Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

n-api: cache Symbol.hasInstance #12246

Closed

Conversation

gabrielschulhof
Copy link
Contributor

@gabrielschulhof gabrielschulhof commented Apr 5, 2017

This improves the performance of napi_instanceof() by retrieving
Symbol.hasInstance from the global object once and then storing a
persistent reference to it in the env.

Re nodejs/abi-stable-node#209
Re nodejs/abi-stable-node#214

Checklist
  • make -j4 test (UNIX), or vcbuild test (Windows) passes
  • commit message follows commit guidelines
Affected core subsystem(s)

N-API

This improves the performance of napi_instanceof() by retrieving
Symbol.hasInstance from the global object once and then storing a
persistent reference to it in the env.

Re nodejs/abi-stable-node#209
Re nodejs/abi-stable-node#214
@nodejs-github-bot nodejs-github-bot added the c++ Issues and PRs that require attention from people who are familiar with C++. label Apr 5, 2017
@mscdex mscdex added the node-api Issues and PRs related to the Node-API. label Apr 5, 2017
}
v8::Isolate* isolate;
v8::Persistent<v8::Value> last_exception;
v8::Persistent<v8::Value> has_instance;
Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

do we need a copy in every env ?

Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

ie could we share one copy across all envs ?

Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

That brings back the question of what prompts the creation of a new env? Currently it is the loading of a module. But should it be a new isolate?

Also, how likely is this to be a large waste of memory? I mean, are we going to have tons of native modules loaded all at once?

I guess when there really start to appear multiple isolates (or whatever keys meaning "this JS world and not the other" because "this v8::Local<v8::Value> is only valid in this world") then we can re-examine our choice of one env per module, because the problem will have been solved upstream and we need only follow suit. We might then introduce a std::map keyed to whatever the engine then provides, or we may have a better mechanism for tacking data onto such a key and we'll once again cast.

... and we can do all this without breaking any ABI, since it's purely within our implementation.

Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I just wondered if the exact same code, except that having has_instance be static would work. The cache would be created the first time an env was created. The only concern was whether this could happen in parallel, but given Node.js generally single thread nature that may not be an issue.

Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

@mhdawson because eventually the env will be keyed to some kind of context, and we are preparing ourselves for that. Otherwise last_exception could also be static - and it used to be, and the env was precisely equal to the isolate. The problem is we can't currently tack onto the isolate because there is no reliable mechanism for that, so we're instantiating an environment per module.

Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

So, yes, it would work perfectly, and it's only our desire to be proactive that prompted us to implement this solution.

Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

ok, got it.

@TimothyGu
Copy link
Member

TimothyGu commented Apr 7, 2017

Alternatively, we can convince V8 to add a v8::Symbol::GetHasInstance(v8::Isolate*) like the ones they already have for some other symbols.

@gabrielschulhof
Copy link
Contributor Author

@TimothyGu why not go for the whole nine yards and ask them to expose the instanceof operator itself?

Copy link
Member

@mhdawson mhdawson left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

LGTM

@mhdawson
Copy link
Member

mhdawson commented Apr 7, 2017

Discussion of getting changes into v8 is a good follow on, but don't think it should block landing this PR.

CI run: https://ci.nodejs.org/job/node-test-pull-request/7270/

@mhdawson
Copy link
Member

mhdawson commented Apr 7, 2017

CI run was green landing.

@mhdawson
Copy link
Member

mhdawson commented Apr 7, 2017

landed as 8fbace1

@mhdawson mhdawson closed this Apr 7, 2017
mhdawson pushed a commit that referenced this pull request Apr 7, 2017
This improves the performance of napi_instanceof() by retrieving
Symbol.hasInstance from the global object once and then storing a
persistent reference to it in the env.

PR-URL: #12246
Reviewed-By: James M Snell <jasnell@gmail.com>
Reviewed-By: Michael Dawson <michael_dawson@ca.ibm.com>
@TimothyGu
Copy link
Member

@gabrielschulhof

why not go for the whole nine yards and ask them to expose the instanceof operator itself?

Indeed it's a much better idea. @hashseed @fhinkel, is Function::HasInstance(Local<Value>) something you would be interested in implementing?

@TimothyGu TimothyGu mentioned this pull request Apr 8, 2017
3 tasks
@mscdex
Copy link
Contributor

mscdex commented Apr 8, 2017

@gabrielschulhof @mhdawson

It appears this PR may be making a n-api test fail on Windows. Here is a CI run of current master: https://ci.nodejs.org/job/node-test-binary-windows/7558/

not ok 332 addons-napi/test_instanceof/test
  ---
  duration_ms: 0.193
  severity: fail
  stack: |-
    
    assert.js:82
      throw new assert.AssertionError({
      ^
    AssertionError: true === undefined
        at assertTrue (c:\workspace\node-test-binary-windows\RUN_SUBSET\3\VS_VERSION\vs2015\label\win2016\test\addons-napi\test_instanceof\test.js:16:17)
        at eval (eval at testFile (c:\workspace\node-test-binary-windows\RUN_SUBSET\3\VS_VERSION\vs2015\label\win2016\test\addons-napi\test_instanceof\test.js:38:3), <anonymous>:29:1)
        at testFile (c:\workspace\node-test-binary-windows\RUN_SUBSET\3\VS_VERSION\vs2015\label\win2016\test\addons-napi\test_instanceof\test.js:38:3)
        at Object.<anonymous> (c:\workspace\node-test-binary-windows\RUN_SUBSET\3\VS_VERSION\vs2015\label\win2016\test\addons-napi\test_instanceof\test.js:42:1)
        at Module._compile (module.js:607:30)
        at Object.Module._extensions..js (module.js:618:10)
        at Module.load (module.js:516:32)
        at tryModuleLoad (module.js:466:12)
        at Function.Module._load (module.js:458:3)
        at Module.runMain (module.js:643:10)

Perhaps something was changed since the last CI run? All other platforms seem to be unaffected.

@Trott
Copy link
Member

Trott commented Apr 8, 2017

Perhaps something was changed since the last CI run?

@mscdex Looking for an n-api change that might have landed after CI was run for this one, the only candidate I'm seeing is 8460284. Maybe that change in conjunction with this one is the problem on Windows but either change by itself is fine? (Totally speculating here, but that's my best guess?)

@gabrielschulhof gabrielschulhof deleted the 209-cache-has-instance branch April 8, 2017 08:18
refack added a commit to refack/node that referenced this pull request Apr 8, 2017
@refack
Copy link
Contributor

refack commented Apr 8, 2017

solved regression by @gabrielschulhof in #12283

@hashseed
Copy link
Member

hashseed commented Apr 10, 2017

I'm considering adding v8::Function::HasInstance to the V8 API. What is napi_instanceof supposed to do? Is it exposing exactly what the instanceof operator does, including overriding behavior through @@hasInstance?

Is this what you need?

@addaleax
Copy link
Member

addaleax commented Apr 10, 2017

Is it exposing exactly what the instanceof operator does, including overriding behavior through @@hasInstance?

Yes, the options were either doing that or calling it something other than napi_instanceof. ;)

Is this what you need?

Yes, looks like it. Thank you! 💚

@hashseed
Copy link
Member

Alright. Will have to add some tests. It will not land before V8 6.0 though because we are cutting the 5.9 branch very soon.

@hashseed
Copy link
Member

I changed the API to v8::Value::InstanceOf so that the rhs can also be a non-v8::Function.

@hashseed
Copy link
Member

The changes for v8::Value::InstanceOf and v8::Symbol::GetHasInstance landed in V8 6.0.

Do we want to open an issue to track so that this can be used as soon as Node.js upgrades to V8 6.0?

@fhinkel
Copy link
Member

fhinkel commented Apr 13, 2017

@hashseed Usually we have an issue for anything related to a V8 upgrade. Might make sense to start that now for 60.

@hashseed
Copy link
Member

Just as a reminder, v8::Value::InstanceOf and v8::Symbol::GetHasInstance can now be used on master.

@hashseed
Copy link
Member

I'm working on a PR to use v8::Value::InstanceOf in napi.

gabrielschulhof pushed a commit to gabrielschulhof/node that referenced this pull request Apr 10, 2018
This improves the performance of napi_instanceof() by retrieving
Symbol.hasInstance from the global object once and then storing a
persistent reference to it in the env.

PR-URL: nodejs#12246
Reviewed-By: James M Snell <jasnell@gmail.com>
Reviewed-By: Michael Dawson <michael_dawson@ca.ibm.com>
MylesBorins pushed a commit that referenced this pull request Apr 16, 2018
This improves the performance of napi_instanceof() by retrieving
Symbol.hasInstance from the global object once and then storing a
persistent reference to it in the env.

Backport-PR-URL: #19447
PR-URL: #12246
Reviewed-By: James M Snell <jasnell@gmail.com>
Reviewed-By: Michael Dawson <michael_dawson@ca.ibm.com>
@MylesBorins MylesBorins mentioned this pull request Apr 16, 2018
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
c++ Issues and PRs that require attention from people who are familiar with C++. node-api Issues and PRs related to the Node-API.
Projects
None yet
Development

Successfully merging this pull request may close these issues.

None yet