Chapter 3 The simple Groovy datatypes
everything as an object and all operators as method calls
Java’s Type SystemJava distinguishes between primitive types (such as int, double) and reference types (such as Object and String).
primitive types have value semantics. fixed set in Java, no method calls
reference types have pointer(reference) semantics. no operators
Groovy’s Type System
Groovy does not use primitive type in Java, use wrapper classes for that.
int value => define an Integer instance
Literals
15, 0x1234FFFF Integer as default
100L, L as Long
12.3F F as Float
1.23D D as Double
123G as BigInteger
1.23, 1.23G G as BigDecimal as default
Autoboxing
primitive types <--> wrapper classes
Optional Typing
explicit static typing int value
implicit dynamic typing def value, Groovy assumes java.lang.Object
duck typing
A rule of thumb
As soon as you think about the static type of a reference, declare it;
when thinking “just an object,” use dynamic typing.
Operator Overriding, Operator Overloading, Operator Implementing
Operators are shorthand for method calls.
a++和++a都是调用a.next(), 一定有办法区分是pre还是post的，next才好返回pre还是post的value
an operation is closed under its type.
coercion
one argument promoted to the more general type.
BigDecimal is more general that Integer. Super Set vs Sub Set
String, GString
String Interpolation, "1 + 2 = ${1+2}"
'c' as char: coerce a string into a character
"... ${...} ..." {} denote a closure
Regular Expression
declaration what, instead of programming how
Groovy's regular expression relies on Java's regex.
A
pattern is a regular language.
不像在理论计算机， 只是有Alphabet和几个Operator组成Pattern， 这里主要有Symbol构成
Uppercase symbols define the complement. d for digit, D for non-digit
greedy match mode and restrictive match mode
Number Coercion
if either operand is Double or Float, then the result is Double
assert (1.1D * 1.2G) instanceof Double
assert (1.1F * 1.2G) instanceof Double
Coercion if exceeds the current range only for **
Division
if any of the arguments is of type Float or Double, then the result is Double
otherwise BigDecimal
1 / 2 ==> BigDecimal
1.intdiv(2) ==> Integer Division
== coerces to the more
general type before comparing.
assert 1.1G == 1.1G
assert (1.1G == 1.1D) == false
1.1 in Binary Radix is infinite,
1.1 can’t be exactly represented as a Double(IEEE Floating Number Format).
but BigDecimal can do that.