Regex incorrectly identify surrogate pair unicode category

[MaceWindu]

This defect makes regex useless when you need to match string using unicode categories and string could contain surrogate pairs:

        [Test]
        public void ClassifySurrogates_Test()
        {
            var value = "𦕒";
            var regexLetters = new Regex(@"^\p{Lo}$");

            var regexSurrogates = new Regex(@"^\p{Cs}{2}$");

            Assert.AreEqual(UnicodeCategory.OtherLetter, char.GetUnicodeCategory(value, 0));
            Assert.AreEqual(UnicodeCategory.OtherLetter, CharUnicodeInfo.GetUnicodeCategory(value, 0));
            Assert.True(regexSurrogates.IsMatch(value));

            // Fails here
            Assert.True(regexLetters.IsMatch(value));
        }

[joshfree]

I suspect this issue exists in desktop and isn't unique to core

[tarekgh]

@Priya91 Unicode Categories looks fine here so the issue looks with regex.

[MaceWindu]

@joshfree, you are right.
I didn't dig into source, but I suspect that code responsible for this uses GetUnicodeCategory(char) instead of GetUnicodeCategory(string, charIndex)

[peltco]

I'll take a stab at this as part of the hackaton

[danmoseley]

@peltco I'll unassign you unless you still have an interest 2 years later 😄