Perform more label name validation

GitHub rejects label names that are entirely composed of emoji characters (not to be confused with label names entirely composed of emoji shortcuts, which are accepted). labelmaker should reject these as well.

It's likely that GitHub enforces further checks, such as prohibiting control characters. Validate for these as well.

Observations so far:

GitHub seems to define an "emoji" as any character with the Emoji, Emoji_Component (except ASCII characters), or Extended_Pictographic property (See TR #51).
- This includes regional indicator symbols (whether paired or alone) and emoji tag sequences for regional flags (e.g., 🏴󠁧󠁢󠁳󠁣󠁴󠁿 — U+1F3F4 U+E0067 U+E0062 U+E0073 U+E0063 U+E0074 U+E007F).
- Emoji keycap sequences (e.g., 0️⃣ — U+0030 U+FE0F U+20E3) are allowed.
- Text presentation sequences (e.g., ⌚︎ — U+231A U+FE0E) are not counted as emoji, but emoji presentation sequences (eg., ⌚️ — U+231A U+FE0F) are.
In addition to space characters, other forms of leading & trailing whitespace including LF, CR, VT, FF and TAB are stripped from label names.
An internal LF is converted to a space, but other internal whitespace characters are not.
Multiple consecutive spaces inside a name are allowed.
The following characters are allowed anywhere in label names and can be used as single-character names:
- Non-whitespace C0 control characters
- C1 control characters
- DEL
- /
- ?
  - Side note: If a label with a ? is created, GitHub will not percent-escape the character when constructing the label's URL for use in API responses, and trying to operate on the this URL gives the wrong results.
- backslash
- %
- U+00AD (soft hyphen; Cf)
- U+0301 (combining acute accent; Mn)
- U+0600 (Arabic number sign; Cf)
- U+200C (zero width non-joiner; Cf)
- U+200E (left-to-right mark; Cf)
- U+200D (right-to-left mark; Cf)
- U+2065 (<reserved>)
- Private-use characters
The following characters are not allowed as single-character label names, but they can be used in longer names (and are not stripped from the ends of names):
- U+00A0 (no-break space; Zs)
- U+1680 (Ogham space mark; Zs)
- U+2000 (en quad; Zs)
- U+200D (zero width joiner; Cf; Emoji_Component)
- U+2028 (line separator; Zl)
- U+2029 (paragraph separator; Zp)
The following characters are not allowed anywhere in label names:
- NUL
- U+200B (zero-width space; Cf)
If a label is created with a name containing an unpaired high surrogate, the surrogate will be silently changed to a ?.
If a user tries to create a label with a name containing an unpaired low surrogate, a 500 error is returned.

jwodder / labelmaker

Perform more label name validation #4