Updated Criware CPK - Tokyo Mirage Sessions

[Ehm]

It looks like the CPK format has been updated for this game. The version in each CPK header is now "CPKMC2.47.04, DLL3.20.00". Here is a sample CPK archive in this new format. Any help with this is much appreciated!

[chrrox]

It crashes current tools on compressed files.
quickbms crashes with this error

it looks like its calculating the offsets incorrectly.
manually setting the correct offsets quickbms works fine.

Code: Select all

-------------------
*EXCEPTION HANDLER*
-------------------
An error or crash occurred:

*EH* ExceptionCode      80000001 guard page violation
*EH* ExceptionFlags     00000000
*EH* ExceptionAddress   00210b00
                        001F0000 + 00020b00 quickbms_4gb_files.exe
*EH* NumberParameters   00000002
*EH*                    00000000
*EH*                    07d8e1ca

*EH* Stack Trace:
0x210b00 : h:\quickbms\quickbms_4gb_files.exe : [unknown file]
0x233683 : h:\quickbms\quickbms_4gb_files.exe : [unknown file]
0x2480e9 : h:\quickbms\quickbms_4gb_files.exe : [unknown file]
0x24f427 : h:\quickbms\quickbms_4gb_files.exe : [unknown file]
0x27b0ed : h:\quickbms\quickbms_4gb_files.exe : [unknown file]
0x27a614 : h:\quickbms\quickbms_4gb_files.exe : [unknown file]
0x273821 : h:\quickbms\quickbms_4gb_files.exe : [unknown file]
0x27a84e : h:\quickbms\quickbms_4gb_files.exe : [unknown file]
0x27afff : h:\quickbms\quickbms_4gb_files.exe : [unknown file]
0x6478d8 : h:\quickbms\quickbms_4gb_files.exe : tripledes_LTX__mcrypt_algorithm_version
0x1f10fd : h:\quickbms\quickbms_4gb_files.exe : [unknown file]
0x77245de3 : C:\WINDOWS\SYSTEM32\ntdll.dll : RtlUnicodeStringToInteger
0x77245dae : C:\WINDOWS\SYSTEM32\ntdll.dll : RtlUnicodeStringToInteger

Last script line before the error or that produced the error:
  154 clog file_name file_offset file_size extract_size

[Ehm]

The files decompress properly if you manually set the offsets? How do you manually set those with QuickBMS?

[brendan19]

CRI Pak Tools managed to extract the CPK.

[Ehm]

I tried Cri Pack Tools initially. The data comes out corrupted. Those files should have an identical, readable header.

[aluigi]

With cpk.bms everything is correct except the offset which is 0x2000 bytes over the correct one.
So I have made a modification to cpk.bms 0.3.1 and now it works but consider it as a work-around:
http://aluigi.org/bms/cpk.bms

[Ehm]

Thank you so much aluigi, it worked perfectly! There was just one archive that didn't extract properly. Maybe the offset is at a different location in this one? It's 2.2GB, so I don't think I can upload the whole thing. Here's the first chunk of data from the top, it should contain all of the header information: http://puu.sh/on0Jb/110c164d25.7z

[aluigi]

With that cutted sample I have an error because the script can't reach offset 0x88a8bc40 which is inside the 2.2Gb of the original archive but not in the sample.

Can you explain better what's the exact problem?
Do you get a crash in quickbms?

[Ehm]

Here's the crash I get in QuickBMS. The same thing happens with the 4GB QuickBMS version.

[aluigi]

Ok, so I guess that the offset wasn't aligned and it tries to decompress data which is not compressed.
Anyway there is nothing else I can do about cpk.bms.

[Ehm]

Alright, thanks anyways!

[Ehm]

Here is the remaining incompatible CPK archive from Tokyo Mirage Sessions. Hopefully having the entire archive will be useful in figuring out what's going wrong.