test/whitespace/cascades.stmt

40 columns                              |
>>> single cascades on same line
"foo"..toString();
<<<
"foo"..toString();
>>> long single cascade forces multi-line
"foo"..toString(argument, argument, argument);
<<<
"foo"
  ..toString(
      argument, argument, argument);
>>> multiple cascades get the same line when the method names are the same
list
  ..add("baz")
  ..add("bar");
<<<
list
  ..add("baz")
  ..add("bar");
>>> cascades indent contained blocks (and force multi-line) multiple cascades get their own line when method names are different
foo..fooBar()..toString();
<<<
foo
  ..fooBar()
  ..toString();
>>> cascaded setters are always multi-line even with the same name
foo..baz = 3..baz=5;
<<<
foo
  ..baz = 3
  ..baz = 5;
>>> cascades indent contained blocks (and force multi-line)
"foo"..toString(() {body;});
<<<
"foo"
  ..toString(() {
    body;
  });
>>> force split if receiver an expression where the precedence isn't obvious
main() async {
  // These are OK.
  a = b..c();
  a += b..c();
  a.b..c();

  // These are unclear.
  a ? b : c..d();
  a ?? b..c();
  a && b..c();
  a || b..c();
  a == b..c();
  a <= b..c();
  a + b..c();
  a / b..c();
  a ^ b..c();
  a << b..c();
  -a..b();
  !a..b();
  --a..b();
  await a..b();
}
<<<
main() async {
  // These are OK.
  a = b..c();
  a += b..c();
  a.b..c();

  // These are unclear.
  a ? b : c
    ..d();
  a ?? b
    ..c();
  a && b
    ..c();
  a || b
    ..c();
  a == b
    ..c();
  a <= b
    ..c();
  a + b
    ..c();
  a / b
    ..c();
  a ^ b
    ..c();
  a << b
    ..c();
  -a
    ..b();
  !a
    ..b();
  --a
    ..b();
  await a
    ..b();
}
>>> omit split if single section on list literal
[1,]..addAll(more);
<<<
[
  1,
]..addAll(more);
>>> omit split if single section on map literal
var map = {1:2,}..addAll(more);
<<<
var map = {
  1: 2,
}..addAll(more);
>>> omit split if single section on record literal
(1,2,)..addAll(more);
<<<
(
  1,
  2,
)..addAll(more);
>>> omit split if single section on trailing comma call
foo(1,)..addAll(more);
<<<
foo(
  1,
)..addAll(more);
>>> omit split if single section on trailing comma constructor
new Foo(1,)..addAll(more);
<<<
new Foo(
  1,
)..addAll(more);
>>> omit split if single section on trailing comma expression call
(foo)(1,)..addAll(more);
<<<
(foo)(
  1,
)..addAll(more);
>>> allow keeping collection on one line and splitting cascade
[1, 2, 3, 4, 5, 6, 7, 8, 9, 10]..cascade();
<<<
[1, 2, 3, 4, 5, 6, 7, 8, 9, 10]
  ..cascade();
>>> don't force cascade to split on collection target if arg splits
[1, 2, 3, 4]..cascade(() {;});
<<<
[1, 2, 3, 4]..cascade(() {
    ;
  });
>>> don't force cascade to split on collection target if target and arg split
[1,]..cascade(() {;});
<<<
[
  1,
]..cascade(() {
    ;
  });
>>> allow same-line cascades to mix null-aware
list
  ?..add("baz")
  ..add("bar");
<<<
list
  ?..add("baz")
  ..add("bar");
>>> mixed
foo?..a()..b()..c();
<<<
foo
  ?..a()
  ..b()
  ..c();
>>> null-aware getter
foo?..baz..baz;
<<<
foo
  ?..baz
  ..baz;
>>> null-aware setter
foo?..baz = 3..baz=5;
<<<
foo
  ?..baz = 3
  ..baz = 5;