In particular, can Emacs deal with cases where the Unicode RTL algorithm gives nonsense, such as editing HTML in Arabic?
(Because HTML tags are made of punctuation, Roman letters, and attribute values that could be in any language, you can end up with LTR segments that may include half a tag and some adjacent text. At that point you can't even see visually what they include; you have to reason out what Unicode is doing. You might even end up with a tag that begins with an open angle bracket that looks like > and ends with a closing angle bracket that also looks like >.)
I ask this not as someone who knows any Arabic, but someone who has to deal with the results of complexities of text such as RTL formatting. To make mixed-direction text editing "perfect", someone would have to put a lot of task-specific design into it.
In particular, can Emacs deal with cases where the Unicode RTL algorithm gives nonsense, such as editing HTML in Arabic?
(Because HTML tags are made of punctuation, Roman letters, and attribute values that could be in any language, you can end up with LTR segments that may include half a tag and some adjacent text. At that point you can't even see visually what they include; you have to reason out what Unicode is doing. You might even end up with a tag that begins with an open angle bracket that looks like > and ends with a closing angle bracket that also looks like >.)
I ask this not as someone who knows any Arabic, but someone who has to deal with the results of complexities of text such as RTL formatting. To make mixed-direction text editing "perfect", someone would have to put a lot of task-specific design into it.