I’m encountering several issues when compiling documents with mixed scripts in Scrivener 3.3.1 (Main script: Persian (RtL), English, Greek, German (LtR)) for .dox
Word Order: In paragraphs and footnotes with mixed Persian and English text (or containing parentheses, Guillemets, Some times even period marks, etc.), word order becomes garbled. In some instances fonts change unexpectedly, and I can only revert them by pasting the text with the “Keep Text” option.
Numerals: Some English numerals are incorrectly converted to Persian numerals, especially those enclosed in parentheses.
It seems both issues have a single cause, difficulty with properly formatting RtL LtR complex scripts.
I’ve tried modifying the compile settings and custom styles, but the issues persist.
Given the length of the text (PhD dissertation), manual fixes are not a practical solution.
Any suggestions for preserving proper formatting (especially for numerals and word order) during compilation?
At a baseline, have you tried using the “Default” compile format, at the top of the list in the left sidebar, to see if you get better results when compiling with minimal transformation (and no transformation to the text)?
That may not be where you want to end up, but it might be a better start if a more heavily designed Format is interfering with paragraph language settings.
Yes, I’ve tried every possible setting and formatting option, but none of them have been helpful. It seems the main reason for the difficulty is the lack of support for complex scripts. However, if you could suggest a solution, it would save me from the laborious task of re-editing my long dissertation.
Given the difficulties with producing a genuinely useful before/after result on my own, would it be possible to send a short sample that demonstrates the problem, and points out precisely the spots that are broken (as I will not be adept at seeing differences).
Meanwhile have you also tried RTF? That is Scrivener’s native format, and so if this is a matter of some esoteric combinations of settings not being understood by the converter, using RTF would dodge that.