,BASE

        mov  ecx,[eax]

        add  ecx,0x21614

        mov  eax,[ecx]

        mov  strt,eax

        mov  eax,[ecx+4]

        mov  end,eax

    }

    RETURN((end-strt)/360)

}

На ERM2 это как-то тоже можно написать...

daemon_n · 05.12.2023 17:31

Archer30,

!?FU(GetTownsCount);
!#VA(result:x);
!!UN:C(GAME_MANAGER)/4/?(gameMgr:y) C(gameMgr)/21618/4/?(townsEnd:y) C(gameMgr)/21614/4/?(towns:y);
!!VR(result):S(townsEnd) -(towns) :360;

Archer30 · 05.12.2023 18:03

Very cool thanks!

But in the end UN:U for finding next town still returns hero if he/she stucks in red

daemon_n · 05.12.2023 19:10

Archer30, mb cause if when it finds hero it checks object below? and it is probably town with new coord?

Archer30 · 16.12.2023 17:30

Hi Berserker, there is a weird bug report from the community. The log is a bit confusing.

Quote:Failed to write data at F8558D0E.
EIP: Era.50076 (AdvErm.GetSlot + 98 in AdvErm.pas on line 1220 offset 2). Code: C0000005

> Registers
EAX: 00000001 (int: 1)
ECX: 0000000A (int: 10)
EDX: 00000009 (int: 9)
EBX: H3era hd.00A5A008 (int: 10854408, pint: 0x00000000)
ESP: 02D359E8 (int: 47405544, pint: 0x02D3617C = 47407484)
EBP: 02D35C00 (int: 47406080, pint: 0x02D36168 = 47407464)
ESI: 00000001 (int: 1)
EDI: 02D35A00 (int: 47405568, pint: 0x02D36001 = 47407105)

> Callstack
H3era hd.007493CC
Era.64AD8 (Erm.ProcessErm + 2456 in Erm.pas on line 5802 offset 11)

Should that be an erm issue?

Download Debug here

According to my test, it has 95% chance to be related to Remagic (Era Scripts option) 102

Berserker · 16.12.2023 17:43

According to write address, seems like Era memory corruption.

Archer30 · 16.12.2023 21:57

Berserker, should that be resolved on scripts or your end?

Berserker · 16.12.2023 22:58

Memory corruption is scripts area. I think some UN:C or plugin write operation could do it. Maybe some buffer overflow in EEF function, but not Era core.

daemon_n · 17.12.2023 00:01

Archer30, try same w/o these suspisious hooks

Threaded Mode \| Linear Mode ERA III
Author	Message