compare special-value symbols against the whole input character

[Ramya-9353]

Repro: with infinity_symbol_ = "Infinity" and nan_symbol_ = "NaN", the uc16 overload accepts non-ASCII text as a special value. U+014E 'a' 'N' and U+FF4E 'a' 'N' both return NaN, U+FF49 'n' 'f' 'i' 'n' 'i' 't' 'y' returns Infinity, and 'N' U+0161 'N' returns NaN, where each should have given junk_string_value with processed_characters_count == 0. Under ALLOW_CASE_INSENSITIVITY with the symbol "nan", U+014E 'a' 'n' matches too.

Cause: the symbols are matched through helpers declared on char (ConsumeFirstCharacter for the leading character, ToLower/Pass inside ConsumeSubStringImpl for the rest), so a 16-bit input character is truncated to its low byte before the comparison and any code point ending in the expected symbol byte aliases it. This is the conversion the C4244 suppression at the top of the file is about.

Fix: compare the unsigned code-unit value of the input character and of the symbol byte. The char overloads are unaffected because both sides pass through unsigned char, and ASCII matching, case-insensitive included, is unchanged.

[google-cla]

Thanks for your pull request! It looks like this may be your first contribution to a Google open source project. Before we can look at your pull request, you'll need to sign a Contributor License Agreement (CLA).

View this failed invocation of the CLA check for more information.

For the most up to date status, view the checks section at the bottom of the pull request.

[floitsch]

Thanks.