Bug 92318
Summary: | [Chromium] WebViewTest.AutoResizeXXX tests are failing | ||
---|---|---|---|
Product: | WebKit | Reporter: | Andrew Wilson <atwilson> |
Component: | Tools / Tests | Assignee: | Dmitry Titov <dimich> |
Status: | RESOLVED WONTFIX | ||
Severity: | Normal | CC: | aa, abarth, allan.jensen, dimich, levin, pkasting |
Priority: | P2 | ||
Version: | 528+ (Nightly build) | ||
Hardware: | Unspecified | ||
OS: | Unspecified |
Andrew Wilson
I'm seeing the following error output for this test in the chromium webkit_unit_tests:
WebViewTest.AutoResizeInBetweenSizes:
/Volumes/data/b/build/slave/webkit-mac-latest-dbg/build/src/third_party/WebKit/Source/WebKit/chromium/tests/WebViewTest.cpp:213: Failure
Value of: client.testData().width()
Actual: 90
Expected: expectedWidth
Which is: 100
/Volumes/data/b/build/slave/webkit-mac-latest-dbg/build/src/third_party/WebKit/Source/WebKit/chromium/tests/WebViewTest.cpp:214: Failure
Value of: client.testData().height()
Actual: 95
Expected: expectedHeight
Which is: 200
It started happening somewhere between 123625 and 123634 (http://build.chromium.org/p/chromium.webkit/builders/Webkit%20Mac10.6%20%28dbg%29/builds/11645) and I'm suspicious of r123626.
Attachments | ||
---|---|---|
Add attachment proposed patch, testcase, etc. |
Allan Sandfeld Jensen
I can't see how r123626 could have that effect. Have you tried running it with r123626 reverted?
Somehow it looks like the minSize is returned now and pageWidth and pageHeight having no effect.
David Levin
This will likely affect places that use autoresize.
Peter Kasting
Disabling this test did not fix the problem. Instead the next test to get run started failing with similar output:
[ RUN ] WebViewTest.AutoResizeOverflowSizes
/Volumes/data/b/build/slave/webkit-mac-latest-dbg/build/src/third_party/WebKit/Source/WebKit/chromium/tests/WebViewTest.cpp:213: Failure
Value of: client.testData().width()
Actual: 90
Expected: expectedWidth
Which is: 200
/Volumes/data/b/build/slave/webkit-mac-latest-dbg/build/src/third_party/WebKit/Source/WebKit/chromium/tests/WebViewTest.cpp:214: Failure
Value of: client.testData().height()
Actual: 95
Expected: expectedHeight
Which is: 300
/Volumes/data/b/build/slave/webkit-mac-latest-dbg/build/src/third_party/WebKit/Source/WebKit/chromium/tests/WebViewTest.cpp:215: Failure
Value of: client.testData().horizontalScrollbarState()
Actual: 0
Expected: expectedHorizontalState
Which is: 1
/Volumes/data/b/build/slave/webkit-mac-latest-dbg/build/src/third_party/WebKit/Source/WebKit/chromium/tests/WebViewTest.cpp:216: Failure
Value of: client.testData().verticalScrollbarState()
Actual: 0
Expected: expectedVerticalState
Which is: 1
[ FAILED ] WebViewTest.AutoResizeOverflowSizes (277 ms)
So it may be that the issue is that the previous test (WebViewTest.AutoResizeFixedHeightAndWidthOverflow) leaves something in a bad state?
I'll try disabling this test as well and see what happens.
Peter Kasting
Disabled in r123957.
Peter Kasting
Sure enough, that merely made the next test start failing.
In r123963 I've re-enabled both these and instead disabled the test before them in hopes of figuring out whether it's leaving something in a bad state.
levin, it looks like you were the last one to touch these tests in a significant way, could you perhaps investigate?
Peter Kasting
Apparently no matter what I do, we successfully run three of the AutoResize tests, then the next one fails. So I just disabled all the tests after the first three :(.
David Levin
Sorry, I'm booked with other work.
Peter Kasting
Well, don't assign to me, I'm just sheriff.
I don't know who else can look at this.
Peter Kasting
dimich, this potentially affects notifications and extension popup autosizing, might you be able to find an owner?
Stephen Chenney
Marked LayoutTest bugs, bugs with Chromium IDs, and some others as WontFix. Test failure bugs still are trackable via TestExpectations or disabled unit tests.