Improve logic that chooses co- vs. contra-variant return type inferences by Andarist · Pull Request #59709 · microsoft/TypeScript

Mateusz Burzyński (Andarist) · 2024-08-21T11:12:11Z

fixes#59656
cc Wesley Wigham (@weswigham) as the reviewer of #57909 by which this got affected

Mateusz Burzyński (Andarist) · 2024-08-21T11:14:11Z

- // and has inferences that would conflict. Otherwise, we prefer the contra-variant inference.
- // Similarly ignore co-variant `any` inference when both are available as almost everything is assignable to it
- // and it would spoil the overall inference.
+ // ideally all inferences would be tried out in a ranked order but that's too costly


I rewritten this comment based on the excellent comment by Wesley Wigham (@weswigham)here. I tried to be concise though as this is already pretty long and Wes' comment was even longer 😉 Feel free to suggest any improvement to this to make it better. I write better in TS than in English.

Mateusz Burzyński (Andarist) · 2024-08-21T11:22:02Z

+ // however, for return types inferences subtyping is used. In those situations, the contravariant inference often comes from the argument of mapping function
+ // and the outer covariant requirement can still be satisfied by the return position of the return type when it's a subtype of that required return type.
+ // this helps to provide better contextual parameter types in scenarios like this:
+ //
+ // declare const obs: Observable<{ a?: string; b?: number }>;
+ // const test = (): Observable<{ a?: string }> => obs.pipe(tap((arg) => {}));


This part of the comment tries to explain the idea behind the fix.
An alternative idea I had was to experiment with a new inference priority to differentiate between covariant inferences made from return positions. However, that's a bigger experiment and what I propose here is much safer at this stage. It's essentially reverting part of #57909 to 5.5 behavior. So it shouldn't behave worse in any case.
But also given that return positions could themselves be within nested functions, I don't have a good intuition on how it could play out. At the moment, there is no concept of "depth" in the inference and the core of this idea conceptually is that a return position might not be more important than a contravariant inference made from the (any?) containing function.
This would have some potential to solve Wes' concern raised here related to inferring from (x: number) => 0 to (x: T) => T.

Jake Bailey (jakebailey) · 2024-08-21T19:53:42Z

TypeScript Bot (@typescript-bot) test it

TypeScript Bot (typescript-bot) · 2024-08-21T19:53:46Z

Starting jobs; this comment will be updated as builds start and complete.

Command	Status	Results
`test top400`	✅ Started	✅ Results
`user test this`	✅ Started	✅ Results
`run dt`	✅ Started	✅ Results
`perf test this faster`	✅ Started	👀 Results

TypeScript Bot (typescript-bot) · 2024-08-21T20:12:55Z

Hey Jake Bailey (@jakebailey), the results of running the DT tests are ready.

Everything looks the same!

You can check the log here.

TypeScript Bot (typescript-bot) · 2024-08-21T20:15:39Z

Jake Bailey (@jakebailey) Here are the results of running the user tests with tsc comparing main and refs/pull/59709/merge:

Everything looks good!

TypeScript Bot (typescript-bot) · 2024-08-21T20:31:39Z

Jake Bailey (@jakebailey)
The results of the perf run you requested are in!

Here they are:

tsc

Comparison Report - baseline..pr

_Metric	_baseline	_pr	_Delta	_Best	_Worst	_p-value
_{Compiler-Unions - node (v18.15.0, x64)}
_Errors	₃₀	₃₀	_~	_~	_~	_{p=1.000 n=6}
_Symbols	_62,153	_62,153	_~	_~	_~	_{p=1.000 n=6}
_Types	_50,242	_50,242	_~	_~	_~	_{p=1.000 n=6}
_{Memory used}	_{193,608k (± 0.92%)}	_{192,994k (± 0.77%)}	_~	_192,340k	_196,046k	_{p=0.936 n=6}
_{Parse Time}	_{1.30s (± 0.58%)}	_{1.31s (± 0.39%)}	_{+0.01s (+ 0.90%)}	_1.31s	_1.32s	_{p=0.020 n=6}
_{Bind Time}	_0.71s	_0.71s	_~	_~	_~	_{p=1.000 n=6}
_{Check Time}	_{9.60s (± 0.91%)}	_{9.57s (± 0.41%)}	_~	_9.52s	_9.61s	_{p=0.467 n=6}
_{Emit Time}	_{2.72s (± 0.63%)}	_{2.72s (± 1.03%)}	_~	_2.66s	_2.74s	_{p=0.560 n=6}
_{Total Time}	_{14.33s (± 0.62%)}	_{14.31s (± 0.29%)}	_~	_14.26s	_14.36s	_{p=0.688 n=6}
_{angular-1 - node (v18.15.0, x64)}
_Errors	₇	₇	_~	_~	_~	_{p=1.000 n=6}
_Symbols	_945,753	_945,753	_~	_~	_~	_{p=1.000 n=6}
_Types	_410,067	_410,067	_~	_~	_~	_{p=1.000 n=6}
_{Memory used}	_{1,222,694k (± 0.00%)}	_{1,222,696k (± 0.00%)}	_~	_1,222,670k	_1,222,728k	_{p=0.936 n=6}
_{Parse Time}	_{6.66s (± 0.36%)}	_{6.65s (± 0.44%)}	_~	_6.63s	_6.71s	_{p=0.684 n=6}
_{Bind Time}	_{1.86s (± 0.40%)}	_{1.86s (± 0.53%)}	_~	_1.85s	_1.88s	_{p=0.858 n=6}
_{Check Time}	_{31.18s (± 0.44%)}	_{31.14s (± 0.35%)}	_~	_30.99s	_31.31s	_{p=0.748 n=6}
_{Emit Time}	_{14.91s (± 0.56%)}	_{14.98s (± 0.44%)}	_~	_14.89s	_15.06s	_{p=0.225 n=6}
_{Total Time}	_{54.60s (± 0.29%)}	_{54.64s (± 0.26%)}	_~	_54.46s	_54.79s	_{p=1.000 n=6}
_{mui-docs - node (v18.15.0, x64)}
_Errors	₀	₀	_~	_~	_~	_{p=1.000 n=6}
_Symbols	_2,517,779	_2,517,779	_~	_~	_~	_{p=1.000 n=6}
_Types	_994,242	_994,242	_~	_~	_~	_{p=1.000 n=6}
_{Memory used}	_{2,442,563k (± 0.00%)}	_{2,442,540k (± 0.00%)}	_~	_2,442,512k	_2,442,577k	_{p=0.230 n=6}
_{Parse Time}	_{9.30s (± 0.35%)}	_{9.29s (± 0.38%)}	_~	_9.22s	_9.32s	_{p=0.366 n=6}
_{Bind Time}	_{2.20s (± 0.62%)}	_{2.19s (± 0.45%)}	_~	_2.18s	_2.20s	_{p=0.673 n=6}
_{Check Time}	_{74.83s (± 0.29%)}	_{74.77s (± 0.51%)}	_~	_74.16s	_75.18s	_{p=1.000 n=6}
_{Emit Time}	_{0.28s (± 3.19%)}	_{0.28s (± 3.53%)}	_~	_0.27s	_0.29s	_{p=0.798 n=6}
_{Total Time}	_{86.60s (± 0.29%)}	_{86.53s (± 0.44%)}	_~	_85.95s	_86.95s	_{p=0.936 n=6}
_{self-build-src - node (v18.15.0, x64)}
_Errors	₀	₀	_~	_~	_~	_{p=1.000 n=6}
_Symbols	_1,231,998	_1,231,999	_{+1 (+ 0.00%)}	_~	_~	_{p=0.001 n=6}
_Types	_264,593	_264,593	_~	_~	_~	_{p=1.000 n=6}
_{Memory used}	_{2,472,512k (± 7.44%)}	_{2,472,419k (± 7.44%)}	_~	_2,353,350k	_2,710,065k	_{p=0.810 n=6}
_{Parse Time}	_{6.04s (± 1.01%)}	_{5.99s (± 0.62%)}	_~	_5.94s	_6.04s	_{p=0.128 n=6}
_{Bind Time}	_{2.26s (± 0.91%)}	_{2.25s (± 1.20%)}	_~	_2.21s	_2.29s	_{p=0.418 n=6}
_{Check Time}	_{40.79s (± 0.79%)}	_{40.76s (± 0.76%)}	_~	_40.23s	_41.03s	_{p=0.689 n=6}
_{Emit Time}	_{4.03s (± 0.87%)}	_{4.01s (± 0.74%)}	_~	_3.96s	_4.04s	_{p=0.335 n=6}
_{Total Time}	_{53.14s (± 0.61%)}	_{53.03s (± 0.56%)}	_~	_52.54s	_53.37s	_{p=0.575 n=6}
_{self-build-src-public-api - node (v18.15.0, x64)}
_Errors	₀	₀	_~	_~	_~	_{p=1.000 n=6}
_Symbols	_1,231,998	_1,231,999	_{+1 (+ 0.00%)}	_~	_~	_{p=0.001 n=6}
_Types	_264,593	_264,593	_~	_~	_~	_{p=1.000 n=6}
_{Memory used}	_{2,428,294k (± 0.02%)}	_{2,428,201k (± 0.02%)}	_~	_2,427,639k	_2,428,911k	_{p=0.689 n=6}
_{Parse Time}	_{6.24s (± 0.45%)}	_{6.26s (± 0.95%)}	_~	_6.19s	_6.34s	_{p=0.689 n=6}
_{Bind Time}	_{2.04s (± 0.67%)}	_{2.04s (± 1.41%)}	_~	_1.99s	_2.07s	_{p=0.808 n=6}
_{Check Time}	_{41.65s (± 0.63%)}	_{41.85s (± 0.59%)}	_~	_41.46s	_42.08s	_{p=0.230 n=6}
_{Emit Time}	_{4.06s (± 0.63%)}	_{4.06s (± 0.76%)}	_~	_4.03s	_4.11s	_{p=0.809 n=6}
_{Total Time}	_{53.99s (± 0.49%)}	_{54.24s (± 0.54%)}	_~	_53.83s	_54.55s	_{p=0.173 n=6}
_{self-compiler - node (v18.15.0, x64)}
_Errors	₀	₀	_~	_~	_~	_{p=1.000 n=6}
_Symbols	_256,992	_256,993	_{+1 (+ 0.00%)}	_~	_~	_{p=0.001 n=6}
_Types	_105,667	_105,667	_~	_~	_~	_{p=1.000 n=6}
_{Memory used}	_{429,722k (± 0.05%)}	_{429,626k (± 0.03%)}	_~	_429,402k	_429,751k	_{p=0.810 n=6}
_{Parse Time}	_{2.79s (± 0.71%)}	_{2.78s (± 0.60%)}	_~	_2.77s	_2.81s	_{p=0.222 n=6}
_{Bind Time}	_{1.08s (± 0.91%)}	_{1.08s (± 1.12%)}	_~	_1.07s	_1.10s	_{p=0.388 n=6}
_{Check Time}	_{15.39s (± 0.35%)}	_{15.36s (± 0.41%)}	_~	_15.28s	_15.43s	_{p=0.521 n=6}
_{Emit Time}	_{1.42s (± 1.07%)}	_{1.40s (± 0.84%)}	_~	_1.38s	_1.41s	_{p=0.071 n=6}
_{Total Time}	_{20.68s (± 0.25%)}	_{20.62s (± 0.39%)}	_~	_20.53s	_20.74s	_{p=0.172 n=6}
_{ts-pre-modules - node (v18.15.0, x64)}
_Errors	₆₈	₆₈	_~	_~	_~	_{p=1.000 n=6}
_Symbols	_225,018	_225,018	_~	_~	_~	_{p=1.000 n=6}
_Types	_94,249	_94,249	_~	_~	_~	_{p=1.000 n=6}
_{Memory used}	_{370,188k (± 0.02%)}	_{370,215k (± 0.02%)}	_~	_370,090k	_370,318k	_{p=0.689 n=6}
_{Parse Time}	_{2.77s (± 0.54%)}	_{2.76s (± 1.16%)}	_~	_2.73s	_2.81s	_{p=0.625 n=6}
_{Bind Time}	_{1.58s (± 1.81%)}	_{1.58s (± 0.74%)}	_~	_1.57s	_1.60s	_{p=0.560 n=6}
_{Check Time}	_{15.75s (± 0.24%)}	_{15.78s (± 0.42%)}	_~	_15.68s	_15.86s	_{p=0.422 n=6}
_{Emit Time}	_0.00s	_0.00s	_~	_~	_~	_{p=1.000 n=6}
_{Total Time}	_{20.10s (± 0.24%)}	_{20.12s (± 0.26%)}	_~	_20.04s	_20.17s	_{p=0.466 n=6}
_{vscode - node (v18.15.0, x64)}
_Errors	₆₆	₆₆	_~	_~	_~	_{p=1.000 n=6}
_Symbols	_3,020,870	_3,020,870	_~	_~	_~	_{p=1.000 n=6}
_Types	_1,039,240	_1,039,240	_~	_~	_~	_{p=1.000 n=6}
_{Memory used}	_{3,140,085k (± 0.00%)}	_{3,140,216k (± 0.00%)}	_{+131k (+ 0.00%)}	_3,140,127k	_3,140,294k	_{p=0.020 n=6}
_{Parse Time}	_{13.98s (± 0.46%)}	_{14.02s (± 0.39%)}	_~	_13.93s	_14.09s	_{p=0.261 n=6}
_{Bind Time}	_{4.35s (± 2.05%)}	_{4.30s (± 0.23%)}	_~	_4.29s	_4.31s	_{p=0.066 n=6}
_{Check Time}	_{80.07s (± 0.31%)}	_{80.27s (± 0.62%)}	_~	_79.66s	_80.86s	_{p=0.378 n=6}
_{Emit Time}	_{20.50s (± 0.62%)}	_{20.52s (± 0.80%)}	_~	_20.34s	_20.70s	_{p=0.936 n=6}
_{Total Time}	_{118.90s (± 0.23%)}	_{119.11s (± 0.45%)}	_~	_118.43s	_119.85s	_{p=0.471 n=6}
_{webpack - node (v18.15.0, x64)}
_Errors	₀	₀	_~	_~	_~	_{p=1.000 n=6}
_Symbols	_275,311	_275,311	_~	_~	_~	_{p=1.000 n=6}
_Types	_112,431	_112,431	_~	_~	_~	_{p=1.000 n=6}
_{Memory used}	_{424,171k (± 0.02%)}	_{424,189k (± 0.02%)}	_~	_424,072k	_424,324k	_{p=0.575 n=6}
_{Parse Time}	_{3.97s (± 0.47%)}	_{3.98s (± 0.61%)}	_~	_3.94s	_4.00s	_{p=0.224 n=6}
_{Bind Time}	_{1.72s (± 0.68%)}	_{1.72s (± 0.88%)}	_~	_1.70s	_1.74s	_{p=0.410 n=6}
_{Check Time}	_{17.51s (± 0.40%)}	_{17.47s (± 0.73%)}	_~	_17.30s	_17.59s	_{p=0.810 n=6}
_{Emit Time}	_0.00s	_0.00s	_~	_~	_~	_{p=1.000 n=6}
_{Total Time}	_{23.19s (± 0.25%)}	_{23.17s (± 0.62%)}	_~	_22.96s	_23.32s	_{p=1.000 n=6}
_{xstate-main - node (v18.15.0, x64)}
_Errors	₀	₀	_~	_~	_~	_{p=1.000 n=6}
_Symbols	_534,554	_534,554	_~	_~	_~	_{p=1.000 n=6}
_Types	_176,379	_176,379	_~	_~	_~	_{p=1.000 n=6}
_{Memory used}	_{479,444k (± 0.01%)}	_{479,474k (± 0.01%)}	_{+30k (+ 0.01%)}	_479,408k	_479,505k	_{p=0.045 n=6}
_{Parse Time}	_{4.26s (± 1.04%)}	_{4.28s (± 0.59%)}	_~	_4.25s	_4.32s	_{p=0.625 n=6}
_{Bind Time}	_{1.54s (± 0.95%)}	_{1.54s (± 0.76%)}	_~	_1.53s	_1.56s	_{p=1.000 n=6}
_{Check Time}	_{22.48s (± 0.24%)}	_{22.49s (± 0.37%)}	_~	_22.37s	_22.60s	_{p=0.810 n=6}
_{Emit Time}	_0.00s	_0.00s	_~	_~	_~	_{p=1.000 n=6}
_{Total Time}	_{28.28s (± 0.29%)}	_{28.31s (± 0.33%)}	_~	_28.18s	_28.46s	_{p=0.521 n=6}

System info unknown

Hosts

node (v18.15.0, x64)

Scenarios

Compiler-Unions - node (v18.15.0, x64)
angular-1 - node (v18.15.0, x64)
mui-docs - node (v18.15.0, x64)
self-build-src - node (v18.15.0, x64)
self-build-src-public-api - node (v18.15.0, x64)
self-compiler - node (v18.15.0, x64)
ts-pre-modules - node (v18.15.0, x64)
vscode - node (v18.15.0, x64)
webpack - node (v18.15.0, x64)
xstate-main - node (v18.15.0, x64)

_Benchmark	_Name	_Iterations
_Current	_pr	₆
_Baseline	_baseline	₆

Developer Information:

Download Benchmarks

TypeScript Bot (typescript-bot) · 2024-08-21T21:39:37Z

Jake Bailey (@jakebailey) Here are the results of running the top 400 repos with tsc comparing main and refs/pull/59709/merge:

Everything looks good!

Jake Bailey (jakebailey) · 2024-08-22T20:31:39Z

TypeScript Bot (@typescript-bot) pack this

TypeScript Bot (typescript-bot) · 2024-08-22T20:31:42Z

Starting jobs; this comment will be updated as builds start and complete.

Command	Status	Results
`pack this`	✅ Started	✅ Results

TypeScript Bot (typescript-bot) · 2024-08-22T20:33:59Z

Hey Jake Bailey (@jakebailey), I've packed this into an installable tgz. You can install it for testing by referencing it in your package.json like so:

{
"devDependencies": {
"typescript": "https://typescript.visualstudio.com/cf7ac146-d525-443c-b23c-0d58337efebc/_apis/build/builds/163412/artifacts?artifactName=tgz&fileId=FC442BACFFC913E8502821A0DD902D0764AF8ADC5017DFEDA553B3332B29087102&fileName=/typescript-5.7.0-insiders.20240822.tgz"
}
}

and then running npm install.

There is also a playground for this build and an npm module you can use via "typescript": "npm:@typescript-deploys/pr-build@5.7.0-pr-59709-8".;

Daniel Rosenwasser (DanielRosenwasser) · 2024-08-22T21:10:40Z

TypeScript Bot (@typescript-bot) test top999

TypeScript Bot (typescript-bot) · 2024-08-22T21:10:43Z

Starting jobs; this comment will be updated as builds start and complete.

Command	Status	Results
`test top999`	✅ Started	✅ Results

TypeScript Bot (typescript-bot) · 2024-08-22T23:38:09Z

Daniel Rosenwasser (@DanielRosenwasser) Here are the results of running the top 999 repos with tsc comparing main and refs/pull/59709/merge:

Everything looks good!

TypeScript Bot (typescript-bot) · 2026-03-24T21:19:53Z

With 6.0 out as the final release vehicle for this codebase, we're closing all PRs that don't fit the merge criteria for post-6.0 patches. If you think this was a mistake and this PR fits the post-6.0 patch criteria, please post to the 6.0 iteration issue with details (specifically, which PR and which patch criteria it satisfies).

Next steps for PRs:

For crash bugfixes or language service improvements, PRs are currently accepted at the typescript-go repo
Changes to type system behavior should wait until after 7.0, at which point mainline TypeScript development will resume in this repository with the Go codebase
Library file updates (lib.d.ts etc) continue to live in this repo or the DOM Generator repo as appropriate

Improve logic that chooses co- vs. contra-variant return type inferences
d776ee9

TypeScript Bot (typescript-bot) added the For Backlog Bug PRs that fix a backlog bug label Aug 21, 2024

Mateusz Burzyński (Andarist) commented Aug 21, 2024
View reviewed changes

Jake Bailey (jakebailey) requested review from Jake Bailey (jakebailey) and Wesley Wigham (weswigham)August 21, 2024 19:53

Jake Bailey (jakebailey) mentioned this pull request Aug 22, 2024
5.6 regression: Incorrect param type inference for type with all optional props #59656
Open

Dan Vanderkam (danvk) mentioned this pull request Sep 26, 2024
Inferred type changes from 5.5 to 5.6 #60077
Closed

TypeScript Bot (typescript-bot) closed this Mar 24, 2026

github-project-automationBot moved this from Not started to Done in PR Backlog Mar 24, 2026

Dhruv Sharma (Dhruv2mars) mentioned this pull request Jul 10, 2026
Google feedback on TS 5.6-beta Dhruv2mars/typescript-relunar-testbed#4318
Open
1 task

Uh oh!

Conversation

Mateusz Burzyński (Andarist) commented Aug 21, 2024

Uh oh!

Mateusz Burzyński (Andarist)Aug 21, 2024

Choose a reason for hiding this comment

Uh oh!

Mateusz Burzyński (Andarist)Aug 21, 2024

Choose a reason for hiding this comment

Uh oh!

Jake Bailey (jakebailey) commented Aug 21, 2024

Uh oh!

TypeScript Bot (typescript-bot) commented Aug 21, 2024• edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

TypeScript Bot (typescript-bot) commented Aug 21, 2024

Uh oh!

TypeScript Bot (typescript-bot) commented Aug 21, 2024

Uh oh!

TypeScript Bot (typescript-bot) commented Aug 21, 2024

tsc

Uh oh!

TypeScript Bot (typescript-bot) commented Aug 21, 2024

Uh oh!

Jake Bailey (jakebailey) commented Aug 22, 2024

Uh oh!

TypeScript Bot (typescript-bot) commented Aug 22, 2024• edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

TypeScript Bot (typescript-bot) commented Aug 22, 2024• edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

Daniel Rosenwasser (DanielRosenwasser) commented Aug 22, 2024

Uh oh!

TypeScript Bot (typescript-bot) commented Aug 22, 2024• edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

TypeScript Bot (typescript-bot) commented Aug 22, 2024

Uh oh!

TypeScript Bot (typescript-bot) commented Mar 24, 2026

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

5 participants

TypeScript Bot (typescript-bot) commented Aug 21, 2024•
edited
Loading

TypeScript Bot (typescript-bot) commented Aug 22, 2024•
edited
Loading

TypeScript Bot (typescript-bot) commented Aug 22, 2024•
edited
Loading

TypeScript Bot (typescript-bot) commented Aug 22, 2024•
edited
Loading