\SetLanguageKeys not documented.

[Udi-Fogiel]

I'm don't fully understand yet the purpose of (which key should be allowed to change), so I don't feel comfortable documenting it.

[jspitz]

I extended the comment. Better for you?

[Udi-Fogiel]

I meant it wasn't documented in the user manual, should it be there?

[jspitz]

No, this command is not meant to be used in documents. It is meant to be used in gloss files.

[Udi-Fogiel]

Ok, so not related to this ticket, but if I understand correctly, any setup is made during the processing of the .ldf files. So I have two questions:

• Should the assignments in \polyglossia@keys_define_lang:n be global? after all, we assume that everything is done during the processing of the .ldf file and it should not happen inside a group.
• Should we use \key_define just for setting up a language? maybe we can use more efficient method

[jspitz]

I'd have to check that. On a general note, I am wondering whether we shouldn't do a release with the bug fixes we have before we start a major code rewrite which always entails the risk of introducing bugs.

[Udi-Fogiel]

Sure, I'll stop with the commits for the time being.

[jspitz]

Thanks. I try do to the release soon. Please check whether your changes deserve change log entries in the manual.

[Udi-Fogiel]

Just one thing. \(lL)ocalnumeral does not require the key lang=, i.e. \localnumeral[lang=hebrew]{12} and \localnumeral[hebrew]{12} are both ok now. I did not see the reason to force users to type lang= if this is the only existing key.

If you think this is a bad Idea, it could be changed back in a few key strokes, otherwise I'll add that to the manual.

[jspitz]

Fine with me

[jspitz]

Sure, I'll stop with the commits for the time being.

1.66 is released, so you can go ahead.