I using an External Text Editor, and
I want to use Japanese characters in comments.
But, PICO-8 automatically converted Japanese characters.
I think, there is no need to change characters in the comments.
In the next version, I want to use Japanese in comments.
Because, my students are not good at English.( and me too )

Can you use romaji?

I don't think a 3x5 font could represent kana. Not even with double-width glyphs. Kanji would be impossible.

yeah or it could display some kind of symbol thing to indicate an unrenderable character, but still leave it unmodified so external editors can continue to read it. but PICO-8 doesn't currently fully support UTF-8, right? isn't it basically its own encoding which consists of ASCII and some Unicode for the symbols? I actually don't know what it does when you load Unicode characters it doesn't support. What do they get "converted" to?

Felice, kittenm4ster

Thank you for your reply!

I tried it variously and research.

First character in Japanese is "0xE3 0x81 0x82"(on UTF-8)
When I write it by external editor, and took this to PICO-8,
so PICO-8 editor displayed 3 characters,

"blank" + "glyph No.81" + "glyph No.82"

and I saved file, so I opened it in external editor.
Japanese character disappeared and it displayed 3 characters instead,

"blank?" + "character like glyph No.81" + "character like glyph No.82"

Then, I looked up more by Binary editor,
and what I understood was
"character like glyph No.81" was "0xE2 0x96 0x92"(UTF-8)
"character like glyph No.82" was "0xF0 0x9F 0x90 0xB1"(UTF-8)

I thought the glyph's number was suspicious,
I entered another KANJI character that does not contain a glyph's number.

"0xE4 0xBA 0x9C"(on UTF-8)

and took this to PICO-8, so PICO-8 editor displayed blank character,
and I saved file, so I opened it in external editor.

Then, the KANJI character was not converted!
and the KANJI was displayed in the external editor as it was.

and I tried more experiments,
Using a binary editor, I entered "0x85" and took this to PICO-8, so PICO-8 editor displayed "glyph No.85" characters.
so I saved it, and I opened it in external editor. so it displayed character like "glyph No.85".
and checked with the binary editor, "character like glyph No.85" was "0xE2 0x9C 0xBD"(UTF-8).

From these investigations,
We can see PICO-8 will automatically convert if the character code contains glyph's numbers.

I think that this is a small bug, don't you?