HEAD | 259d061 | Claim to support testdriver in servodriver | 23 December 2018, 05:25:51 UTC |
refs/heads/MultipartContent | f7651d3 | Simplify MultipartContent. The class is only used once, so change it to simplify that case. | 26 September 2018, 11:55:47 UTC |
refs/heads/annevk-navigate-to-weird-schemes | 9b373c2 | test about URLs | 31 May 2016, 13:13:51 UTC |
refs/heads/annevk/EventSource-U+0000 | d474c2a | EventSource: test U+0000 in ID better See https://bugzilla.mozilla.org/show_bug.cgi?id=1387355 for context. | 16 November 2018, 12:38:42 UTC |
refs/heads/annevk/autofocus | 4f4d424 | HTML: remove autofocused element before its task runs See https://bugzilla.mozilla.org/show_bug.cgi?id=1463563 for context. We'll need to update -2 to figure out the correct pass condition. Given Chrome and Safari, I guess we should expect input2 and then we should update the specification to note that if the element is no longer connected you just abort the algorithm, meaning no flags get set. | 28 May 2018, 14:47:32 UTC |
refs/heads/annevk/bc-closing | 072ffc9 | BroadcastChannel: make closing return early See https://github.com/whatwg/html/issues/3528. | 28 November 2018, 11:14:36 UTC |
refs/heads/annevk/content-length | 858dccf | no hex or oct digits | 15 October 2018, 09:37:30 UTC |
refs/heads/annevk/document-domain | ecc1a8f | WIP test for document.domain | 17 February 2017, 10:44:29 UTC |
refs/heads/annevk/document-open-load-event | 904da22 | correct test | 03 May 2018, 08:59:36 UTC |
refs/heads/annevk/document-open-steps-readiness | 40dc6ae | Capture readyState changes before document.open too | 31 July 2018, 19:10:11 UTC |
refs/heads/annevk/document-write-reloading | 1e7d904 | add variants where .close() is called | 11 May 2018, 06:10:04 UTC |
refs/heads/annevk/dom-listener-order | 4ff3821 | DOM: test listener invocation order For https://github.com/whatwg/dom/pull/686. | 17 September 2018, 13:51:33 UTC |
refs/heads/annevk/encodeInto | 60bfa1e | test passing buffer directly and detached buffer | 14 December 2018, 10:25:18 UTC |
refs/heads/annevk/fetch-fragments | 37c10e9 | Update tests | 12 April 2018, 11:07:55 UTC |
refs/heads/annevk/h1-message-parsing | 0589207 | HTTP: message parsing with CR See https://github.com/httpwg/http-core/issues/31 for context. | 15 October 2018, 14:50:01 UTC |
refs/heads/annevk/html-ancestororigins | 48a297b | semicolons for everyone | 05 February 2018, 15:48:20 UTC |
refs/heads/annevk/http-versioning | 7606bc9 | Test HTTP parsing Closes #5064. | 09 March 2017, 10:59:34 UTC |
refs/heads/annevk/legacy-pre-activation-behavior-phase | 7c18f59 | DOM: eventPhase during legacy-pre-activation behavior For https://github.com/whatwg/dom/pull/686. | 17 September 2018, 13:30:38 UTC |
refs/heads/annevk/meta-global-annevk | 4b97ac7 | Merge remote-tracking branch 'origin/master' into annevk/meta-global | 25 January 2018, 12:55:18 UTC |
refs/heads/annevk/navigate-redirect | bf0b0da | WIP | 11 August 2016, 12:41:28 UTC |
refs/heads/annevk/navigate-source | d424efc | "Source browsing context" tests Tests to determine what "source browsing context" actually is. (Hint: it's not a browsing context.) | 20 May 2016, 08:59:32 UTC |
refs/heads/annevk/opaque-redirect-handling | da25c99 | move service worker into resources too | 06 June 2018, 09:55:14 UTC |
refs/heads/annevk/origin-header | 7d1437d | WIP: figuring out Origin outside of CORS | 27 November 2018, 15:40:44 UTC |
refs/heads/annevk/simplify-license | 0ebe89e | Use the The 3-Clause BSD License from opensource.org Fixes #11009. | 28 May 2018, 13:26:49 UTC |
refs/heads/annevk/sw-redirect-fragment | add926c | add more tests | 27 August 2018, 16:32:58 UTC |
refs/heads/annevk/template-content-node-document | a1b9c9f | clear cookies | 29 November 2016, 17:33:31 UTC |
refs/heads/annevk/url-host-parser | a41d3a4 | run these faster | 09 February 2017, 14:46:05 UTC |
refs/heads/annevk/window.opener | a967a92 | add some tests (don't address all lint nits just yet) | 05 February 2018, 09:22:44 UTC |
refs/heads/annevk/www-authenticate-parsing | 29ad6b8 | Fetch/HTTP: test WWW-Authenticate parsing For https://github.com/httpwg/http-core/issues/136. | 24 September 2018, 14:41:07 UTC |
refs/heads/annevk/x-frame-options | a337d72 | Move x-frame-options/ into content-security-policy/ The plan of record is for CSP to define X-Frame-Options (with small parts of it in HTML). Moving this directory there to reduce top-level directories (without owner). Helps with #5915. | 24 April 2018, 09:35:09 UTC |
refs/heads/annevk/xhr-h2-statusText | 70e1336 | this is it | 15 August 2018, 14:25:47 UTC |
refs/heads/annevk/xhr-no-headers-received | 03de239 | XHR: no headers received event for a network error Closes #8049. | 04 October 2018, 15:11:13 UTC |
refs/heads/annevk/xhr-responseurl-with-fragment | fcd7eab | XHR: responseURL with fragment For https://github.com/whatwg/xhr/pull/200. | 24 August 2018, 13:30:16 UTC |
refs/heads/appveyor | a619744 | Only run Windows CI for tools/ changes | 09 October 2018, 17:34:30 UTC |
refs/heads/ayg/document-named-item | 9ee63dc | Rewrite document named property tests Incorperates https://github.com/whatwg/html/issues/2930 and https://github.com/whatwg/html/pull/3013. | 06 September 2017, 14:01:46 UTC |
refs/heads/bburg-unidentified-keys | 0c61d8d | webdriver: fix expected key code for "Unidentified" keys Per the UI Events specification (https://w3c.github.io/uievents-code/#key-legacy), keys such as \u00e0 (i.e., à) should have a code of "Unidentified". Some tests in key.py expect an empty value, which seems to be an oversight. | 16 July 2018, 23:05:19 UTC |
refs/heads/billing-address-redactions | 4155e54 | Check that billingAddress gets nulled and redacted | 28 November 2018, 05:11:04 UTC |
refs/heads/canvas/getContext-second-argument | d45dfd0 | more clearly indicate where these tests originate (needed in particular for those stuffed in /html/ somewhere) | 25 January 2018, 10:51:01 UTC |
refs/heads/chrome_experimental | a3e01e5 | Pass --enable-experimental-web-platform-features to Chrome instability checker | 13 February 2017, 10:16:48 UTC |
refs/heads/chromium-export-cl-1013322 | e7c6b01 | [Layered API] Add wpt tests TODO: - More coverage for import statement. - Update relative path behavior and tests - fallback_cors.js is not needed - Maybe rename directory name from layered_api to layered-api? - <link rel="modulepreload"> Bug: 829084 Change-Id: I4ce168dd024a9ee52c3d9318066451ee3d5477bd | 04 May 2018, 20:45:01 UTC |
refs/heads/chromium-export-cl-1032554 | fef5160 | Remove catch from WPT destination tests As commented on https://chromium-review.googlesource.com/c/chromium/src/+/1029858, the `catch()` statements in the destination WPT tests are spurious. This CL removes them. Change-Id: Id3e3ab4e4cfb90ace1fa3ffcbfa8d68c3081c9fa | 27 April 2018, 10:30:26 UTC |
refs/heads/chromium-export-cl-1118750 | 937e5a7 | Enable analyser tests to run with Chrome This CL just connects the output of the ScriptProcessorNode to the destination. Chrome won't run the ScriptProcessorNode process unless it is connected. This connection doesn't change the test. Also, a ScriptProcessorNode with 0 outputs is not well-defined in the Web Audio specification, so make it have a single output. It's unlikely Chrome will fix these issues with ScriptProcessorNode because the node is deprecated. Bug: 626703 Change-Id: I2995c892c1739700adc2c9868788c78f2dba7c20 | 28 June 2018, 17:31:35 UTC |
refs/heads/chromium-export-cl-1128385 | d825776 | [css-tables] Test for table width when cell is sized between its min/max Note this is a test only, the bug persists. Bug: 860084 Change-Id: I89d0cabda6018beaef2ec33ee8eb96b44a86e1c8 | 06 July 2018, 23:21:12 UTC |
refs/heads/chromium-export-cl-1136543 | 10c4433 | [css-text] New value 'break-spaces' for the white-space property Finally the CSS WG decided [1] to move back the 'break-spaces' value to the 'white-space' property. This makes the parsing logic easier than the previous approach of using the 'overflow-wrap' property. This new value prevents the white-space sequence to collapse and gives breaking opportunities after every preserved white-space. https://drafts.csswg.org/css-text-3/#valdef-white-space-break-spaces Additionally, unlike 'pre-wrap', non-collapsible spaces or tabs at the end of a line cannot be hung or visually collapsed, since we want them to be preserved and broken. [1] https://github.com/w3c/csswg-drafts/pull/2841 Bug: 767634 Change-Id: I55e888d4472de11c64c4b14e2710c6e3d1832e67 | 22 December 2018, 22:34:40 UTC |
refs/heads/chromium-export-cl-1144133 | 8c228b8 | Try to automate u2f authentication, with vendor-specific implementations of what 'authenticate_u2f' would be. Inspired by https://chromium-review.googlesource.com/c/chromium/src/+/1070778 Change-Id: I5aff81d918a2e243dbe4b05c5143417a0763297f | 20 July 2018, 17:32:08 UTC |
refs/heads/chromium-export-cl-1146644 | b43aa57 | Correctly propagate request initiator for 'noopener' Bug: 830091 Change-Id: I726f57f582dd8b14d0328b2914940695b2674f21 | 24 July 2018, 12:41:46 UTC |
refs/heads/chromium-export-cl-1154225 | 3dadc0c | HTML/events/beforeunload - respect preventDefault(). An existing web-platform-test was enhanced to make sure the preventDefault only (or combined) cases work. Added many layout tests (due to the lack of web-platform-tests APIs for handling beforeunload alert) for most of the intricacies of beforeunload. Bug: 866818 Change-Id: I35c27cee1f71a4e4d10331d3451e0a36b9db684d | 03 November 2018, 12:16:45 UTC |
refs/heads/chromium-export-cl-1188018 | b7f7c26 | Modify containing block logical height for positioned in flipped blocks writing mode scrollbars are excluded from viewport size while calculating container block logical height for positioned items, but scrollbars should not be excluded in flipped blocks writing mode as currently scrollbar is on the right Bug: 833167 Change-Id: Iec1b94d917ca5ca7521d7b030300caaf0bd6e333 | 31 August 2018, 12:26:43 UTC |
refs/heads/chromium-export-cl-1188643 | 23ecf31 | document.open(): Align history/URL behavior with spec This change aligns document.open() with the current HTML Standard, which mandates that in addition to setting the document's URL to the last-entered document's, the document's current history item's URL must also be updated à la history.replaceState(). To accomplish that, this CL reuses the logic in History::StateObjectAdded(), including the throttling behavior, for Document::open() as well. The update steps are run unconditionally, no matter what the document's current URL is, in order to have consistent behavior for other things in the history entry like POST form data, which document.open() now erases. This also means that document.open() now also counts as a navigation, just like history.replaceState(). Several browsertests are updated as such. In this CL, we also enables some WPTs that were previously disabled; in particular, reload.window.html has been enabled to converge to WebKit's behavior. Bug: 68833, 866274 Change-Id: Iea6d665fd97bcaee44bcfaa45f8e92c356003d8a | 19 October 2018, 20:46:11 UTC |
refs/heads/chromium-export-cl-1199810 | f6f3c2a | Remove ExtendedTextMetrics flag to launch canvas TextMetrics. This change removes the flag ExtendedTextMetrics and the code that references it in order to ship the full set of Text Metrics returned by canvas.measureText. Currently chrome (as well as safari and edge) is only returning the width of the measured text. This change exposes the full TextMetrics object defined by the spec: https://html.spec.whatwg.org/multipage/canvas.html#textmetrics Intend-to-ship: https://groups.google.com/a/chromium.org/d/msg/blink-dev/KQW4thKogkk/6fFtjZWvAwAJ Bug: 277215 Change-Id: I03a6c149f86998cc27abf33816d1a02f7900c492 | 06 September 2018, 14:38:08 UTC |
refs/heads/chromium-export-cl-1200863 | 9cef0de | Align resource timing buffer full processing to spec PR 163 This change implements the processing model from PR 163[1], when it comes to setResourceTimingBufferSize(), clearResourceTimings() and the firing of the resourcetimingbufferfull event. [1] https://github.com/w3c/resource-timing/pull/163 Change-Id: I3fd901fa411dd128a723e77cd120f50974a775a8 | 20 November 2018, 21:30:06 UTC |
refs/heads/chromium-export-cl-1212270 | e15d2c1 | Sub-sample accurate start for ABSN Implement sub-sample accurate start for AudioBufferSourceNode. Previously, if the start time was between sample boundaries, we would start the output at the frame before the start time. This is actually incorrect because we haven't actually started yet. We should start the output at the next boundary, and interpolating the value based on the true start time and the sample boundary. Many tests needed to be updated. Basically for each test that needed to be changed, the sample rate is set to a power of two and all ABSN sources are updated to make sure the source starts exactly on a frame boundary. We also took the opportunity to adjust the error thresholds for the tests in case lower values could be used. Some additional notes for the tests: audiobuffersource-playbackrate-zero.html: Add new test to make sure sub-sample accurate start handles a zero playback rate. audiobuffersource-loop-points.html: add some code to save the actual output. This is needed because a new reference file is needed since the sample rate has changed. Also manually tested all of the modified tests with Firefox nightly. They all pass still (except for the new sub-sample test because Firefox doesn't do sub-sample accurate start/stop). Bug: 876917 Test: the-audiobuffersourcenode-interface/sub-sample-scheduling.html Change-Id: Ib13ba30eaa160cfd10739feabac961bf074ee309 | 20 September 2018, 21:42:05 UTC |
refs/heads/chromium-export-cl-1221696 | d0e5b82 | Add experimental support for enterKeyHint. enterKeyHint allows customization of the virtual keyboard enter key. Enabled behind a flag since it isn't approved to ship yet. Spec: https://github.com/whatwg/html/pull/3538 BUG=634330 Cq-Include-Trybots: luci.chromium.try:linux_layout_tests_layout_ng Change-Id: I32b510d8157e0c676300a38bb4c7fa2bbced0b3d | 12 September 2018, 17:49:57 UTC |
refs/heads/chromium-export-cl-1222913 | 21723c6 | Cookie Store: Test cross-origin frames + service workers. Ensure cookieStore changes made in cross-origin iframe propagate to serviceWorkers Bug:729800 Change-Id: I447973d2ed886e434a2db0587290dda883f4eb89 | 13 December 2018, 18:38:01 UTC |
refs/heads/chromium-export-cl-1249391 | f45c992 | WIP: Fix testdriver clicks to make bluetooth wpt pass with UAv2. Blocker: testdriver doens't let click inside iframes! https://cs.chromium.org/chromium/src/third_party/WebKit/LayoutTests/external/wpt/resources/testdriver.js?rcl=5d1bd990e4c21787e6371ab8db5b4fd711eaa7c3&l=102 Bug: 860713 Change-Id: I83b80908b7adc2a65446e5b27000dc67e37d4b9a | 22 November 2018, 14:48:26 UTC |
refs/heads/chromium-export-cl-1264165 | ab57809 | Revert "Implement DTMF [[ToneBuffer]] in the blink layer" This reverts commit 3f90035e2fac160d23309a96ef2cf465b29cebe4. Reason for revert: The original problem was solved using another approach, and this CL just adds complexity with no purpose. Original change's description: > Implement DTMF [[ToneBuffer]] in the blink layer > > This CL makes the Blink layer keep a copy of the tone buffer > and update it on insertDTMF and ontonechange events only; this > makes it possible to expose the state of the tone buffer at the > time the event is fired to the Javascript callback. > > It removes a queueing step inside the DTMF sender, because > that queueing step destroyed the consistency. > > Bug: chromium:816475 > Change-Id: I5aa68396299a67d6cea1e8a17d364f553514c291 > Reviewed-on: https://chromium-review.googlesource.com/1213084 > Reviewed-by: Guido Urdaneta <guidou@chromium.org> > Reviewed-by: Kentaro Hara <haraken@chromium.org> > Commit-Queue: Harald Alvestrand <hta@chromium.org> > Cr-Commit-Position: refs/heads/master@{#589910} TBR=hta@chromium.org,haraken@chromium.org,hbos@chromium.org,guidou@chromium.org,foolip@chromium.org # Not skipping CQ checks because original CL landed > 1 day ago. Bug: chromium:816475 Change-Id: Iaee9fefa8a37e3c6c7256dacac0855f426601e0d | 05 October 2018, 09:29:42 UTC |
refs/heads/chromium-export-cl-1264520 | d3d5233 | Test unicode: ABC~¤•★星🌟星★•¤~XYZ ABC~¤•★星🌟星★•¤~XYZ This is to test Unicode handling of Chromium wpt-exporter. DO NOT MERGE! Change-Id: If4cc639f727c24085f18c15c16a632f0ec200cad | 05 October 2018, 16:01:01 UTC |
refs/heads/chromium-export-cl-1270699 | bbafe29 | HTML parser: Fix parser to close the p tag According to spec, if the start tag is a dialog tag and the stack of open elements has a p element in button scope then closes a p element. Also if the end tag is a dialog tag and the stack of close elements have a p element in button scope, then it should close a p element. reference: https://html.spec.whatwg.org/multipage/parsing.html#parsing-main-inbody html5lib-tests PR: https://github.com/html5lib/html5lib-tests/pull/91 Bug: 704534 Change-Id: I5b90243eda3dfca7b14bf53c5767af65fc2b7dc5 | 10 October 2018, 07:29:59 UTC |
refs/heads/chromium-export-cl-1272339 | dc95d6e | Convert blocks.dat to html5lib_blocks.html The HTML parser should close the p for </dialog>. To test it we need to convert a html5lib test into the wpt test. Bug: 704534 Change-Id: If7c1302825eca5b681d70c4562e63dac79175376 | 10 October 2018, 08:14:18 UTC |
refs/heads/chromium-export-cl-1277666 | c538030 | Compute azimuth correctly according to the spec The azimuth calculation is incorrect if one of the vectors is the zero vector. The spec handles this case correctly by just using the dot product alone with "normalized" vectors instead of trying to divide the dot product by the lengths. Bug: 893983 Test: panner-azimuth.html Change-Id: I7f890ae3353a71ffccfc3138038572a8a18b2379 | 11 October 2018, 21:41:30 UTC |
refs/heads/chromium-export-cl-1288573 | 4f05c01 | [Background Fetch] Support access to active fetches. This bug removes the check from renderer that prevents this access, and updates idl files accordingly. It also adds a WPT test to check that incomplete fetches provide access to downloaded content. Bug: 875201 Change-Id: I54ab33180f4d8ff49718b502bdea10740195c45c | 18 October 2018, 15:50:21 UTC |
refs/heads/chromium-export-cl-1291229 | ec12e6a | NEL: Add test cases for redirects/cache Browsers should create NEL reports for all steps of a redirect chain, and should NOT create reports when requests are serviced out of the local cache. (Though if the cached response requires validation, the validation request DOES get reported.) This patch adds WPT test cases for all of these examples. In Chrome, these tests are all currently failing, because we are generating NEL reports in the wrong place in the //net stack. Bug: 895823 Change-Id: If52119acc3a03cc44a4a889eaf642ef6468241e3 | 19 October 2018, 12:57:35 UTC |
refs/heads/chromium-export-cl-1293050 | e1321f5 | Add SameObject attribute to navigator xr with wpt Navigator's xr attribute in third_party/blink/renderer/modules/vr/navigator_vr.idl doesn't contain SameObject attribute unlike WebXR Device API spec. I added SameObject wpt on Navigator and its test is passed. Bug: 897486 Change-Id: I38d3826f200f224107391625499ee270889c2ec2 | 22 October 2018, 16:31:16 UTC |
refs/heads/chromium-export-cl-1293256 | 1497ab6 | wpt: move tests for value/min/max/low/high/optimum IDL attributes of METER Since reflection-forms.html does not folow the spec of METER element, we should update the reflection-forms.html for METER element. Also, IDL attributes of METER are not reflect attribute, I decide to move tests of unset attribute into the-meter-element/meter.html. Bug: 894661 Change-Id: Ic002406f4da84d246cf854d0550ee4f7a1d7cf20 | 26 October 2018, 10:12:54 UTC |
refs/heads/chromium-export-cl-1294874 | 23f93ed | Add wpt test for font feature precedence for 'dlig'. The change to Lato-Medium is to add a mapping from U+fb06 to the 'st' ligature glyph, which was previously only accessible through the 'dlig' feature. This tests https://github.com/w3c/csswg-drafts/pull/3430 Change-Id: Iaf0ff1f99f3ec282bacb0e3cb44ede27d47fac5a | 11 December 2018, 21:09:19 UTC |
refs/heads/chromium-export-cl-1295618 | 12b6163 | Change <area> element's style to display:none Element <area>'s default visiblity is hidden[display: none] per spec, but [display: inline] currently. SPEC: https://html.spec.whatwg.org/multipage/rendering.html#hidden-elements Bug: 689361 Change-Id: I3d4c38713e0a9f3e879815dadbc1cf8a212c0d7a | 28 November 2018, 11:15:33 UTC |
refs/heads/chromium-export-cl-1306337 | 27f4433 | Check canvas color correction pipeline to be HDR This change list adds the proper layout test to examine if the out of range (HDR) sRGB values are preserved in color managed canvas code. Bug:732904 Change-Id: Id0b01050b8dd53ef0e5958e05b0a5e30ffd093ca | 06 December 2018, 19:35:02 UTC |
refs/heads/chromium-export-cl-1306967 | 4becbdb | ServiceWorker: Add the WPT tests to catch an error when fetch a script. This tests makes sure to catch a network error when fetch a script. Implementing CL: https://chromium-review.googlesource.com/c/chromium/src/+/1304274 Bug: 824647 Change-Id: I3d7b3df7ae6a6ca39cd3d1dbf4e8e86817a9ffe6 | 30 October 2018, 09:19:07 UTC |
refs/heads/chromium-export-cl-1319616 | 8d877aa | Recognize 'optimizespeed' and 'optimizequality' for 'image-rendering' In css_properties.json5 we list these keywords but we didn't recognize them when parsing. Synchronize the two lists. CSS Images L3 says the following about these two values: "This property previously accepted the values optimizeSpeed and optimizeQuality. These are now deprecated; a user agent must accept them as valid values but must treat them as having the same behavior as pixelated and auto respectively, and authors must not use them." (https://drafts.csswg.org/css-images-3/#propdef-image-rendering) Bug: 901669 Change-Id: Iecc4f38e711b14fa7b6fd4beb550abdaf99d5b46 | 07 November 2018, 09:00:38 UTC |
refs/heads/chromium-export-cl-1322260 | d21dccb | Let css keyframes name accept string A @keyframes block has a name given by the <custom-ident> or <string> in its prelude. The two syntaxes are equivalent in functionality; the name is the value of the ident or string[1]. [1]: https://drafts.csswg.org/css-animations/#typedef-keyframes-name Bug: 882286 Change-Id: I2e9531beb17e38699dd0c535ae03f0bb16638227 | 08 November 2018, 07:44:53 UTC |
refs/heads/chromium-export-cl-1338461 | 7da4e3f | WIP to add native YUV support for lossy WebP without alpha in the GPU rasterization non-OOPR case. For RGBX, the typical path is WEBPImageDecoder::DecodeSingleFrame WEBPImageDecoder::Decode ImageDecoder::DecodeFrameBufferAtIndex ImageDecoderWrapper::Decode ImageFrameGenerator::DecodeandScale DecodingImageGenerator::GetPixels PainImage::DecodeFromGenerator PaintImage::Decode GpuImageDecodeCache::DrawAndScaleImage GpuImageDecodeCache::DecodeImageIfNecessary GpuImageDecodeCache::DecodeImageInTask Now for YUV, the typical path will become WEBPImageDecoder::DecodeSingleFrameToYUV() WEBPImageDecoder::DecodeSingleFrame() WEBPImageDecoder::Decode() WEBPImageDecoder::DecodeToYUV() WEBPImageDecoder::DecodeToYUV() ImageFrameGenerator::DecodeToYUV() DecodingImageGenerator::GetYUVA8Planes() PaintImage::DecodeFromGenerator() PaintImage::Decode() GpuImageDecodeCache::DrawAndScaleImage() GpuImageDecodeCache::DecodeImageIfNecessary() GpuImageDecodeCache::DecodeImageInTask() Change-Id: I3bc7b5470b240e11eebd90d7b4a2ede04d5b3d93 | 21 December 2018, 02:41:02 UTC |
refs/heads/chromium-export-cl-1340567 | eb313d6 | Snap after layout change As the spec mentions, if the content or layout of the document changes (e.g. content is added, moved, deleted, resized) such that the content of a snapport changes, the UA must re-evaluate the resulting scroll position, and re-snap if required. If the scroll container was snapped before the content change and that same snap position still exists (e.g. its associated element was not deleted), the scroll container must be re-snapped to that same snap position after the content change. This patch adds SnapForAllContainers() right after UpdateAllSnapContainerData() during PerformPostLayoutTasks() and adds tests for that. Some scroll snap tests assume that a scroller's initial scroll position is (0,0) even with one and only non-zero snap position. This assumption won't hold true after this change since the scroller would snap at the snap position at the beginning. So we add an anchor snap element at (0,0) to ensure the scroller always starts at (0,0) for the tests. This patch also fixes a bug in VisualViewport::MaximumScrollOffset(). Bug: 866127 Change-Id: Ib050ef91b04d4391a846eda3dbf667e4d412cb08 | 19 December 2018, 15:38:18 UTC |
refs/heads/chromium-export-cl-1344261 | d6ebd71 | Remove the multiview attribute from XRWebGLLayer Bug: 907148 Change-Id: Ie701648b5c79a3fcb63b4ec5d11f30a0db57ec16 | 20 November 2018, 17:24:18 UTC |
refs/heads/chromium-export-cl-1345090 | 374f409 | Async Clipboard: ReadImage/WriteImage Proof of Concept (experimental). API Goal: - Async Clipboard functionality to ReadImage() and WriteImage() - JS side will send Blobs with MIME type "image/png" - Images will be "safe", as we'll re-encode to strip potentially dangerous metadata and protect from arbitrary code execution. Only image types supported by the Skia library will be supported for this initial POC. - First release into experimental with this ReadImage()/WriteImage() - Later integrate into Read()/Write() and remove ReadImage()/WriteImage() Features: - FileReader used to read Blobs asynchronously. - Image Decoding on a separate thread. - Image read/write accomplished, where Images are represented by Blobs holding MIME type "image/png" - Manual test requires user to copy the green image and then refresh. Future work: - Image encoding on a separate thread as well. - Improve test to be automatic instead of manual, and to not require user copy. - Consider use of ImageBitmapFactory for increased types available when accessing the async clipboard, and reduction in potentially duplicate code, at the cost of a potential for increased complexity. Bug: 150835 Change-Id: I4380116642b4720bb9773185eb2e28757688c5ec | 22 December 2018, 17:56:00 UTC |
refs/heads/chromium-export-cl-1351490 | f941f71 | Idle Detection API: Prototype https://github.com/inexorabletash/idle-detection Just a webification of the chrome.idle.* APIs. Does not yet include: - permission checks - android integration - final API Bug: 878979 Change-Id: Ie45b34d5c29e04dfe65f8ff5127c6be34850d68f | 18 December 2018, 20:56:38 UTC |
refs/heads/chromium-export-cl-1355923 | afa7bc9 | [css-contain][css-grid] Support size containment on grid containers We were always sizing grid containers with "contain: size" as 0x0. However this is wrong as discussed on the following GitHub issue: https://github.com/w3c/csswg-drafts/issues/2804 To do this we add a new method GridContainerSizeWithoutItems() that returns the size based only on the grid container style. As it should has the same size than the same grid container without contents. For widths the change is pretty simple in LayoutGrid::ComputeIntrinsicLogicalWidths(). For heights we don't have a phase to compute the intrinsic size so apart from the changes in LayoutGrid::UpdateBlockLayout() we need some extra checks in LayoutBox too. There is a minor change in test contain-size-grid-002.html, because after this patch it was not passing yet due to some overflow that is unrelated to the purpose of the test. BUG=855261 TEST=external/wpt/css/css-contain/contain-size-grid-002.html TEST=external/wpt/css/css-contain/contain-size-grid-003.html Change-Id: I4d0fd417183068518070721afde84efdbfe1fcb4 | 30 November 2018, 22:57:59 UTC |
refs/heads/chromium-export-cl-1356966 | a3f9298 | Revert "Enforce active lifetime of XRFrame objects" This reverts commit ff451c7bfc7010d9bea16d7b7036be389afa1db5. Reason for revert: Obvious culprit for WebXrVrBrowserTestStandard.TestPresentationPoses failures on the GPU FYI waterfall Windows builders. Original change's description: > Enforce active lifetime of XRFrame objects > > Deactivates XRFrame objects once the relevant callback returns, which > causes future calls to the object methods to throw an exception. > > Bug: 906842 > Change-Id: I192453f9f23b1b4d9a404daca352ab80a5393753 > Reviewed-on: https://chromium-review.googlesource.com/c/1343065 > Commit-Queue: Brandon Jones <bajones@chromium.org> > Reviewed-by: Klaus Weidner <klausw@chromium.org> > Cr-Commit-Position: refs/heads/master@{#612887} TBR=bajones@chromium.org,klausw@chromium.org # Not skipping CQ checks because original CL landed > 1 day ago. Bug: 906842 Change-Id: I3ad87a50a85dffb6eb3b20a6230eee607bf46f0e | 03 December 2018, 10:18:30 UTC |
refs/heads/chromium-export-cl-1363178 | 7d92dbd | Trusted Types: URLs should be USVString, not DOMString. This patch adjusts `ScriptURLString` to be a union including `USVString`, not `DOMString`. The advice in [WebIDL][1] isn't exactly clear, but it boils down to @domenic's notes in [heycam/webidl#84][2] and [w3ctag/design-principles#93][3]. Long story short, URLs are `USVString`. This patch adjusts our implementation to match. [1]: https://heycam.github.io/webidl/#idl-USVString [2]: https://github.com/heycam/webidl/issues/84#issuecomment-300857185 [3]: https://github.com/w3ctag/design-principles/issues/93#issuecomment-379816152 Change-Id: I9bf1240b421287d7d9c291b13d887ca981a66231 | 06 December 2018, 10:19:53 UTC |
refs/heads/chromium-export-cl-1363848 | 6881c71 | [CI] Use LayoutBox::ShouldClipOverflow() for LayoutReplaced Previously video with object-fit:none or object-fit:cover was not clipped when the contents overflowed the replaced content box. 1. Combine overflow handling of LayoutReplaced subclasses and LayoutBox by overriding ComputeShouldClipOverflow() method. Previously the similar logic was in NeedsOverflowClipForReplacedContents() in paint_property_tree_builder.cc. 2. Remove the dependency of image overflow clip on compositing state. 3. ReplacedPainter will emit paint chunk with overflow clip depending on paint property only. No longer depends on other conditions. This applies to all replaced elements. Bug: 912357 Change-Id: Ie173551aab723d9bb28a2057af78a579c93b1e51 | 20 December 2018, 18:18:36 UTC |
refs/heads/chromium-export-cl-1363971 | 6a33b24 | [LayoutNG] New value 'break-spaces' for the white-space property The 'white-space' CSS propery has a new value 'break-spaces' which makes any sequence of preserved to take space and can be broken at any position. https://www.w3.org/TR/css-text-3/#valdef-white-space-break-spaces Bug: 767634 Change-Id: Id00599c0f113f5895085a8cf243155b4c84455ec | 18 December 2018, 22:50:09 UTC |
refs/heads/chromium-export-cl-1365235 | 9ac1564 | [Picture-in-Picture] Don't require user gesture if capturing user media https://github.com/WICG/picture-in-picture/pull/109 Change-Id: Ib8fab10b713d9ed7df1147045a75abfc974e866b | 06 December 2018, 10:51:49 UTC |
refs/heads/chromium-export-cl-1365984 | 7196fc9 | [LayoutNG] Remove NGLayoutResult::PositionedFloats. This patch does a few (related) things: 1) Makes floats children of their appropriate inline-level container. Previously we weren't doing this, which was making OOF positioned children skip their correct containing block. Now they'll correctly get captured, during NGContainerFragmentBuilder::AddChild. 2) Now that float children are in the "correct" position within the fragment tree, we can now paint them. This changes the ng_box_fragment_painter.cc to paint inline-level floats and makes sure we set the NGPaintFragment correctly. 3) Lots of subtle painting layer changes. As floats are now in the "correct" fragment tree position, self-painting layers of inline-level elements (spans, etc) need to know that they are able to paint floats, determine position correctly, etc. The major behaviour changes are: <div> text <span style="position: relative;">abc <div style="float: left;"> <div style="position: absolute;"></div> </div> def</span> </div> The OOF positioned child will be contained within the <span> <div> <span style="position: relative; top: 100px;"> <div style="float: left;"></div></span> </div> The float will also shift by 100px, which previously wasn't the case. Additionally filter: blur(2px); etc, will be applied to the float. Change-Id: I61a2638cb1472f1a22f5994688d69fb6fffb8b88 | 22 December 2018, 23:06:09 UTC |
refs/heads/chromium-export-cl-1366318 | b6110e0 | WebBluetooth Scanning API Spec: https://webbluetoothcg.github.io/web-bluetooth/scanning.html Design Doc: https://docs.google.com/document/d/1YdjdjQxl7acSQ3EeuMVroLfVKuoGjrtig6FbjvPQ7pg/edit#heading=h.xzptrog8pyxf Bug: 897312 Binary-Size: Increase needed to implement Web Bluetooth Scanning. Change-Id: Ib8aa1f213a6864609facba0f5964a7e09050aecd Reviewed-on: https://chromium-review.googlesource.com/c/1366318 Commit-Queue: Doug Turner <dougt@chromium.org> Reviewed-by: Daniel Cheng <dcheng@chromium.org> Reviewed-by: Philip Jägenstedt <foolip@chromium.org> Reviewed-by: Reilly Grant <reillyg@chromium.org> Reviewed-by: Kinuko Yasuda <kinuko@chromium.org> Reviewed-by: Scott Violet <sky@chromium.org> Cr-Commit-Position: refs/heads/master@{#618076} | 20 December 2018, 02:04:09 UTC |
refs/heads/chromium-export-cl-1366767 | e016933 | Move visual overflow computation into the descendant-dependent tree walk. Visual overflow is different than layout overflow, and is used in: * Pre-CAP (aka SPv1) compositing (to determine bounds of composited layers) * Paint/raster invalidation (to determine whether drawn areas on the screen change). Tested: There were multiple layout or unit tests which failed due to the absence of each of the parts of this CL. We appear to have quite good test coverage of this logic. Bug: 892347 Change-Id: I1fadc5d600bda50992d9aba203319ed2d5059aac | 24 December 2018, 21:55:55 UTC |
refs/heads/chromium-export-cl-1367910 | 1ed5ba2 | [WebLocks]: Modifying weblocks algos to be O(1) The behaviour of the request/release operations of web locks are modified to be O(1) instead of their currently O(n) worst case runtime. Additionally the query-order wpt is modified to reflect the new spec requirement that the state returned by navigator.locks.query need only respect ordering for requested locks per resource. Bug: 913014 Change-Id: I819f8c27c995cb698a7c8b2c75ee80d32c744f07 Spec: https://wicg.github.io/web-locks/#algorithms | 14 December 2018, 23:57:08 UTC |
refs/heads/chromium-export-cl-1368826 | 66f1ce9 | [EventTiming] Ship First Input Timing Intent to ship: https://groups.google.com/a/chromium.org/forum/?utm_medium=email&utm_source=footer#!msg/blink-dev/LN92hkaKCzw/Y-ryVxu5AgAJ Bug: 823744, 841224 Change-Id: Ic4d3d67facd0d717b929b9798650eb502057fea6 | 12 December 2018, 18:54:14 UTC |
refs/heads/chromium-export-cl-1371186 | d9f55b3 | MojoInterfaceInterceptor tests for the IdleManager. starts writing a mojo interceptor Change-Id: Iccbd6a394ec8bb2ac3e56411d7ecb5f25c50b5d5 | 15 December 2018, 00:47:49 UTC |
refs/heads/chromium-export-cl-1372556 | aa5f597 | Snap the snapped_dest_rect for backgrounds For some reason the snapped_dest_rect for background image geometry calculations was not actually snapped. Snapping it fixes cases where we were not filling to the border without regressing any tests (in fact improving some). Probably fixes most of the background bugs resulting from the previous big refactoring. R=fmalita@chromium.org BUG=912423,889426 Change-Id: Ieba65130e7f8c0e95db7375ab3a8188bfc78d8d5 Reviewed-on: https://chromium-review.googlesource.com/c/1372556 Commit-Queue: Stephen Chenney <schenney@chromium.org> Reviewed-by: Florin Malita <fmalita@chromium.org> Cr-Commit-Position: refs/heads/master@{#616014} | 12 December 2018, 20:34:28 UTC |
refs/heads/chromium-export-cl-1373848 | 40aed2b | history.back() called twice must not navigate back twice. history.back() navigates back relatively to the current document. So calling it twice should give the same result as if it was called once. This fixes bug 869710. The specification says the following: --- If there is an ongoing attempt to navigate specified browsing context that has not yet matured (i.e. it has not passed the point of making its Document the active document), then cancel that attempt to navigate the browsing context. --- Specification: https://html.spec.whatwg.org/multipage/history.html#traverse-the-history-by-a-delta Bug: 869710 Change-Id: I7a9b451e11cf298919b446d3000cbd99bf672c18 | 17 December 2018, 08:59:26 UTC |
refs/heads/chromium-export-cl-1374113 | fe38104 | Support multiple mouse pointers inputs In order to support multiple mouse pointers, we need to save all the mouse actions into a two dimensional arrays, and combine them into a single array with the correct order to dispatch the events. We should dispatch events by the order of ticks, all the events in the same tick should be dispatched together. Bug: chromedriver:1897 Change-Id: If13c98285faa19286f6ca3a919a0f8ac46c91c73 | 15 December 2018, 01:20:34 UTC |
refs/heads/chromium-export-cl-1374988 | 7f3f384 | Regression test for crbug.com/626744 Bug: 626744 Change-Id: I4244c3f0316ceb55bc69655212f6677345b0a6cc Reviewed-on: https://chromium-review.googlesource.com/c/1374988 Reviewed-by: Stephen Chenney <schenney@chromium.org> Commit-Queue: Fredrik Söderquist <fs@opera.com> Cr-Commit-Position: refs/heads/master@{#616311} | 13 December 2018, 14:54:19 UTC |
refs/heads/chromium-export-cl-1375714 | 30b74c4 | Trusted Types: Store TrustedURL and TrustedScriptURL contents as string Bug: 739170, 913180 Change-Id: I01391891d89aeb55e387059ed4c4a4b92c6dcd7b | 14 December 2018, 11:04:45 UTC |
refs/heads/chromium-export-cl-1376317 | 2b8a5cc | Revert "Implement correct opacity and filter handling within backdrop-filter" This reverts commit 7e7e4790f28e6339dd2de044e851b43de5028cbc. Reason for revert: This broke the background blur of the ChromeOS launcher Original change's description: > Implement correct opacity and filter handling within backdrop-filter > > Prior to this CL, opacity and filters applied to the backdrop-filtered > element did not get applied to the backdrop-filtered backdrop. Now, the > opacity and filters get applied directly to the backdrop-filtered > image directly, which is an approximation but should be very close. > > Also fixed up a few of the backdrop-filter tests to account for this, > and removed a few lines from TestExpectations that now work. > > Bug: 497522, 524689 > Change-Id: If671bdf65ac0ae04cf175143eb8949ff4be0837b > Reviewed-on: https://chromium-review.googlesource.com/c/1357626 > Reviewed-by: vmpstr <vmpstr@chromium.org> > Commit-Queue: Mason Freed <masonfreed@chromium.org> > Cr-Commit-Position: refs/heads/master@{#614383} TBR=vmpstr@chromium.org,masonfreed@chromium.org # Not skipping CQ checks because original CL landed > 1 day ago. Bug: 497522, 524689 Change-Id: Ia9174298ca78dd8fd3766cabc2e6fb0b74659e40 | 13 December 2018, 18:21:43 UTC |
refs/heads/chromium-export-cl-1377129 | 31de4a6 | [UserTiming] Fix mark-measure-return-null test The test was incorrect per L2 spec: a void method 'returns' an undefined value, not a null object. This CL changes the test to show how to perform feature detection and compares L2 and L3 behavior. Bug: 914441 Change-Id: Id7173b2693d5bd42013dcdd25266d64fa8956e6a Reviewed-on: https://chromium-review.googlesource.com/c/1377129 Commit-Queue: Nicolás Peña Moreno <npm@chromium.org> Reviewed-by: Yoav Weiss <yoav@yoav.ws> Reviewed-by: Timothy Dresser <tdresser@chromium.org> Cr-Commit-Position: refs/heads/master@{#616712} | 14 December 2018, 17:11:11 UTC |
refs/heads/chromium-export-cl-1377711 | 7c998af | Replace XRCoodinateSystem/FrameOfReference with XRSpace/XRReferenceSpace Updated to the newer types described by the latest WebXR Device API spec updates. Bug: 915050 Change-Id: I14f07f922ff2eee046664bd9a6ecf1cf4e954312 | 21 December 2018, 18:45:11 UTC |
refs/heads/chromium-export-cl-1378805 | cc2ddff | Reland "Add quota usage details tests for all other storage backends." This is a reland of 548e7b5e622ad79155ff95ef26738d7a40fe9397 Original change's description: > Add quota usage details tests for all other storage backends. > > IndexedDB test was included in parent CL. > > Bug: 904000 > Test: > Change-Id: Icc3462f13d0dce0197a8ec54f22d5ad794a51292 > Reviewed-on: https://chromium-review.googlesource.com/c/1334659 > Reviewed-by: Victor Costan <pwnall@chromium.org> > Reviewed-by: Joshua Bell <jsbell@chromium.org> > Commit-Queue: Jarryd Goodman <jarrydg@chromium.org> > Cr-Commit-Position: refs/heads/master@{#616398} Bug: 904000 Change-Id: I507ec17eabc8e323a7c6418075d4be4787d4c5df Reviewed-on: https://chromium-review.googlesource.com/c/1378805 Reviewed-by: Joshua Bell <jsbell@chromium.org> Commit-Queue: Jarryd Goodman <jarrydg@chromium.org> Cr-Commit-Position: refs/heads/master@{#617285} | 17 December 2018, 23:45:56 UTC |
refs/heads/chromium-export-cl-1381144 | b5ef37f | Support pen inputs in web platform tests Because pen inputs are very similar to mouse inputs, for now, we use mouse events to represent pen events with pointer type of "pen". Bug:chromedriver:1897 Change-Id: I0c830f65a1f1bdf1bc6d4a5ef042fc333d18d4ca | 20 December 2018, 02:33:33 UTC |
refs/heads/chromium-export-cl-1381240 | 622b6fd | Return resolved-value for ComputedStyle line-height Change getComputedStyle to return the resolved value for line-height per the CSSOM specification [1]. Specifically for 'normal' or invalid values this will return the computed line-height instead of the specified ones. 1) https://drafts.csswg.org/cssom/#resolved-value Bug: 383409 Test: wpt/css/cssom/line-height-resolved-value.html Change-Id: Id58fdc700605779a7f34ccbefb62290e3eea1117 | 18 December 2018, 22:56:53 UTC |
refs/heads/chromium-export-cl-1383297 | 3b558a9 | Add tentative WPT tests for stale while revalidate handling. Add test to ensure that handling the fetch doesn't trigger stale while revalidate loading. Add test to ensure that scripts loaded trigger a stale while revalidate cache hit and resource timing entries are correct. BUG=348877 Change-Id: Ib07b98d0d2595b6b99857161f830343bf7516518 | 18 December 2018, 21:43:49 UTC |
refs/heads/chromium-export-cl-1384639 | 42a9828 | Fix table cell background painting under zoom and min-width Sub-pixel min-width calculations were rounding up always, as that ensures they can contain their contents. However, background image painting always uses a different rounding method (pixel snapping), as does the rest of the rendering pipeline. This causes problems under zoom with auto background size and images chosen to fit the cell exactly. So round the min cell size instead. A half pixel size reduction is highly unlikely to hide content, but gaps in backgrounds are really noticeable. R=fs@opera.com Bug: 873400 Change-Id: I07c52e5f1cf1d00637123aad46da89628ac87ff9 | 20 December 2018, 17:40:54 UTC |