Add operators support

[MaximSokolov]

Fixes #114, fixes #123

// sizeof
sizeof unary_expression
sizeof(int)

// arithmetic
+a
-a

a + b
a - b
a * b
a / b
a % b

i++
++i
i--
--i

// comparison
a < b
a > b
a >= b
a <= b
a == b
a != b

// bitwise
~a
a & b
a | b
a ^ b
a << b
a >> b

// logical
!a
a && b
a || b

// assignment
a = b
a += b
a -= b
a *= b
a /= b
a %= b
a &= b
a ^= b
a |= b
a <<= b
a >>= b

// pointer
*a
&a

// ternary
a ? b : c

[winstliu]

maybe keyword.operator.bitwise.c?

[winstliu]

.logical.bitwise?

[MaximSokolov]

It's not possible to distinguish an indirection operator and a multiplication operator

a * b
*a

Same for bitwise NOT and address-of operators

&a
a & b

I didn't add bitwise subscope for |, ~, ^ for consistency with &. I made an exception for shift operators(see 2a5fafe)

[winstliu]

Good point. This isn't something that the grammar can handle right now.

[winstliu]

Can we throw in a sizeof$ or a sizeof_ check somewhere?

[winstliu]

Very nice work as usual @MaximSokolov. Left some pretty minor comments.

Also, what happened with your concern about overloaded operators? Is that still relevant, or did you realize that it didn't matter?

[hgdsraj]

Overloaded operators are probably still relevant but I'm uncertain why this code would not account for them?

[MaximSokolov]

The only issue with overloading operators:

cout << "test"; // insertion operator '<<' would have a `bitwise.shift` subscope

Fixed in 80db385:

~Foo() {}

I didn't fix this:

vector<vector<int> > a;

template <typename T>

[winstliu]

So while there are some incorrect scopes introduced with this PR, I think that the benefits of at least being able to syntax highlight operators outweighs the downsides of said incorrect scopes.

@MaximSokolov would you be willing to create new issues for your last two examples?