[compiler] Preserve TSNonNullExpressions #33097
Conversation
This is a first step toward teaching the compiler that TS non-null expressions mean a variable may be nullable and that we can't take dependencies past the non-null assertion. To start, we add a new instruction type and add support from lowering through codegen. [ghstack-poisoned]
This is a first step toward teaching the compiler that TS non-null expressions mean a variable may be nullable and that we can't take dependencies past the non-null assertion. To start, we add a new instruction type and add support from lowering through codegen. ghstack-source-id: 730319c60e173f347705e8d29f12deca44bfaf88 Pull Request resolved: #33097
There was a problem hiding this comment.
This might not be relevant for Forget, but keep in mind that the ! operator interacts with optional chaining, so having full support for it will probably require more than this, either in lowering or in HIR. E.g. assuming a and a.b are nullable, a?.b!.c is nullable despite the presence of the ! operator within the chain (with the null from a, but not from a.b), while (a?.b)!.c is not nullable and neither is a?.b?.c!. Mentioning this because I just discovered this and was surprised by it recently lol.
|
@mvitousek Interesting! I definitely want to understand the rules behind that, some of those aren't intuitive. In either case I was planning to hold off landing this until we explore building the nullability inference on top. |
This comment was marked as off-topic.
This comment was marked as off-topic.
Sorry, something went wrong.
Stack from ghstack (oldest at bottom):
This is a first step toward teaching the compiler that TS non-null expressions mean a variable may be nullable and that we can't take dependencies past the non-null assertion - see #33054.
To start, we add a new instruction type and add support from lowering through codegen.