No support for Tales of Graces F Remastered

[nekkk]

The game presents somehow a new .zarc compression that seems similar to TLZC but not quite the same. Would it be possible to add support to decompress/compress this archive, both for PC and NX?

I'm trying to find the game's dialogues for translation purpose.

[AdmiralCurtiss]

This is a really weird archive format and I'm not entirely sure what to do with it. I've pushed some attempt at unpacking this but this is definitely wrong for some files. If you can figure out wtf some of those header fields mean I'd be very happy to improve this.

https://github.com/AdmiralCurtiss/HyoutaTools/blob/master/HyoutaToolsLib/Tales/Graces/ZARC/ZARC.cs

[nekkk]

This is a really weird archive format and I'm not entirely sure what to do with it. I've pushed some attempt at unpacking this but this is definitely wrong for some files. If you can figure out wtf some of those header fields mean I'd be very happy to improve this.

https://github.com/AdmiralCurtiss/HyoutaTools/blob/master/HyoutaToolsLib/Tales/Graces/ZARC/ZARC.cs

Hey. Thank you very much for your efforts. I'll try to help too, I have more than 1 .zarc sample to compare, if that helps.

[nekkk]

Quick info that may be helpful of some sort: nx eur zarc and nx jp zarc are slightly different.

[Digote]

I’m trying to see if I can help with something, but I don’t have a deep understanding of programming.

I made a change to the script so it doesn’t decompress the files. Looking at file 13185, it seems like there’s an issue with the file splitting logic, as the same file has LZMA compression applied twice and then TLZC.

This repeated value of hexadecimal EE 72 times is not from the LZMA file; it seems to be related to some file splitting, but I’m not sure.

[Sora3100]

I figured out that Unknown1 in FileInfo is the file's hash
togf_usrdir/btl/acf/pcc003.acf -> 0xF08E949CB6CBDE2A
(pcc003 is Asbel's Honorary Knight costume)

I have also made code in lua to generate this hash

hash_gen_zarc.txt

[Sora3100]

I have figured out more

Unknown4 is multiplied by 2, then added to 0xC4BE8 (this is the start of data after the SingleFileInfo headers)
The result of this is the address of data containing the file length (2 bytes). The amount is based on BlockCount + 1, and they are in order. If the data is 0 here, you refer to the original FileLength.

[AdmiralCurtiss]

Thank you, that helped a lot. I think all the files should extract correctly now, though you need to manually do the tlzc decompression (HyoutaToolsCLI tlzc -d filename) after the unpacking for those files.

[AdmiralCurtiss]

I've also added an option where you can provide a list of filenames and it will match the hashes against that and extract as the correct filename. Maybe we can figure out a complete list somehow.

[nekkk]

I've also added an option where you can provide a list of filenames and it will match the hashes against that and extract as the correct filename. Maybe we can figure out a complete list somehow.

I was looking for the filenames and the only reference I've found so far is inside the GameNative.dll (PC) or GameNative.nro (NX) file. Couldn't figure it out entirely, though.

[DaZombieKiller]

This format is a doozy, I just finished adding extraction support for it in TLTool, you can view the code here.
I also have an ImHex pattern file for the format available here. Feel free to use or reference any of the code!

[Digote]

I tested it after the latest changes, and the files seem to be properly divided. I also tried a simple repack, and that worked too.

[Sora3100]

Here is a text file containing the names of files in the ZARC file.
You can use this in combination with HyoutaTools or TLTool

Currently, 7271 names are still missing out of 33576.
cleaned_up_graces_names.txt