More specific inference for constrained 'infer' types in template literal types #48094
Conversation
This file contains 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
Author: Team
For Uncommitted Bug
Could a new intrinsic like type ParseInt<AString extends string, Radix extends number = 10> = intrinsic;solve that problem? The reason I ask is that a |
I'd considered that approach, however the |
|
Hello, thank you very much for your work. Would the following behavior be correct with this PR? type NumToBigint<T extends number> = `${T}` extends `${Is<infer $BN, bigint>}` ? $BN : never;
type asdf = NumToBigint<23>; // 23n
type asdf2 = NumToBigint<23.5>; // never
type adsf5 = NumToBigint<-2>; // -2n |
|
@typescript-bot pack this |
|
Hey @rbuckton, I've packed this into an installable tgz. You can install it for testing by referencing it in your and then running |
Yes, that is correct. |
ahejlsberg
requested changes
Mar 22, 2022
|
Additionally, might be good to have some tests that check inference to constrained type parameters of functions. For example: declare function test<T extends string | number>(s: `**${T}**`): T;
const x = test('**123**'); // Should produce "123" | 123 |
This doesn't seem to work. We correctly produce inference candidates of |
|
Ah, I may just need to add a new inference priority to indicate an inference to a template type placeholder. |
rbuckton
commented
Mar 22, 2022
|
I've added a new |
|
@typescript-bot perf test |
|
Heya @rbuckton, I've started to run the perf test suite on this PR at df23e95. You can monitor the build here. Update: The results are in! |
|
@rbuckton Here they are:Comparison Report - main..48094
System
Hosts
Scenarios
Developer Information: |
|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
The user suite test run you requested has finished and failed. I've opened a PR with the baseline diff from master. |
rbuckton
force-pushed
the
templateLiteralTypeInferToNonString
branch
from
21c053c to
00c4b26
Compare
|
@ahejlsberg can you take another look? @weswigham I put in a check that will hopefully work with #47050, or at least be a start. |
|
Will this also support floating point numbers? or just integers? |
|
Oh, so we're not far from computing a sum of two numeric literal types! type Test = Sum<1099123, 9901123>
type ToString<A extends number> = `${A}`
type Reverse<A extends string> =
A extends `${infer H}${infer T}`
? `${Reverse<T>}${H}`
: A
type HT<T> = T extends `${infer H}${infer T}` ? {H: H, T: T} : {H: '0', T: ''}
type Head<T> = T extends `${infer H}${any}` ? H : '0'
type Tail<T> = T extends `${string}${infer T}` ? T : ''
type R0 = Tail<''>
type Concat<A, B> = A extends string ? B extends string ? `${A}${B}` : never : never
type Nums = ['0', '1', '2', '3', '4', '5', '6', '7', '8', '9']
type Shuffle<T extends any[]> = T extends [infer H, ...infer T] ? [...T, H] : never
type MakeTable<T extends any[], S extends any[]> = T extends [any, ...infer T] ? [S, ...MakeTable<T, Shuffle<S>>] : []
type SumTable = MakeTable<Nums, Nums>
type SumD = {0: SumTable, 1: Shuffle<SumTable>}
type Index<B, A> = A extends keyof B ? B[A] : never
type SumDR<A, B, C> = Index<Index<Index<SumD, A>, B>, C>
type Carries = [0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1];
type CarryTable = MakeTable<[0, ...Nums], Carries>
type SumC = {0: CarryTable, 1: Shuffle<CarryTable>}
type Lookup<T, C, A, B> = Index<Index<Index<T, C>, Head<A>>, Head<B>>
type SumS<A extends string, B extends string, C = 0> = Concat<A, B> extends ''
? C extends 1 ? '1' : ''
: Concat<SumS<Tail<A>, Tail<B>, Lookup<SumC, C, A, B>>, Lookup<SumD, C, A, B>>
type Sum<A extends number, B extends number> = SumS<Reverse<ToString<A>>, Reverse<ToString<B>>> |
ahejlsberg
reviewed
May 24, 2022
rbuckton
force-pushed
the
templateLiteralTypeInferToNonString
branch
from
80877c2 to
9ccca98
Compare
ahejlsberg
approved these changes
May 27, 2022
6 tasks
Closed
2 tasks
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
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.
This modifies inference in template literal types when inferring to a constrained type variable by leveraging the constraint of the type variable to infer a more specific type:
The primary motivation for this feature is converting numeric string index keys from a tuple type into their numeric literal equivalents. While you can use something like a mapping table for a subset of numbers (i.e., as many as you are willing to manually define), that can run into issues when that might then generate a too-large union. Rather than introducing a feature specific to numeric string index keys, this instead focuses on a more general purpose solution that can apply to any valid template literal type placeholder (i.e.,
string,number,bigint,boolean,null, andundefined).Please note that this approach cannot be used to convert an arbitrary numeric string into a
numberorbigintliteral type, as the inferred type must still satisfy the extends condition in the conditional type. This means that a string such as"0x10"cannot be converted to anumber, since that value would not round-trip: If you attempt to infer a number literal type from"0x10", the result would be16, however when16is substituted back into the string literal type it becomes"16", which is not a supertype of"0x10".Also note that this PR does not introduce any new mechanism for establishing a constraint for an
infer Ttype. Instead, a helper type likeIs(above) might be necessary to enforce the constraint. A syntactic mechanism to constraininfer Twould be valuable, but is out of scope for this PR.Addendum: If we decide to take #48112, the above example could be rewritten using
infer T extends ...instead of theIstype alias:Fixes #47141