Nana Subete wa Daimaou no Omichibiki hangup when starting a match

[daniel229]

seems to never work before.
info log infinitely throw out these

user_main    W[KERNEL]: acf56f986e98e7c9\Core/HLE/sceKernel.h:460 Kernel: Bad object handle -1 (ffffffff)
user_main    E[KERNEL]: HLE\sceKernelMbx.cpp:510 sceKernelPollMbx(-1, 09fff66c): invalid mbx id

when it hang debug log infinitely throw out these

snd_thread_m D[ME]: HLE\sceAtrac.cpp:718 80630024=sceAtracDecodeData(4, 08aef340, 09fbda68[00000000], 09fbda64[00000001], 09fbda60[0])

debug log (rename jpg to rar)

[sum2012]

JPCSP fix in-game hang in
http://code.google.com/p/jpcsp/source/detail?r=3467
Not sure PPSSPP need same change

[sum2012]

Seem same problem in sceatrac in #3552

[sum2012]

I have fixed it.

fixed #3552 too.

@daniel229 Does this change break other games ?
sum2012@e7ae993

[sum2012]

@daniel229 Sorry,no need test,this is wrong

[sum2012]

PPSSPP sceatrac log
https://drive.google.com/file/d/0B3OaSdeV0L8kMGZtZ0RKSTdEc1E/edit?usp=sharing
Real PSP log:
https://gist.github.com/sum2012/a9093cacac79843de945

In Real PSP , do not return 80630024 (ATRAC_ERROR_ALL_DATA_DECODED)
edit:I find this one from real PSP log but PPSSPP do not have it
16:17:21.041003 snd_thread_main - <- sceAtracSetLoopNum 0x1, 0xFFFFFFFF = 0x80630021
Edit2:Just no logging in PPSSPP

[unknownbrackets]

Well, all those funcs really need tests. In the short run it'll break stuff but in the long run more things will work.

-[Unknown]

[sum2012]

I test from JPCSP emulator.I find that PPSSPP this check is too fast.
JPCSP also hang with this check first.

    if (atrac->currentSample >= atrac->endSample && atrac->loopNum == 0) {
        *SamplesNum = 0;
        *finish = 1;
        *remains = 0;
        ret = ATRAC_ERROR_ALL_DATA_DECODED;

[sum2012]

I have another idea,at least make sense to me.
edit:studenly compiler error :/

[sum2012]

@unknownbrackets sorry,ask you a question ,why comment this have compile error ?
sum2012@0612f95

[unknownbrackets]

You're doing this:

/*
if (x) {
   y();
} else {
*/
   z();
}

That won't work. You need to also comment the trailing brace, or leave it uncommented. Either of these will work:

/*
if (x) {
   y();
} else*/ {
   z();
}

/*
if (x) {
   y();
} else {
*/
   z();
/* } */

-[Unknown]

[sum2012]

@unknownbrackets Thanks ,this is a bit complex to me.

[sum2012]

@unknownbrackets Does this change make sense ?
sum2012@c6850a4
Modify log:
https://gist.github.com/sum2012/f2acddc7c1c3f34fb856

[unknownbrackets]

Hmm. I think the effect of that is that it will overwrite everything but ret. Theoretically the only difference should be atrac->getRemainFrames(). I'm assuming this improves it? Without tests I don't know, but it's possible it's correct.

You can try the original code and change *remains = 0; to *remains = atrac->getRemainFrames();.

-[Unknown]

[sum2012]

@unknownbrackets your change same with offical version
Hang due to call ATRAC_ERROR_ALL_DATA_DECODED
log:
https://drive.google.com/file/d/0B3OaSdeV0L8kRGd4MVhueXc4SnM/edit?usp=sharing

[unknownbrackets]

Well, I know that the first decode is supposed to return a smaller amount of data. If this game only works when it to take exactly X frames or something, that will be a problem.

So, your previous change made it work? It wasn't returning ATRAC_ERROR_ALL_DATA_DECODED?

-[Unknown]

[sum2012]

@unknownbrackets my previous change work,It only call ATRAC_ERROR_ALL_DATA_DECODED(80630024) once
edit: I mean second change (Move "passed the end" 's test to later)

[sum2012]

@daniel229 Does this change break any games (sound problem or hang) ? Thanks
sum2012@c6850a4

[daniel229]

quick test 30 games seems no problem
.

[sum2012]

Thanks @daniel229

[unknownbrackets]

Fixed by #6935, yay.

-[Unknown]