Cyclomatic Complexity Rule #1464
Conversation
|
|
||
| export class Rule extends Lint.Rules.AbstractRule { | ||
|
|
||
| public static ANONYMOUS_FAILURE_STRING = "The cyclomatic complexity of is higher than the threshold"; |
|
Looking good @bencoveney! Just a few small changes needed before this can merge |
|
Anticipating a CI failure, I think there is some rule additions that are causing the linting to fail on unrelated code (#1508). I can re-merge and update once they are fixed. |
|
Sorry about that @bencoveney, I believe #1511 should have fixed the issue you were seeing |
|
So I believe that's all the comments so far addressed. The only one I have deviated from slightly is the recommendation to up the minimum threshold from 1 to 5 or similar. It makes sense to disallow 1 as this would not allow any branches in the code however 5 might be slightly too high in my opinion as there would be logical values which would be prevented for example my tests use a limit of 3. With this considered I have increased the limit from 1 to 2 which would be the lowest reasonable value AFAICT. Whether anyone would actually want to enforce a limit of 2 on real code would probably be debatable... |
|
Looking good @bencoveney, and apologies for the slowness in getting back to this. I'm thinking the failure messages should display the measured complexity and the threshold complexity - this would make it easier for users to adjust their code. |
|
Error messages now take the following formats: The function has a cyclomatic complexity of 4 which is higher than the threshold of 3 The function myNamedFunction has a cyclomatic complexity of 4 which is higher than the threshold of 3 |
|
Looks good, thanks @bencoveney! |
Resolves #1411
Roughly matches the corresponding ESLint and JSHint rule with the exception that
&&is now considered a code branch.