convert js-test.js to testharness.js in LayoutTest/scrollbars/* folder |
||||||
Issue descriptionconvert js-test.js to testharness.js in LayoutTest/scrollbars/* folder
,
Aug 11 2016
The following revision refers to this bug: https://chromium.googlesource.com/chromium/src.git/+/54a8f9a2dd2cdc8768220d60e6dd369ff05378c5 commit 54a8f9a2dd2cdc8768220d60e6dd369ff05378c5 Author: sataya.m <sataya.m@samsung.com> Date: Thu Aug 11 08:04:15 2016 convert LayoutTest/scrollbars/* js-test.js tests to testharness.js based tests. BUG= 636285 Review-Url: https://codereview.chromium.org/2234773003 Cr-Commit-Position: refs/heads/master@{#411289} [delete] https://crrev.com/aeb28a9fcae8f3b72990c192761ae3f65796f693/third_party/WebKit/LayoutTests/scrollbars/drag-rtl-resizer-expected.txt [modify] https://crrev.com/54a8f9a2dd2cdc8768220d60e6dd369ff05378c5/third_party/WebKit/LayoutTests/scrollbars/drag-rtl-resizer.html [delete] https://crrev.com/aeb28a9fcae8f3b72990c192761ae3f65796f693/third_party/WebKit/LayoutTests/scrollbars/scrollable-iframe-click-gets-focus-expected.txt [modify] https://crrev.com/54a8f9a2dd2cdc8768220d60e6dd369ff05378c5/third_party/WebKit/LayoutTests/scrollbars/scrollable-iframe-click-gets-focus.html
,
Aug 11 2016
The following revision refers to this bug: https://chromium.googlesource.com/chromium/src.git/+/3bf34c3dac6f4d91594c22c2bec6bb4cf63a65c1 commit 3bf34c3dac6f4d91594c22c2bec6bb4cf63a65c1 Author: sataya.m <sataya.m@samsung.com> Date: Thu Aug 11 10:17:53 2016 convert LayoutTest/scrollbars/rtl/* js-test.js tests to testharness.js based tests. BUG= 636285 Review-Url: https://codereview.chromium.org/2229393002 Cr-Commit-Position: refs/heads/master@{#411295} [delete] https://crrev.com/983625d272c1b35a6297dd02c51fd1ee0fce3230/third_party/WebKit/LayoutTests/scrollbars/rtl/div-absolute-expected.txt [modify] https://crrev.com/3bf34c3dac6f4d91594c22c2bec6bb4cf63a65c1/third_party/WebKit/LayoutTests/scrollbars/rtl/div-absolute.html [delete] https://crrev.com/983625d272c1b35a6297dd02c51fd1ee0fce3230/third_party/WebKit/LayoutTests/scrollbars/rtl/div-horizontal-expected.txt [delete] https://crrev.com/983625d272c1b35a6297dd02c51fd1ee0fce3230/third_party/WebKit/LayoutTests/scrollbars/rtl/div-horizontal-with-vertical-scrollbar-expected.txt [modify] https://crrev.com/3bf34c3dac6f4d91594c22c2bec6bb4cf63a65c1/third_party/WebKit/LayoutTests/scrollbars/rtl/div-horizontal-with-vertical-scrollbar.html [modify] https://crrev.com/3bf34c3dac6f4d91594c22c2bec6bb4cf63a65c1/third_party/WebKit/LayoutTests/scrollbars/rtl/div-horizontal.html [delete] https://crrev.com/983625d272c1b35a6297dd02c51fd1ee0fce3230/third_party/WebKit/LayoutTests/scrollbars/rtl/div-vertical-expected.txt [modify] https://crrev.com/3bf34c3dac6f4d91594c22c2bec6bb4cf63a65c1/third_party/WebKit/LayoutTests/scrollbars/rtl/div-vertical.html [delete] https://crrev.com/983625d272c1b35a6297dd02c51fd1ee0fce3230/third_party/WebKit/LayoutTests/virtual/prefer_compositing_to_lcd_text/scrollbars/rtl/div-absolute-expected.txt
,
Aug 11 2016
@rbyers / @foolip: Do we want to be blanket-converting layout tests to testharness.js? It makes sense to do this for tests that we intend to upstream, but has questionable value otherwise, especially for tests that use internals / eventSender. I think there is significant risk of introducing bugs or flakiness in the conversion, as well as some cost to reviewing the patches, so we should perhaps discuss whether this is a road we want to go down.
,
Aug 12 2016
I agree that there is risk, and that if it isn't a step towards upstreaming to w-p-t then it's not very useful. Of course, even tests using internals sometimes become more readable by converting to testharness.js, so in those cases if the CL author and reviewer wants to spend the time, that seems OK. For a big conversion project, I think it'd make the most sense to do it in the context of actually upstreaming to web-platform-tests, in a cycle roughly like this: 1. Convert a group of tests locally. 2. Upstream those to w-p-t 3. Import them and verify that they're identical to local and passing reliably 4. Delete the local tests Having that focus will naturally avoid the cases where we're not able to upstream.
,
Aug 18 2016
gentle ping for conclusion !!!
,
Aug 18 2016
I think that it's the person who would review these changes that should decide on whether it's worthwhile. Who is that?
,
Aug 18 2016
I have been receiving the reviews so far, which prompted the concerns I raised in #4. The process in #5 sounds good to me but I'm afraid I don't have time to properly review all of the patches for a large conversion effort.
,
Aug 19 2016
If you guys are ok, i can review them internally before submiting for final review.
,
Aug 22 2016
Skobes@, can I defer these test to srirama/foolip/fs ?
,
Aug 22 2016
I agree with follip@ that it's unlikely to be worth the cost unless we're also upstreaming the tests to web-platform-tests (adding coverage where there wasn't previously coverage there). In that case, the non-trivial review will come as part of the upstreaming process - web-platform-test reviewers will want to verify that you're actually testing standard behavior, not just blindly copying blink implementation tests for things that may or may not be standardized. "scrollbars" in particular might be a poor candidate for this - scrollbar behavior tends not to be well standardized. But perhaps there are some tests that can be upstreamed?
,
Dec 6 2016
,
Dec 6 2016
,
Mar 7 2017
,
Sep 13 2017
discarding this work. Future tests will be covered under testharness. |
||||||
►
Sign in to add a comment |
||||||
Comment 1 by bugdroid1@chromium.org
, Aug 11 2016