Check property existance before accessing it #243
Conversation
|
@rhys-vdw This change will require modifications to many tests (63 in total), so I would like to confirm that this approach is acceptable before proceeding with those corrections. I have made corrections and implementations for several key tests. |
| typeof typedObj === "object" || | ||
| typeof typedObj === "function") && | ||
| ( !("foo" in typedObj) || | ||
| "foo" in typedObj && |
There was a problem hiding this comment.
!("foo" in typedObj) || "foo" in typedObj && // ...This check seems redundant, given that the previous code handled this case just fine.
|
Hey @Mojashi, not sure if this change to existing tests makes sense. They are now much more complicated but they were passing before. Is this avoidable?
Hm. I wonder if we should be avoiding |
Closes #213
As described in the issue above,
in cases like the one shown, the following code is generated, which violates the noImplicitAny rule when accessing key1.
a: number|undefinedanda?: numberare different.It's possible to implement a special case only for scenarios where a union is used, but I believe this method to be simpler.