|
|
|
|
|
by curveship
3348 days ago
|
|
Sort of. Strict null checking now lets you distinguish between the types Foo[] and (Foo | undefined)[]. The first will type with the assumption that there are no out of bounds indices (which, as you say, could be wrong). The second will require you to always check the result of array dereferences to see if they're undefined before you reference the object. var foos1 = [] as Foo[],
foos2 = [] as (Foo | undefined)[];
// ... later
var foo1 = foos1[i],
foo2 = foos2[i];
foo1.bar(); // types fine, TS assumes non-undefined
foo2.bar(); // TS type error, as you haven't checked undef
if (foo2) foo2.bar(); // types fine, undef excluded
|
|
E.g. if I missed a bounds check, and access past the end of a `Foo[]`, it would be typed as `Foo` (and not `Foo | undefined`), but really be `undefined.
Consider this code.
foos[1] will be undefined (since it's past the end of the array) but will typecheck as a Foo.Obviously it's bad if I forget a bounds check, but it's similarly bad if I forgot a null check.
Of course, I understand why this issue exists, but it is an issue.