Documentation

Identifiers in Nix are essentially strings. They can be constructed (and viewed) with the ident isomorphism. For the sake of convenience, Identifiers are an instance of the IsString class.

Reasonable people restrict themselves to identifiers of the form [a-zA-Z_][a-zA-Z0-9_'-]*, because these don't need quoting. The methods of the Text class can be used to parse and pretty-print an identifier with proper quoting:

>>> pPrint (ident # "test") test >>> pPrint (ident # "foo.bar") "foo.bar" 

\str -> Just (ident # str) == parseM "Ident" (quote str)

\i -> Just (i :: Identifier) == parseM "Ident" (prettyShow i)

An isomorphism that allows conversion of Identifier from/to the standard String type via review.

\str -> fromString str == ident # str

\str -> set ident str undefined == ident # str

\str -> view ident (review ident str) == str

Helper function to quote a given identifier string if necessary.

>>> putStrLn (quote "abc") abc >>> putStrLn (quote "abc.def") "abc.def" 

Checks whether a given string needs quoting when interpreted as an Identifier. Simple identifiers that don't need quoting match the regular expression ^[a-zA-Z_][a-zA-Z0-9_'-]*$.

Parsec parser for simple identifiers, i.e. those that don't need quoting.

ReadP parser for quoted identifiers, i.e. those that do need quoting.