New rule: dont_create_a_return_var

BAD:

void x() {
  final y = 1;

  return y;
}

GOOD:

void x() {
  return 1;
}

Is this one a pretty usual? I feel like maybe it's not worth the effort - wdyt?

Issue #92, new rule `dont_create_a_return_var`:

Issue #92: added rule in analysis_options.yaml

Issue #92: added new test case

Issue #92: added naive implementation for dont_create_a_return_var_rule

Issue #92: fixed conflicting lints in test

Issue #92: added test case where return var defined from mutable, but…

Issue #92: implemented advanced rule with visitor, checking where var…

Issue #92: fixed formatting with 'dart format'

@solid-dmytrotkachuk

During conversation with @solid-dmytrotkachuk the following criteria were decided for now:

it is Bad, if no statements are between declaration and return,
it is Bad, if declaration initializer is an immutable expression (all parameters are const or final),
it is Good, if variable is used more than in return statement itself,
it is Good, if declaration initializer contains mutable arguments, since it would be hard to predict if this wasn't used for valid caching.

Issue #92: changed if chain to switch statement (more readable)

Issue #92: renaming rule to avoid_unnecessary_return_variable

Issue #92: small readability improvements

Issues #92, #148: refactor to use common code between two rules

Issues #92, #148: fixed code formatting

Issues #92, #148: simplified to use expr.accept(visitor)

Issues #92, #148: removed unnecessary super call

Issues #92, #148: refactor to use common code between two rules (#221)

Navigation Menu