Add unchecked-type-assertion #889
Conversation
Dirk007
force-pushed
the
rule/unchecked_dynamic_cast
branch
from
652dc0a to
de3b67b
Compare
|
@denisvmedia |
|
@chavacava I can, but I think this should be a part of the tests ;) |
I totally agree! Unfortunately I am a bit short of time this week but will work on the suggested ideas ASAP. |
|
Just one thing came to my mind now, probably this is still left out (sorry, can't check it now): |
good idea.. let me write a test 👍 |
|
|
||
| ```go | ||
| foo, _ := bar(.*Baz). | ||
| // ^ |
There was a problem hiding this comment.
It could be interesting to list all the cases the rule is capable to detect (switch, channel, ...)
| u.Lock() | ||
| defer u.Unlock() | ||
|
|
||
| if len(arguments) == 0 { |
There was a problem hiding this comment.
We could add a configured field to the rule's struct to trace the rule's configuration was already processed and then quit ASAP once the rule was configured a first time.
| if len(arguments) == 0 { | |
| if len(arguments) == 0 || u.configured { |
Closes #888
Unchecked type assertions results can (and will) result in an unexpected panic and should be a lint.
This new rule covers this with configurable non-default option to ignore
foo, _ := bar.(Baz)- cases where theokvalue is ignored.See tests.