4.5.12 - [meta] republish broken 4.5.11 4.5.11 - [Fix] sync Object.keys excluded list from object-keys (#456) - [Dev Deps] update `eslint`, `@ljharb/eslint-config`, `jasmine-node`, `replace`, `semver` - [Tests] pin jasmine-node to ~1.13 4.5.10 - [Fix] Safari 11 throws on `.sort({})`, but not on `.sort(null)` - [Fix] ensure the shimmed parseInt throws with Symbols (#450) - [Fix] skip over `localStorage`, which can’t be accessed on file:// - [Fix] Accessing window.top.{constructor|prototype} throws error in iOS (#445) - [Fix] avoid `width` and `height` on `window`, to prevent reflow (https://github.com/ljharb/object-keys/issues/31) - [Fix] ensure minified literal of `1000000000000000128` stays as that literal (#441) - [Robustness] always prefer `String(x)` over `x.toString()` - [Tests] up to `node` `v9.3`, `v8.9`, `v7.10`, `v6.12`, `v5.12`, `v4.8`; improve test matrix; use `nvm install-latest-npm`; comment out OS X builds; pin included builds to LTS - [Tests] `parseInt`: add another test for NaN parsing (#433) - [Dev Deps] `uglify-js`: add `--support-ie8` and peg to v2.7.x since it doesn’t follow semver - [Dev Deps] update `eslint`, `@ljharb/eslint-config`, `jscs`, `uglify-js`, `semver`; remove `concurrently` (#421) 4.5.9 - [Fix] parseInt and parseFloat should both accept null/undefined (#402) - [Tests] up to `node` `v6.2` - [Dev Deps] update `eslint`, `@ljharb/eslint-config`, `jscs`, `concurrently` 4.5.8 - [Fix] Check if property descriptor is configurable before re-defining it (#394, #373) - [Dev Deps] update `eslint`, `jscs`, `@ljharb/eslint-config` - [Tests] up to `node` `v5.10`, `v4.4` - [Tests] Use `concurrently` instead of `parallelshell` - [Tests] use `pretest` to run the linter 4.5.7 - [Fix] `bind` in IE 8: Update `is-callable` implementation to v1.1.3 (#390) 4.5.6 - [Fix] `new Date(new Date())` should work in IE 8 (#389) - [Tests] on `node` `v5.7` - [Dev Deps] update `uglify-js` 4.5.5 - [Fix] Adobe Photoshop’s JS engine bizarrely can have `+date !== date.getTime()` (#365) - [Dev Deps] update `eslint` - [Refactor] Update `is-callable` implementation to match latest - [Dev Deps] update `eslint`, `@ljharb/eslint-config`, `jscs` 4.5.4 - [Fix] careless error from 5cf99aca49e59bae03b5d542381424bb1b13ec42 4.5.3 - [Fix] Saturday is a day in the week (#386) - [Robustness] improve Function#bind (#381) - [Tests] on `node` `v5.6`, `v4.3` - [Tests] use json3 (#382) - [Dev Deps] update `eslint`, `@ljharb/eslint-config` - [Docs] add note about script order (#379) 4.5.2 - [shim: fix] use `Array#slice`, not `String#slice`, on `String#split` output (#380) 4.5.1 - [Fix] Make sure preexisting + broken `Array#` methods that throw don’t break the runtime (#377) - [Dev Deps] update `jscs` 4.5.0 - [New] `parseFloat('-0')` should return -0 in Opera 12 (#371) - [New] Provide and replace Date UTC methods (#360) - [Robustness] cache `Date` getUTC* methods so that `Date#toISOString` doesn’t observably look them up on the receiver - [Robustness] use a cached and shimmed `String#trim` - [Tests] up to `node` `v5.5` - [Tests] add `parallelshell` and use it in a few tasks - [Refactor] rename cached methods to avoid linter warnings - [Dev Deps] update `eslint`, `jscs`, '@ljharb/eslint-config' - [Docs] Update license year to 2016 (#374) 4.4.2 - [shim: fix] use `Array#slice`, not `String#slice`, on `String#split` output (#380) 4.4.1 - [Fix] ensure that IE 11 in compatibility mode doesn't throw (#370) - [Docs] add missing shimmed things 4.4.0 - [New] Detect and patch `RegExp#toString` in IE 8, which returns flags in the wrong order (#364) - [Fix] Patch `Array#sort` on {Chrome, Safari, IE < 9, FF 4} that throws improperly, per ES5 (#354) - [Fix] In IE 6, `window.external` makes `Object.keys` throw - [Fix] `Array#slice`: boxed string access on IE <= 8 (#349) - [Fix] `Array#join`: fix IE 6-8 join called on string literal (#352) - [Fix] Ensure that `Error#message` and `Error#name` are non-enumerable (#358) - [Fix: sham] `Object.getOwnPropertyDescriptor`: In Opera 11.6, `propertyIsEnumerable` is a nonshadowable global, like `toString` - [Robustness] Use a bound form of `Array#slice.call` - [Tests] Properly check for descriptor support in IE <= 8 - [Tests] on `node` `v5.1` - [Tests] Add `Array#slice` tests (#346) - [Dev Deps] update `uglify-js`, `eslint`, `jscs`, `uglify-js`, `semver` - [Docs] Fix broken UMD links (#344) 4.3.2 - [shim: fix] use `Array#slice`, not `String#slice`, on `String#split` output (#380) 4.3.1 - [Fix] `String#split`: revert part of dcce96ae21185a69d2d40e67416e7496b73e8e47 which broke in older browsers (#342) - [Dev Deps] update `eslint`, `@ljharb/eslint-config`, `jscs` - [Tests] Firefox allows `Number#toPrecision` values of [1,100], which the spec permits 4.3.0 - [New] `Array#push`: in IE <= 7, `Array#push` was not generic (#336) - [New] `Array#push` in Opera `10.6` has a super weird bug when pushing `undefined` - [New] `Array#join`: In IE <= 7, passing `undefined` didn't use the default separator (#333) - [New] `Error#toString`: prints out the proper message in IE 7 and below (#334) - [New] `Number#toPrecision`: IE 7 and below incorrectly throw when an explicit `undefined` precision is passed (#340) - [Fix] `String#lastIndexOf`: ensure the correct length in IE 8 - [Fix] ensure `parseInt` accepts negative and plus-prefixed hex values (#332) - [Robustness] Use a bound `Array#push` instead of relying on `Function#call` - [Dev Deps] update `eslint`, `@ljharb/eslint-config`, `jscs` - [Tests] Add some conditionals to avoid impossible-to-fix test failures in IE 6-8, due to it being unable to distinguish between `undefined` and an absent index (#114) - [Tests] Fix false negatives in IE 6-8 with jasmine comparing arrays to arraylikes (#114) - [Tests] add additional `Array#shift` tests (#337) - [Tests] Add additional `Array#splice` tests (#339) - [Tests] Add `Array#pop` tests, just in case (#338) - [Tests] include `global` tests in HTML test files - [Tests] Make sure the HTML tests run with the right charset - [Tests] ensure `node` `v0.8` tests stay passing. - [Tests] Prevent nondeterminism in the tests - this sometime produced values that are one ms off - [Tests] on `node` `v5.0` - [Tests] fix npm upgrades for older nodes 4.2.1 - [shim: fix] use `Array#slice`, not `String#slice`, on `String#split` output (#380) 4.2.0 - [shim: new] Overwrite `String#lastIndexOf` in IE 9, 10, 11, and Edge, so it has proper unicode support. - [Dev Deps] update `eslint`, `jscs` 4.1.16 - [shim: fix] use `Array#slice`, not `String#slice`, on `String#split` output (#380) 4.1.15 - [shim: fix] new Date + Date.parse: Fix a Safari 8 & 9 bug where the `ms` arg is treated as a signed instead of unsigned int (#329) - [shim: fix] add 'frame' to blacklisted keys (#330) - [Dev Deps] update `eslint`, `@ljharb/eslint-config`, `jscs`, `uglify-js` - [Tests] on `node` `v4.2` - [Tests] Date: prevent nondeterminism in the tests - this sometime produced values that are one ms off 4.1.14 - [shim: fix] Wrap more things in a try/catch, because IE sucks and sometimes throws on [[Get]] of window.localStorage (#327) - [Refactor] Use `ES.ToUint32` instead of inline `>>>` - [Tests] up to `node` `v4.1` - [Dev Deps] update `eslint`, `@ljharb/eslint-config`, `semver`, `jscs` 4.1.13 - [shim: fix] Fix a bug where `Date(x)` threw instead of equalling `String(Date(x))` (#326) 4.1.12 - [fix] Make sure uglify doesn't remove function names - [shim: fix] Use `is-arguments` implementation; don't call down legacy code path in modern engines (#325) - [Tests] up to `io.js` `v3.3` - [Dev Deps] update `eslint`, `@ljharb/eslint-config` 4.1.11 - [shim: fix] Object.keys in Safari 9 has some bugs. (Already fixed in Webkit Nightly) - [shim: fix] Omit !Date.parse check in the if statement (#323) - [sham: fix] Fix Object.create sham to not set __proto__ (#301) - [sham: fix] Add a typeof check to Object.getPrototypeOf (#319, #320) - [Tests] up to `io.js` `v3.1` - [Tests] Make sure `Object.getPrototypeOf` tests don't fail when engines implement ES6 semantics - [Docs] Switch from vb.teelaun.ch to versionbadg.es for the npm version badge SVG (#318) - [Dev Deps] Update `eslint`, `uglify-js`, `jscs`; use my personal shared `eslint` config 4.1.10 - [Fix] Fix IE 8 issue with `window.frameElement` access in a child iframe (#322) - [Tests] Consolidating `Date.parse` extended year tests - [Tests] Account for a `Date.parse` precision variance in Safari 8 - [Tests] DRY up some Date.parse tests - [Tests] Don't create globals in Date tests - [Tests] Better failure output when an invalid date's toJSON throws - [Tests] Also compare lengths of array-likes - [Tests] Extra check on Object.keys(arguments) - [Tests] Skip appropriate tests when objects aren't extensible/freezeable/sealable 4.1.9 - [Fix] Remove "extended", add "unicode" in `String#split` shim, to match ES6 - [Fix] Object.keys: Preserve the IE 8 dontEnum bugfix, and the automation equality bugfix. - [Fix] Object.keys: Prevent a deprecation message from showing up in Chrome. - [Performance] Speed up blacklisted key check for Object.keys automation equality bug. - [Tests] Test on `io.js` `v2.4` - [Dev Deps] Update `eslint`, `semver` 4.1.8 - [Fix] Fix an `Object.keys` IE 8 bug where `localStorage.prototype.constructor === localStorage` would throw (#275) - [Fix] Shimmed `Object.defineProperty` should not throw for an empty descriptor (#315) - [Fix] Fix `Date#toISOString` in Safari 5.1 (#243) - [Fix] Use `Object#propertyIsEnumerable` to default the initial "enumerable" value in `Object.getOwnPropertyDescriptor` sham (#289) - [Fix] Fix `Array#splice` with large sparse arrays in Safari 7/8, and Opera 12.15 (#295) - [Robustness] Safely use and reference many builtins internally (also see #313) - [Tests] Add `Date#{getUTCDate,getUTCMonth}` tests to expose Opera 10.6/11.61/12 `Date` bugs - [Dev Deps] Update `eslint` 4.1.7 - Make sure `Date.parse` is not enumerable (#310) 4.1.6 - Support IE 8 when `document.domain` is set (#306, #150) - Remove version from `bower.json` (#307) 4.1.5 - Add a failing runtime check for Safari 8 `Date.parse` - Update `eslint`, `semver` - Test on `io.js` `v2.2` 4.1.4 - Make sure copied `Date` properties remain non-enumerable. - Using a more reliable check for supported property descriptors in non-IE ES3 - Fix 'constructor' in Object.defineProperties sham in ES3 (#252, #305) - Use a reference to `Array#concat` rather than relying on the runtime environment's `concat`. - Test on `io.js` `v2.1` - Clean up `Array.prototype` iteration methods 4.1.3 - Update `license` in `package.json` per https://docs.npmjs.com/files/package.json#license - Update `uglify-js`, `eslint` 4.1.2 - In IE 6-8, `Date` inside the function expression does not reference `DateShim` (#303) - Date: Ensure all code paths have the correct `constructor` property - Date: Don't copy non-own properties from original `Date` - Test up to `io.js` `v2.0.0` - Simplify `isPrimitive` check. - Adding sanity check tests for ES5 `Number` constants. - Update `uglify-js`, `eslint`, `semver` 4.1.1 - Fix name of `parseInt` replacement. - Update copyright year - Update `eslint`, `jscs` - Lock `uglify-js` down to v2.4.17, since v2.4.18 and v2.4.19 have a breaking change. - All grade A-supported `node`/`iojs` versions now ship with an `npm` that understands `^`. - Run `travis-ci` tests on latest `node` and `iojs`; speed up builds; allow 0.8 failures. - Ensure some Object tests don't fail in ES6 - Make sure `Date` instances don't have an enumerable `constructor` property, when possible. 4.1.0 - Update `eslint` - Improve type checks: `Array.isArray`, `isRegex` - Replace `isRegex`/`isString`/`isCallable` checks with inlined versions from npm modules - Note which ES abstract methods are replaceable via `es-abstract` - Run `travis-ci` tests on `iojs`! 4.0.6 - Update `jscs`, `uglify-js`, add `eslint` - es5-sham: fix Object.defineProperty to not check for own properties (#211) - Fix Array#splice bug in Safari 5 (#284) - Fix `Object.keys` issue with boxed primitives with extra properties in older browsers. (#242, #285) 4.0.5 - Update `jscs` so tests pass 4.0.4 - Style/indentation/whitespace cleanups. - README tweaks 4.0.3 - Fix keywords (#268) - add some Date tests - Note in README that the es5-sham requires the es5-shim (https://github.com/es-shims/es5-shim/issues/256#issuecomment-52875710) 4.0.2 - Start including version numbers in minified files (#267) 4.0.1 - Fix legacy arguments object detection in Object.keys (#260) 4.0.0 - No longer shim the ES5-spec behavior of splice when `deleteCount` is omitted - since no engines implement it, and ES6 changes it. (#255) - Use Object.defineProperty where available, so that polyfills are non-enumerable when possible (#250) - lots of internal refactoring - Fixed a bug referencing String#indexOf and String#lastIndexOf before polyfilling it (#253, #254) 3.4.0 - Removed nonstandard SpiderMonkey extension to Array#splice - when `deleteCount` is omitted, it's now treated as 0. (#192, #239) - Fix Object.keys with Arguments objects in Safari 5.0 - Now shimming String#split in Opera 10.6 - Avoid using "toString" as a variable name, since that breaks Opera - Internal implementation and test cleanups 3.3.2 - Remove an internal "bind" call, which should make the shim a bit faster - Fix a bug with object boxing in Array#reduceRight that was failing a test in IE 6 3.3.1 - Fixing an Array#splice bug in IE 6/7 - cleaning up Array#splice tests 3.3.0 - Fix Array#reduceRight in node 0.6 and older browsers (#238) 3.2.0 - Fix es5-sham UMD definition to work properly with AMD (#237) - Ensure that Array methods do not autobox context in strict mode (#233) 3.1.1 - Update minified files (#231) 3.1.0 - Fix String#replace in Firefox up through 29 (#228) 3.0.2 - Fix `Function#bind` in IE 7 and 8 (#224, #225, #226) 3.0.1 - Version bump to ensure npm has newest minified assets 3.0.0 - es5-sham: fix `Object.getPrototypeOf` and `Object.getOwnPropertyDescriptor` for Opera Mini - Better override noncompliant native ES5 methods: `Array#forEach`, `Array#map`, `Array#filter`, `Array#every`, `Array#some`, `Array#reduce`, `Date.parse`, `String#trim` - Added spec-compliant shim for `parseInt` - Ensure `Object.keys` handles more edge cases with `arguments` objects and boxed primitives - Improve minification of builds 2.3.0 - parseInt is now properly shimmed in ES3 browsers to default the radix - update URLs to point to the new organization 2.2.0 - Function.prototype.bind shim now reports correct length on a bound function - fix node 0.6.x v8 bug in Array#forEach - test improvements 2.1.0 - Object.create fixes - tweaks to the Object.defineProperties shim 2.0.0 - Separate reliable shims from dubious shims (shams). 1.2.10 - Group-effort Style Cleanup - Took a stab at fixing Object.defineProperty on IE8 without bad side-effects. (@hax) - Object.isExtensible no longer fakes it. (@xavierm) - Date.prototype.toISOString no longer deals with partial ISO dates, per spec (@kitcambridge) - More (mostly from @bryanforbes) 1.2.9 - Corrections to toISOString by @kitcambridge - Fixed three bugs in array methods revealed by Jasmine tests. - Cleaned up Function.prototype.bind with more fixes and tests from @bryanforbes. 1.2.8 - Actually fixed problems with Function.prototype.bind, and regressions from 1.2.7 (@bryanforbes, @jdalton #36) 1.2.7 - REGRESSED - Fixed problems with Function.prototype.bind when called as a constructor. (@jdalton #36) 1.2.6 - Revised Date.parse to match ES 5.1 (kitcambridge) 1.2.5 - Fixed a bug for padding it Date..toISOString (tadfisher issue #33) 1.2.4 - Fixed a descriptor bug in Object.defineProperty (raynos) 1.2.3 - Cleaned up RequireJS and