First index of hard line break

[HueHue44]

I want to segment runs based on direction change, script change and hard line breaks. But I have a problem when it comes to the segmentation of line breaks. The reported break.Position points to the character after the characters that caused the line break. For example, the string A\r\nB will report the first hard line break at index 3, which is the 'B', but I want the segmentation to only include 'A' in this case. How should one know where the hard line break started?

This may be my lack of understanding how this works. Really appreciate the work on this library!

[JimmyLefevre]

Line breaks are reported this way because this is how Unicode defines them. Hard line breaks can be caused by Unicode line break classes BK, CR, LF and NL, which cover the characters \n, '\r', 0xB, 0xC, 0x85, 0x2028 and 0x2029. This is probably okay to hardcode if you specifically want to trim hard line breaks and nothing else. For anything more involved, you would probably be better served with looking at the Unicode properties directly, using something like utf8proc.

We do have our own Unicode tables that we use internally, but most of them are the results of pre-computation based off of the original Unicode properties. As such, they don't make a lot of sense to use outside of the library itself, and so exposing them would probably not be very useful.

Obviously, this whole situation sucks for binary size...