Why is `x && y` not parsed as `x & (&y)`?

Question

I'm implementing a C compiler and found a curious issue. Since & has higher precedence than &&, it seems reasonable to consider it as a binary-and of the first operand with the address of the second:

x && y = (x) & ( &(y) )

The syntax overview of the C specification seems to allow this interpretation. I'm probably missing or misreading something?

My understanding of the syntax:

andExpression := equalityExpression | (andExpression '&' equalityExpression) | ...
...
unaryExpression := postfixExpression | ( ('&' | '*' | '+' | '-' | '~' | '!') castExpression ) | ...

Answer 1

C operator expressions are parsed through something known as "maximal munch" ¹⁾, meaning that from left to right, the compiler goes for the longest chunk of symbols that can form a valid token. Since x && is longer than x &, the compiler picks the former.

This is why code like x+++1 compiles, +++x does not, but + ++x does.

---

¹⁾C11 §6.4 Lexical elements ¶4:

If the input stream has been parsed into preprocessing tokens up to a given character, the next preprocessing token is the longest sequence of characters that could constitute a preprocessing token.