Fixes the segment logic from splitting surrogates in half and not encoding correctly

[vitorpamplona]

Basically, the combineSurrogates logic doesn't run when the two characters of a surrogate are exactly at the end of one segment and the beginning of another.

Fixes #1473

[cowtowncoder]

note: I think src/main/java/com/fasterxml/jackson/core/json/WriterBasedJsonGenerator.java wouldn't be affected as it's up to underlying java.io.Writer to handle escaping.

[cowtowncoder]

Looks good overall, happy to get the fix!

One process thing: we'd need CLA (unless you have sent one earlier). It's here:

https://github.com/FasterXML/jackson/blob/main/contributor-agreement.pdf

and the usual way is to print, fill & sign, scan/photo, email to cla at fasterxml dot com.

Once I get it I can do final review, approve and merge PR.

I think this is also safe enough to go in 2.20 patch (I can cherry-pick the fix).

[vitorpamplona]

Just sent the CLA

[cowtowncoder]

Will move to test class that verifies #223 handling (where COMBINE_UNICODE_SURROGATES_IN_UTF8 implemented).

[cowtowncoder]

I think there's a bug (not using offset) and infinite-loop edge case (len == 1) to address.

[cowtowncoder]

LGTM

[cowtowncoder]

Looks good, hoping to sanity check later, merge.

[cowtowncoder]

Merged into 2.x, backported in 2.20 (for 2.20.1).

Thank you @vitorpamplona ! This is very useful fix (even if not common bug).