Permalink
Commits on Jan 11, 2019
  1. Automated g4 rollback of changelist 227940964.

    nreid260 committed Jan 10, 2019
    *** Reason for rollback ***
    
    Rollforward with EventType fix. Instead of checking useragent, check for the existence
    of the [prefix]requestFullscreen methods and use that to determine the event type.
    Also reorders the other methods so that requestFullscreen is before the webkit version.
    
    RELNOTES[NEW]: Add support for FullscreenOptions in requestFullscreen utility
    
    *** Original change description ***
    
    Automated g4 rollback of changelist 226359674.
    
    *** Reason for rollback ***
    
    I believe this CL has resulted in a fullscreen regression in our app: http://b/122101460
    
    It also looks like there is a logical error in this CL. The ordering change means that Chrome will now be calling requestFullscreen() instead of webkitRequestFullscreen(). However, goog.dom.fullscreen.EventType was not updated, so users referencing EventType.CHANGE will still be listening for `webkitfullscreenchange` not `fullscree...
    
    ***
    
    -------------
    Created by MOE: https://github.com/google/moe
    MOE_MIGRATED_REVID=228764798
  2. Automated g4 rollback of changelist 228530531.

    nreid260 committed Jan 10, 2019
    *** Reason for rollback ***
    
    Breakage
    
    *** Original change description ***
    
    Safe alternative to creating an image element from a blob.
    
    RELNOTES[NEW]: Added a helper function that guarantees revoking the object url created from createObjectUrl when creating an Image element from a blob with MIME type image/.*.
    
    Also whitelists use of createObjectUrl in goog.dom.safe with MANUALLY_REVIEWED
    
    ***
    
    -------------
    Created by MOE: https://github.com/google/moe
    MOE_MIGRATED_REVID=228618069
Commits on Jan 9, 2019
  1. Safe alternative to creating an image element from a blob.

    nreid260 committed Jan 9, 2019
    RELNOTES[NEW]: Added a helper function that guarantees revoking the object url created from createObjectUrl when creating an Image element from a blob with MIME type image/.*.
    
    Also whitelists use of createObjectUrl in goog.dom.safe with MANUALLY_REVIEWED
    
    -------------
    Created by MOE: https://github.com/google/moe
    MOE_MIGRATED_REVID=228530531
  2. RELNOTES[NEW]: [goog.dom.asserts] Added missing assertIsHTMLTextAreaE…

    nreid260 committed Jan 8, 2019
    …lement() and assertIsHTMLCanvasElement().
    
    -------------
    Created by MOE: https://github.com/google/moe
    MOE_MIGRATED_REVID=228281054
  3. Enable recording broadcasted events.

    nreid260 committed Jan 8, 2019
    This is step 1.
    
    Step 2 will be to try to flip this to the default by finding all tests that fail. But in the meantime, people can use this method.
    
    RELNOTES: n/a
    
    -------------
    Created by MOE: https://github.com/google/moe
    MOE_MIGRATED_REVID=228274939
  4. Automated g4 rollback of changelist 226359674.

    nreid260 committed Jan 5, 2019
    *** Reason for rollback ***
    
    I believe this CL has resulted in a fullscreen regression in our app: http://b/122101460
    
    It also looks like there is a logical error in this CL. The ordering change means that Chrome will now be calling requestFullscreen() instead of webkitRequestFullscreen(). However, goog.dom.fullscreen.EventType was not updated, so users referencing EventType.CHANGE will still be listening for `webkitfullscreenchange` not `fullscreenchange`.
    
    *** Original change description ***
    
    Add options to requestFullscreen closure utility function
    
    Also reorders the checks so that "requestFullscreen" is first. The other prefixes
    should be fallbacks only when requestFullscreen is not available.
    
    RELNOTES[NEW]: Add support for FullscreenOptions in requestFullscreen utility
    
    ***
    
    -------------
    Created by MOE: https://github.com/google/moe
    MOE_MIGRATED_REVID=227940964
  5. Fix numeric formatting to use quantity value of negative zero for uni…

    nreid260 committed Jan 3, 2019
    …ts in the past, positive zero for future, e.g., "0 minutes ago" vs. "in 0 minutes."
    
    RELNOTES: Fix goog.i18n.RelativeDateTimeFormat to use quantity value of negative zero to format zero units as past rather than future.
    
    In the previous release, a quantity with value zero gives a future format, e.g., "in 0 minutes". With this update, the caller may specify a negative zero value which will select the past form, e.g., "0 minutes ago".
    
    This change allows the caller to choose the specific direction of time when a value of zero is needed.
    
    The two values of zero are part of the floating point standard IEEE 754. Since regular comparison in JavaScript does not distinguish the two signed zero values, it is necessary to divide positive infinity by the zero value. A result of negative infinity indicates negative zero, otherwise the value was positive zero.
    
    The change applies to all units and all locales.
    
    -------------
    Created by MOE: https://github.com/google/moe
    MOE_MIGRATED_REVID=227722147
Commits on Dec 19, 2018
  1. Fix $waitAndVerifyAll for LooseMock with $times

    nreid260 committed Dec 18, 2018
    RELNOTES: same
    
    Bug was introduced in cl/225756459 (4e4b35e) that broke $waitAndVerifyAll, code worked for StrictMock but not LooseMock.
    
    -------------
    Created by MOE: https://github.com/google/moe
    MOE_MIGRATED_REVID=226025436
  2. Eliminates prototypal inheritance errors related to class-side inheri…

    nreid260 committed Dec 18, 2018
    …tance.
    
    This CL is a preparatory step to prevent build errors and test failures for an upcoming change to typecheck prototypal inheritance of classes. The goal is to unblock progress by the JS Infrastructure team with the smallest possible overhead.
    
    This is the list of errors used to generate this Rosie: https://docs.google.com/spreadsheets/d/1a9QMRl6hGfQ03tfH0-CWGKLVJTB050-V4-ATTjisuTo/edit#gid=758793470 It may go out of date, but should contain any errors in these files. To confirm your fixes please patch CL/215669793 and build the affected targets. It's probably wise to build from head as well ("--define JSC_AT_HEAD=true").
    
    This is the list of errors used to generate this Rosie: https://docs.google.com/spreadsheets/d/1a9QMRl6hGfQ03tfH0-CWGKLVJTB050-V4-ATTjisuTo/edit#gid=758793470 It may go out of date, but should contain any errors in these files. To confirm your fixes please patch CL/215669793 and build the affected targets. It's probably wise to build from head as well ("--define JSC_AT_HEAD=true").
    
    -------------
    Created by MOE: https://github.com/google/moe
    MOE_MIGRATED_REVID=225935227
  3. Fix $never expectations and $waitAndVerifyAll() for Mocks.

    nreid260 committed Dec 17, 2018
    RELNOTES: same
    
    -------------
    Created by MOE: https://github.com/google/moe
    MOE_MIGRATED_REVID=225756459
  4. RELNOTES: Add YEAR_MONTH_SHORT skeleton to Closure date time and date…

    nreid260 committed Dec 13, 2018
    … interval formatters.
    
    -------------
    Created by MOE: https://github.com/google/moe
    MOE_MIGRATED_REVID=225389303
Commits on Dec 5, 2018
  1. Fix `goog.net.Xhrio::getRequestHeaders` to better handle whitespace a…

    nreid260 authored and shicks committed Dec 2, 2018
    …nd colons.
    
    There are still significant deviations from the header parsing spec, but those are ignored for now.
    
    Closes #946
    
    RELNOTES: Headers containing only whitespace, or multiple colons are handled better.
    
    -------------
    Created by MOE: https://github.com/google/moe
    MOE_MIGRATED_REVID=223709859
Commits on Nov 28, 2018
  1. Setting spellcheck on the HSVPalette's input element to false to supp…

    nreid260 committed Nov 27, 2018
    …ress the red underline in the input field (i.e. b/119617714).
    
    DELTA=15 (13 added, 0 deleted, 2 changed)
    DELTA_BY_EXTENSION=js=15
    TESTED=on web
    RELNOTES: Setting spellcheck on the HSVPalette's input field to false to suppress the red underline in the input field.
    
    -------------
    Created by MOE: https://github.com/google/moe
    MOE_MIGRATED_REVID=222980811
  2. Allow ES6 Promises to be returned by labs.Environment setUp() and set…

    nreid260 committed Nov 26, 2018
    …UpPage()
    
    RELNOTES: same
    
    -------------
    Created by MOE: https://github.com/google/moe
    MOE_MIGRATED_REVID=222851113
  3. Allow required input in the JS HTML sanitizer.

    nreid260 committed Nov 26, 2018
    RELNOTES: The HTML sanitizer should allow the "required" attribute (which is defined on "input" elements).
    
    -------------
    Created by MOE: https://github.com/google/moe
    MOE_MIGRATED_REVID=222851078
  4. Cleanup Fusion XJS final touch

    nreid260 committed Nov 19, 2018
    RELNOTES: Some user-visible change.
    
    -------------
    Created by MOE: https://github.com/google/moe
    MOE_MIGRATED_REVID=222095217
  5. Closure: Fix browser UA match to include Firefox iOS.

    nreid260 committed Nov 19, 2018
    RELNOTES[INC]: Identify iOS Firefox browser (FxiOS useragent) as firefox in goog.labs.userAgent.browser.isFirefox instead of Safari.
    
    -------------
    Created by MOE: https://github.com/google/moe
    MOE_MIGRATED_REVID=222091793
Commits on Nov 6, 2018
  1. Fix some tests to conform with Safari 12.0 / Forge on Mac.

    nreid260 committed Nov 5, 2018
    RELNOTES: n/a
    
    -------------
    Created by MOE: https://github.com/google/moe
    MOE_MIGRATED_REVID=220148468
  2. A safe URL assertions should report the URL when it fails.

    nreid260 committed Nov 5, 2018
    RELNOTES: n/a
    
    -------------
    Created by MOE: https://github.com/google/moe
    MOE_MIGRATED_REVID=220100934
  3. Automated g4 rollback of changelist 217609072.

    nreid260 committed Nov 5, 2018
    *** Reason for rollback ***
    
    Roll-forward, moving the logic to payment's checkbox.js. The tabindex=-1 isn't being set in this particular case because the element is not visible (when the control is made visible, the focusable index is set then), and there's some interaction with needing that tabindex attribute set with the collapsing-form's expand, the exact logic of which still eludes me, but the one failing test passes with this change now.
    
    That move of the logic down is in cr/218527300.
    
    *** Original change description ***
    
    Automated g4 rollback of changelist 217571029.
    
    *** Reason for rollback ***
    
    Confirmed the affected test passes with this rollback: http://sponge/5eef0949-7c33-4291-b753-e43da8de4028
    
    [AUTO] This CL breaks the following TAP projects:
    payments_web [http://test/ui#p=payments_web&d=buildcop]
    
    TAP links to tests broken by this CL:
    http://test/ui#d=targetZoomout&flags=CAMQArABAQ==&t=//javatests/com/google/commerce/payments/business/widgets/usermanagement/browser:browser2_tests_chrome-linux
    
    *** Original change description ***
    
    Remove the Checkbox#setEnabled override that additionally sets the tabindex on the control element.
    
    Removing the tabindex fix here should be sufficient now since Control#setEnabled calls renderer_.setFocusable:
    http://google3/javascript/closure/ui/control.js?l=878&rcl=199159062
    
    The default renderer for a Checkbox is a CheckboxRenderer, which doesn't override setFocusable, so it results in ControlRenderer#setFocusable:
    http://google3/javascript/closure/ui/controlrenderer.js?l=522&rcl=184578348...
    
    RELNOTES: Restores Checkbox#setEnabled (rolling back previous change)
    
    -------------
    Created by MOE: https://github.com/google/moe
    MOE_MIGRATED_REVID=220095602
  4. Change possibly untyped document.all reference to document['all'] in …

    nreid260 committed Nov 5, 2018
    …closure/base.js
    
    This change makes base.js compile cleanly when the BanUnknownTypedClassPropsReferences conformance check is enabled.
    
    document.all is an IE extension. It is defined in the IE externs at http://google3/javascript/externs/ie_dom.js?l=966&rcl=209453504, but these externs aren't always loaded.
    
    RELNOTES: Change possibly untyped document.all reference to document['all'] in closure/base.js
    
    -------------
    Created by MOE: https://github.com/google/moe
    MOE_MIGRATED_REVID=220026426
  5. When we need to fallback to positive module loading, also tell the mo…

    nreid260 committed Nov 1, 2018
    …dule manager that all the loaded modules have already been loaded. Also modified the test to verify this behavior.
    
    RELNOTES: new API to moduleinfo object.
    
    -------------
    Created by MOE: https://github.com/google/moe
    MOE_MIGRATED_REVID=219722000
  6. Remove deprecated goog.net.cookies.MAX_COOKIE_LENGTH constant.

    nreid260 committed Nov 1, 2018
    Fixes build warnings:
    WARNING: javascript/closure/net/cookies.js:380: WARNING - Property MAX_COOKIE_LENGTH never defined on goog.net.Cookies
    goog.net.cookies.MAX_COOKIE_LENGTH = goog.net.Cookies.MAX_COOKIE_LENGTH;
    
    RELNOTES: Removes deprecated constant goog.net.cookies.MAX_COOKIE_LENGTH.
    
    -------------
    Created by MOE: https://github.com/google/moe
    MOE_MIGRATED_REVID=219665287
  7. Replaced broken goo.gl links with links to the official documentation…

    nreid260 committed Nov 1, 2018
    … on GitHub.
    
    RELNOTES: n/a
    
    -------------
    Created by MOE: https://github.com/google/moe
    MOE_MIGRATED_REVID=219651560
  8. JS Tests - Group output logs.

    nreid260 committed Oct 31, 2018
    Use console.group to group logs together for better readability.
    JSUnit: https://screenshot.googleplex.com/COcdqpMYaLm.png
    Karma: https://screenshot.googleplex.com/XNa9ngjAKZO.png
    
    RELNOTES: Group test logs together for better output readability.
    
    -------------
    Created by MOE: https://github.com/google/moe
    MOE_MIGRATED_REVID=219431308
Commits on Oct 8, 2018
  1. RELNOTES: Fix a loss-of-precision issue in `NumberFormat` that caused…

    nreid260 authored and jplaisted committed Oct 1, 2018
    … some formattings to appear less accurate.
    
    See #916.
    
    This change introduces the `decimalShift_` function and applies it as a primitive throughout the class. This function is expected to be inefficient, but it more accurately describes the intent of the callsites and corrects for possible edge cases related to floating-point numbers. The loss of performance shouldn't be an issue given the intended applications of `NumberFormat`.
    
    Closes #916
    
    -------------
    Created by MOE: https://github.com/google/moe
    MOE_MIGRATED_REVID=215245799
Commits on Jul 17, 2018
  1. Convert more "js_library" targets in Closure to "js_lib"s and fix the…

    nreid260 authored and jplaisted committed Jul 3, 2018
    … uncovered errors.
    
    RELNOTES: Tightened some types.
    
    -------------
    Created by MOE: https://github.com/google/moe
    MOE_MIGRATED_REVID=203139335