前に転送した egreg さんのメールに対する私の返信です。 -- Hironobu Yamashita ---------- Forwarded message ---------- From: Hironobu Yamashita <h.y.acetaminophen@gmail.com> Date: Wed, 16 Jan 2019 23:25:36 +0900 Subject: Re: regexpatch: error on (u)pTeX with Japanese char To: Enrico Gregorio <enrico.gregorio@univr.it> Hi, The problem seems to be solved, but just for your information:
If there is a sane way to check whether a character token will give a character code illegal with \catcode, then I guess a further check can be added when the engine is epTeX.
I think you can detect whether a character token is JP char or non-JP char by following code, for example. By design, the behavior of \kcatcode is different between pTeX and upTeX (the latter one allows flexible changing of JP / non-JP classification), so different code should be used here. \ifx\enablecjktoken\@undefined \def\@ifisjp#1{% pTeX version \ifnum`#1<255\relax \expandafter\@secondoftwo % non-JP \else \expandafter\@firstoftwo % JP \fi} \else \def\@ifisjp#1{% upTeX version \ifnum\kcatcode`#1=15\relax \expandafter\@secondoftwo % non-JP \else \expandafter\@firstoftwo % JP \fi} \fi The test case is shown at https://gist.github.com/aminophen/154cb16cfe989dd605b0983f3e3bd20b Best regards, Hironobu