Enhance Lint/SelfAssignment to check attribute assignment and key assignment
#12415
Conversation
| case node.method_name | ||
| when :[]= | ||
| handle_key_assignment(node) if node.arguments.size == 2 | ||
| when /=\z/ | ||
| handle_attribute_assignment(node) if node.arguments.size == 1 | ||
| end |
There was a problem hiding this comment.
Can you tweak to the below?
| case node.method_name | |
| when :[]= | |
| handle_key_assignment(node) if node.arguments.size == 2 | |
| when /=\z/ | |
| handle_attribute_assignment(node) if node.arguments.size == 1 | |
| end | |
| if node.method?(:[]=) | |
| handle_key_assignment(node) if node.arguments.size == 2 | |
| elsif node.assignment_method? | |
| handle_attribute_assignment(node) if node.arguments.size == 1 | |
| end |
| value_node = node.arguments[1] | ||
|
|
||
| if value_node.send_type? && value_node.method?(:[]) && | ||
| node.receiver == value_node.receiver && |
There was a problem hiding this comment.
Can you add a test case for the receiver condition?
| first_argument = node.first_argument | ||
|
|
||
| if first_argument.send_type? && | ||
| node.receiver == first_argument.receiver && |
fatkodima
force-pushed
the
enhance-lint-self_assignment
branch
from
e13c3f9
to
f973340
Compare
|
Updated. |
| foo.bar = foo.baz | ||
| foo.bar | ||
| bar.foo = baz.foo | ||
| foo.bar = foo.bar + 1 |
There was a problem hiding this comment.
Can you split the it blocks for each testing content?
There was a problem hiding this comment.
foo.bar = foo.bar + 1 is the same as L155. Is it left out intentionally?
There was a problem hiding this comment.
Copy paste errors 🤦 Updated
| foo["bar"] = foo["baz"] | ||
| bar["foo"] = baz["foo"] | ||
| foo["bar"] | ||
| foo["bar"] = foo["bar"] + 1 |
fatkodima
force-pushed
the
enhance-lint-self_assignment
branch
from
f973340
to
0e9e54e
Compare
| it 'does not register an offense when using []= assignment with different receivers' do | ||
| expect_no_offenses(<<~RUBY) | ||
| bar["foo"] = baz["foo"] | ||
| foo["bar"] = foo["bar"] + 1 |
There was a problem hiding this comment.
foo["bar"] = foo["bar"] + 1 is the same as L181. Is it left out intentionally?
fatkodima
force-pushed
the
enhance-lint-self_assignment
branch
from
0e9e54e
to
4c4b4cf
Compare
|
|
||
| if value_node.send_type? && value_node.method?(:[]) && | ||
| node.receiver == value_node.receiver && | ||
| node.first_argument == value_node.first_argument |
There was a problem hiding this comment.
would this pick up cases where the hash index is non-deterministic, which (while extremely rare) would also render this cop unsafe? i.e.
hash = {}
hash[rand] = hash[rand]
###
{ 0.03256159230386502=>nil } There was a problem hiding this comment.
Yeah, this would pick up such cases. That is indeed extremely rare and I am not sure if it is worth marking this cop as unsafe because of this.
There was a problem hiding this comment.
Good catch. To ensure safety and certainty, it would be better to register an offense only when the key is some kind of literal or variable. Method calls carry the risk of side effects.
There was a problem hiding this comment.
Updated with this suggestion.
There was a problem hiding this comment.
It could be beneficial to include in the documentation the reason for not registering offenses when the key is a method call.
fatkodima
force-pushed
the
enhance-lint-self_assignment
branch
from
4c4b4cf
to
d1d7939
Compare
|
|
||
| if value_node.send_type? && value_node.method?(:[]) && | ||
| node.receiver == value_node.receiver && | ||
| !node.first_argument.call_type? && |
There was a problem hiding this comment.
Can you add a test for safe navigation method calls?
| foo["bar"] = foo["bar"] | ||
| ^^^^^^^^^^^^^^^^^^^^^^^ Self-assignment detected. | ||
| RUBY | ||
| end |
There was a problem hiding this comment.
Can you add test cases for other literal types and variables are used as keys?
fatkodima
force-pushed
the
enhance-lint-self_assignment
branch
from
d1d7939
to
639f370
Compare
|
Updated. |
| foo[Foo] = foo[Foo] | ||
| ^^^^^^^^^^^^^^^^^^^ Self-assignment detected. | ||
| foo[:bar] = foo[:bar] | ||
| ^^^^^^^^^^^^^^^^^^^^^ Self-assignment detected. |
There was a problem hiding this comment.
Can you please separate the tests for each type of literal?
| foo[1] = foo[2] | ||
| foo[1.2] = foo[2.2] | ||
| foo[Foo] = foo[Bar] | ||
| foo[:bar] = foo[:baz] |
| foo[@@var] = foo[@@var] | ||
| ^^^^^^^^^^^^^^^^^^^^^^^ Self-assignment detected. | ||
| foo[$var] = foo[$var] | ||
| ^^^^^^^^^^^^^^^^^^^^^ Self-assignment detected. |
There was a problem hiding this comment.
Can you please separate the tests for each type of variable?
| foo[var1] = foo[var2] | ||
| foo[@var1] = foo[@var2] | ||
| foo[@@var1] = foo[@@var2] | ||
| foo[$var1] = foo[$var2] |
fatkodima
force-pushed
the
enhance-lint-self_assignment
branch
from
639f370
to
f45ae10
Compare
|
Separated tests. |
|
Thanks! |
This is an implementation of #12410 for
Lint/SelfAssignment.