problem with casing change for ngerman

[u-fischer]

I wanted to demonstrate the change of the casing to eszett for german. With traditional german it works with (and without) ldf-files without problems:

\documentclass[]{article}
\usepackage[german]{babel}
\babelprovide[casing=eszett]{german} 
\usepackage{fontspec}
\setmainfont{dejavusans}
\begin{document}
\MakeUppercase{grüße}
\end{document}

But the same fails if the language is ngerman:

\documentclass[]{article}
\usepackage[ngerman]{babel}
\babelprovide[casing=eszett]{ngerman} %german doesn't work either
\usepackage{fontspec}
\setmainfont{dejavusans}
\begin{document}
\MakeUppercase{grüße}
\end{document}

So how should it be done there? (It should naturally also work if more language like french are loaded with ldf-files.)

Side remark: I do find it highly confusing that "german" sometimes mean "new german" (in babelprovide) and sometimes "german-traditional" and would not mind if german generally means ngerman.

[jbezos]

At a quick glance, I’ve found some code in \extrasngerman resetting the value of the tag to de. It doesn’t belong to the babel core (which sets correctly the tag to de-x-eszett), so I have to investigate, but I’d say there is some issue in ngerman.ldf.

On the side remark. Agreed. Currently almost all languages names has been normalized to the Unicode CLDR.

[u-fischer]

Ok looking at the documentation of babel-german I found that this works:

\documentclass[]{article}
\usepackage[ngerman.capsz]{babel}
\usepackage{fontspec}
\setmainfont{dejavusans}
\begin{document}
\MakeUppercase{grüße}
\end{document}

But I think that should be handled in a better way. Now we have \babelprovide[casing=eszett]{german} for german-traditional (hoping that this doesn't active ngerman hyphenation patterns) , \usepackage[ngerman.capsz]{babel} for ngerman, \MakeUppercase[locale=de-x-eszett]{grüße} for local adjustments and \setmainlanguage[capitaleszett]{german} with polyglossia. @jspitz, @josephwright, @FrankMittelbach

On the side remark. Agreed. Currently almost all languages names has been normalized to the Unicode CLDR.

Well I think this should be done here too. While trying out different variants I wasn't even able to tell which hyphenation patterns and german variant I actually got at the end. german and ngerman should be synonyms everywhere and whoever wants traditional writing should use german-traditional.

[jbezos]

@u-fischer:

Ok looking at the documentation of babel-german I found that this works:

\documentclass[]{article}
\usepackage[ngerman.capsz]{babel}
...

These have also been my findings. The casing option was introduced into babel in December June 2023, and it was functional at that time with ngerman, but a few week months after the modifier capsz was added to ngerman and casing has become no-op. @jspitz, please, could you revise the code in ngerman so that casing is still honored and the tag isn’t reset when the language is selected? Of course, having a modifier for this task is fine, but it shouldn’t interfere with a standard babel functionality.

Well I think this should be done here too.

Many styles have evolved and the language name kept – russian, spanish and polish are examples. I think the same should apply to any other language, indeed. There is another problematic language, namely serbian, because the ‘correct’ one is serbianc (and there is also swissgerman, which is a separate language with tag gsw).

[jbezos]

I’m closing this issue because the problem is in the language style (reported to the maintainer).