Improve narrowing of generic types in control flow analysis#43183
Uh oh!
There was an error while loading. Please reload this page.
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 Milestone Bug 
MemberAuthor
ahejlsberg commented Mar 10, 2021
@typescript-bot test this |
Collaborator
Heya @ahejlsberg, I've started to run the extended test suite on this PR at a0fbf5c. You can monitor the build here. |
Collaborator
Heya @ahejlsberg, I've started to run the perf test suite on this PR at a0fbf5c. You can monitor the build here. Update: The results are in! |
Collaborator
Heya @ahejlsberg, I've started to run the parallelized Definitely Typed test suite on this PR at a0fbf5c. You can monitor the build here. |
Collaborator
Heya @ahejlsberg, I've started to run the abridged perf test suite on this PR at a0fbf5c. You can monitor the build here. Update: The results are in! |
Collaborator
Heya @ahejlsberg, I've started to run the parallelized community code test suite on this PR at a0fbf5c. You can monitor the build here. |
Collaborator
typescript-bot commented Mar 10, 2021
@ahejlsberg Here they are:Comparison Report - master..43183
System
Hosts
Scenarios
Developer Information: | |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Collaborator
typescript-bot commented Mar 10, 2021
The user suite test run you requested has finished and failed. I've opened a PR with the baseline diff from master. |
Collaborator
typescript-bot commented Mar 10, 2021
@ahejlsberg Here they are:Comparison Report - master..43183
System
Hosts
Scenarios
Developer Information: | |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
src/compiler/checker.ts Outdated
| function getConstraintForReference(type: Type, reference: Identifier | ElementAccessExpression | PropertyAccessExpression | QualifiedName, checkMode: CheckMode | undefined){ | ||
| // When the type of a reference is or contains an instantiable type with a union type constraint, and | ||
| // when the reference is in a constraint position (where it is known we'll obtain the apparent type) or | ||
| // has a contextual type containing no instantiables (meaning constraints will determine assignability), |
There was a problem hiding this comment.
Suggested change
| // has a contextual type containing no instantiables (meaning constraints will determine assignability), | |
| // has a contextual type containing no top-level instantiables (meaning constraints will determine assignability), |
or
Suggested change
| // has a contextual type containing no instantiables (meaning constraints will determine assignability), | |
| // has a contextual type containing no immediate instantiables (meaning constraints will determine assignability), |
Member
DanielRosenwasser commented Mar 10, 2021
@typescript-bot pack this |
Collaborator
Heya @DanielRosenwasser, I've started to run the tarball bundle task on this PR at 1a695e9. You can monitor the build here. |
Collaborator
Hey @DanielRosenwasser, I've packed this into an installable tgz. You can install it for testing by referencing it in your and then running There is also a playground for this build and an npm module you can use via |
Member
DanielRosenwasser commented Mar 10, 2021
@ahejlsberg might be worth adding in the following test cases - interfaceBox<T>{item: T;}declarefunctionisBox(x: any): x is Box<unknown>;declarefunctionisUndefined(x: unknown): x is undefined;declarefunctionunbox<T>(x: Box<T>): T;functionf1<TextendsBox<T>|undefined>(x: T){if(isBox(x)){unbox(x);}}functionf2<TextendsBox<T>|undefined>(x: T){if(!isUndefined(x)){unbox(x);}}functionf3<TextendsBox<T>|undefined>(x: T){if(!isBox(x)){// errorunbox(x);}}functionf4<TextendsBox<T>|undefined>(x: T){if(isUndefined(x)){// errorunbox(x);}} |
MemberAuthor
ahejlsberg commented Mar 11, 2021
@DanielRosenwasser Will do. |
MemberAuthor
Performance looks to be unaffected. RWC test suite looks good too, only change is a couple of error messages that now show constraint types instead of generic types (as expected). |
5 tasks
RaynalHugo added a commit to labelflow/labelflow that referenced this pull request Jul 23, 2021
13 tasks
chriskrycho added a commit to true-myth/true-myth that referenced this pull request Apr 22, 2022
TS 4.7 catches a class of error which 4.6 and earlier did not, courtesy of microsoft/TypeScript#43183: if the type passed does not implement `toString()`, it notices, because it no longer defaults to falling back to `{}`. Fixes#330
chriskrycho added a commit to true-myth/true-myth that referenced this pull request Apr 22, 2022
TS 4.7 catches a class of error which 4.6 and earlier did not, courtesy of microsoft/TypeScript#43183: if the type passed does not implement `toString()`, it notices, because it no longer defaults to falling back to `{}`. To resolve this, *loosen* the constraints on what is acceptable as input to `toString`, by making the implementation itself more robust. Doing so also fixes a bug with the output for `toString` when working with strings: previously if you had `Maybe("a string")`, the `toString` output would be `'Maybe(a string)'`; it is now `'Maybe("a string")'`. Fixes#330
chriskrycho added a commit to true-myth/true-myth that referenced this pull request Apr 22, 2022
TS 4.7 catches a class of error which 4.6 and earlier did not, courtesy of microsoft/TypeScript#43183: if the type passed does not implement `toString()`, it notices, because it no longer defaults to falling back to `{}`. To resolve this, *loosen* the constraints on what is acceptable as input to `toString`, by making the implementation itself more robust. Doing so also fixes a bug with the output for `toString` when working with strings: previously if you had `Maybe("a string")`, the `toString` output would be `'Maybe(a string)'`; it is now `'Maybe("a string")'`. Fixes#330
chriskrycho added a commit to true-myth/true-myth that referenced this pull request Apr 22, 2022
(cherry picked from commit 25adb1a) --- Original message: Add support for TypeScript 4.7 TS 4.7 catches a class of error which 4.6 and earlier did not, courtesy of microsoft/TypeScript#43183: if the type passed does not implement `toString()`, it notices, because it no longer defaults to falling back to `{}`. To resolve this, *loosen* the constraints on what is acceptable as input to `toString`, by making the implementation itself more robust. Doing so also fixes a bug with the output for `toString` when working with strings: previously if you had `Maybe("a string")`, the `toString` output would be `'Maybe(a string)'`; it is now `'Maybe("a string")'`. Fixes#330
4 tasks
Sign up for freeto subscribe to this conversation on GitHub. Already have an account? Sign in.
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.
In #15576 we improved control flow analysis for references of generic types with nullable constraints when those references occur in constraint positions (locations where it is known we'll fetch the constraint).
This PR expands the concept further. Specifically, when the type of a reference is or contains a generic type with a union type constraint, and when the reference is in a constraint position or has a contextual type containing no generic types (meaning constraints will determine assignability), we substitute constraints for all generic types in the type of the reference to give control flow analysis an opportunity to narrow it further. For example:
Above, in
x.lengthwe substitutestring | undefinedforTin the type ofxbecausexoccurs in a constraint position. Similarly inreturn xwe substitutestring | undefinedforTin the type ofxbecausexhas the contextual typestring(due to the return type annotation on the containing function) which contains no generic types. Following substitution, control flow analysis is then able to narrowstring | undefinedtostringbased on the truthiness check in the containingifstatement.Some additional examples:
Fixes#13995.
Fixes#20375.
Fixes#42939.