Fix mojibake in Japanese file names when moving to and from The Outside World

[seishun]

I can think of two ways to handle MacJapanese in the emulator:

• Encode and decode it manually like MacRoman. The code would be more complex than for MacRoman because MacJapanese has both one-byte and two-byte characters, plus there are many of them (see http://www.unicode.org/Public/MAPPINGS/VENDORS/APPLE/JAPANESE.TXT).
• Pretend it's Shift JIS and use iconv. It seems to work fine in one direction (a more recent version of Emscripten is needed for the other one).

[mihaip]

Doing it as Shift JIS seems like a reasonable start.

There's already some encoding metadata on the Python side (

infinite-mac/scripts/disks.py

Line 14 in e49b327

stickies_encoding: str = "mac_roman"

) because we need to take it into account when generating the Stickies file. Instead of having to duplicate it on the TypeScript side, can you make the Python code that generates the chunked file JSON def (

infinite-mac/scripts/import-disks.py

Lines 470 to 476 in e49b327

	json.dump(
	{
	"name": os.path.splitext(image.name)[0],
	"totalSize": total_size,
	"chunks": chunks,
	"chunkSize": CHUNK_SIZE,
	},

) also include that (the TypeScript side reads in that data via the type at

infinite-mac/src/emulator/emulator-common.ts

Lines 136 to 143 in e49b327

	export type EmulatorChunkedFileSpec = {
	name: string;
	baseUrl: string;
	totalSize: number;
	chunks: string[];
	chunkSize: number;
	prefetchChunks: number[];
	};

)

[seishun]

Setting name_encoding based on stickies_encoding seems wrong because they're kinda orthogonal. Setting both based on a "common" value (e.g. encoding) makes more sense, but it's tricky: in case of KanjiTalk, we want to pass "shift_jis" to str.encode() and something that corresponds to MacJapanese (probably a number) to the emulator. Can you think of a simple way to do this?

[gingerbeardman]

something that corresponds to MacJapanese (probably a number)

The ancient and official numbering scheme probably makes most sense:

Encoding	Enum
MacRoman	0
MacJapanese	1
MacChineseTrad	2
MacKorean	3
...	...

Sources for the full list:

• https://forensics.wiki/mac_os_x/
• http://man.hubwiz.com/docset/Xojo.docset/Contents/Resources/Documents/docs.xojo.com/TextConverter.html
• https://github.com/libyal/libfshfs/blob/main/documentation/Hierarchical%20File%20System%20(HFS).asciidoc
• https://developer.apple.com/documentation/corefoundation/cfstringencodings/macjapanese