[MC-20675] pack.mcmeta uses system character encoding instead of UTF-8

Current implementation pack.mcmeta uses system character encoding.
E.g. on windows 7 Japanese edition, writing description in mcmeta in SHIFT_JIS will show correct letters.

For easy distributing, any kind of metadatas should always use UTF-8 (or other Unicode-based character encoding).

Attachments

Comments 6

Erik Broes 2013-07-02T07:43:47Z

pack.mcmeta: ASCII text

Not really sure what you are talking about. There should be no unicode characters in the whole file.

ANBO Motohiko 2013-07-02T08:27:19Z

description (generic resources), name and region (language packs) will contain non-ASCII letters.

ANBO Motohiko 2013-07-02T08:32:38Z

Example screenshot: my own language pack for 1.6.1 with Japanese description on Windows 7 64bit/Japanese.

ANBO Motohiko 2013-07-02T08:40:18Z

Same resource pack on Mac OS 10.6 with English locale setting.
Shows garbage characters.

Erik Broes 2013-07-02T08:46:45Z

In json you are supposed to escape any non-ascii letters like you'd do in java. \uXXXXXX.

ANBO Motohiko 2013-07-02T16:38:42Z

Thank you for closing. Users can refer this ticket as 'specification' of .mcmeta.

This maybe belongs to the 'request for enchancement', so please keep this issue closed.

According to RFC 4627:
> JSON text SHALL be encoded in Unicode. The default encoding is UTF-8.

And .lang files are written in UTF-8. Translated end.txt also uses UTF-8 (same as with credits.txt).
Some files are written in UTF-8, and some files are not... this makes us very confusing.

To avoid confusing, Unicode based character encoding should be used in (pseudo) JSON files.