Widen boolean literals when contextual type is full boolean type

[Anders Hejlsberg (ahejlsberg)]

Fixes#48363.

Also fixes issue mentioned here because boolean literals contextually typed by boolean are now widened.

[Anders Hejlsberg (ahejlsberg)]

TypeScript Bot (@typescript-bot) test this
TypeScript Bot (@typescript-bot) user test this inline
TypeScript Bot (@typescript-bot) run dt
TypeScript Bot (@typescript-bot) perf test faster

[TypeScript Bot (typescript-bot)]

Heya Anders Hejlsberg (@ahejlsberg), I've started to run the extended test suite on this PR at b54de54. You can monitor the build here.

[TypeScript Bot (typescript-bot)]

Heya Anders Hejlsberg (@ahejlsberg), I've started to run the abridged perf test suite on this PR at b54de54. You can monitor the build here.

Update: The results are in!

[TypeScript Bot (typescript-bot)]

Heya Anders Hejlsberg (@ahejlsberg), I've started to run the parallelized Definitely Typed test suite on this PR at b54de54. You can monitor the build here.

[TypeScript Bot (typescript-bot)]

Heya Anders Hejlsberg (@ahejlsberg), I've started to run the diff-based community code test suite on this PR at b54de54. You can monitor the build here.

Update: The results are in!

[TypeScript Bot (typescript-bot)]

Anders Hejlsberg (@ahejlsberg)
The results of the perf run you requested are in!

Here they are:

Comparison Report - main..48368

Metric	main	48368	Delta	Best	Worst
Angular - node (v14.15.1, x64)
Memory used	332,421k (± 0.00%)	332,455k (± 0.00%)	+34k (+ 0.01%)	332,422k	332,499k
Parse Time	2.01s (± 0.38%)	2.04s (± 0.36%)	+0.02s (+ 1.14%)	2.02s	2.06s
Bind Time	0.86s (± 0.43%)	0.87s (± 0.51%)	+0.00s (+ 0.35%)	0.86s	0.88s
Check Time	5.58s (± 0.55%)	5.60s (± 0.45%)	+0.02s (+ 0.45%)	5.53s	5.65s
Emit Time	6.28s (± 0.71%)	6.32s (± 0.80%)	+0.04s (+ 0.59%)	6.21s	6.45s
Total Time	14.74s (± 0.42%)	14.82s (± 0.47%)	+0.09s (+ 0.58%)	14.70s	15.01s
Compiler-Unions - node (v14.15.1, x64)
Memory used	192,805k (± 0.59%)	193,835k (± 0.60%)	+1,030k (+ 0.53%)	191,837k	195,140k
Parse Time	0.85s (± 0.40%)	0.85s (± 0.56%)	0.00s ( 0.00%)	0.84s	0.86s
Bind Time	0.56s (± 0.71%)	0.56s (± 0.93%)	0.00s ( 0.00%)	0.55s	0.57s
Check Time	7.48s (± 0.74%)	7.43s (± 0.37%)	-0.05s (- 0.70%)	7.36s	7.48s
Emit Time	2.51s (± 1.50%)	2.48s (± 0.54%)	-0.03s (- 1.35%)	2.45s	2.51s
Total Time	11.41s (± 0.65%)	11.32s (± 0.28%)	-0.09s (- 0.78%)	11.24s	11.38s
Monaco - node (v14.15.1, x64)
Memory used	325,401k (± 0.01%)	325,363k (± 0.00%)	-38k (- 0.01%)	325,340k	325,393k
Parse Time	1.58s (± 0.62%)	1.58s (± 0.61%)	-0.00s (- 0.19%)	1.55s	1.59s
Bind Time	0.78s (± 0.85%)	0.77s (± 0.47%)	-0.01s (- 0.77%)	0.77s	0.78s
Check Time	5.49s (± 0.45%)	5.46s (± 0.40%)	-0.03s (- 0.47%)	5.41s	5.51s
Emit Time	3.33s (± 0.78%)	3.29s (± 0.55%)	-0.04s (- 1.11%)	3.26s	3.34s
Total Time	11.17s (± 0.37%)	11.10s (± 0.33%)	-0.07s (- 0.64%)	11.01s	11.18s
TFS - node (v14.15.1, x64)
Memory used	288,859k (± 0.01%)	288,857k (± 0.01%)	-2k (- 0.00%)	288,801k	288,904k
Parse Time	1.34s (± 1.22%)	1.32s (± 0.96%)	-0.02s (- 1.64%)	1.29s	1.35s
Bind Time	0.73s (± 1.04%)	0.73s (± 0.71%)	-0.00s (- 0.54%)	0.72s	0.74s
Check Time	5.11s (± 0.55%)	5.11s (± 0.35%)	+0.00s (+ 0.04%)	5.08s	5.16s
Emit Time	3.51s (± 2.21%)	3.52s (± 1.89%)	+0.01s (+ 0.26%)	3.38s	3.63s
Total Time	10.69s (± 0.86%)	10.68s (± 0.77%)	-0.02s (- 0.15%)	10.49s	10.83s
material-ui - node (v14.15.1, x64)
Memory used	453,485k (± 0.08%)	453,590k (± 0.07%)	+104k (+ 0.02%)	452,376k	453,798k
Parse Time	1.86s (± 0.54%)	1.85s (± 0.33%)	-0.02s (- 0.86%)	1.84s	1.86s
Bind Time	0.71s (± 0.97%)	0.70s (± 1.18%)	-0.00s (- 0.00%)	0.69s	0.72s
Check Time	20.97s (± 0.59%)	20.49s (± 0.84%)	-0.48s (- 2.28%)	20.25s	20.93s
Emit Time	0.00s (± 0.00%)	0.00s (± 0.00%)	0.00s ( NaN%)	0.00s	0.00s
Total Time	23.55s (± 0.55%)	23.05s (± 0.76%)	-0.50s (- 2.11%)	22.78s	23.48s
xstate - node (v14.15.1, x64)
Memory used	535,220k (± 0.00%)	535,233k (± 0.01%)	+13k (+ 0.00%)	535,184k	535,298k
Parse Time	2.58s (± 0.61%)	2.57s (± 0.43%)	-0.01s (- 0.35%)	2.55s	2.59s
Bind Time	1.15s (± 1.24%)	1.13s (± 0.77%)	-0.02s (- 1.48%)	1.12s	1.15s
Check Time	1.51s (± 0.65%)	1.51s (± 0.73%)	+0.00s (+ 0.13%)	1.49s	1.54s
Emit Time	0.07s (± 4.13%)	0.07s (± 0.00%)	-0.00s (- 2.78%)	0.07s	0.07s
Total Time	5.31s (± 0.27%)	5.29s (± 0.25%)	-0.03s (- 0.49%)	5.26s	5.32s

System

Machine Name	ts-ci-ubuntu
Platform	linux 4.4.0-210-generic
Architecture	x64
Available Memory	16 GB
Available Memory	4 GB
CPUs	4 × Intel(R) Core(TM) i7-4770 CPU @ 3.40GHz

Hosts

• node (v14.15.1, x64)

Scenarios

• Angular - node (v14.15.1, x64)
• Compiler-Unions - node (v14.15.1, x64)
• Monaco - node (v14.15.1, x64)
• TFS - node (v14.15.1, x64)
• material-ui - node (v14.15.1, x64)
• xstate - node (v14.15.1, x64)

Benchmark	Name	Iterations
Current	48368	10
Baseline	main	10

Developer Information:

Download Benchmark

[TypeScript Bot (typescript-bot)]

Anders Hejlsberg (@ahejlsberg)
Great news! no new errors were found between main..refs/pull/48368/merge

[Anders Hejlsberg (ahejlsberg)]

Apart from lots of false and true types changing to boolean (as expected), there are really only two notable changes in the baselines. One is in excessPropertyCheckWithMultipleDiscriminants.ts, simplified to:

letx: {kind: false,x: number}|{kind: boolean,x: string}={kind: false,x: 42};

Previously this would match against the first variant and succeed. It now matches against the second variant (because the kind: false in the object literal is given type boolean as it is contextually typed by boolean). It's possible to get the old behavior by writing kind: false as const. I think this change is acceptable.

The other change is in inferFromGenericFunctionReturnTypes3.ts:

declarefunctionfoldLeft<U>(z: U,f: (acc: U,t: boolean)=>U): U;letres: boolean=foldLeft(true,(acc,t)=>acc&&t);// Was error, now ok

Previously the true literal wasn't widened and therefore acc and t would have type true, which doesn't work out. We now widen and the example works as expected.

[Anders Hejlsberg (ahejlsberg)]

TypeScript Bot (@typescript-bot) run dt

[TypeScript Bot (typescript-bot)]

Heya Anders Hejlsberg (@ahejlsberg), I've started to run the parallelized Definitely Typed test suite on this PR at af5bd9e. You can monitor the build here.

[Bruce Pascoe (fatcerberus)]

Someone brought this up in Discord:

letfoo: boolean=false;

With the change in this PR, does this no longer narrow-on-assignment to false?

[Anders Hejlsberg (ahejlsberg)]

Several Definitely Typed packages are adversely affected. Closing this PR and putting up #48380 with less impact.