What’s new in babel 3.90

2023-06-26

Casing

The option casing introduced experimentally in version 3.88 should now work. Here is again the explanation.

There is a new option in \babelprovide to set the default casing rules in a few languages. They are (from interface3.pdf):

• Armenian: yiwn, which maps U+0587 to capital ech and yiwn on uppercasing.
• German: eszett, which maps the lowercase Eszett to a großes Eszett.
• Greek: iota, which converts the ypogegrammeni (subscript muted iota) to capital iota when uppercasing.

For example:

\usepackage[greek]{babel}
\babelprovide[casing=iota]{greek}

New in 3.90 is a couple of macros devised as a high level interface for declaring case mappings, based on the locale name as declared by or with babel. They are the equivalent of \DeclareUppercaseMapping and \DeclareLowercaseMapping (see usrguide.pdf):

• \BabelUppercaseMapping{<locale name>}{<codepoint>}{<output>}
• \BabelLowercaseMapping{<locale name>}{<codepoint>}{<output>}

(A variant for titlecasing will follow soon.)

The purpose is twofold:

• A user-friendly way to declare them, because often BCP 47 tags are not known (and sometimes can be complex).
• If for some reason the tag changes (eg, you decide to tag english as en-001 instead of en-US), the new mappings will be still assigned to that language.

For example:

\BabelUppercaseMapping{medievallatin}{`u}{V}

There are still some rough edges when declaring a mapping with the x extension, or when two babel languages share the same BCP 47 tag. These issues will be sorted out in future releases.

Improvements in Arabic justification

Kashida are built by adding the Unicode ARABIC TATWEEL character. Now a limit can be set with a parameter as illustrated:

\directlua{  Babel.arabic.kashida_limit = 3 }

0 means ‘add nothing’ and -1 means ‘no limit’.

The width of the glyph can vary greatly with the font, and so does the exact value of this parameter (which is currently global).

Locale names

• The names classiclatin and ecclesiasticlatin are now discouraged because classicallatin and ecclesiasticallatin are more correct.
• Since farsi is a well established name for persian, it’s now a valid synonymous, although the preferred name is still the latter (as per the CLDR).
• vietnam was deprecated and now finally has been removed.

Fixes

• \normalsfcodes had a wrong value in heads and foots, which meant the text was sometimes badly spaced after dots, colons and the like (#240).
• Line breaking was sometimes prevented when switching the text direction (#213). Thanks to Udi Fogiel.