@LB-- I'm wondering whether this should be added to the v1.1.1 milestone. It looks like Unicode strings are not read correctly.

@bogbert It would be great if a test could be added. File tests/valid-0012.json already contains UTF-16 surrogate pairs as far as I can understand. If the bug does not show with the existing examples, could you add one this 𠜱 example?

"CJK Unified Ideograph-20731 (U+20731)": "\ud841\udf31",

I've added some basic string comparison tests in 174e5a7 as a good starting point for this. It seems like this change does in fact fix the issue, but I'll want to get more comparisons in the tests first to make sure nothing else gets broken.

See for example:

• Programming with Unicode
• unicode utf-16 surrogate pair print problem
• Unicode - UTF-16(UCS-2)

I have forked the bogbert/masterbranch into Dimitri/Papadopoulos/bogbert to rebase it and attempt to add more tests.