Don't use reserved word 'volatile' for non-cached properties by rykov · Pull Request #957 · emberjs/ember.js

rykov · 2012-06-06T06:13:36Z

The verdict online seems mixed on whether volatile is a reserved word in Javascript, but Closure Compiler seems to think that it is and spew many errors when I try to use it as my minifier.

This commit replaces the volatile function with uncached.

wycats · 2012-06-06T07:29:08Z

Thanks for the pull request. I'm confused about why closure compiler thinks it's reserved. See http://es5.github.com/#x7.6.1.2 for a complete list of the reserved words in ES5.

travisbot · 2012-06-06T15:30:39Z

This pull request passes (merged fca01d8 into 96791e2).

rykov · 2012-06-06T16:28:57Z

Did some digging -- volatile is not reserved in ES5, but it is in ES3. Closure Compiler supports both and defaults to ES3. And you're right, in ES5 mode, Closure Compiler doesn't think it's reserved and finishes the compression. This does the trick in Rails 3.1+:

Closure::Compiler::DEFAULT_OPTIONS[:language_in] = 'ECMASCRIPT5'

Source: http://www.ecma-international.org/publications/standards/Ecma-262-arch.htm

wagenet · 2012-06-09T02:39:43Z

@wycats What's the verdict on this? Merge or no?

wagenet · 2012-06-09T02:40:13Z

If we don't merge, we should definitely document the solution somewhere.

rykov · 2012-06-09T04:33:25Z

I've submitted a pull request to closure-compiler gem to make ES5 the default here: documentcloud/closure-compiler#21

rykov · 2012-06-09T04:38:37Z

@wagenet To answer your question, no merge in favor of something like this added to ember-rails

if defined?(Closure::Compiler)
  Closure::Compiler::DEFAULT_OPTIONS[:language_in] = 'ECMASCRIPT5'
end

rykov · 2012-06-09T15:33:43Z

Should I go ahead with the ember-rails suggestion?

wagenet · 2012-06-09T15:40:17Z

@rykov, sounds good to me.

rykov · 2012-06-09T16:05:06Z

Submitted here: emberjs/ember-rails#61

Adds the `@ember/scheduler` package proposed by RFC 0957: - `render`, `layout`, `composite`, `next` and `idle` phase functions, each returning a promise that resolves according to the registered scheduling strategy - `registerStrategy`, for providing the scheduling strategy when defining the Application - `@ember/scheduler/strategy`, the default strategy implementation, which flushes the render/layout/composite phases in order via ordered requestAnimationFrame callbacks within a single frame, prior to paint The deprecations of @ember/runloop and RSVP described by the RFC are left to follow-up work; this is the additive API surface. Fix Safari flake and idle() starvation in the default strategy The "composite while composite is flushing" test asserted ordering across two independent channels: a setTimeout task scheduled during frame 1 versus frame 2's requestAnimationFrame callbacks. The HTML spec does not order pending timer tasks against the next rendering opportunity, and Safari 15.6 runs the next frame's rAF callbacks first. The test now anchors entirely to the rAF channel, using a raw requestAnimationFrame registered ahead of the rescheduled phase windows as the frame-2 boundary. idle() also armed requestIdleCallback without a timeout; fully-idle or backgrounded pages can starve rIC indefinitely, leaving the promise unresolvable. Cap the wait with { timeout: 500 }.

This is the all-optimizations branch as one commit on main f693f24. The branch had 3 merge commits with conflict resolutions in them, and a commit-by-commit rebase drops those resolutions. So this commit has the end state of the branch (c09efdc) merged with main. The old history is on the branch all-optimizations-pre-rebase. Changes that main needed, beyond the 8 conflicting files: - route.ts, router.ts: main's new runloop calls use the same microtask helpers as the rest of the spike (queueMicrotask, scheduleMethodOnce) - classic route manager: native Promise instead of RSVP - ArrayIterator#nextInto: follows main's next(), which reads the live array length - runloop shim: a typed overload for run(fn), so run returns the callback's type - rollup: rsvp stays a hidden dependency. Main removed the hidden entry and the spike removed the exposed one. - package.json: main's version, without backburner.js Checks: the type-check has the same 10 errors as c09efdc and no new ones. pnpm build passes. The browser test suite stops early in the debug render tree tests, on c09efdc and on this commit. Commits squashed from c09efdc: - a9f197d Implement RFC emberjs#957: Render Aware Scheduler Interface - ed3e2ab Skip unchanged {{#each}} item subtrees during updates - d9f1250 SPIKE: stack all dbmon-class optimizations to measure the ceiling - 2607466 SPIKE: delete legacy read-path support from _getProp - 5530b80 SPIKE: allocation hygiene in the updating VM (fps-neutral) - b7a4e38 Fix Safari flake and idle() starvation in the default strategy - e562c08 Adaptive tri-mode flush scheduler on clean main (scheduler lever only) - 3a3b9e2 Diet the flush machinery; keep macrotask leg alive under rAF starvation - a80bce4 Fix double-flushing and stand-down starvation in the chain leg - f8b4d0c End state: remove backburner from the render path entirely - 908f68b Drop imports orphaned by the legacy read-path removal - 7c0b543 Clock semantics: one render per tick, bounded settle rounds - 233200e Adopt Angular's microtask-window tick classifier - 1a0a834 Make the renderer clock the default @ember/scheduler strategy - c93641e Diet the per-set path: notify once per tick, one-hop tracked cells - d6be883 Fix the two hazards the per-set diet shipped with - 7e5e61e Remove the runloop and RSVP from the framework entirely - 739bf03 Sync lockfile and renamed-modules with the runloop/RSVP removal - 2b703ac Known-tag references and pooled tracking frames - 3e9fc3c Pool the per-tick transaction and updating VM - 5d5c603 Revert "Pool the per-tick transaction and updating VM" - 39c062e Remove the tracking-frame Tracker pool as well - 17ac4c3 Delete backburner; the runloop becomes a dependency-free compat shim - 5916d3e Restore trailing newline in package.json - a967e87 Restore _backburner compat stub and boot-deferred assignability - f69f089 Drain scheduled destroys synchronously when an engine instance is destroyed - 7077507 Report pending work through the _backburner stub; renderSettled resolves via a flush - 60f0fc9 Replace the _backburner stub with isRenderPending on @ember/renderer - c09efdc Report render settledness as edges, not a pollable level Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>

Adds the `@ember/scheduler` package proposed by RFC 0957: - `render`, `layout`, `composite`, `next` and `idle` phase functions, each returning a promise that resolves according to the registered scheduling strategy - `registerStrategy`, for providing the scheduling strategy when defining the Application - `@ember/scheduler/strategy`, the default strategy implementation, which flushes the render/layout/composite phases in order via ordered requestAnimationFrame callbacks within a single frame, prior to paint The deprecations of @ember/runloop and RSVP described by the RFC are left to follow-up work; this is the additive API surface. Fix Safari flake and idle() starvation in the default strategy The "composite while composite is flushing" test asserted ordering across two independent channels: a setTimeout task scheduled during frame 1 versus frame 2's requestAnimationFrame callbacks. The HTML spec does not order pending timer tasks against the next rendering opportunity, and Safari 15.6 runs the next frame's rAF callbacks first. The test now anchors entirely to the rAF channel, using a raw requestAnimationFrame registered ahead of the rescheduled phase windows as the frame-2 boundary. idle() also armed requestIdleCallback without a timeout; fully-idle or backgrounded pages can starve rIC indefinitely, leaving the promise unresolvable. Cap the wait with { timeout: 500 }.

rykov added 2 commits June 5, 2012 23:09

Use 'uncached' instead of reserved 'volatile'

ef28db4

Merge branch 'master' of git://github.com/emberjs/ember.js

fca01d8

wagenet closed this Jun 9, 2012

rykov mentioned this pull request Jun 9, 2012

Default to ES5 for Closure Compiler emberjs/ember-rails#61

Merged

jankopriva mentioned this pull request Jul 23, 2012

Another question to volatile() #1202

Closed

pangratz mentioned this pull request Sep 29, 2012

.volatile causes clojure compiler to fail #1421

Closed

krumpi mentioned this pull request Dec 17, 2012

Put old ES3 reserved words behind a flag mozilla/rhino#93

Closed

Navigation Menu

Repository navigation

Don't use reserved word 'volatile' for non-cached properties - #957

Don't use reserved word 'volatile' for non-cached properties#957
rykov wants to merge 2 commits into
emberjs:masterfrom
rykov:master

rykov commented Jun 6, 2012

Uh oh!

wycats commented Jun 6, 2012

Uh oh!

travisbot commented Jun 6, 2012

Uh oh!

rykov commented Jun 6, 2012

Uh oh!

wagenet commented Jun 9, 2012

Uh oh!

wagenet commented Jun 9, 2012

Uh oh!

rykov commented Jun 9, 2012

Uh oh!

rykov commented Jun 9, 2012

Uh oh!

rykov commented Jun 9, 2012

Uh oh!

wagenet commented Jun 9, 2012

Uh oh!

rykov commented Jun 9, 2012

Uh oh!

Reviewers

Uh oh!

Assignees

Labels

Projects

Milestone

Development

Uh oh!

4 participants

Conversation

rykov commented Jun 6, 2012

Uh oh!

wycats commented Jun 6, 2012

Uh oh!

travisbot commented Jun 6, 2012

Uh oh!

rykov commented Jun 6, 2012

Uh oh!

wagenet commented Jun 9, 2012

Uh oh!

wagenet commented Jun 9, 2012

Uh oh!

rykov commented Jun 9, 2012

Uh oh!

rykov commented Jun 9, 2012

Uh oh!

rykov commented Jun 9, 2012

Uh oh!

wagenet commented Jun 9, 2012

Uh oh!

rykov commented Jun 9, 2012

Uh oh!

Reviewers

Uh oh!

Assignees

Labels

Projects

Milestone

Development

Uh oh!

4 participants