#62 · closed · 2 comments
Hi, First of all, this is a really great tool! Would it be possible to add support for proper RTL punctuation handling? For example, when translating from English to an RTL language, such as Hebrew or Arabic, the question mark can remain on the left side at the beginning of the sentence instead of appearing correctly at the end. It would be great if punctuation marks and other non-letter characters could be handled according to the target language's text direction. Thanks!
Thanks for the report — I dug into this properly. The translated data was already in correct logical order. So nothing is reordered in the file. What you're seeing is the Unicode bidi algorithm resolving the paragraph direction: sentence-final punctuation is bidi-neutral, so in a paragraph that resolves LTR it gets pushed to the opposite end of the sentence. What I changed: the result and preview panes were hard-coded to dir="ltr". They now use dir="auto", which on a <textarea> means unicode-bidi: plaintext — direction resolved per line from the first strong character. In an SRT the index and timecode lines stay LTR while the Hebrew/Arabic subtitle lines flip to RTL and the punctuation lands correctly. Same for the live translation rows, the bilingual review panel, the failed-line list and the glossary fields.
@rockbenben Just checked it, and it works great! Thank you again for all your effort—this is a fantastic tool! 🙌