Hi Eric,

On Jan 29, 2008, at 4:34 PM, Eric Seidel wrote:

Hixie is nearly done with his Acid3 test. He's said he has space for 2 more tests, but is otherwise "ready".

Acid3 has certainly found some bugs in WebKit. WebKit has also found bugs in Acid3 (I'm sure we'll find more).

Thanks for classifying the failures. This is really useful! Standards compliance and long-term interoperability are important to the WebKit project. Acid tests are a fun and productive way to update corners of our web standards support that might otherwise not get enough love, with some confidence that other browsers will also improve.

Working on Acid3 fixes is a great thing to do, and a fun way to get into WebKit hacking for those who have been thinking about dipping their toes in.

 - Maciej

Here is a list of all of the failures we currently have, and their associated Bugzilla bugs:

Failed 35 tests.
Test 0: expected: pre-wrap, got: normal - found unexpected computed style
4812 - :last-child bug
Test 1: method [object NodeIterator].nextNode() didn't forward exception
4714 - NodeIterator forwarding exceptions
Test 2: expected: [object HTMLElement], got: null - expectation 13 failed Test 4: expected: null, got: [object HTMLIFrameElement] - expectation 21 failed Test 6: expected: [object HTMLTitleElement], got: null - failed to handle regrafting correctly
3492 - TreeWalker brokenness
Test 8: Null value
16766 - Range doesn't update on tree modifications
Test 11: when trying to surround two halves of comment: wrong exception raised; code = 3
16749 - Range.surroundContents throws wrong exception
Test 13: collapsed is wrong after deletion
16766 - Range doesn't update on tree modifications
Test 18: expected: 10, got: 8 - DOCTYPE nodeType wrong
12751 - DocType nodes aren't part of the document
Test 23: no exception for createElementNS('null', ':div')
16833 - createElementNS doesn't throw enough exceptions
Test 25: failed to raise exception
16693 - createDocumentType does not throw enough exceptions
Test 26: e1 - parent element doesn't exist after looping
Test 27: e1 - parent element doesn't exist after waiting
17076 - Document tree does not survive long enough
Test 35: expected: 0, got: 1 - :first-child still applies to element that was previously a first child
Test 36: expected: 1, got: 0 - last child did not match :last-child
Test 37: expected: 1, got: 0 - :only-child did not match only child
5468 - broken :first-child, :last-child, :only-child selectors
Test 38: expected: 0, got: 1 - removing all children didn't make the element match :empty
11387 - :empty applies to all elements
Test 39: expected: 1, got: 0 - :nth-child(odd) failed with child 0
6248 - implement :nth-child selectors
Test 40: expected: 1, got: 0 - part 2:7
5468 - broken :first-child, :last-child, :only-child selectors
Test 42: expected: 1, got: 0 - rule did not start matching after change
11384 - CSS2/3 child selectors aren't dynamic
Test 44: expected: 0, got: 1 - misparsed selectors
16751 - html* .test {} is parsed as html .test
Test 59: expected: submit, got:  - <button> doesn't have type=submit
16798 - HTMLButtonElement.type should default to "submit"
Test 64: expected: ./test.html, got: test.html - object elements didn't resolve URIs correctly
16799 - HTMLObjectElement.data should cannonicalize urls
Test 66: expected: null, got: #text - wrong localName
17060 - Text.localName should be null
Test 69: expected: 33, got: 0 - SVGSVGTextElement.getNumberOfChars() incorrect
17062 - SVGTextElement.getNumberOfChars returns incorrect value
Test 71: UTF-8 encoded XML document with invalid character did not have a well-formedness error
17079 - XML document sent with wrong encoding should fail
Test 72: expected: TITLE (first test), got: TITLE - undefined
17080 - WebKit fails Acid3 HTML parsing test
Test 73: Undefined value
17082 - WebKit fails Acid3 modification of style text block test
Test 74: expected: 10, got: 1 - click event handler called the wrong number of times
17083 - Acid3 expects nested event dispatch support
Test 75: getSVGDocument failed for <iframe> referencing an svg document.
17061 - HTMLIFrameElement should have a getSVGDocument() method
Test 76: Value undefined (result of expression anim.beginElement) is not object. Test 77: expected: 0, got: 100 - Incorrect animVal value after svg animation.
17077 - SVG Animation is broken (and thus turned off)
Test 78: expected: 4776, got: 5550 - getComputedTextLength failed.
17078 - SVGTextElement.getComputedTextLength does not work as expected
Test 86: calling setMilliseconds() with no arguments didn't result in NaN
16753 - Date.set* methods with no arguments should result in NaN
Test 92: Function object's prototype's constructor was ReadOnly
15003 - Function object's prototype's constructor should not be ReadOnly
Elapsed time: 4.03s (debug build)

There are also some visual failures:
16760 - HTML Object elements should not render 404 content
16770 - Acid3 expects :visited styles to update after iframe load
17085 - Acid3 Ahem test font no longer loads properly
17086 - Positioned X is a few pixels off

Bugs discovered during testing, but not required to pass Acid3:
8961 - document.write() does not flush plain text (shift-click on the A to see failure)
14256 - Inspector hits assertion while inspecting image in Acid3
17004 - Assertion clicking on n/100 in Acid3 results

This is certainly not a full list of all the bugs which will need to be fixed before Acid3 will pass 100% in WebKit-based browsers. However, this is all the bugs that I know to be visible at this time.

All Acid3 bugs have Acid3 in the title, you can find them using this query:

Your comments (and help fixing them) are certainly appreciated.


webkit-dev mailing list

webkit-dev mailing list

Reply via email to