Fix Unicode character error building Japanese PDF documents

[atsuoishimoto]

Fix Unicode character error building Japanese PDF documents.

See python/python-docs-ja#31, python/python-docs-ja#35

• Use LuaTex instead of pLatex.
• use ltjsbook and ltjsarticle as document class.

[ezio-melotti]

Suggested change

	PLATEX_DEFAULT = (
	LUALATEX_DEFAULT = (

Should the variable name be updated too?

[atsuoishimoto]

Thanks, I updated in #149.

[ezio-melotti]

Suggested change

	Language("ja", "ja", "Japanese", True, PLATEX_DEFAULT, html_only=True),
	Language("ja", "ja", "Japanese", True, LUALATEX_DEFAULT, html_only=True),

Also later in the file there's a piece of code that prints the platex version, that should be updated to print the lualatex version instead if platex is no longer used anywhere.

[JulienPalard]

If I undertand correctly this PR is stuck with the newlines in the preamble?

[JulienPalard]

As a「 « "temporary" » 」fix, would running:

sed -i $'s/\xef\xbf\xbd/\xef\xa3\xbd/g' Doc/**/*.rst

help? It could be easily added to docsbuild script inside an if language.tag == "ja" and with a big comment explaining why.

[atsuoishimoto]

Sorry, I had to start over with a new branch. Please take a look at #149.

[atsuoishimoto]

sed -i $'s/\xef\xbf\xbd/\xef\xa3\xbd/g' Doc/**/*.rst

With this change, the pdf is fine. But unfortunately, the HTML document will be messed up. I gave up to support U+FFFD with luatex(for now). Please refer #149.