Hello Jens, yes, it does. I have switched from arabxetex to polyglossia, needed to fix a few font specifications, and now the results look correct (says out Egyptian intern). The automatic RTL-LTR switching depending on the Unicode script information works perfectly, with the special case that we will use non-breakable spaces u00a0 in some places to keep the sequence of non-arabic text fragments as needed.
Thank you for your help! Hartmut Von: XeTeX <xetex-bounces+hartmut.niemann=siemens....@tug.org> Im Auftrag von Jens Bakker Gesendet: Dienstag, 11. Juni 2024 16:47 An: XeTeX (Unicode-based TeX) discussion. <xetex@tug.org> Betreff: Re: [XeTeX] Typesetting arabic and european mix encoded in utf8 Hello Hartmut Niemann, may be that the XeLaTex-package polyglossia could serve your purposes better much better. You could use many languages in one document, also Arabic and other RTL text. Best wishes and best regards, Jens Bakker Am 11.06.2024 um 12:06 schrieb Niemann, Hartmut via XeTeX <xetex@tug.org<mailto:xetex@tug.org>>: Hello! In my current project I use XeLaTeX to typeset PDF files from texts in different languages held in a separate database. (This is done with a generator that is language-unaware, generating lines like \long\def\msgtext{عطل في التهيئة البنيوماتية GS} Into a .inc file and a manually written, language dependent, frame document that defines \msgtext{} I typeset a (mostly) Arabic document using XeLaTeX and \usepackage{arabxetex}[utf] Arabxetex supports encoding Arabic in ASCII, and this interferes with the fact, that our texts have latin characters, like English abbreviations, location IDs and such. The documented solution would be enclosing these latin characters which are to be typeset verbally into \text{LR}, which is rather hard if the text comes from a database. Does anybody how to switch off arabxetex’s ASCII-to-arabic conversion completely? Or is there a package that supports Arabic (with Arabic typographic conventions) but made for pure Unicode sources? With best regards Hartmut Niemann
