Require space after -- to start single line comment in MySQL

[hansott]

• Docs https://dev.mysql.com/doc/refman/8.4/en/ansi-diff-comments.html
• Lexer https://github.com/mysql/mysql-server/blob/trunk/sql/sql_lex.cc#L1465-L1472
• Char checks

static constexpr uint8_t MY_CHAR_CTR =  040; /* Control character */

inline bool my_iscntrl(const CHARSET_INFO *cs, char ch) {
  return ((cs->ctype + 1)[static_cast<uint8_t>(ch)] & MY_CHAR_CTR) != 0;
}

static constexpr uint8_t MY_CHAR_SPC =  010; /* Spacing character */

inline bool my_isspace(const CHARSET_INFO *cs, char ch) {
  return ((cs->ctype + 1)[static_cast<uint8_t>(ch)] & MY_CHAR_SPC) != 0;
}

It's not super clear which characters belong to my_iscntrl and my_isspace but we cannot use char.is_whitespace() as it includes '\n'?

[iffyio]

Suggested change

	fn requires_whitespace_to_start_comment(&self) -> bool {
	fn requires_single_line_comment_whitespace(&self) -> bool {

thinking something like this to flag that this is only for the -- comment style

[iffyio]

Suggested change

	/// Returns whether it's the start of a single line comment
	/// Returns true if this dialect requires a whitespace after `--` for single line comments.

[iffyio]

Since this is already duplicated at the definition of the dialect method, we could consider not including it (The link to the doc on the other hand is useful in any case)

[iffyio]

Can we move the example to the doc string in the dialect method? That would help clarify the behavior to folks without having to dig deeper in the docs or the parser code.
Then for the rest of the comment here we can probably skip since we've already documented as much in the dialect method

[iffyio]

Suggested change

	match chars.peekable.clone().nth(1) {
	Some(' ') => (),
	_ => is_comment = false,
	}
	is_comment = Some(' ') == chars.peekable.clone().nth(1);

looks like this can be simplified?

[iffyio]

would it be possible to have this remain as a match statement, calling chars.peek() only once? i.e.

match chars.peek() {
    Some('-') => { ... }
    Some('>') => { ... }
 }

[iffyio]

can we use mysql().verified_stmt() for the tests?

[hansott]

@iffyio Addressed all your comments, great feedback.

[hansott]

@iffyio doing a check for which characters are allowed to start comment (using a script against mysql)

[iffyio]

LGTM! Thanks @hansott!
cc @alamb