fix receiver on calls of imported and exported functions

[Klaus Meinhardt (ajafff)]

fixes: #35420

Ron Buckton (@rbuckton) for review

[Jordan Harband (ljharb)]

I see a lot of fixture changes but none of the test cases from #35420 added - specifically, a this-sensitive exported function being both called bare, and also .called with an explicit receiver.

[Klaus Meinhardt (ajafff)]

Jordan Harband (@ljharb) the vast majority of the tests only verifies the correct transpilation result by comparing with the baselines. You can see from the changed baselines that imported functions are now called indirectly, which fixes the issue with the receiver.
IMO it's not really necessary to verify the runtime behavior with a test that evaluates the transpiled code. However, if you or someone else has a strong opinion on that, I'll add a test that does it.

[Jordan Harband (ljharb)]

It's certainly not up to me, but the important thing is the correct runtime semantics of the code, not the specifics of what gets output, so it seems like those tests would be the important thing :-)

[Jordan Harband (ljharb)]

Nathan Shively-Sanders (@sandersn) 3.8.1 has come and gone (#35420 is listed as in that milestone), and this important bugfix hasn't gone out yet. any update?

[Nathan Shively-Sanders (sandersn)]

Ron Buckton (@rbuckton) This is now scheduled for 4.0. Can you review this fix during the beta?

[Ron Buckton (rbuckton)]

The substitution for the identifier should be invoked by the printer, you shouldn't have to call it directly.

[Jordan Harband (ljharb)]

Ping again - i know this needs a rebase, but it’s a pretty critical bug fix.

[Ron Buckton (rbuckton)]

I've rebased this PR and made a few small updates. We may also want to consider doing the same thing for namespace exports:

namespaceX{exportfunctionfoo(this: unknown){console.log(this===X);}}namespaceX{foo();// compiles to `X.foo()`, so would print `true`}

Daniel Rosenwasser (@DanielRosenwasser) we should discuss this emit change at the next design meeting.

[Ron Buckton (rbuckton)]

I've updated this to support namespace exports as well. as its essentially a bug there too and I'd like to see how this change effects compiler performance (because we call a lot of functions on the ts namespace).

[Ron Buckton (rbuckton)]

TypeScript Bot (@typescript-bot) perf test

[TypeScript Bot (typescript-bot)]

Heya Ron Buckton (@rbuckton), I've started to run the perf test suite on this PR at b788c09. You can monitor the build here.

Update: The results are in!

[TypeScript Bot (typescript-bot)]

Ron Buckton (@rbuckton)
The results of the perf run you requested are in!

Here they are:

Comparison Report - master..35877

Metric	master	35877	Delta	Best	Worst
Angular - node (v10.16.3, x64)
Memory used	345,618k (± 0.02%)	345,624k (± 0.02%)	+6k (+ 0.00%)	345,493k	345,769k
Parse Time	1.92s (± 0.40%)	1.91s (± 0.35%)	-0.01s (- 0.31%)	1.90s	1.93s
Bind Time	0.82s (± 0.60%)	0.82s (± 0.41%)	-0.00s (- 0.12%)	0.82s	0.83s
Check Time	4.97s (± 0.41%)	4.94s (± 0.52%)	-0.04s (- 0.72%)	4.88s	4.99s
Emit Time	5.28s (± 0.97%)	5.55s (± 0.59%)	+0.27s (+ 5.11%)	5.50s	5.65s
Total Time	12.99s (± 0.38%)	13.22s (± 0.39%)	+0.23s (+ 1.77%)	13.14s	13.36s
Compiler-Unions - node (v10.16.3, x64)
Memory used	214,523k (± 0.06%)	215,489k (± 0.16%)	+966k (+ 0.45%)	214,155k	215,876k
Parse Time	0.78s (± 0.67%)	0.77s (± 0.67%)	-0.01s (- 1.03%)	0.76s	0.78s
Bind Time	0.50s (± 1.34%)	0.50s (± 0.98%)	+0.01s (+ 1.41%)	0.50s	0.52s
Check Time	10.65s (± 0.40%)	10.66s (± 0.69%)	+0.01s (+ 0.05%)	10.52s	10.80s
Emit Time	2.35s (± 1.67%)	2.51s (± 1.37%)	+0.17s (+ 7.16%)	2.46s	2.58s
Total Time	14.27s (± 0.38%)	14.44s (± 0.64%)	+0.17s (+ 1.19%)	14.27s	14.64s
Monaco - node (v10.16.3, x64)
Memory used	355,251k (± 0.02%)	355,289k (± 0.03%)	+38k (+ 0.01%)	355,060k	355,589k
Parse Time	1.55s (± 0.58%)	1.55s (± 0.57%)	-0.00s (- 0.06%)	1.53s	1.57s
Bind Time	0.72s (± 0.55%)	0.73s (± 1.30%)	+0.01s (+ 0.83%)	0.71s	0.75s
Check Time	5.09s (± 0.57%)	5.11s (± 0.66%)	+0.02s (+ 0.39%)	5.04s	5.20s
Emit Time	2.79s (± 0.64%)	2.87s (± 1.12%)	+0.08s (+ 3.05%)	2.83s	2.97s
Total Time	10.15s (± 0.32%)	10.26s (± 0.41%)	+0.11s (+ 1.08%)	10.19s	10.37s
TFS - node (v10.16.3, x64)
Memory used	308,114k (± 0.02%)	308,161k (± 0.01%)	+47k (+ 0.02%)	308,123k	308,249k
Parse Time	1.20s (± 0.63%)	1.20s (± 0.63%)	+0.00s (+ 0.08%)	1.19s	1.22s
Bind Time	0.68s (± 0.49%)	0.68s (± 0.66%)	-0.00s (- 0.29%)	0.67s	0.69s
Check Time	4.60s (± 0.57%)	4.59s (± 0.63%)	-0.01s (- 0.22%)	4.54s	4.65s
Emit Time	2.94s (± 0.83%)	3.01s (± 0.51%)	+0.07s (+ 2.28%)	2.97s	3.04s
Total Time	9.42s (± 0.38%)	9.48s (± 0.31%)	+0.05s (+ 0.58%)	9.42s	9.56s
material-ui - node (v10.16.3, x64)
Memory used	495,626k (± 0.02%)	495,544k (± 0.02%)	-81k (- 0.02%)	495,311k	495,683k
Parse Time	1.98s (± 0.39%)	1.97s (± 0.53%)	-0.01s (- 0.30%)	1.95s	2.00s
Bind Time	0.65s (± 0.95%)	0.65s (± 0.72%)	-0.00s (- 0.15%)	0.64s	0.66s
Check Time	13.95s (± 0.92%)	13.93s (± 0.59%)	-0.03s (- 0.19%)	13.77s	14.11s
Emit Time	0.00s (± 0.00%)	0.00s (± 0.00%)	0.00s ( NaN%)	0.00s	0.00s
Total Time	16.59s (± 0.77%)	16.55s (± 0.51%)	-0.04s (- 0.22%)	16.39s	16.74s
Angular - node (v12.1.0, x64)
Memory used	323,177k (± 0.03%)	323,238k (± 0.03%)	+61k (+ 0.02%)	323,024k	323,441k
Parse Time	1.90s (± 0.66%)	1.90s (± 0.87%)	+0.00s (+ 0.26%)	1.86s	1.93s
Bind Time	0.80s (± 0.86%)	0.80s (± 0.60%)	+0.00s (+ 0.13%)	0.79s	0.81s
Check Time	4.85s (± 0.63%)	4.87s (± 1.32%)	+0.02s (+ 0.39%)	4.76s	5.10s
Emit Time	5.40s (± 0.54%)	5.75s (± 0.43%)	+0.36s (+ 6.62%)	5.70s	5.81s
Total Time	12.94s (± 0.47%)	13.32s (± 0.67%)	+0.38s (+ 2.95%)	13.12s	13.60s
Compiler-Unions - node (v12.1.0, x64)
Memory used	199,902k (± 0.05%)	201,074k (± 0.06%)	+1,173k (+ 0.59%)	200,776k	201,288k
Parse Time	0.77s (± 0.67%)	0.77s (± 0.80%)	-0.00s (- 0.26%)	0.76s	0.78s
Bind Time	0.50s (± 0.60%)	0.50s (± 0.44%)	+0.00s (+ 0.60%)	0.50s	0.51s
Check Time	9.72s (± 1.02%)	9.70s (± 1.16%)	-0.02s (- 0.16%)	9.55s	10.07s
Emit Time	2.34s (± 1.01%)	2.56s (± 0.96%)	+0.22s (+ 9.26%)	2.51s	2.62s
Total Time	13.33s (± 0.73%)	13.54s (± 0.87%)	+0.21s (+ 1.57%)	13.37s	13.93s
Monaco - node (v12.1.0, x64)
Memory used	337,476k (± 0.01%)	337,500k (± 0.02%)	+25k (+ 0.01%)	337,357k	337,652k
Parse Time	1.53s (± 0.87%)	1.53s (± 0.75%)	+0.00s (+ 0.20%)	1.50s	1.55s
Bind Time	0.70s (± 0.63%)	0.70s (± 0.70%)	+0.00s (+ 0.28%)	0.69s	0.71s
Check Time	4.91s (± 0.48%)	4.91s (± 0.45%)	0.00s ( 0.00%)	4.85s	4.94s
Emit Time	2.86s (± 0.57%)	2.91s (± 0.67%)	+0.05s (+ 1.64%)	2.87s	2.95s
Total Time	10.00s (± 0.32%)	10.06s (± 0.33%)	+0.06s (+ 0.58%)	10.00s	10.13s
TFS - node (v12.1.0, x64)
Memory used	292,351k (± 0.01%)	292,332k (± 0.02%)	-19k (- 0.01%)	292,224k	292,454k
Parse Time	1.22s (± 0.69%)	1.23s (± 0.50%)	+0.01s (+ 0.66%)	1.21s	1.24s
Bind Time	0.65s (± 0.34%)	0.66s (± 0.88%)	+0.01s (+ 0.92%)	0.65s	0.67s
Check Time	4.50s (± 0.40%)	4.51s (± 0.34%)	+0.01s (+ 0.18%)	4.49s	4.55s
Emit Time	2.96s (± 0.91%)	3.02s (± 0.91%)	+0.06s (+ 2.20%)	2.95s	3.09s
Total Time	9.33s (± 0.37%)	9.42s (± 0.43%)	+0.09s (+ 0.92%)	9.33s	9.51s
material-ui - node (v12.1.0, x64)
Memory used	472,776k (± 0.01%)	472,809k (± 0.01%)	+33k (+ 0.01%)	472,724k	472,892k
Parse Time	1.99s (± 0.51%)	2.00s (± 0.60%)	+0.01s (+ 0.70%)	1.97s	2.03s
Bind Time	0.64s (± 0.35%)	0.64s (± 1.18%)	+0.00s (+ 0.47%)	0.62s	0.66s
Check Time	12.49s (± 0.70%)	12.49s (± 0.72%)	-0.00s (- 0.04%)	12.33s	12.72s
Emit Time	0.00s (± 0.00%)	0.00s (± 0.00%)	0.00s ( NaN%)	0.00s	0.00s
Total Time	15.12s (± 0.58%)	15.13s (± 0.59%)	+0.01s (+ 0.07%)	14.98s	15.35s
Angular - node (v14.15.1, x64)
Memory used	321,828k (± 0.01%)	321,910k (± 0.01%)	+82k (+ 0.03%)	321,851k	321,993k
Parse Time	1.90s (± 0.55%)	1.90s (± 0.54%)	+0.00s (+ 0.05%)	1.88s	1.92s
Bind Time	0.84s (± 0.43%)	0.84s (± 0.81%)	+0.00s (+ 0.12%)	0.84s	0.87s
Check Time	4.83s (± 0.38%)	4.86s (± 0.43%)	+0.03s (+ 0.60%)	4.83s	4.91s
Emit Time	5.48s (± 0.77%)	5.83s (± 0.51%)	+0.35s (+ 6.46%)	5.79s	5.93s
Total Time	13.05s (± 0.49%)	13.44s (± 0.39%)	+0.39s (+ 2.96%)	13.35s	13.63s
Compiler-Unions - node (v14.15.1, x64)
Memory used	199,870k (± 0.38%)	200,674k (± 0.10%)	+804k (+ 0.40%)	200,098k	200,928k
Parse Time	0.79s (± 0.62%)	0.79s (± 0.43%)	-0.00s (- 0.13%)	0.79s	0.80s
Bind Time	0.53s (± 0.42%)	0.53s (± 0.56%)	-0.00s (- 0.56%)	0.52s	0.53s
Check Time	9.68s (± 0.56%)	9.68s (± 0.81%)	-0.00s (- 0.01%)	9.54s	9.93s
Emit Time	2.34s (± 1.04%)	2.54s (± 1.48%)	+0.20s (+ 8.49%)	2.48s	2.63s
Total Time	13.35s (± 0.48%)	13.55s (± 0.53%)	+0.20s (+ 1.48%)	13.39s	13.76s
Monaco - node (v14.15.1, x64)
Memory used	336,849k (± 0.01%)	336,929k (± 0.00%)	+80k (+ 0.02%)	336,901k	336,961k
Parse Time	1.55s (± 0.58%)	1.55s (± 0.64%)	+0.00s (+ 0.13%)	1.53s	1.57s
Bind Time	0.73s (± 0.71%)	0.73s (± 0.64%)	+0.00s (+ 0.14%)	0.72s	0.74s
Check Time	4.85s (± 0.61%)	4.85s (± 0.48%)	-0.00s (- 0.08%)	4.80s	4.91s
Emit Time	2.91s (± 0.41%)	3.00s (± 0.86%)	+0.08s (+ 2.85%)	2.96s	3.08s
Total Time	10.05s (± 0.39%)	10.12s (± 0.45%)	+0.08s (+ 0.78%)	10.06s	10.28s
TFS - node (v14.15.1, x64)
Memory used	291,545k (± 0.00%)	291,547k (± 0.01%)	+2k (+ 0.00%)	291,518k	291,588k
Parse Time	1.24s (± 0.99%)	1.24s (± 1.23%)	+0.00s (+ 0.24%)	1.22s	1.29s
Bind Time	0.69s (± 0.64%)	0.69s (± 0.84%)	+0.00s (+ 0.14%)	0.68s	0.71s
Check Time	4.50s (± 0.62%)	4.50s (± 0.41%)	+0.01s (+ 0.16%)	4.45s	4.54s
Emit Time	3.03s (± 0.46%)	3.13s (± 0.70%)	+0.09s (+ 3.13%)	3.08s	3.18s
Total Time	9.46s (± 0.30%)	9.57s (± 0.37%)	+0.11s (+ 1.17%)	9.51s	9.66s
material-ui - node (v14.15.1, x64)
Memory used	471,401k (± 0.06%)	471,503k (± 0.01%)	+102k (+ 0.02%)	471,437k	471,554k
Parse Time	2.06s (± 0.78%)	2.05s (± 0.71%)	-0.00s (- 0.19%)	2.02s	2.08s
Bind Time	0.70s (± 0.74%)	0.70s (± 0.64%)	-0.00s (- 0.43%)	0.69s	0.71s
Check Time	12.63s (± 0.83%)	12.55s (± 1.53%)	-0.07s (- 0.59%)	12.38s	13.31s
Emit Time	0.00s (± 0.00%)	0.00s (± 0.00%)	0.00s ( NaN%)	0.00s	0.00s
Total Time	15.38s (± 0.75%)	15.30s (± 1.31%)	-0.08s (- 0.54%)	15.11s	16.09s

System

Machine Name	ts-ci-ubuntu
Platform	linux 4.4.0-198-generic
Architecture	x64
Available Memory	16 GB
Available Memory	7 GB
CPUs	4 × Intel(R) Core(TM) i7-4770 CPU @ 3.40GHz

Hosts

• node (v10.16.3, x64)
• node (v12.1.0, x64)
• node (v14.15.1, x64)

Scenarios

• Angular - node (v10.16.3, x64)
• Angular - node (v12.1.0, x64)
• Angular - node (v14.15.1, x64)
• Compiler-Unions - node (v10.16.3, x64)
• Compiler-Unions - node (v12.1.0, x64)
• Compiler-Unions - node (v14.15.1, x64)
• Monaco - node (v10.16.3, x64)
• Monaco - node (v12.1.0, x64)
• Monaco - node (v14.15.1, x64)
• TFS - node (v10.16.3, x64)
• TFS - node (v12.1.0, x64)
• TFS - node (v14.15.1, x64)
• material-ui - node (v10.16.3, x64)
• material-ui - node (v12.1.0, x64)
• material-ui - node (v14.15.1, x64)

Benchmark	Name	Iterations
Current	35877	10
Baseline	master	10

[Ron Buckton (rbuckton)]

TypeScript Bot (@typescript-bot) pack this

[TypeScript Bot (typescript-bot)]

Heya Ron Buckton (@rbuckton), I've started to run the tarball bundle task on this PR at b788c09. You can monitor the build here.

[TypeScript Bot (typescript-bot)]

Hey Ron Buckton (@rbuckton), 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/95810/artifacts?artifactName=tgz&fileId=F8AD257B669FDE703D7FF39DE9A54CE25327D7236435DF39341C2F15BD09C00802&fileName=/typescript-4.3.0-insiders.20210212.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@4.3.0-pr-35877-13".;

[Ron Buckton (rbuckton)]

Per discussion with Ryan Cavanaugh (@RyanCavanaugh) today, I'm going to revert the change for namespaces for the time being.

[Ron Buckton (rbuckton)]

Hmm. build is failing for a unit test that doesn't exist anymore...

Ah, that's a typo in the externalModules.ts test.

[Ron Buckton (rbuckton)]

It looks like we will need to revert this change until TS 4.4 while we investigate the significant performance regression this causes during emit. In the worst case, we may need to land this behind a flag to opt-in to the change, though ideally I can find a mitigation for the performance regression so that we can land this without a flag.

[Jordan Harband (ljharb)]

Reverting this seems like it will restore the active breakage on a number of my shims like Promise.allSettled, where the module.exports function explicitly relies on having an undefined receiver.

Why does a performance regression matter when compared to runtime breakage?

[Jordan Harband (ljharb)]

Ron Buckton (@rbuckton) ping, TS is severely broken without this fix; any update on restoring it?

[Daniel Rosenwasser (DanielRosenwasser)]

We re-re-discussed this at #44555.

We've always tried to balance reasonable/canonical-looking emit and general UX with exact semantics of the spec. We're not ready to sacrifice that and also impose an emit overhead on everyone, especially when performance is something we're really working to improve. If we can find some optimizations that make this "free-ish", we'd feel better about this feature; otherwise, we'll need to put it behind a flag (like --downlevelIteration, --useDefinedForClassFields, etc.).

Why does a performance regression matter when compared to runtime breakage?

The bigger risk here is actually imposing runtime breakage on existing users who accidentally came to rely on the existing behavior. Given that, the safer thing might actually be to put it behind a flag.

[Jordan Harband (ljharb)]

Daniel Rosenwasser (@DanielRosenwasser) so you’re worried about users who expect that, defying the way it works everywhere else in the JS exosystem, an exported function’s receiver will be the module namespace object when called as a bare function? I’d love to understand what code is relying on this behavior.

[Ryan Cavanaugh (RyanCavanaugh)]

so you’re worried about users who expect that, defying the way it works everywhere else in the JS exosystem

Yes. Improved spec compliance has caused problems before, e.g. when we made exports immutable/nonenumerable (jasmine/jasmine#1817, #38568).

[Jordan Harband (ljharb)]

Alright - can it at least be landed behind a flag, so i can tell users how to unbreak their applications?

[Ryan Cavanaugh (RyanCavanaugh)]

We're definitely going to figure something out, just trying to work out what exactly that is

[Ron Buckton (rbuckton)]

We were able to re-land this in #44624 in a way that does not significantly impact emit performance.

[Jordan Harband (ljharb)]

That's amazing, great to hear!