fix mcman detection of unformatted card insertion

[israpps]

Before this change, every time mcdetectcard was called from libmc over an unformatted card, the mcsync return value was -2

It should have been -2 if a new unformatted card was inserted since las call, and 0 if the same card remains plugged

fixes #785

Only tested on emulator so far...

For refence, here you have other instances of this code:

Soulcalibur2 DONGLEMAN.IRX:

int mcman_probePS2Card(int param_1,int param_2)

{
  undefined4 uVar1;
  int r;
  int iVar2;
  
  WaitSema(sema_hakama);
  mcman_55_retval = 0xff;
  r = mcman_cardchanged(param_1,param_2);
  if ((r == 0) && (r = McGetFormat(param_1,param_2), r != 0)) {
    if (param_1 == 0) {
      uVar1 = mcman_getcnum(0,param_2);
      SecrAuthDongle(2,param_2,uVar1);
    }
    SignalSema(sema_hakama);
    r = 0;
  } else {
...........
  }
  return r;

SCPH-50001/N rom0:XMCMAN:

int mcman_probePS2Card(int port,int slot)
{
  int iVar1;
  int iVar2;
  undefined4 uVar3;
  
  iVar1 = mcman_cardchanged();
  if ((iVar1 == 0) && (iVar1 = McGetFormat(port,slot), iVar1 != 0)) {
    return 0;
  }
  iVar2 = mcman_resetauth(port,slot);
  iVar1 = -0xb;
  ......

[sp193]

Could this be linked to this change within commit 8161284?

8161284#diff-1ec248a428f847eadfae186e29b3ecfe33fa22910461227cac5db73fc604430dL856-R866

[israpps]

seems to be....

@uyjulian

[sp193]

So anyway. If the official code is always like what you've found, I think there's no need to comment out the code; just make the change with confidence and Git will track the changes. But perhaps you might want to restore the original behaviour that:

1. Logged the outcome.
2. Returned sceMcResSucceed instead of 0.

Thanks for looking into this.

[uyjulian]

Lgtm