Avoid rewriting homomorphic mapped types with homomorphic instantiations #53215
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
typescript-bot
added
the
For Uncommitted Bug
weswigham
requested changes
Mar 14, 2023
src/compiler/checker.ts
Outdated
| return isMappedTypeWithKeyofConstraintDeclaration(type) | ||
| && !(getModifiersTypeFromMappedType(type).flags & TypeFlags.TypeParameter); | ||
| && !(getModifiersTypeFromMappedType(type).flags & TypeFlags.TypeParameter) | ||
| && !!type.aliasSymbol; |
Member
There was a problem hiding this comment.
This shouldn't have anything to do with the alias symbol as by the time we're trying to print with this function, the alias symbol has already been discarded as an option for printing the mapped type. Instead, what we care about is if the type will be picked up as a homomorphic type variable (or not). Type parameters always are, so were a conservative initial choice - it's generally safe to wrap anything else which is still recognized as a homomorphic type variable up and present it as a type parameter, too - all you get is unnecessarily verbose printback sometimes. Instead of checking the alias symbol here (which seems pretty unrelated), you want to check, rather than if the modifiers type is still a type parameter, if an index on the modifiers type still produces a homomorphic type variable. In this case, an index of a mapped type just returns the inner mapped type's constraint type, which can then be unwrapped, so it is recognized as homomorphic. So, instead of the simple, conservative !(getModifiersTypeFromMappedType(type).flags & TypeFlags.TypeParameter) check, you want something to more closely mirror getHomomorphicTypeVariable's conditions like
let index: Type;
return isMappedTypeWithKeyofConstraintDeclaration(type)
&& (index = getIndexType(getModifiersTypeFromMappedType(type)), !(index.flags & TypeFlags.Index && (index as IndexType).type.flags & TypeFlags.TypeParameter)))
Contributor
Author
There was a problem hiding this comment.
I applied the suggested change but it reverted the change to the emit of this thing (TS playground):
export declare type Omit<T, K extends keyof T> = Pick<T, Exclude<keyof T, K>>;
export declare type PartialProperties<T, K extends keyof T> = Partial<Pick<T, K>> & Omit<T, K>;
export function doSomething_Actual<T extends {
prop: string;
}>(a: T) {
const x: { [P in keyof PartialProperties<T, "prop">]: PartialProperties<T, "prop">[P]; } = null as any;
return x;
}Does this one have to be rewritten? 🤔
Member
There was a problem hiding this comment.
Probably not, since, in that case, I don't think PartialProperties is homomorphic, so the wrapping mapped type shouldn't be either. Right now we're using isMappedTypeWithKeyofConstraintDeclaration as a proxy for was-this-declared-as-homomorphic - we probably need to swap that out for something a bit more accurate, too. Potentially the same check we're doing on type but on type.target (and negated)?
Member
There was a problem hiding this comment.
Indeed, it's not actually homomorphic -
doSomething_Actual<{prop: string, propb: string} | {prop: string, propa: string}>(null as any);ends up with type
{
prop?: string | undefined;
}so we just need to tighten up that side of the check, too.
Contributor
Author
There was a problem hiding this comment.
This particular mapped type doesn't have a .target property - I might have misunderstood your recommendation for the proposed changes but for now, I've chosen to just avoid rewriting in case of !type.target. I think this is correct (it yields the expected result for this case and the rest of the test suite) but maybe I'm missing some further details.
Member
There was a problem hiding this comment.
Actually, that's perfect - if the mapped type doesn't have a .target, then leaving it as-is is always correct - it's still it's declared form, uninstantiated.
Andarist
changed the title
weswigham
requested changes
Mar 16, 2023
Member
weswigham
left a comment
weswigham
left a comment
There was a problem hiding this comment.
Yeah, let's replace the isMappedTypeWithKeyofConstraintDeclaration check with something that more accurately checks if the declared type was homomorphic (eg, by doing the same check we do to see if type is homomorphic, but on type.target) - that should be the last piece needed here, I think.
weswigham
requested changes
Mar 20, 2023
src/compiler/checker.ts
Outdated
Comment on lines
6627
to
6632
| function isHomomorphicMappedTypeWithNonHomomorphicInstantiation(type: MappedType) { | ||
| return isMappedTypeWithKeyofConstraintDeclaration(type) | ||
| && !(getModifiersTypeFromMappedType(type).flags & TypeFlags.TypeParameter); | ||
| if (!type.target || !isMappedTypeWithKeyofConstraintDeclaration(type)) { | ||
| return false; | ||
| } | ||
| const index = getIndexType(getModifiersTypeFromMappedType(type)); | ||
| return !(index.flags & TypeFlags.Index && (index as IndexType).type.flags & TypeFlags.TypeParameter); |
Member
There was a problem hiding this comment.
So I was thinking something like
function isMappedTypeHomomorphic(type: MappedType) {
return !!getHomomorphicTypeVariable(type);
}
function isHomomorphicMappedTypeWithNonHomomorphicInstantiation(type: MappedType) {
return !type.target ? false : isMappedTypeHomomorphic(type.target) && !isMappedTypeHomomorphic(type);
}Looks easier to reason about a bit, too, by replacing the complex check I proposed before with a direct getHomomorphicTypeVariable call on the type and its' target.
Contributor
Author
There was a problem hiding this comment.
nice, this is definitely is easier to reason about 👍 pushed out the suggested change
…e-rewriting-in-declarations
weswigham
approved these changes
Mar 20, 2023
Member
weswigham
left a comment
weswigham
left a comment
There was a problem hiding this comment.
let's just double check the tests...
Member
|
@typescript-bot test this |
Collaborator
|
Heya @weswigham, I've started to run the diff-based top-repos suite on this PR at 4b6afba. You can monitor the build here. Update: The results are in! |
Collaborator
|
Heya @weswigham, I've started to run the extended test suite on this PR at 4b6afba. You can monitor the build here. |
Collaborator
|
Heya @weswigham, I've started to run the parallelized Definitely Typed test suite on this PR at 4b6afba. You can monitor the build here. Update: The results are in! |
Collaborator
|
Heya @weswigham, I've started to run the perf test suite on this PR at 4b6afba. You can monitor the build here. Update: The results are in! |
Collaborator
|
@weswigham Here they are:
CompilerComparison Report - main..53215
System
Hosts
Scenarios
TSServerComparison Report - main..53215
System
Hosts
Scenarios
StartupComparison Report - main..53215
System
Hosts
Scenarios
Developer Information: |
|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Collaborator
|
@weswigham Here are the results of running the top-repos suite comparing Everything looks good! |
Collaborator
|
Hey @weswigham, it looks like the DT test run failed. Please check the log for more details. |
Sign up for free
to subscribe to this conversation on GitHub.
Already have an account?
Sign in.
3 participants
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
fixes #53109
I'm not 100% sure of this change but I can't find why a non-aliased mapped type would have to be rewritten. The PR that introduced this rewriting also mentions explicitly that it's about "inlined mapped types" (PR here) so I think that the ones that are already inline can be just left as-is.