Dot follows chapter name in a Hungarian header

[magyarkuti]

Polyglossia provides the correct order of the chapter counter and the chapter name in case of the Hungarian setup. When we write 1. fejezet. in the running head of a document, then the Hungarian typography requires a dot after the chapter name (fejezet). Thus: 1. fejezet. is required, but 1. fejezet is the output now. The solution can be to change appropriate \chapapps to \chapapps. between row 227 and row 261--where the running headers are adjusted--of the file gloss-hungarian.ldf.

[jspitz]

Thanks, merged. Although I am a bit reluctant as this changes existing documents.

[magyarkuti]

Yes, yes, no doubt, it is incompatible with the possible earlier patches of this mistake. For example, I also defined a special header of my memoir based document exactly for that reason. Is this what makes you reluctant?

[jspitz]

Yes. If a polyglossia update changes existing documents' texts (here: adds a dot to the header) "behind" the back of users, this is always problematic, even of the change is more correct.

[magyarkuti]

Agreed. The problem is even more serious. Hungarian typography requires a lots of small similar changes. Just two of randomly chosen problems: 1) Sections \section{Bla bla} ouputs 1.1 Bla bla today, but the correct is 1.1. Bla bla. Completely the same as the russian typography, and gloss-russian.ldf does it correctly. 2) Theorem like environments The required output is 1.1. tétel (Pythagoras). Again, the key is the dot after theorem counter.

In my opinion 1) belongs to the category what polygossia should be able to manage. Perhaps it is good idea if forcing a dot is not a parameter, as in the case of Russian setup, but polyglossia provides a function, for compatibility reasons, which rejects the forcing of the dot after the section counters. I think the default behavior of polyglossia must be the typographically correct output.

With respect to problem 2) I do not recommend similar patch, even if babel does it. The theroem packages, for axample amsthm are flexible enough to be able to define a correct theorem style. I think the Hungarian users must understand the problem.

But the fundamental question is here. Do we want to develop gloss-hungarian.ldf until the code fits all of Hungarian typographic conventions? In an ideal word a Hungarian math related tex code if lualatex or xelatex is the compiler can be the following: \documentclass{memoir} \usepackage{polyglossia}\setdefaultlanguage{hungarian} \usepackage{hyperref, amsthm} \begin{document} ... \end{document} This is what we want, or we accept that some user intervention will always be required. The answer of babel is definitely yes. I really do not know. I could argue, and I could accept the negative and the positive answer as well. My personal opinion is closer to the negative answer, but a comprehensive document with all of requirements of Hungarian typography including polyglossia based solutions is warmly welcome. Perhaps it can be the part of the polyglossia documentation. What do you think?

[jspitz]

I think we should provide an option for that, similar to Russian's forceheadingpunctuation. Though it should be false by default for backwards compatibility reasons. The option should also cover your recent punctuation to running headers.

[jspitz]

So I tend to revert https://github.com/reutenauer/polyglossia/commit/ff4654fa85076d0b863c12c5ab23f045ca270a57 until we have a more general solution with an option. Agreed?

[magyarkuti]

Agreed.