but … FT is a plain text (utf-8 unicode) editor, and while utf-8 does a good job of representing RTL characters, I’m not sure whether it can encode and retain information about which paragraphs in a text are RTL and which are LTR (in the way that, for example, RTF can).
It would be ideal to have a FoldingText equivalent of the direction button/command in Mellel
But I notice that whereas once can round-trip out of and back into Mellel retaining the direction of each paragraph thru RTF, the same journey in UTF-8 preserves the characters but loses the line-specific direction information. It looks as if the editor might have to check the character types in each line and make line by line guesses about direction.
At the moment, exporting the document above to UTF-8 text, and then opening it in Mellel or in FT gives us all LTR.