PlainPrinter

Usual target for Annotation#toText, overridden in RefinedPrinter

Textual representation of symbol's declaration

Textual representation of single denotation's declaration

Textual representation of all symbols in given list, using dclText for displaying each.

A description of sym's location

The fully qualified name of the symbol

String representation of symbol's kind.

Textual representation of symbol and its location

If symbol's owner is a printable class C, the text "in C", otherwise ""

The name, possibly with with namespace suffix if debugNames is set: /L for local names, /V for other term names, /T for type names

The name of the given symbol. If !settings.debug, the original name where expansions of operators are translated back to operator symbol. E.g. $eq => =. If settings.uniqid, adds id.

A plain printer without any embellishments

The name as a text

Textual representation of type

Textual representation, including symbol's kind e.g., "class Foo", "method Bar". If hasMeaninglessName is true, uses the owner's name to disambiguate identity.

Textual representation of denotation

Textual representation of constant

Textual representation of annotation

Textual representation of lambda param

Textual representation of all definitions in a scope using dclText for each

Textual representation of tree

Textual representation of source position

Textual representation of implicit search result

Textual representation of info relating to an import clause

Textual representation of a constraint

The string representation of this type used as a prefix, including separator

The string representation of this type used as a prefix

Generate text using op, assuming a given precedence level prec.

atPrec vs changePrec

This is to be used when changing precedence inside some sort of parentheses: for instance, to print T[A]usetoText(T) ~ '[' ~ atPrec(GlobalPrec) { toText(A) } ~ ']'`.

If the presence of the parentheses depends on precedence, inserting them manually is most certainly a bug. Use changePrec instead to generate them exactly when needed.

Generate text using op, assuming a given precedence level prec. If new level prec is lower than previous level, put text in parentheses.

atPrec vs changePrec

To pretty-print A op B, you need something like changePrec(parsing.precedence(op, isType)) { toText(a) ~ op ~ toText(b) } // BUGGY that will insert parentheses around A op B if, for instance, the preceding operator has higher precedence.

But that does not handle infix operators with left- or right- associativity.

If op and op' have the same precedence and associativity, A op B op' C parses as (A op B) op' C if op and op' are left-associative, and as A op (B op' C) if they're right-associative, so we need respectively

val isType = ??? // is this a term or type operator?
val prec = parsing.precedence(op, isType)
// either:
changePrec(prec) { toText(a) ~ op ~ atPrec(prec + 1) { toText(b) } } // for left-associative op and op'
// or:
changePrec(prec) { atPrec(prec + 1) { toText(a) } ~ op ~ toText(b) } // for right-associative op and op'

The current precedence level. When pretty-printing arguments of operator op, currentPrecedence must equal op's precedence level, so that pretty-printing expressions using lower-precedence operators can insert parentheses automatically by calling changePrec.

Render elements alternating with sep string

Render elements within lowest precedence

Render element within lowest precedence

Render elements within highest precedence

Render element within highest precedence