Line and position context to Lexer/Parser

[srfrog]

This PR adds line counts to the lexer so that parsers can make use of it. Also, adds line counters for the GQL and RDF parsers (LF or CR). This information is tracked in the lexer struct and returned with lexer errors, along with the position where the error happened.

Example error ouput:

# line 2 position 48
line 2:48: while lexing {set { _:a1 <reg> \"a1 content\" . }}\n\t\tsomething: Invalid operation type: something

Ref: #2815

---

This change is 

[manishrjain]

Can you add some tests? Ensure that they catch the line number correctly?

Reviewable status: 0 of 3 files reviewed, 1 unresolved discussion (waiting on @srfrog and @manishrjain)

---

lex/lexer.go, line 148 at r1 (raw file):

	l.items = append(l.items, Item{
		Typ: ItemError,
		Val: fmt.Sprintf("line %d:%d: while lexing %v: "+format,

We can't really tell the character on that line, because l.Pos does NOT take into account the line separation. It is global to the input. So, maybe just do it at the line number level.

[srfrog]

Looks like the counts wont work correctly without some extra tracking. We miss the toplevel and the op block counts.

Reviewable status: 0 of 3 files reviewed, 1 unresolved discussion (waiting on @srfrog)