64 bugs in Internet Explorer 8 for Windows

1- MSIE 8 requests automatically favicon.ico. Internet Explorer 8 should stop automatically requesting for favicon.ico in the root directory - a file which may not exist - and instead look for (fetch, HTTP request) an icon only and only if it's explicitly linked to from the webpage. The first time a user visits your Web page, Internet Explorer automatically searches for this [favicon.ico] file. Otherwise, when the favicon.ico does not exist, the error.log shows favicon.ico (hundreds of times) as "404 file not found". From a web standards and HTTP perspective, it does not make sense to request a file that has not been explicitly linked to begin with. This bug has been reported by several web authors since 1997!:

Firefox 1.x, Firefox 2.x, Firefox 3, Opera 9.x, Safari 2, Safari 3.0.2, Seamonkey 1.x, Seamonkey 2.x, K-meleon 1.x, Konqueror 3.5.8, Galeon 2, Epiphany 2 all pass this test.

2- Favicon standardisation at Wikipedia, HTML 4 specification, rel attribute value (link-types) and the W3C Quality Assurance tip, How to Add a Favicon to your Site indicate that rel attribute must contain a space-delimited list of link types. Internet Explorer requires the rel attribute value "shortcut icon" when just "icon" is good enough for other browsers. So then, what does "shortcut" rel attribute value actually refers to then? rel="shortcut icon" will not be understood correctly (will be ignored in fact) by conforming web browsers. This issue was also mentioned and explained by Dan Tobias in his Dan's Web Tips: Titles, META Tags, LINK tags, and Search Engine Robots as follows: "MSIE looks only for 'shortcut icon'. Technically, a link attribute with multiple words is saying that the link is of multiple types, with each word being a different type identifier, but I think the Microsoft programmers were just clueless enough to think that spaces could be used within type names, and that 'shortcut icon' is what they think the type is."
'Shortcut icon' instead of 'icon', now that was definitely a simple oversight., from Eric Lawrence, IE program manager, at Jeff Davis' blog, why doesn't the favicon for my site appear in IE7?

Firefox 2.0.0.8, Firefox 3, Opera 9.50, Safari 3.0.3, Seamonkey 1.x, Seamonkey 2.0a1pre, Konqueror 3.5.8, K-meleon 1.x, NS 7.0, NS 7.2 all pass this test.

This bug has been reported at connect's IE beta feedback as bug 370225 by "ShadowChaser".

3- document.getElementById(SelectList[SelectIterator]).add(objOption, null);
add(objOption, null) does not work under Internet Explorer 6, Internet Explorer 7 and Internet Explorer 8. Firefox 1.5, Firefox 2, Firefox 3, Opera 9.27, Opera 9.50, Safari 3.1, Konqueror 3.5.8, Seamonkey 1.x Seamonkey 2.x, NS 7.0, NS 7.2, K-meleon 1.x, Galeon 2.0.4, Epiphany 2.22 all pass this test. This bug has been reported at connect's IE beta feedback as bug 388710.

4- DOM nodeType constant values are not listable, are not accessible Firefox 2.0.0.9, Opera 9.50, Safari 3.0.4, Konqueror 3.5.8, Seamonkey 1.x, Seamonkey 2.x, NS 7.2, K-meleon 1.x, Galeon 2.0.4, Epiphany 2.22 all pass this test. This bug has been reported at connect's IE beta feedback as bug 339307.

5- Support for cssFloat Firefox 2.0.0.9, Opera 9.50, Safari 3.0.4, Konqueror 3.5.8, Hv3 TKHTML alpha 16, Seamonkey 1.x, Seamonkey 2.x, NS 7.2, K-meleon 1.x, Galeon 2.0.4, Epiphany 2.22 all pass this test.

6- Many DOM attributes are wrong This bug has been reported at connect's IE beta feedback as bug 348537.

7- Formatted alternate text is not implemented Alternate text for nonexistent images should be rendered as specified by markup code. If alternate text is to be rendered as inline, then it should inherit from its parent the inheritable properties, including font-size. This bug will still occur even if the Internet Explorer 7 or Internet Explorer 8 user has the Tools/Internet Options.../Advanced tab/Accessibility section/Always expand ALT text for images checkbox checked. Firefox 2.0.0.4, Opera 9.21, Konqueror 3.5.8, Amaya 9.55, Seamonkey 1.x, Seamonkey 2.x, NS 6.2, NS 7.0, NS 7.2, K-meleon 1.x, Galeon 2, Epiphany 2 all pass this test.

8- The value of usemap must match the value of the name attribute of the associated map element. This comes directly from the HTML 3.2 and HTML 4.01 spec. Firefox 2.0.0.5, Safari 3.0.2, Konqueror 3.5.8, NS 6.2, NS 7.0, NS 7.2, Hv3 TKHTML alpha 16, Seamonkey 1.x, Seamonkey 2.x, K-meleon 1.x, Galeon 2, Epiphany 2 all pass this test.

9- Rendering ALT text for images as inline and expanding the image placeholder dimensions as required by the alt text. When an image can not be displayed - for whatever the reasons -, then the alt text should be rendered as inline and image frame/placeholder dimensions should be ignored. Firefox 2.0.0.6, Amaya 9.55, Seamonkey 1.x, Seamonkey 2.x, K-meleon 1.x, Galeon 2, Epiphany 2, NS 6.2, NS 7.0, NS 7.2 all pass this test.

10- Images should be underlined in the following testcases:

Firefox 2.0.0.6, Safari 2.0.4, Safari 3.0.3, Konqueror 3.5.8, Seamonkey 1.x, Seamonkey 2.0a1pre, K-meleon 1.x, Galeon 2, Epiphany 2, NS 7.2 all pass these 3 tests.

This bug 10 has been closed because it turns out it is INVALID for CSS 2.1 while CSS 3 text module (latest Working Draft: 6 March 2007 does not yet identify such expected result.

11- Internet Explorer 6, Internet Explorer 7 and Internet Explorer 8 do not correctly support Number.toFixed() for numbers in the ranges {(-0.94,-0.5], [0.5,0.94)}: Number.toFixed() demo Firefox 2.0.0.8, Opera 9.50, Safari 3.0.3, Konqueror 3.5.x, Seamonkey 1.x, Seamonkey 2.x, K-meleon 1.x, Galeon 2, Epiphany 2, NS 6.1, NS 6.2, NS 7.0, NS 7.2, Hv3 TKHTML alpha 16 all pass this test.

12- Internet Explorer 6, Internet Explorer 7 and Internet Explorer 8 do not support favicon of type image/png (or of other types of image like gif): favicon of type image/png demo. A favicon (or web site icon) image in png (a PNG is an official W3C web standard) or in gif image format are not supported by Internet Explorer for specifying favicon; only the .ico file format for specifying a favicon is supported by Internet Explorer. This is clearly insufficient.
Jeff Davis, IE project manager in his why doesn't the favicon for my site appear in IE7? states "It must actually be a .ico (an Icon) file. Bitmaps, pngs, gifs, etc, will not work.".
How to Add a Shortcut Icon to a Web Page at MSDN states "To create the icon, use an icon editor, such as the one included in Microsoft Visual Studio, and save the icon in the .ico file format. Internet Explorer will ignore icons that are not in .ico format, and will instead display the default Internet Explorer shortcut icon.".
Firefox 2.0.0.8, Firefox 3, Opera 9.50, Safari 3.0.3, Konqueror 3.5.8, Seamonkey 1.x, Seamonkey 2.0a1pre, K-meleon 1.x, NS 7.0, NS 7.2, Hv3 TKHTML alpha 16 all pass this test. This bug has been reported at connect's IE beta feedback as bug 370228 by "ShadowChaser".

13- Image nested inside a label element with a for attribute unexpectedly does not transfert focus to its associated, targeted form control in Internet Explorer 7 and Internet Explorer 8. This bug applies to several form input controls and has been testcase-ed for the following elements: <input type="text">, <textarea> and <input type="checkbox">. Firefox 2.x, Firefox 3, Opera 9.26, Opera 9.50, Safari 3.0.4, Safari 3.1, Konqueror 3.5.8, Seamonkey 1.x, Seamonkey 2.x, NS 7.2, K-meleon 1.x, Galeon 2.0.4, Epiphany 2.22 all pass this test. Credit must go to Batiste Bieler for discovering, reporting this bug and for providing a testcase on this precise bug. This bug has been reported at connect's IE beta feedback as bug 348575.

14- DOM 3 Core textContent support. Firefox 1.5, Firefox 2.x, Firefox 3, Opera 9.26, Opera 9.50, Safari 3.1, Konqueror 4.1, Seamonkey 1.x, Seamonkey 2.x, NS 7.2, K-meleon 1.x, Galeon 2.0.4, Epiphany 2.22 all pass this test. This bug has been reported at connect's IE beta feedback as bug 334438. This bug has been closed by the IE team and has not been reopened.

15- Closing non-script-initiated single-tab windows and tab instances According to MSDN, non-script-initiated single-tab windows and tab instances should not be closed by javascript without first getting an explicit consent from the user via a prompt confirmation window (dialog box). Firefox 2.x, Firefox 3.x, Konqueror 4.5.4, Seamonkey 1.x, Seamonkey 2.x, K-meleon 1.x, Epiphany 2.22, Galeon 2.0.4 all pass this test by refusing to close the window and by notifying so in the error console. This bug has been reported at connect's IE beta feedback as bug 379310.

16- event.offsetX, event.offsetY are mouse coordinates relative to padding-box of target node, not relative to content-box of target node. Internet Explorer 6 and Internet Explorer 7 were calculating event.offsetX and event.offsetY coordinates relative to the padding-box of target node. These coordinate references are also what the W3C CSSOM View Module Reference specify for event.offsetX and event.offsetY. Internet Explorer 6, Internet Explorer 7, Safari 3.1, Konqueror 3.5.8 all pass this test. This is a regression. This bug has been reported at connect's IE beta feedback as bug 362142.

17- getAttribute("style") on an element returns extended form of parsed CSS rules in IE8, not the string of content attribute or string of current textual value As noted by Peter-Paul Koch in his DOM compatibility table for attributes on getAttribute() and as noted by many web authors, when querying getAttribute("style") on a node, Internet Explorer 6, Internet Explorer 7 return an object and not a string. In IE 8 final release, though, things have improved. IE 8 final release will return a string which is the extended form of the textual parsable CSS rule from the author. Browsers currently do not stringnify internally CSS rules the same way: this is particularly true in cases when shorthand form is used. DOM 2 CSS, section 2.3 Interface CSS2Properties states "When dealing with CSS 'shorthand' properties, the shorthand properties should be decomposed into their component longhand properties as appropriate, and when querying for their value, the form returned should be the shortest form exactly equivalent to the declarations made in the ruleset." Firefox 1.5, Firefox 2.x, Firefox 3, Opera 9.50, Safari 3.1, Konqueror 3.5.8, NS 7.0, NS 7.2, Seamonkey 1.x, Seamonkey 2.x, K-meleon 1.x, Galeon 2.0.4, Epiphany 2.22 all pass this test. This bug has been reported at connect's IE beta feedback as bug 427773.

18- rules="all" attribute specification not supported in border-collapse: separate tables There is nothing in the HTML 4.01 reference that prevents or hampers a border-collapse: separate table from rendering table rules around table cells when rules="all" is specified. Internet Explorer 6, Internet Explorer 7, Firefox 2.x, Firefox 3, Opera 9.26, Opera 9.50, Safari 2.0.4, Safari 3.1, Konqueror 3.5.8, Seamonkey 1.x, Seamonkey 2.x, NS 7.2, K-meleon 1.x, Galeon 2.0.4, Epiphany 2.22 all pass this test. This is a regression. This bug has been reported at connect's IE beta feedback as bug 409470. This bug has been closed by the IE team and has not been reopened.

19- An unclosed <p> immediately preceding a form element should be implicitly closed by IE when parsing and building DOM tree. This bug report is very much similar to the implicit closing on an unclosed <p> element when it encounters a <table>. Firefox 2.0.0.13, Opera 9.27, Opera 9.50, Safari 3.1, Konqueror 3.5.8, Amaya 11 build 20081216, Seamonkey 1.x, Seamonkey 2.x, NS 7.2 all pass this test. Credit must go to Jukka Korpela for discovering, figuring out and reporting this bug thanks to his Unclosed <P> before a form webpage testcase. This bug has been reported at connect's IE beta feedback as bug 338580.

20- @import-ing a style sheet for "print" fails in IE 7 and in IE 8. Firefox 2, Firefox 3, Opera 9.27, Opera 9.50, Safari 3.1.1, Konqueror 3.5.8, Seamonkey 1.x, Seamonkey 2.x, NS 7.0, NS 7.2 all pass this test. This bug has been reported at connect's IE beta feedback as bug 354316.

21- Inherited margin with inherit keyword: the parent's computed value should be inherited. Internet Explorer 8 does not use the margin-left computed value of the parent. Firefox 2.0.0.14, Firefox 3, Opera 9.27, Opera 9.50, Safari 3.1.1, Konqueror 3.5.9, Hv3 TKHTML alpha 16, Seamonkey 1.x, Seamonkey 2.x, NS 7.2, K-meleon 1.x, Galeon 2.0.4, Epiphany 2.22 all pass this test. This bug has been reported at connect's IE beta feedback as bug 344787.

22- An HTML comment before doctype declaration and after xml declaration triggers backward-compatible rendering mode in IE 7 and IE 8. This bug has also been reported in A comment before the doctype triggers quirksmode in IE7 by Alan Gresley. Firefox 1.5, Firefox 2, Firefox 3, Opera 9.x, Safari 3.1.2, Safari 4.0, Konqueror 3.5.9, Hv3 TKHTML alpha 16, Amaya 10.0.1, Seamonkey 1.x, Seamonkey 2.x, NS 7.2, K-meleon 1.x, Galeon 2.x, Epiphany 2.x all pass this test. Credit must go to Ingo Turski for discovering and reporting this bug at IE blog on March 6th 2008 and to Dean Edwards for discovering and reporting this bug at IE blog on September 15th 2005. This bug has been reported at connect's IE beta feedback as bug 354956.

23- objOption.defaultSelected = true is not supported by Internet Explorer 6, Internet Explorer 7 and Internet Explorer 8. Mozilla 1.x, Firefox 1.x, Firefox 2.x, Firefox 3, Netscape 7.x, Seamonkey 1.x, Seamonkey 2.x, Safari 2.x, Safari 3.1, Safari 4.0, Konqueror 3.5.9, Opera 8.x, Opera 9.27, Opera 9.50, K-meleon 1.x, Galeon 2.x, Epiphany 2.x all pass this test. This bug has been reported at connect's IE beta feedback as bug 365833.

5 more complex Ian "Hixie" Hickson CSS 2.1 testpages (which all failed in MSIE 8)
link to actual test page IE 7 IE 8 Firefox 3.0 Opera 9.50 Safari 3.1.2 Konq 4.1.3 Other browsers which
pass the test page
link to actual test page IE 7 IE 8 Firefox 3.0 Opera 9.50 Safari 3.1.2 Konq 4.1.3 Other browsers which
pass the test page
24 border-width: inherit Fails Fails Passes Passes Passes Passes
25 Tall table with wide L-shaped floats Fails Fails Passes Passes Passes Passes NS 7.2
26 Stacking order: Floats and Blocks in Inlines Fails Fails Passes Passes Passes Passes NS 7.2
27 Cascade: Ordering test with various import techniques Fails Fails Passes Passes Fails Fails NS 7.2
28 CSS inheritance: border-color Fails Fails Passes Passes Fails Passes NS 7.2, Hv3 TKHTML alpha 16, Safari 4.0

29- Creating and inserting a command button before defining its type fails in Internet Explorer 7 and Internet Explorer 8. Firefox 1.5, Firefox 2, Firefox 3, Opera 9.27, Opera 9.51, Safari 3.1.2, Safari 4.0, Konqueror 4.1, Seamonkey 1.x, Seamonkey 2.x, NS 7.2, K-meleon 1.x, Galeon 2.0.4, Epiphany 2.22 all pass this test.

30- font shorthand and inherit are incorrectly parsed and rendered in Internet Explorer 7 and Internet Explorer 8. "Shorthand properties take a list of subproperty values or the value 'inherit'. One cannot mix 'inherit' with other subproperty values as it would not be possible to specify the subproperty to which 'inherit' applied." coming from CSS 2.1, Appendix C. Changes, C.3.1 Shorthand properties. Firefox 2, Firefox 3, Seamonkey 1.x, Seamonkey 2.x, NS 7.2, K-meleon 1.x, Galeon 2.0.4, Epiphany 2.22 all pass this test. This bug has been reported at connect's IE beta feedback as bug 361182.

31- Incorrect parsing and rendering of font-size when involving system font in Internet Explorer 6, Internet Explorer 7 and Internet Explorer 8. It seems that subsequent CSS declaration can not override one font subproperty of font shorthand when styled with a font system reserved name such as message-box. Nevertheless, CSS 2.1, section 15.8 Shorthand font property: the 'font' property states "System fonts may only be set as a whole; that is, the font family, size, weight, style, etc. are all set at the same time. These [font subproperty] values may then be altered individually if desired." Firefox 2, Firefox 3.0.1, Shiretoko 3.1, Opera 9.52, Safari 3.1.2, Konqueror 3.5.9, Amaya 11 build 20081216, Seamonkey 1.x, Seamonkey 2.x, NS 7.2, K-meleon 1.x, Galeon 2.0.4, Epiphany 2.22.3 all pass this test. This bug has been reported at connect's IE beta feedback as bug 361181.

32- Unquoted font-family names must not contain unescaped parentheses ( ), unescaped square brackets [ ], unescaped curly braces { }, unescaped exclamation marks !, etc. Firefox 2, Firefox 3.0.1, Shiretoko 3.1, Safari 3.1.2, Konqueror 4.1, Seamonkey 1.x, Seamonkey 2.x, NS 7.2, K-meleon 1.x, Galeon 2.0.4, Epiphany 2.22.3 all pass this test. Opera 9.52 fails 4 tests out of 16. This bug has been reported at connect's IE beta feedback as bug 361953.

33- <object> does not support natively PNG image if content-type (MIME type) is not specified; another testcase for bug 389191 Object-as-Image: (...) If no type is given, then Internet Explorer must try to determine what the object is based on any additional attributes provided (classid, codetype, etc.) or solely by the URL if no additional attributes are provided (by getting MIME type information from the server). (quote from HTML Improvements and Acid2: Internet Explorer 8 Beta 1 for Developers, page 4, available from Internet Explorer 8 Beta 1 Whitepapers). This testcase comes from Antti "sairwas" Näyhä's Object, Images test suite (1999). Firefox 2.x, Firefox 3.0.3, Opera 9.60, Safari 3.1.2, Konqueror 4.1.1, Amaya 11 build 20081216, Seamonkey 2.x, NS 7.2, K-meleon 1.x all pass this test. This bug has been reported at connect's IE beta feedback as bug 389191.

34- Dynamically moving a DHTML layer with unitless offset position values should fail IE 8 must not honor unitless offset position values when in web standards compliant rendering mode (when document.compatMode == CSS1Compat). All modern browsers (even IE 6) do not honor unitless offset position values when specified declaratively. So I am submitting that unitless offset position values should not be honored when specified dynamically via script: such values should be ignored. Firefox 3.0.3, Safari 3.1.2, Safari 4.0, Konqueror 4.1.2, K-meleon 1.x, NS 7.2, Seamonkey 2.x all pass this test. This bug has been reported at connect's IE beta feedback as bug 381347.

35- An HTML comment before the doctype declaration will trigger IE 7 and IE 8 into backward-compatible "quirks" rendering mode when it should not do so. Firefox 1.5, Firefox 2.x, Firefox 3.0.4, Opera 9.62, Safari 3.2, Safari 4.0, Konqueror 3.5.10, Konqueror 4.1.1, Hv3 TKHTML alpha 16, Amaya 10.0.1, Seamonkey 1.x, Seamonkey 2.x, NS 7.2, K-meleon 1.x, Galeon 2.0.7, Epiphany 2.22 all pass this test. This bug has been reported at connect's IE beta feedback as bug 382800.

36- Inaccurate CSS parsing of cursor: url() declarations Firefox 1.5, Firefox 2.x, Firefox 3.x, Safari 3.0.4, Safari 3.2, Konqueror 3.5.8 all pass this test. This bug has been reported at connect's IE beta feedback as bug 407963.

37- Dynamic magnification of an image makes content grow outside padding-box and outside border-box Additionally, the border-width unexpectedly increases or decreases while it should remain 2px. This is a regression. Internet Explorer 7, Firefox 1.5, Firefox 2.x, Firefox 3.x, Opera 9.x, Safari 3.0.4, Safari 3.2, Konqueror 3.5.8, NS 7.2, Seamonkey 2.x all pass this test.

38- SelectElement.value should return the value which would be submitted, if the user had clicked the submit button Another testcase for this bug is initial value of the 'value' attribute of <option> is not set to the contents of the element by Nisse Engström. Credit must go to Kyle Mulin for explaining this bug and for providing an initial, working testcase along with useful explanations, details in IE Blog in January 20th 2006. Firefox 1.x, Firefox 2.x, Firefox 3.0.5, Opera 9.63, Safari 2, Safari 3.2, Konqueror 4.1.2, Seamonkey 1.x, Seamonkey 2.x, NS 7.2, K-meleon 1.x, Galeon 2, Epiphany 2 all pass this test. This bug has been reported at connect's IE beta feedback as bug 389742.

39- Submit button value not posted with form submission A form with 1 input for text and 1 input for a submit button will only post the input text value and not the submit value when the user hits the Enter key from the input text. Credit must go to Mikey for defining and explaining this bug in IE Blog on December 18th 2008. Firefox 1.x, Firefox 2.x, Firefox 3.0.5, Opera 9.63, Safari 2, Safari 3.2, Konqueror 4.1.2, Seamonkey 1.x, Seamonkey 2.x, NS 7.2, K-meleon 1.x, Galeon 2, Epiphany 2 all pass this test. This bug has been reported at connect's IE beta feedback as bug 389736. This bug had been previously filed and reported as bug 362726 but was unfortunately closed (not reproducible) and was not reactivated.

40- Ignore font sizes can affect accessibility to content None of this happens in all other non-IE browsers: resizing font size or ignoring font sizes specified on webpages does not affect, should not affect or should not interfere with accessibility to content. This bug has been reported at connect's IE beta feedback as bug 399794.

41- Horizontal formatting of horizontal line Text-align should not affect horizontal formatting of <hr> while margin-left should in Internet Explorer 8. Firefox 2.0.0.8, Firefox 3.0.5, Safari 3.0.3, Konqueror 3.5.2, Konqueror 4.1.3, Seamonkey 1.x, Seamonkey 2.x, K-meleon 1.x, Epiphany 2.20, NS 7.0, NS 7.2, Hv3 TKHTML alpha 16, Amaya 9.55 all pass both of these tests. This bug has been reported at connect's IE beta feedback as bug 408027.

42- Bullet (filled disc) list-markers are misrendered when using MS Sans Serif font A normal bullet (filled disc) should be rendered for list items. Instead, IE 8 RC1 build 18372 renders a bullet list-marker as a thin vertical bar. This is a regression. This bug seems to arise from a mismapping of the bullet character entity in some MS fonts. Internet Explorer 7, Firefox 3.0.6, Opera 9.63, Safari 3.2, Konqueror 4.1.3, Seamonkey 2.x all pass this test. Credit must go to Philip A. Hagen for originally discovering and reporting this bug. This bug has been reported at connect's IE beta feedback as bug 413039. This bug has been closed by the IE team and has not been reopened.

43- clientLeft DHTML object property value is incorrect in rtl overflowed content box Steps: load the URL, then click on the "right to left" radio button. Actual results in IE 8 RTW build 18702: clientLeft value remains at 24. Expected results: clientLeft value should change from 24 to 41. This is a regression. Internet Explorer 7, Firefox 3.0.6, Seamonkey 2.x all pass this test. This bug has been reported at connect's IE beta feedback as bug 413852.

44- CPU hangs (infinite loop) when examining read-only properties in Developer tools Initial condition: First, have the Windows Task Manager opened (right-click on the task bar and then select Task Manager). Steps to reproduce: 1- Load Mark "Tarquin" Wilton-Jones' page IE 8 and Object Tags 2- Do Tools/Developer Tools (F12) 3- In the HTML tab, click the [+] in front of <html>, <body>, <div class="bodyarea">, then click the [+] in front of <p class="demo"> and then click the <object> to select it, to highlight it. You should have navigated to <object data="tick.html" width="80%" height="55" type="text/html" altHtml="Your browser could not display the image, so it is displaying this text instead"/> 4- In the right pane, click the Attributes button 5- At the bottom, check the "Show Read-Only Attributes" checkbox. Actual results in IE 8 RC1 build 18372: %tage of CPU activity for iexplore.exe process is maximized and remains maximized. This is a serious issue. This bug has been reported at connect's IE beta feedback as bug 414807.

45- border-style: inset | outset | ridge | groove not rendered as expected for black color. Currently, border-style-inset-001, border-style-inset-002, border-style-outset-001, border-style-outset-002 fail. This bug had been already reported 12 years ago by richinstyle.com: 3D styles rendered as solid when border-color is dark and 3D border styles rendered as solid when color is dark. This bug has been reported at connect's IE beta feedback as bug 414815. This bug has been closed by the IE team and has not been reopened.

46- Dynamically setting clip: auto does not reset a clipped container overflowed by its block child content appropriately Credit must go to Vadikom-Vasil Dinkov "vasko_dinkov" for originally discovering and reporting this bug. This bug has been reported at connect's IE beta feedback as bug 415011.

47- Embedded HTML document in an <object> unexpectedly creates a vertical scrollbar by Mark "Tarquin" Wilton-Jones: "for reasons best known to itself, [IE 8 RC1 build 18372] choses to display scrollbars in the demo on this page." This bug has been reported at connect's IE beta feedback as bug 415028.

48- Embedded HTML document in an <iframe> unexpectedly creates a dimmed, inactive vertical scrollbar. 1 example where this occurs: test 6772 This bug has been reported at connect's IE beta feedback as bug 415037.

49- Defining font-size with ex length unit must refer to computed font-size of the parent element Firefox 2.x, Firefox 3.0.11, Opera 9.64, Safari 4.0, Konqueror 3.5.8, Konqueror 4.2.4, Seamonkey 1.x, Seamonkey 2.x, NS 7.2, K-meleon 1.x, Galeon 2.0.4, Epiphany 2.22 all pass this test.

50- visibility: inherit applied to command button fails in IE 8 This bug was originally discovered by Fredrik "DXter" Jonsson and reported at launchpad's bug 345777. This is a regression. Internet Explorer 7, Firefox 2.x, Firefox 3.0.11, Opera 9.64, Safari 4.0, Konqueror 3.5.8, Konqueror 4.2.4, Seamonkey 1.x, Seamonkey 2.x, NS 7.2, K-meleon 1.x, Galeon 2.0.4, Epiphany 2.22 all pass this test.

51- Text-align should only format inline content of a list-item's principal block box Text-align: center or text-align: right should only affect inline content of a list item's principal block box and not the position of its list marker when the list-item's list-style-position is set to outside. Safari 4.0 and Konqueror 4.2.4 pass this test.

52- A floated image in a list-item overlaps or masks or removes its list-marker Firefox 3.5 and NS 7.2 pass all 3 tests; Opera 9.64, Safari 4 and Konqueror 4.2.4 pass only the 3rd test.

53- Overlapped background-image with transparent animated gif does not get repainted accordingly This bug was originally reported by ahacks.com web developer and explained in IE Blog on July 3rd 2009. Firefox 3.5, Opera 9.64, Safari 4, Konqueror 4.2.4 and NS 7.2 pass this test.

54- Block-level, replaced elements with width: auto;, margin-left: auto; and margin-right: auto; should be centered within their respective parent block-level elements This bug was originally reported by stusmith at Stackoverflow.com Blog on March 19th and 29th 2009. This is a regression. Internet Explorer 6, Internet Explorer 7, Firefox 3.5.2, Opera 9.64, Safari 4.0.3 and Konqueror 4.2.4 pass both of these 2 tests.

55- CRASH in IE 8 with findText() in a text range This bug was originally reported by Anoakie at htmlbugs.com blog on April 9th 2007 as affecting IE 6 and IE 7. Now, we can see this bug happens as well in IE 8. What is surprising and upsetting is that a perfectly valid HTML 4.01 markup code with a declared strict DTD webpage of 440 bytes can crash like that. It's quite possible that the original bug report from Anoakie at htmlbugs.com was not investigated by IE Team in the process of developing IE 8.

56- Incorrect default attribute selector parsing. Firefox 3.5.2, Opera 10.0, Safari 4.0.3, Konqueror 4.3.1, NS 7.2 and Seamonkey 2.x pass this test.

57- Descendants of a "visibility: hidden" ancestor element must be visible if they have "visibility: visible". This is a regression. Internet Explorer 7, Firefox 3.5.3, Opera 10.0, Safari 4.0.3, Konqueror 4.3.1, NS 7.2 and Seamonkey 2.x pass this test.

58- Intra-table navigation thanks to fragment identifier fails. Navigating from a link into a table to a linked table-row with an id attribute should succeed. This is a regression. Credit must go to BootNic for initially, originally reporting this bug in alt.html thread "IE 8 Table Row Anchor" and to Andreas Prilop in comp.infosystems.www.authoring.html thread "ID attribute for table row". Internet Explorer 7, Firefox 3.5.5, Opera 10.01, Safari 4.0.4, Konqueror 4.3.3, NS 7.2 and Seamonkey 2.x pass this test.

59- Margin-top of block-level element is incorrectly substracted when calculating clearance. Margin-top of block-level element is incorrectly substracted when calculating clearance for a preceding empty content block on which "clear" is set. Other variations of this bug are: Margin-top of block-level element is incorrectly substracted when calculating clearance with a right-floated preceding sibling block and Margin-top of block-level element is incorrectly substracted when calculating clearance with 2 left-and-right-floated preceding sibling blocks. Credit must go to Matt Bradley from InventPartners for originally discovering and reporting this bug in Internet Explorer 8: Top margin missing from first element after floated elements and clearing <div> Firefox 3.6, Opera 10.10, Safari 4.0.4 and Seamonkey 2.x pass this test.

60- Quoted -ms-filter Alpha opacity value interferes with float rendering layout. When quoting -ms-filter value, floats are misrendered. This bug happens on other -ms-filter values, not just with -ms-filter: progid:DXImageTransform.Microsoft.Alpha(opacity=). This bug was originally reported at eggheadcafe's post Bug: IE8 Standards Mode Float Mode on december 29th 2009. Firefox 3.6, Opera 10.10, Safari 4.0.4, Konqueror 4.3.4 and Seamonkey 2.x pass this test.

61- shrink-to-fit width of floating, non-replaced elements are incorrectly rendered This bug is originally from Daniel Schattenkirchner. Shrink-to-fit width calculation is an algorithm which can be involved in many situations: float non-replaced elements, absolutely positioned non-replaced elements, inline-block non-replaced elements. Firefox 3.6.8, Opera 10.60, Safari 5.0.1, Chrome 6.0.472.55, Konqueror 4.5.5, NS 7.2 and Seamonkey 2.x pass this test.

62- Inherit applied on padding set with em unit. Firefox 3.6.9, Opera 10.62, Safari 5.0.2, Chrome 6.0.472.55, Konqueror 4.4.5, NS 7.2 and Seamonkey 2.x pass this test.

63- border-width: inherit applied on a border-style: hidden. Firefox 3.6.9, Opera 10.62, Safari 5.0.2, Chrome 6.0.472.55, Konqueror 4.4.5 and Seamonkey 2.x pass this test.

64- outline-style: none and inherited outline-width Safari 5.0.2, Chrome 6.0.472.55, Konqueror 4.4.5 pass this test.

Specific MSIE 8 bug collection sites

Individual testcases or webpages that fail in MSIE 8

Browser conformance test suites


171 bugs fixed in Internet Explorer 8

with 10 bugs invalid or not fixable for various reasons

3- Last character non-text-selectable with mouse if ponctuation character Exclamation mark, period, comma, etc. are not text selectable (highlightable) with mouse-dragging. Internet Explorer 7, Firefox 2.x, Opera 9.x, Safari 2, Safari 3.0.2, Konqueror 3.5.8, NS 6.1, NS 6.2, NS 7.0, NS 7.1, NS 7.2, Hv3 TKHTML alpha 16 all pass this test. This bug has been FIXED in Internet Explorer 8 beta 2 (build 6001.18241)

4- Background shorthand with rgb() and background-image is buggy and Background shorthand with rgb() is buggy Credit must go to Nicole's IE8 Beta - Broken RGB Color. The problem has also been reported by Patrick Blackstone with his CSS Background Properties Shorthand with RGB Color Bug testcase and by Jeremy Smith's IE8 Background repeat-x CSS Bug. This bug has been reported at connect's IE beta feedback as bug 339082. This bug has been FIXED in Internet Explorer 8 beta 2 (build 6001.18241)

5- Table header cells should be vertically aligned in the middle, not at the top. This is a regression. This bug has been FIXED in Internet Explorer 8 beta 2 (build 6001.18241)

6- A disabled button should not fire onclick events Similarly, disabled attribute is not supported, not implemented. This is a regression. This bug has been reported at connect's IE beta feedback as bug 333934. This bug has been FIXED in Internet Explorer 8 beta 2 (build 6001.18241)

7- defaultChecked for radio button is not implemented. This is a regression. This bug has been reported at connect's IE beta feedback as bug 339083. This bug has been FIXED in Internet Explorer 8 beta 2 (build 6001.18241)

8- alt attribute for image is displayed in a tooltip testcase by Simon Pieters. Internet Explorer implements incorrectly the alt attribute (versus the title attribute). This bug has been reported and documented for well over 10 years now:

This bug has been reported at connect's IE beta feedback as bug 334225. This bug has been FIXED in Internet Explorer 8 beta 2 (build 6001.18241)

9- <q> element is not supported. Q element testcase from (DRAFT) HTML4 Test Suite
"Visual user agents must ensure that the content of the Q element is rendered with delimiting quotation marks."
coming from HTML 4.01 spec., section 9.2.2 Quotations, Rendering quotations. This absence of support has also been explained and testcase-ed by Aaron Gustafson in his Support Q in IE.next Wishlist This bug has been reported at connect's IE beta feedback as bug 339078. This bug has been FIXED in Internet Explorer 8 beta 2 (build 6001.18241)

10- Non-defined value of option is not set to the contents of the OPTION element Firefox 1.x, Firefox 2.x, Opera 9.x, Safari 2, Safari 3.0.2, Konqueror 3.5.8, Seamonkey 1.x, Seamonkey 2.x, NS 7.2, K-meleon 1.x, Galeon 2, Epiphany 2 all pass this test. Another interesting testcase for this bug is initial value of the 'value' attribute of <option> is not set to the contents of the element by Nisse Engström. This bug has been reported at connect's IE beta feedback as bug 348515. This bug has been FIXED in Internet Explorer 8 beta 2 (build 6001.18241)

11- Id versus name when using getElementById is incorrectly implemented. This bug has been reported by many people, including Peter-Paul Koch and this bug has now been reported on Internet Explorer 6, Internet Explorer 7 and Internet Explorer 8. This bug has been reported at connect's IE beta feedback as bug 333979 and has also been reported at Web Bug Track bug 152: getElementById returns incorrect objects in IE. This bug has been FIXED in Internet Explorer 8 beta 2 (build 6001.18241)

12- When HTML 4 rules attribute is defined for rows, cols or groups, then the border-collapse value must be collapse in CSS 2.1 Firefox 2.x, Firefox 3, Opera 9.x, Safari 3.0.4, Konqueror 3.5.8, NS 7.2, Seamonkey 1.x, Seamonkey 2.x, K-meleon 1.x, Galeon 2, Epiphany 2 all pass these 3 tests. This bug has been reported at connect's IE beta feedback as bug 351855.

This bug has been closed because it turns out it is INVALID. From the beginning, I assumed that currentStyle (and then, later, runtimeStyle) equated/was for IE the W3C equivalent of getComputedStyle. My mistake.

13- tbody {height: 150px; overflow: auto;} does not work

This bug has been closed because it turns out it is INVALID. From the beginning, I assumed that overflow could apply to table-row-group objects (like <tbody>) but the overflow property, as defined by CSS 2.1 specification, section 11.1.1, does not apply to table-row-group objects.

18- Support of comma as separator in CSS 2.1 property clip: rect(valOffsetTop, valOffsetRight, valOffsetBottom, valOffsetLeft) Firefox 1.5.0.7, Firefox 2.0, Opera 9.02, Konqueror 3.5.4, Safari 2.0, Safari 3.0.2, NS 6.2, NS 7.0, NS 7.1, NS 7.2 and MSIE 5.5 all pass this test.
"Internet Explorer takes their usual route and ignores the required comma separated values and only supports the optional comma-less method. So, even though using commas is the correct method, you'll need to drop the commas in order for the clip property to work in Internet Explorer (even in IE7)." taken from Misunderstood CSS Clip by Nicholas Gagne. This is also reported in the tutorial Creating Thumbnails Using the CSS Clip Property by seifi.org in the following manner:
"The W3C recommendation suggests using commas between the coordinates, however this is broken in Internet Explorer. Strangely using the commas in IE does not work when in standards-compliant mode, but it does work when in quirks mode.". This bug has been FIXED in Internet Explorer 8 beta 2 (build 6001.18241)

19- Margin-top of cleared block does not collapse with top margin of the parent block as defined by CSS 2.1, section 9.5.2 In the rules that govern the positioning of a cleared element's top border edge within its parent block, it is said that "Computing the clearance of an element on which 'clear' is set is done by first determining the hypothetical position of the element's top border edge within its parent block. This position is determined after the top margin of the element has been collapsed with previous adjacent margins (including the top margin of the parent block).(...) Note: The clearance can be negative." This bug has also been identified by Bruno Fassino in his testing webpage Margin-top on elements with clear and precisely under the test 1.4. Firefox 3.0.4, Opera 9.62, Safari 3.2, Safari 4.0, Konqueror 4.1.1, SeaMonkey/2.0a2 all pass this test. This bug has been reported at connect's IE beta feedback as bug 383063. This bug has been FIXED in Internet Explorer 8 pre-RC1 (build 6001.18344)

20- Dynamically DOM-created and DOM-inserted form controls are not DOM-fetchable, not DOM-queryable This bug was also reported by Aaron Gustafson in Death to bad DOM Implementations in september 2005. "I just encountered a DOM implementation issue in IE which took about three hours to solve (and like a year off my life)." This bug is also mentioned, explained at Semicolon's "Setting the 'name' attribute in Internet Explorer" by Bennett McElwee. This bug has been FIXED in Internet Explorer 8 beta 2 (build 6001.18241)

21- Vertical-align: middle is incorrectly rendered, inaccurately implemented in Internet Explorer 5.x, 6, 7 and 8 beta 2. Additional vertical-align: middle testcase vertical-align is a CSS 1 property. NS 6.1, NS 6.2, NS 7.0, NS 7.1, NS 7.2, Firefox 1.5.x, Firefox 2.0.0.3, Firefox 3, Safari 2.0.4, Safari 3.0.4, Konqueror 3.5.5, Konqueror 4.1, Seamonkey 1.x, Seamonkey 2.x, K-meleon 1.x, Epiphany 2.14 all pass this test. Opera 9.26 and Opera 9.50 do not pass this test. This bug has been reported at connect's IE beta feedback as bug 365792. This bug has been FIXED in Internet Explorer 8 RC1 (build 6001.18372)

22- Links with display: block do not react to hover over their non-text part if previous block-level element has a negative margin-top. NS 6.1, NS 6.2, NS 7.0, NS 7.1, NS 7.2, Seamonkey 1.x, K-meleon 1.x, Firefox 1.x, Firefox 2.x, Safari 2.0, Safari 3.0.2, Opera 9.x, Konqueror 3.5.8, Hv3 TKHTML alpha 16, Galeon 2, Epiphany 2 all pass this test.
A similar problem/issue was explained in Sliding doors of CSS, part II in "A list apart" article, in October 2003. This bug has been reported at connect's IE beta feedback as bug 332625. This bug has been FIXED in Internet Explorer 8 beta 2 (build 6001.18241)

23- CSS columnar layout that fails in Internet Explorer 7 and Internet Explorer 8: 1st demo: resizing text size breaks this CSS columnar layout in MSIE 7. Firefox 2.0.0.4, Safari 3.0.2, Opera 9.21, Konqueror 3.5.8, NS 7.0, NS 7.2, Amaya 9.55, Hv3 TKHTML alpha 16, K-meleon 1.x, Seamonkey 1.x, Seamonkey 2.x, Galeon 2, Epiphany 2 all pass this test. This bug has been reported at connect's IE beta feedback as bug 339308. This bug has been FIXED in Internet Explorer 8 beta 2 (build 6001.18241)

24- Miswritten system font reserved names are honored: Internet Explorer 7 and Internet Explorer 8 do not parse correctly miswritten system font reserved names. Firefox 2.0.0.4, Safari 2, Safari 3.0.2, Opera 9.21, NS 6.2, NS 7.0, NS 7.2, Hv3 TKHTML alpha 16, K-meleon 1.x, Seamonkey 1.x, Seamonkey 2.x, Galeon 2, Epiphany 2 all pass this test. This bug has been reported at connect's IE beta feedback as bug 348925. This bug has been FIXED in Internet Explorer 8 pre-RC1 (build 6001.18343)

25- hasAttributes() method is not supported. Internet Explorer 7 and Internet Explorer 8 do not support DOM 2 Core hasAttributes() method. Firefox 2.0.0.4, Safari 2, Safari 3.0.2, Opera 9.21, Konqueror 3.5.8, NS 6.2, NS 7.0, NS 7.2, K-meleon 1.x, Galeon 2, Epiphany 2 all pass this test. This bug has been FIXED in Internet Explorer 8 beta 2 (build 6001.18241)

26- DOM 1 Core and DOM 2 Core attributes NamedNodeMap implementation Internet Explorer 6, Internet Explorer 7 and Internet Explorer 8 do not implement correctly the attributes attribute as a list of specified attributes. Firefox 2.0.0.4, Safari 2, Safari 3.0.2, Opera 9.21, Konqueror 3.5.8, NS 6.2, NS 7.0, NS 7.2, K-meleon 1.x, Galeon 2, Epiphany 2 all pass this test. This bug has been FIXED in Internet Explorer 8 beta 2 (build 6001.18241)

27- Javascript reserved keyword const is not supported Internet Explorer 7 and Internet Explorer 8 do not support the javascript reserved keyword const used to declare and to define constants. Firefox 2.0.0.4, Opera 9.21, Safari 2, Safari 3.0.2, NS 6.2, NS 7.0, NS 7.2, K-meleon 1.x, Galeon 2, Epiphany 2 all pass this test.

Update november 21st 2008: August 26th 2008 ECMAScript 3.1 proposal suggests object.freeze instead to "sets the writable attribute to false for every own data property of the object.". const as a reserved word has been deleted from ECMAScript 3.1 proposals and drafts.

28- Absence of user interface for choosing an alternate stylesheet This absence of support has been reported for now 10 years by webstandards.org in its IE 5 Top 10 CSS Problems.
CSS 1, section 1.1 states
"The 'LINK' element references alternative style sheets that the reader can select".
CSS 2.1, 3.2 UA Conformance states
"5. If the source document comes with alternate style sheet sets (such as with the 'alternate' keyword in HTML 4), [then] the UA must allow the user to select which style sheet set the UA should apply."
HTML 4.01, section 14.1 states
"User agents should give users the opportunity to select from among alternate style sheets or to switch off style sheets altogether."
HTML 4.01, section 14.3.1 and section 14.3.2 states
"Authors may specify a number of mutually exclusive style sheets called alternate style sheets. Users may select their favorite among these depending on their preferences. (...) User agents should provide a means for users to view and pick from the list of alternate styles."
W3C User Agent Accessibility Guidelines 1.0, guideline 4.14, provision 1 states
"Allow the user to choose from and apply alternative author style sheets (such as linked style sheets)."
Apply alternative author style sheets test from WAI and the UAAG test suite
This issue has also been reported by Mark "Tarquin" Wilton-Jones in his No, Internet Explorer did not handle it properly: Alternate Stylesheets and by Paul Sowden in his November 2001 article Alternative Style: Working With Alternate Style Sheets. Firefox 1.5, Firefox 2, Firefox 3, Opera 9.27, Opera 9.50, Konqueror 3.5.8, Seamonkey 1.x, Seamonkey 2.x, NS 7.2, K-meleon 1.x, Galeon 2.0.4, Epiphany 2.22 all pass this test. This bug has been FIXED in Internet Explorer 8 beta 2 (build 6001.18241)

30- A relatively positioned child is not clipped, hidden by parent with overflow: hidden. This bug has also been reported and described by Alan Gresley's Overflow hidden with relatively positioned child bug report and in connect's bug 332028 CSS overflow test by JackChapple. Firefox 2.0.0.5, Opera 9.22, Safari 3.0.2, Konqueror 3.5.8, NS 6.2, NS 7.0, NS 7.2, Hv3 TKHTML alpha 16, Seamonkey 1.x, Seamonkey 2.x, K-meleon 1.x, Galeon 2, Epiphany 2 all pass this test. This bug has been FIXED in Internet Explorer 8 beta 2 (build 6001.18241)

32- CSS columnar layout that fails in MSIE 8: 3rd demo: because of its left: auto miscalculation offset bug, the navigation menu links at the top are incorrectly offset. Firefox 2.0.0.6, Safari 3.0.3, Opera 9.22, Konqueror 3.5.8, Konqueror 4.1, NS 6.2, NS 7.0, NS 7.2, Seamonkey 1.x, Seamonkey 2.x, K-meleon 1.x, Galeon 2.x, Epiphany 2.x all pass this test. Hv3 TKHTML alpha 16 positions navigation menu links at the top with the correct offset but it does not render the whole webpage as expected. This bug has been reported at connect's IE beta feedback as bug 339084. This bug has been FIXED in Internet Explorer 8 beta 2 (build 6001.18241)

33- Links in <ul> styled with display: block do not trigger :hover when hovering the mouse cursor over their non-text part. Additionally, a duplicate (phantom) list-marker bullet appears unexpectedly on the previous list-item when hovering over a list-item. Firefox 2.0.0.6, Opera 9.23, Opera 9.50, Safari 3.0.3, Konqueror 3.5.8, NS 6.2, NS 7.0, NS 7.2, Hv3 TKHTML alpha 16, Seamonkey 1.x, Seamonkey 2.x, K-meleon 1.1.2, Galeon 2.x, Epiphany 2.x all pass this test. This sort of bug happens often in CSS columnar webpage templates (e.g.: Bridge CSS columnar template and is reported often in newsgroups (e.g.: http://www.itzehoer.de/ in alt.html newsgroup forum discussion on May 4th 2007 under the subject line Link-Problem with IE7) and is one reason why amateur web authors resort to display: table-cell, bizarre hacks, ugly and non-forward-compatible hacks or/and even table-based layout design. This bug has been reported at connect's IE beta feedback as bug 332625.

An example of regression for this type of bug: this basic layout with 3 columns by Gabriele Romanato. The links in the leftmost column are reactive even in the non-text parts of the links in Internet Explorer 7 but not so in Internet Explorer 8. This is a regression. This bug has been FIXED in Internet Explorer 8 beta 2 (build 6001.18241)

34- Vertical-align should NOT and must NOT apply to <ul> with list-style-image. Firefox 2.0.0.6, Opera 9.23, Opera 9.50, Safari 3.0.3, Konqueror 3.5.8, Seamonkey 2.x, Konqueror 3.5.2, Seamonkey 1.x, K-meleon 1.x, Galeon 2, Epiphany 2, NS 6.2, NS 7.0, NS 7.2 all pass this test. This bug has been FIXED in Internet Explorer 8 beta 2 (build 6001.18241)

37- Internet Explorer 6, Internet Explorer 7 and Internet Explorer 8 fail to correctly position an image with background-position when it is defined with em: MSIE 7 background position bug demo. Firefox 2.0.0.7, Opera 9.50, Safari 3.0.3, Konqueror 3.5.5, Seamonkey 1.x, Seamonkey 2.0a1pre, K-meleon 1.x, Galeon 2, Epiphany 2, NS 6.2, NS 7.0, NS 7.2 all pass this test. Credit must go to Tim Snadden for his Internet Explorer Background Position Bug demo. This bug has been reported at connect's IE beta feedback as bug 348564. This bug has been FIXED in Internet Explorer 8 pre-RC1 (build 6001.18343)

38- Internet Explorer 5, Internet Explorer 6, Internet Explorer 7 and Internet Explorer 8 beta 2 do not increase or decrease correctly nested block-level elements when they use font-size: 1em: EM text resizing bug demo. This is a widely known, very frequently encountered and reported bug in Internet Explorer 5, Internet Explorer 6 and Internet Explorer 7. Basically, the problem with setting the font-size in EM is that increasing or decreasing text size will grow or shrink fonts out of proportion (exponentially) for nested block-level elements. Firefox 2.0.0.8, Opera 9.50, Safari 3.0.3, Konqueror 3.5.5, Seamonkey 1.x, Seamonkey 2.0a1pre, K-meleon 1.x, Galeon 2, Epiphany 2, NS 6.1, NS 6.2, NS 7.0, NS 7.2, Hv3 TKHTML alpha 16, Amaya 9.55 all pass this test. This bug has been reported at connect's IE beta feedback as bug 339085. This IE 5+ bug has been reported many times, in many newsgroup discussion forums on web authoring, in various webpages, blogs from web developers. It's really worth mentioning a few of them and quoting some highly relevant testimonials describing accurately the effects of this bug:

"IE gets its text resizing horribly wrong when just using ems. (...) bump the text size up or down, and you'll notice that IE scales ems far more drastically than percentages, resulting in too big / too small text."
from Patrick H. Lauke, May 18th 2004, Weblog comments: How to size text using ems

"text sized in ems can be resized across all browsers. However IE6 and IE7 unacceptably exaggerate the smallness and largeness of the resized text"
from Richard Rutter, November 20th 2007, How to Size Text in CSS

"However, (gasp) IE has a problem with ems. Resizing text from medium (default) to large in IE5/6 would lead to a huge increase in font size rather than the gradual one expected"
from Jon Tan, September 27th 2007, The Incredible Em & Elastic Layouts with CSS

"If for instance you set the root element (either <body> or <html>) to font-size:1em, then just setting View > Text Size to 'smaller' can cause the text to become unreadably small."
from css-discuss.incutio.com, 2007, UsingEms, A word of caution concerning Internet Explorer

"When I tried font-size:1em, my IE6 browser 'text size' will toggle from very tiny to very huge.
....Not acceptable for accessibility!"
from Krystof, August 29th 2005, Weblog comments: How to size text using ems

"Simply put, IE for Windows' relative font size inheritance cascade is broken. (...) If you find it necessary to size through cascade, you'll find % works more reliably than ems."
from InternetExplorerWinBugs, IE Font Sizing Bugs: Inheritance at css-discuss.incutio.com

Only 2 people explained and documented this bug in detail (with screenshots, examples, demos):

  1. em font-resizing bug in IE5 - IE7 by Gunlaug & Georg, November 20th 2005.
  2. Bug avec les unités relatives em et Internet Explorer by Florent Verschelde, 2005

This bug has been reported at connect's IE beta feedback as bug 339085. This bug has been FIXED in Internet Explorer 8 pre-RC1 (build 6001.18343)

41- The last (in order of appareance) rule with same media, same importance, same origin, same specificity must prevail demo. According to CSS 2.1, section 6.4.1 Cascading order, when 2 css rules have the same media, same importance, same origin and same specificity, then the declarations of the last css rule must prevail: (...) if two declarations have the same weight, origin and specificity, the latter specified wins. Firefox 1.5, Firefox 2.x, Safari 2.0.4, Safari 3.0.4, Opera 9.25, Opera 9.50, Konqueror 3.5.8, Hv3 TKHTML alpha 16, Seamonkey 1.x, Seamonkey 2.x, K-meleon 1.x, NS 7.0, NS 7.2, Galeon 2, Epiphany 2 all pass this test. This bug has been reported at connect's IE beta feedback as bug 344800. This bug has been FIXED in Internet Explorer 8 beta 2 (build 6001.18241)

42- Top and bottom borders not rendered. Internet Explorer 6, Firefox 1.5, Firefox 2.x, Safari 2.0.4, Safari 3.0.4, Opera 9.25, Opera 9.50, Konqueror 3.5.8, Hv3 TKHTML alpha 16, Seamonkey 1.x, Seamonkey 2.x, K-meleon 1.x, NS 7.0, NS 7.2, Galeon 2, Epiphany 2 all pass this test. This bug has been reported at connect's IE beta feedback as bug 344801. This bug has been FIXED in Internet Explorer 8 beta 2 (build 6001.18241)

43- Padding and border are rendered beneath body node. Since the blue-bordered inline-elements seem to have a lower stacking position than the red bordered body node in Internet Explorer 7 and Internet Explorer 8, then the border and padding areas are not visible, not viewable, seem to be truncated. This is most likely the same bug as the Disappearing border-top bug reported by Ben Hollis and the same bug that is affecting this 1997 Web Standards Project basic testcase. Firefox 1.x, Firefox 2.x, Safari 2.0.4, Safari 3.0.4, Opera 9.25, Opera 9.50, Konqueror 3.5.8, Konqueror 4.1, Hv3 TKHTML alpha 16, Seamonkey 1.x, Seamonkey 2.x, K-meleon 1.x, NS 6.2, NS 7.0, NS 7.2, Galeon 2, Epiphany 2 all pass this test.

CSS 2.1, Section 9.9.1 states
"Each stacking context consists of the following stacking levels (from back to front):
1. the background and borders of the element forming the stacking context.
2. the stacking contexts of descendants with negative stack levels.
3. a stacking level containing in-flow non-inline-level descendants.
4. a stacking level for floats and their contents.
5. a stacking level for in-flow inline-level descendants. (...)"
CSS 2.1, Section 9.9.1 Layered presentation; Specifying the stack level: the 'z-index' property

This testcase is actually rendered worse in Internet Explorer 8 beta 1 as the left padding area is not reactive on an hovering of the mouse.

This bug has been reported at connect's IE beta feedback as bug 338799. This bug has been FIXED in Internet Explorer 8 beta 2 (build 6001.18241)

44- Caption box should be rendered outside of table box: Caption box should be rendered outside of table box when border-collapse value is separate, Internet Explorer 8 renders the caption box inside (not outside) the table box. Internet Explorer 7, Firefox 2, Firefox 3, Opera 9.x, Safari 3, Konqueror 3.5.8, Seamonkey 1.x, Seamonkey 2.x, K-meleon 1.x, NS 7.x, Galeon 2.14, Epiphany 2.20, Hv3 TKHTML alpha 16 all pass this test. This bug has been FIXED in Internet Explorer 8 beta 2 (build 6001.18241)

45- Text selection is weird, sometimes difficult, often inaccurate. Here's the latest discovery with text selection. Visit this advanced layout 5 with floats by Gabriele Romanato and then (steps to reproduce) select/highlight the words "Site.com" and (actual results) you'll see the horse background-image (expected results) when you should not. This bug has been FIXED in Internet Explorer 8 beta 2 (build 6001.18241)

46- Text underlining is not always redrawn when scrolling. This is a regression. This bug has been reported at connect's IE beta feedback as bug 339309. This bug has been FIXED in Internet Explorer 8 RC1 (build 6001.18372)

48- Top and bottom margins unexpectedly not rendered in Internet Explorer 8; 1st additional testcase: bottom margin not rendered, 2nd additional testcase: bottom margin not rendered. Firefox 2.x, Firefox 3, Opera 9.26, Opera 9.50, Safari 3.1, Konqueror 3.5.8, Seamonkey 1.x, Seamonkey 2.x, NS 7.0, NS 7.2, K-meleon 1.x, Epiphany 2.22, Galeon 2.0.4, Hv3 TKHTML alpha 16, Amaya 10.0 build 20080229 all pass these 3 testcases. Credit must go to Dorayme for posting this issue in comp.infosystems.www.authoring.stylesheets forum discussion newsgroup the IE 7 Margin Problem message and for his No Expected margin bottom in IE 7 testcase. This bug has been reported at connect's IE beta feedback as bug 344804. This bug has been FIXED in Internet Explorer 8 pre-RC1 (build 6001.18343)

49- Archive webstandards.org testcase still not passed after 11 years!. For various reasons, every single version of IE fails to pass this testcase: that's over 11 years! In IE 8 beta 1, the padding area of the inline element is incorrectly painted. A white strip on top of the words appears when aqua color should be painted uniformly. This bug has been FIXED in Internet Explorer 8 beta 2 (build 6001.18241)

50- Nested Q element (nested short quotations) incorrectly rendered in Internet Explorer 6, Internet Explorer 7 and Internet Explorer 8 beta 1: this testcase directly comes from the HTML 4.01 example of nested quotations with the Q element from HTML 4.01, section 9.2.2 example on Rendering nested quotations. This absence of support has also been explained and testcase-ed by Aaron Gustafson in his Support Q in IE.next Wishlist. Firefox 2.x, Firefox 3, Opera 10.00 alpha 1 build 1139, Amaya 11 build 20081216, Seamonkey 1.x, Seamonkey 2.x, K-meleon 1.x, Epiphany 2.22, Galeon 2.0.4 all pass this test. This bug has been reported at connect's IE beta feedback as bug 339078. This bug has been FIXED in Internet Explorer 8 beta 2 (build 6001.18241)

51-60 10 testcases from Ian "Hixie" Hickson evil testsuite still fail in IE 8 beta 1. Those 10 testcases had been explicitly reported already back in september 2006 and have been published more than 6 years ago:

10 Ian "Hixie" Hickson CSS 2.1 testpages which all failed in MSIE 7 and all failed in MSIE 8 beta 1
link to actual test page IE 7 IE 8 Firefox 1.5 Opera 9 Other browsers which
pass the test page
link to actual test page IE 7 IE 8 Firefox 1.5 Opera 9 Other browsers which
pass the test page
51 Shorthand border property Fails Passes Passes Passes NS 6.2, NS 7.0, Amaya 9.53
52 At-rules Fails Passes Passes Passes NS 6.2, NS 7.0
53 :visited link Invalid test in my opinion
54 Dynamic pseudo-class :focus Fails Passes Passes ?Passes?
55 DOM mutation on :hover effect Fails Passes Passes Fails
56 border-width-right Fails Passes Passes Passes NS 6.2, NS 7.0
57 thin bottom border Fails Passes Passes Passes NS 6.2, NS 7.0, Amaya 9.54
58 thin left border Fails Passes Passes Passes NS 6.2, NS 7.0, Amaya 9.54
59 border-width Fails Passes Passes Passes NS 6.2, NS 7.0, Amaya 9.5
60 border-color Fails Passes Passes Passes NS 6.2, NS 7.0

63- Padding values are not applied to HTML document root element Firefox 2.x, Firefox 3, Opera 9.26, Opera 9.50, Safari 3.1, Konqueror 3.5.8, Seamonkey 1.x, Seamonkey 2.x, NS 7.2, K-meleon 1.x, Galeon 2.0.4, Epiphany 2.22 all pass this test. This bug has also been reported by Mark "Tarquin" Wilton-Jones in his Padding on the HTML element is ignored demo and by James Hopkins in his IE8 Beta 1 Regression Bug: specifying a padding value on html demo. This bug has been reported at connect's IE beta feedback as bug 360383. This bug has been FIXED in Internet Explorer 8 beta 2 (build 6001.18241)

64- Border values are not applied to HTML document root element Firefox 2.x, Firefox 3, Opera 9.26, Opera 9.50, Safari 3.1, Konqueror 3.5.8, Seamonkey 1.x, Seamonkey 2.x, NS 7.2, K-meleon 1.x, Galeon 2.0.4, Epiphany 2.22 all pass this test. This bug has been reported at connect's IE beta feedback as bug 360386. This bug has been FIXED in Internet Explorer 8 beta 2 (build 6001.18241)

67- Table cells should expand vertically to fill a table height if required. CSS 2.1, section 17.5.3 Table height algorithms states that Cell boxes that are smaller than the height of the row receive extra top or bottom padding. Internet Explorer 6, Internet Explorer 7, Firefox 1.5, Firefox 2.x, Firefox 3, Opera 9.26, Opera 9.50, Safari 2.0.4, Safari 3.1, Konqueror 3.5.8, NS 7.0, NS 7.2, Seamonkey 1.x, Seamonkey 2.x, K-meleon 1.x, Galeon 2.0.4, Epiphany 2.22 all pass this test. This is a regression. This bug has been FIXED in Internet Explorer 8 beta 2 (build 6001.18241)

68- Counting nested tables with a counter fails Internet Explorer 8 beta 1 counts 12(!) nested tables when there are only 4. Firefox 2.x, Firefox 3, Opera 9.26, Opera 9.50, Safari 3.1, Seamonkey 1.x, Seamonkey 2.x, K-meleon 1.x, Galeon 2.0.4, Epiphany 2.22 all pass this test. This bug has been reported at connect's IE beta feedback as bug 344806. This bug has been FIXED in Internet Explorer 8 pre-RC1 (build 6001.18343)

70- Activating form controls thanks to implicit labels and accesskey fails. Internet Explorer 6, Internet Explorer 7 and Internet Explorer 8 beta 1 fail 8 tests out of 10. Firefox 2.x, Firefox 3, Opera 9.26 (fails 3 tests out of 10), Opera 9.50, Safari 3.1, Konqueror 3.5.8, Seamonkey 1.x, Seamonkey 2.x, NS 7.2, K-meleon 1.x, Galeon 2.0.4, Epiphany 2.22 all pass these 10 tests. This bug has been reported at connect's IE beta feedback as bug 338796. This bug has been FIXED in Internet Explorer 8 beta 2 (build 6001.18241)

71- Margin-bottom of child in overflowed block is not rendered This is a spec violation regarding CSS 2.1, section 10.6.3. Firefox 2.x and Internet Explorer 7 pass both tests. This bug has been reported at connect's IE beta feedback as bug 344808.

This bug has been closed by the IE team and will not be fixed for the release of IE 8 final. Apparently, what IE 8 does (and the same for Firefox 3, Opera 9.27, Opera 9.50, Safari 3.1.1) complies with CSS 2.1, section 10.6.3. The specification is far from being explicit and clear as to why the margin-bottom of the last inflow content child (which does not collapse) may not be and should not be rendered. A clarification of CSS 2.1, section 10.6.3 from W3C would be welcomed.

72- Dynamically adding <option>s into a select with DOM 1/2 HTML add() method is broken in Internet Explorer. Firefox 1.5, Firefox 2, Firefox 3, Opera 9.27, Opera 9.50, Safari 3.1, Konqueror 3.5.8, Seamonkey 1.x, Seamonkey 2.x, NS 7.0, NS 7.2, K-meleon 1.x, Galeon 2, Epiphany 2 all pass this test. This bug has been reported at connect's IE beta feedback as bug 338795. This bug has been FIXED in Internet Explorer 8 beta 2 (build 6001.18241)

73- Top: auto offset miscalculation in Internet Explorer 8 beta 1. Internet Explorer 7, Firefox 2, Firefox 3, Opera 9.27, Opera 9.50, Safari 3.1, Konqueror 4.1.3, Seamonkey 2.x, NS 7.2 all pass this test. This bug has been reported at connect's IE beta feedback as bug 338792. This bug has been FIXED in Internet Explorer 8 pre-RC1 (build 6001.18343)

75- Support for <option disabled> in Internet Explorer. Lack of support for <option disabled> in a select (drop-down list) has been mentioned, explained, utterly documented, testcase-ed by many in all kinds of forums, webpages, public lists of bugs, web authoring forum discussions over the last 5 years. Firefox 1.5, Firefox 2, Firefox 3, Opera 9.27, Opera 9.50, Safari 3.1, Seamonkey 1.x, Seamonkey 2.x, NS 7.0, NS 7.2, K-meleon 1.x, Galeon 2 and Epiphany 2 all pass this test. This bug has been reported at connect's IE beta feedback as bug 336685. This bug has been FIXED in Internet Explorer 8 beta 2 (build 6001.18241)

76- Absolutely positioned box that is a child of a relatively positioned box is mispositioned: the CSS 2.1, section 9.8.4 example. This first test comes directly from CSS 2.1, section 9.8.4. There is also another abs. pos. box child of a rel. pos. box test, a demo for CSS tooltip, which IE 8 beta 1 fails. Firefox 2, Firefox 3, Opera 9.27, Opera 9.50, Safari 2.0.4, Safari 3.1.1, Konqueror 4.1, Seamonkey 1.x, Seamonkey 2.x, NS 7.2, K-meleon 1.x, Galeon 2.0.4, Epiphany 2.22 all pass these 2 tests. Konqueror 3.5.8 passes only the 2nd test; Konqueror 4.1 passes both tests. This bug has been reported at connect's IE beta feedback as bug 338788. This bug has been FIXED in Internet Explorer 8 beta 2 (build 6001.18241)

78- Hovering floated links makes adjacent sibling move up and down (1st test). Firefox 2, Firefox 3, Opera 9.27, Opera 9.50, Safari 3.1.1, Konqueror 3.5.8, Seamonkey 1.x, Seamonkey 2.x, NS 7.0, NS 7.2 all pass this test. This bug has been reported at connect's IE beta feedback as bug 341844. This bug has been FIXED in Internet Explorer 8 beta 2 (build 6001.18241)

79- Hovering floated links makes adjacent sibling move up and down (2nd test). Firefox 2, Firefox 3, Opera 9.27, Opera 9.50, Safari 3.1.1, Konqueror 3.5.8, Seamonkey 1.x, Seamonkey 2.x, NS 7.0, NS 7.2 all pass this test. This bug has been reported at connect's IE beta feedback as bug 341844. This bug has been FIXED in Internet Explorer 8 beta 2 (build 6001.18241)

80- Internet Explorer 8 beta 2 does not clear accordingly. In the same test, we can also notice that Internet Explorer 8 beta 2 still fails to choose the correct color for a border. Firefox 2.0.0.14, Firefox 3, Opera 9.27, Opera 9.50, Safari 3.1.1, Konqueror 3.5.9, Hv3 TKHTML alpha 16, Seamonkey 1.x, Seamonkey 2.x, NS 7.2, K-meleon 1.x, Galeon 2.0.4, Epiphany 2.22 all pass this test. This bug has been FIXED in Internet Explorer 8 RC1 (build 6001.18372)

82- Rel attribute values for alternate stylesheet are incorrectly parsed. Firefox 2.0.0.14, Firefox 3, Opera 9.27, Opera 9.50, Safari 3.1.1, Konqueror 3.5.9, Hv3 TKHTML alpha 16, Seamonkey 1.x, Seamonkey 2.x, NS 7.2, K-meleon 1.x, Galeon 2.0.4, Epiphany 2.22 all pass this test. This bug has been FIXED in Internet Explorer 8 beta 2 (build 6001.18241)

83- z-index and stacking levels. Internet Explorer 8 beta 1 gets this wrong: this is a regression as Internet Explorer 7 had this right. Firefox 2.0.0.14, Firefox 3, Opera 9.27, Opera 9.50, Safari 3.1.1, Konqueror 3.5.9, Hv3 TKHTML alpha 16, Seamonkey 1.x, Seamonkey 2.x, NS 7.2, K-meleon 1.x, Galeon 2.0.4, Epiphany 2.22 all pass this test. This bug has been reported at connect's IE beta feedback as bug 334427. This bug has been FIXED in Internet Explorer 8 beta 2 (build 6001.18241)

84- document.body.offsetParent does not return null. According to W3C CSSOM View Module on the offsetParent attribute, document.body.offsetParent should return null. This is a regression as Internet Explorer 7 had this right. Firefox 2.0.0.14, Firefox 3, Opera 9.27, Opera 9.50, Safari 3.1.1, Konqueror 4.1, Seamonkey 1.x, Seamonkey 2.x, NS 7.2, K-meleon 1.x, Galeon 2.0.4, Epiphany 2.22 all pass this test. This bug has been reported at connect's IE beta feedback as bug 354410. This bug has been FIXED in Internet Explorer 8 beta 2 (build 6001.18241)

85- tableHeaderCellRef.offsetParent does not return its own table element. According to W3C CSSOM View Module on the offsetParent attribute, when the computed value of the position property of <th> is static, then its offsetParent attribute must be its own table element. This is a regression as Internet Explorer 7 had this right. When querying this offsetParent attribute, Internet Explorer 8 beta 1 gets a script Runtime Error which ultimately indicates that document.getElementById("ThInsideTable").offsetParent returns {...} which means is not implemented. Firefox 2.0.0.14, Firefox 3, Opera 9.27, Opera 9.50, Safari 3.1.1, Seamonkey 1.x, Seamonkey 2.x, NS 7.2, K-meleon 1.x, Galeon 2.0.4, Epiphany 2.22 all pass this test. This bug has been FIXED in Internet Explorer 8 beta 2 (build 6001.18241)

86- <object> still does not support native image loading of type image/png. Firefox 2.0.0.14, Firefox 3, Opera 9.27, Opera 9.50, Safari 3.1.1, Amaya 11 build 20081216, Seamonkey 1.x, Seamonkey 2.x, NS 7.2, K-meleon 1.x, Galeon 2.0.4, Epiphany 2.22 all pass this test.

Update september 7th 2008: I am still investigating this bug and this testcase as people have notified me that such testcase does not work optimally in Internet Explorer 8 beta 2. Apparently, an ActiveX warning (via the yellow information bar) prompts the user to allow to run the "Microsoft HTML viewer...." when trying this testcase. The other browsers do not ask for/require such permission. This bug has been FIXED in Internet Explorer 8 pre-RC1 (build 6001.18344)

87- <object> fallback for graceful degradation purpose does not occur when image content-type is not supported. This absence of fallback support happens in all TIFF and PNG file format tests at Antti "sairwas" Näyhä's Object, Images test suite (1999). Firefox 2.0.0.14, Firefox 3, Opera 9.27, Opera 9.50, Amaya 11 build 20081216, Seamonkey 1.x, Seamonkey 2.x, NS 7.2, K-meleon 1.x, Galeon 2.0.4, Epiphany 2.22 all pass this test. This bug has been FIXED in Internet Explorer 8 beta 2 (build 6001.18241)

88- <object> and image map support. Internet Explorer 7, Firefox 2.0.0.14, Firefox 3, Opera 9.27, Opera 9.50, Seamonkey 1.x, Seamonkey 2.x, NS 7.2, K-meleon 1.x, Galeon 2.0.4, Epiphany 2.22 all pass this test. This bug has been FIXED in Internet Explorer 8 beta 2 (build 6001.18241)

89- Correct line-height support on <br> elements. Setting a <br>'s line-height happens in the design and coding of sidebar vertical menu. Internet Explorer 7, Firefox 2.0.0.14, Firefox 3, Opera 9.27, Opera 9.50, Safari 3.1.1, Konqueror 4.1, Seamonkey 1.x, Seamonkey 2.x all pass this test. This bug has been FIXED in Internet Explorer 8 RC1 (build 6001.18372)

90- Generated content numbering bug. Firefox 2.0.0.14, Firefox 3, Opera 9.27, Opera 9.50, Safari 3.1.1, Konqueror 3.5.9, Seamonkey 1.x, Seamonkey 2.x all pass this test. This bug has been FIXED in Internet Explorer 8 beta 2 (build 6001.18241)

91- Empty elements should not generate line boxes. Internet Explorer 6, Internet Explorer 7, Firefox 2.0.0.14, Firefox 3, Opera 9.27, Opera 9.50, Safari 3.1.1, Konqueror 3.5.9, Hv3 TKHTML alpha 16, Seamonkey 1.x, Seamonkey 2.x, NS 7.2, K-meleon 1.x, Galeon 2.0.4, Epiphany 2.22 all pass this test. This bug has been reported at connect's IE beta feedback as bug 350086. This bug has been FIXED in Internet Explorer 8 beta 2 (build 6001.18241)

92- Misaligned floated links. Firefox 2.0.0.14, Firefox 3, Opera 9.27, Opera 9.50, Safari 3.1.1, Konqueror 4.1, Hv3 TKHTML alpha 16, Amaya 11 build 20081216, Seamonkey 1.x, Seamonkey 2.x, NS 7.2, K-meleon 1.x, Galeon 2.0.4, Epiphany 2.22 all pass this test. This bug has been reported at connect's IE beta feedback as bug 354367. This bug has been FIXED in Internet Explorer 8 beta 2 (build 6001.18241)

93- Background-position declaration involves requirements when at least one value is not a keyword (test 1). The order of the 2 values in the declaration for background-position matters decisively when at least one value is not a keyword. Firefox 1.5, Firefox 2.x, Firefox 3, Opera 9.27, Opera 9.50, Konqueror 4.1.2, Seamonkey 1.x, Seamonkey 2.x, NS 7.2, K-meleon 1.x, Galeon 2.0.4, Epiphany 2.22 all pass this test. This bug has been reported at connect's IE beta feedback as bug 354408. This bug has been FIXED in Internet Explorer 8 pre-RC1 (build 6001.18343)

94- Background-position declaration involves requirements when at least one value is not a keyword (test 2). The order of the 2 values in the declaration for background-position matters decisively when at least one value is not a keyword. Firefox 1.5, Firefox 2.x, Firefox 3, Opera 9.27, Opera 9.50, Konqueror 4.1.2, Hv3 TKHTML alpha 16, Seamonkey 1.x, Seamonkey 2.x, NS 7.2, K-meleon 1.x, Galeon 2.0.4, Epiphany 2.22 all pass this test. This bug has been reported at connect's IE beta feedback as bug 354408. This bug has been FIXED in Internet Explorer 8 pre-RC1 (build 6001.18343)

95- Text selection with mouse-drag can make content move unexpectedly, shift up and down. Internet Explorer 7, Firefox 2.0.0.4, Firefox 3, Safari 3.1.2, Opera 9.27, Opera 9.50, Konqueror 3.5.8, Amaya 9.55, Hv3 TKHTML alpha 16, Seamonkey 1.x, Seamonkey 2.x, NS 7.0, NS 7.2, K-meleon 1.x, Galeon 2, Epiphany 2 all pass this test. This bug has been reported at connect's IE beta feedback as bug 353254. This bug has been FIXED in Internet Explorer 8 beta 2 (build 6001.18241)

96- automatic counters and numbering. Firefox 3, Opera 9.50, Safari 3.1.2, Konqueror 3.5.9, Seamonkey 2.x all pass this test. This bug has been FIXED in Internet Explorer 8 RC1 (build 6001.18372)

97- Reactive image border color does not inherit from link style color. Firefox 1.5, Firefox 2, Firefox 3, Opera 9.27, Opera 9.50, Safari 3.1.2, Konqueror 3.5.9, Hv3 TKHTML alpha 16, Seamonkey 1.x, Seamonkey 2.x, NS 7.2, K-meleon 1.x, Galeon 2.0.4, Epiphany 2.22, Amaya 10 (partial success) all pass this test. Credit must go to Rob Novak for originally discovering and reporting this bug in IE bug concerning CSS and IMG tag post. This bug has been reported at connect's IE beta feedback as bug 354312. This bug has been FIXED in Internet Explorer 8 pre-RC1 (build 6001.18343)

98- vertical-align: text-top incorrectly implemented. Firefox 1.5, Firefox 2, Firefox 3, Opera 9.27, Opera 9.50, Safari 3.1.2, Safari 4.0, Konqueror 3.5.2, Seamonkey 1.x, Seamonkey 2.x, NS 7.2, K-meleon 1.x, Galeon 2.x, Epiphany 2.x all pass this test. Hv3 TKHTML alpha 16 partly pass this test. This bug has been FIXED in Internet Explorer 8 beta 2 (build 6001.18241)

100- *:first-child+html selector hack in IE 7 and IE 8 beta 1. Firefox 1.5, Firefox 2, Firefox 3, Opera 9.27, Opera 9.50, Safari 3.1.2, Safari 4.0, Konqueror 3.5.9, Hv3 TKHTML alpha 16, Seamonkey 1.x, Seamonkey 2.x, NS 7.2, K-meleon 1.x, Galeon 2.x, Epiphany 2.x all pass this test. Credit must go to Ingo Turski for discovering and reporting this bug at IE blog on March 6th 2008. This bug has been reported at connect's IE beta feedback as bug 350523. This bug has been FIXED in Internet Explorer 8 beta 2 (build 6001.18241)

101- Vertical padding of nested inline elements rendered under wrapping inline element. Internet Explorer 6, Internet Explorer 7, Firefox 1.5, Firefox 2, Firefox 3, Opera 9.27, Opera 9.50, Safari 3.1.2, Safari 4.0, Konqueror 3.5.9, Konqueror 4.1, Hv3 TKHTML alpha 16, Amaya 10.0.1, Seamonkey 1.x, Seamonkey 2.x, NS 7.2, K-meleon 1.x, Galeon 2.x, Epiphany 2.x all pass this test. This bug has been FIXED in Internet Explorer 8 beta 2 (build 6001.18241)

23 more complex Ian "Hixie" Hickson CSS 2.1 testpages (which all failed in MSIE 8 beta 1)
link to actual test page IE 7 IE 8 Firefox 3.0 Opera 9.50 Safari 3.1.2 Konq 4.1.3 Other browsers which
pass the test page
link to actual test page IE 7 IE 8 Firefox 3.0 Opera 9.50 Safari 3.1.2 Konq 4.1.3 Other browsers which
pass the test page
103 border-color shorthand (basic) Fails Passes Passes Passes Passes Passes NS 7.2, Amaya 10.0.1 (almost)
104 border-color shorthand (basic) Fails Passes Passes Passes Passes Passes NS 7.2, Amaya 10.0.1
106 Absolutely positioned boxes in inlines Passes Passes Passes Passes (almost) Passes Passes NS 7.2
107 Absolutely positioned boxes in inlines Passes Passes Passes Passes (almost) Passes Passes
108 Absolutely positioned boxes in inlines Passes Passes Passes Passes (almost) Passes Passes
109 CSS: z-index 0, -1 and backgrounds (child) Passes Passes Passes Passes Passes Passes Hv3 TKHTML alpha 16, Amaya 10.0.1
110 CSS table model: nested tables Fails Passes Passes Passes Passes Passes
111 Table near 2 floats Passes Passes Passes Passes Passes Passes Hv3 TKHTML alpha 16
112 Table with wide width near two floats Passes Passes Passes Passes Passes Passes Hv3 TKHTML alpha 16
link to actual test page IE 7 IE 8 Firefox 3.0 Opera 9.50 Safari 3.1.2 Konq 4.1.3 Other browsers which
pass the test page
115 Nested floats: Basic Passes Passes Passes Passes Passes Passes NS 7.2, Amaya 10.0.1, Hv3 TKHTML alpha 16
116 Positioning of floats in tables: auto-width floats with negative outer width Fails Passes Passes Passes Passes (almost) Passes (almost) NS 7.2
117 Clear cases Fails Passes Passes Passes Passes Passes Hv3 TKHTML alpha 16 (almost)
118 Box clipping Passes Passes Passes Passes Passes Passes NS 7.2
119 Box clipping Passes Passes Passes Passes Passes Passes NS 7.2, Hv3 TKHTML alpha 16
120 Box clipping ?Passes? Passes Passes Passes Passes Passes
121 Box clipping Fails Passes Passes Passes Passes Passes NS 7.2, Hv3 TKHTML alpha 16
link to actual test page IE 7 IE 8 Firefox 3.0 Opera 9.50 Safari 3.1.2 Konq 4.1.3 Other browsers which
pass the test page
123 Generated Content via Universal Selectors Fails Passes Passes Passes Passes Passes NS 7.2, Hv3 TKHTML alpha 16
125 Inheritance of em units Fails Passes Passes Passes Passes (almost) Passes NS 7.2, Hv3 TKHTML alpha 16
126 Stylesheet encodings: ISO-8859-1 Fails Passes Passes Passes Passes Passes NS 7.2, Hv3 TKHTML alpha 16
127 Stylesheet encodings: ISO-8859-11 Fails Passes Passes Passes Fails Passes NS 7.2, Safari 4.0
128 Escaping and spaces with declarations Fails Passes Passes Passes Passes Passes NS 7.2, Amaya 10.0.1, Hv3 TKHTML alpha 16
129 CSS background position: parsing Fails (partial) Passes Passes Passes Fails (partial) Fails (partial) NS 7.2; Konqueror 4.1.3 fails 1 out of 50; Hv3 TKHTML alpha 16 fails 2 out of 50; IE 7 and Safari 3.1.2 fail 12 out of 50
130 CSS background position: parsing Fails (partial) Passes Passes Passes Fails (partial) Passes NS 7.2; Hv3 TKHTML alpha 16 fails 12 out of 50; IE 7 and Safari 3.1.2 fails 13 out of 50

131- List-item styled as inline elements are not accurately rendered. Firefox 1.5, Firefox 2.x, Firefox 3, Opera 9.27, Opera 9.51, Safari 3.1.2, Safari 4.0, Konqueror 3.5.8, Amaya 10.0.1, Hv3 TKHTML alpha 16, Seamonkey 1.x, Seamonkey 2.x, NS 7.2, K-meleon 1.x, Galeon 2.0.4, Epiphany 2.22 all pass this test. This bug has been reported at connect's IE beta feedback as bug 365838. This bug has been FIXED in Internet Explorer 8 pre-RC1 (build 6001.18343)

135- TextRange.pasteHTML() will cause IE 8 beta 1 to CRASH. Internet Explorer 7, Firefox 2, Firefox 3.0.1, Shiretoko 3.1, Opera 9.52, Safari 3.1.2, Konqueror 3.5.9, Seamonkey 1.x, Seamonkey 2.x, NS 7.2, K-meleon 1.x, Galeon 2.0.4, Epiphany 2.22.3 all pass this test. This bug has been reported at connect's IE beta feedback as bug 360912. This bug has been FIXED in Internet Explorer 8 beta 2 (build 6001.18241)

136- /*/*/ CSS filter in IE 8 beta 1. Internet Explorer 7, Firefox 2, Firefox 3.0.1, Shiretoko 3.1, Opera 9.52, Safari 3.1.2, Konqueror 3.5.9, Seamonkey 1.x, Seamonkey 2.x, NS 7.2, K-meleon 1.x, Galeon 2.0.4, Epiphany 2.22.3 all pass this test. This bug has been reported at connect's IE beta feedback as bug 361018. This bug has been FIXED in Internet Explorer 8 beta 2 (build 6001.18241)

137- text-align: justify and background bug in IE 8 beta 1. Internet Explorer 7, Firefox 2, Firefox 3.0.1, Shiretoko 3.1, Opera 9.52, Safari 3.1.2, Konqueror 3.5.9, Seamonkey 1.x, Seamonkey 2.x, NS 7.2, K-meleon 1.x, Galeon 2.0.4, Epiphany 2.22.3 all pass this test. This bug has been reported at connect's IE beta feedback as bug 361028. This bug has been FIXED in Internet Explorer 8 beta 2 (build 6001.18241)

138- New lines in modified text nodes should be ignored. Internet Explorer 7, Firefox 2, Firefox 3.0.1, Shiretoko 3.1, Opera 9.52, Safari 3.1.2, Konqueror 4.1, Seamonkey 1.x, Seamonkey 2.x, NS 7.2, K-meleon 1.x, Galeon 2.0.4, Epiphany 2.22.3 all pass this test. This bug has been reported at connect's IE beta feedback as bug 361033. This bug has been FIXED in Internet Explorer 8 beta 2 (build 6001.18241)

139- Parent should wrap around floated block child. Parent container will not wrap around a floated child followed by a clear. Longer lines will neutralize this bug behavior: so resizing the browser window viewport actually can trigger or neutralize this bug. Firefox 2, Firefox 3.0.1, Shiretoko 3.1, Opera 9.52, Safari 3.1.2, Konqueror 3.5.9, Konqueror 4.1, Amaya 10.0.1, Hv3 TKHTML alpha 16, Seamonkey 1.x, Seamonkey 2.x, NS 7.2, K-meleon 1.x, Galeon 2.0.4, Epiphany 2.22.3 all pass this test. This bug has been reported at connect's IE beta feedback as bug 361179. This bug has been FIXED in Internet Explorer 8 beta 2 (build 6001.18241)

140- Dynamically updating HTML 4 button value attribute should not change the text of such button. Firefox 2, Firefox 3.0.1, Shiretoko 3.1, Opera 9.52, Safari 3.1.2, Konqueror 3.5.9, Seamonkey 1.x, Seamonkey 2.x, NS 7.2, K-meleon 1.x, Galeon 2.0.4, Epiphany 2.22.3 all pass this test. Credit must go to Garrett Smith for originally discovering, figuring out and reporting this bug. This bug has been FIXED in Internet Explorer 8 beta 2 (build 6001.18241)

142- Adjacent sibling selector and comment bug in Internet Explorer 7 and Internet Explorer 8 beta 1. An HTML comment incorrectly gets counted as an adjacent sibling node. Firefox 2, Firefox 3.0.1, Shiretoko 3.1, Opera 9.52, Safari 3.1.2, Konqueror 4.1, Seamonkey 1.x, Seamonkey 2.x, NS 7.2, K-meleon 1.x, Galeon 2.0.4, Epiphany 2.22.3 all pass this test. Credit must go to Andreas Prilop for his original webpage on IE 7 bug. This bug has been reported at connect's IE beta feedback as bug 362106. This bug has been FIXED in Internet Explorer 8 beta 2 (build 6001.18241)

143- margin-left: auto should not apply to table cells. Declaring td {margin-left: auto;} should be ignored but, in IE 8 beta 1, table cells will be mispositioned (huge offset). Firefox 2, Firefox 3.0.1, Shiretoko 3.1, Opera 9.52, Safari 3.1.2, Konqueror 4.1, Seamonkey 1.x, Seamonkey 2.x, NS 7.2, K-meleon 1.x, Galeon 2.0.4, Epiphany 2.22.3 all pass this test. This bug has been reported at connect's IE beta feedback as bug 362541. This bug has been FIXED in Internet Explorer 8 beta 2 (build 6001.18241)

144- Squeezed central column between a floated left div and a floated right div is pushed under, below. This CSS columnar webpage template is from Florent Verschelde. This is another CSS columnar webpage template (a fluid page with 3 columns) which fails: presumably, possibly because of a rounding up of a fraction of pixel. Steps: maximize the viewport, load provided URL, optionally reload it again; restore down the viewport, load provided URL, optionally reload it again. The failure to show appropriately the central "Explications" column is reproducible almost 100% on my monitor screen with a 1024x768 screen resolution. Another set of steps to reproduce or another way to trigger the bug is to select/highlight some text in the right-most brown column. Push the central column by text-selecting/text-highlighting with the mouse the words "CSS code" in the right-most brown column. Make the central column appear up again between the floated div by text-selecting/text-highlighting with the mouse the words "contexte de formatage" in the <h1> header. Firefox 2, Firefox 3.0.1, Shiretoko 3.1, Opera 9.52, Safari 3.1.2, Seamonkey 1.x, Seamonkey 2.x, NS 7.2, K-meleon 1.x, Galeon 2.0.4, Epiphany 2.22.3 all pass this test. This bug has been reported at connect's IE beta feedback as bug 362553. Credit must also go to Bruno Fassino for creating a reduced testcase enlightening this bug. This bug has been FIXED in Internet Explorer 8 beta 2 (build 6001.18241)

145- outline: 0px solid red still renders a red outline. Firefox 2, Firefox 3.0.1, Shiretoko 3.1, Opera 9.52, Safari 3.1.2, Konqueror 4.1, Seamonkey 1.x, Seamonkey 2.x, NS 7.2, K-meleon 1.x, Galeon 2.0.4, Epiphany 2.22.3 all pass this test. This bug has been reported at connect's IE beta feedback as bug 362838. This bug has been FIXED in Internet Explorer 8 beta 2 (build 6001.18241)

146- Misused comma in font shorthand is ignored (CSS parsing issue). Misuse of commas in font shorthand declarations are honored instead of leading to appropriate rejection of involved CSS rules in Internet Explorer 7 and Internet Explorer 8 beta 2. Shiretoko 3.1, Opera 9.52, Safari 3.1.2, Safari 4.0, Konqueror 4.1, Hv3 TKHTML alpha 16, Seamonkey 2.0a1pre (rv: 1.9.1b1pre) all pass these 2 tests. This bug has been FIXED in Internet Explorer 8 pre-RC1 (build 6001.18343)

147- CSS tooltip only reacting once to hover of mouse in IE 8 beta 2. Toggling a DHTML layer from display: none to display: inline thanks to an :hover declaration only succeeds ONCE, not more in Internet Explorer 8 beta 2. Firefox 2.x, Firefox 3, Opera 9.60, Safari 3.1.2, Safari 4.0, Konqueror 4.1, Shiretoko 3.1, Hv3 TKHTML alpha 16 (although mispositioned), Seamonkey 1.x, Seamonkey 2.x, NS 7.2, K-meleon 1.x, Galeon 2.0.4, Epiphany 2.22.3 all pass this test. This bug has been reported at connect's IE beta feedback as bug 365747. This bug has been FIXED in Internet Explorer 8 pre-RC1 (build 6001.18343)

148- Hovering a link several consecutive times does not trigger :hover more than once in IE 8 beta 2. Hovering a single link styled with :hover succeeds only ONCE, not more in Internet Explorer 8 beta 2. Hover the mouse cursor over a single link (over and out and over consecutively) fails. Firefox 1.5.x, Firefox 2.x, Firefox 3, Opera 9.60, Safari 3.1.2, Safari 4.0, Konqueror 4.1, Shiretoko 3.1, Hv3 TKHTML alpha 16, Seamonkey 1.x, Seamonkey 2.x, NS 7.2, K-meleon 1.x, Galeon 2.0.4, Epiphany 2.22.3 all pass this test. This bug has been reported at connect's IE beta feedback as bug 365776. This bug has been FIXED in Internet Explorer 8 RC1 (build 6001.18372)

149- objColumn.style.display = "none" should do nothing. Setting dynamically a table column to display as "none" should do nothing but it will make the whole webpage become/turn blank in IE 8 beta 2. Firefox 2.x, Firefox 3, Opera 9.60, Safari 3.1.2, Safari 4.0, Konqueror 4.1, Shiretoko 3.1, Seamonkey 2.x, NS 7.2, K-meleon 1.x, Galeon 2.0.4, Epiphany 2.22.3 all pass this test. This bug has been reported at connect's IE beta feedback as bug 365820. This bug has been FIXED in Internet Explorer 8 pre-RC1 (build 6001.18343)

150- Background-color of link with display: block does not change on :hover in IE 8 beta 2. Hovering the mouse cursor over links styled with display: block over their text part or over their non-text part do not change their background-color in IE 8 beta 2. This bug is very important to fix as it breaks lots of webpages, menus, webpage designs and it overall affects navigation functionality and accessibility. Firefox 2.x, Firefox 3, Opera 9.60, Safari 3.1.2, Safari 4.0, Konqueror 4.1, Shiretoko 3.1, Seamonkey 2.x, NS 7.2, K-meleon 1.x, Galeon 2.0.4, Epiphany 2.22.3 all pass this test. This bug has also been reported in Block links sometimes not fully clickable/hoverable testcase by Bruno Fassino and in IE: block-level hyperlinks lack full-clickable area bug report by Rowan Wigginton. This bug has been reported at connect's IE beta feedback as bug 365927. This bug has been FIXED in Internet Explorer 8 final release (build 6001.18702)

151- Precise text selection within floats very difficult or impossible to perform in Internet Explorer 8 beta 2 Another good webpage to test this is the CSS1 Test Suite: 4.1.4 Floating Elements webpage. Firefox 3, Opera 9.60, Safari 3.1.2, Konqueror 4.1, Shiretoko 3.1, Seamonkey 2.x, NS 7.2, K-meleon 1.x all pass this test. This bug has been FIXED in Internet Explorer 8 RC1 (build 6001.18372)

152- Left offset of absolutely positioned inline elements with writing-mode: tb-lr is wrongly determined because the proper containing block is incorrectly determined This bug involves writing-mode: tb-lr and proper determination of what is the containing block for absolutely positioned inline elements. Internet Explorer 6 and Internet Explorer 7 pass this test; so this is a regression bug. The original webpage where the bug happens is Document all DHTML object properties. This bug has been FIXED in Internet Explorer 8 RC1 (build 6001.18372)

35 CSS 2.1 Conformance Test Suite tests which all failed in IE 8 beta 2
link to actual test page IE 7 IE 8 Firefox 3.0.3 Opera 9.60 Safari 3.1.2 Konq 4.1.3 Other browsers which
pass the test page
link to actual test page IE 7 IE 8 Firefox 3.0.3 Opera 9.60 Safari 3.1.2 Konq 4.1.3 Other browsers which
pass the test page
154 Containing block for absolute positioning Passes Passes Passes Passes Passes Passes NS 7.2, Amaya 10.0.1
155 Containing block for absolute positioning Passes Passes Passes Passes Passes Passes NS 7.2, Amaya 10.0.1
156 bidirection box model - margin-left on inline in normal block Passes Passes Passes Passes Passes Passes
157 bidirection box model - margin-left on inline in embed block Passes Passes Passes Passes Passes Passes
158 bidirection box model - margin-right on inline in normal block Passes Passes Passes Fails Passes Passes
159 bidirection box model - margin-right on inline in embed block Passes Passes Passes Fails Passes Passes
160 bidirection box model - padding-left on inline in normal block Passes Passes Passes Passes Passes Passes
161 bidirection box model - padding-left on inline in embed block Passes Passes Passes Passes Passes Passes
162 bidirection box model - padding-right on inline in normal block Passes Passes Passes Fails Passes Passes
163 bidirection box model - padding-right on inline in embed block Passes Passes Passes Fails Passes Passes
164 first-letter and inherent rtl Passes Passes Passes Passes Passes Passes NS 7.2
165 :first-letter and direction:rtl - inherent ltr text Passes Passes Passes Passes Passes Passes NS 7.2
166 :first-letter in the middle of the line because of bidi reordering Passes Passes Passes Passes Passes Passes NS 7.2
167 unicode-bidi: bidi-override - glyph mirroring Passes Passes Passes Passes Passes Passes NS 7.2
link to actual test page IE 7 IE 8 Firefox 3.0.3 Opera 9.60 Safari 3.1.2 Konq 4.1.3 Other browsers which
pass the test page
168 dynamic changes to 'counter-increment' Fails Passes Passes Passes Fails Passes
169 dynamic changes to 'counter-increment' Fails Passes Passes Passes Fails Passes Safari 4.0
170 Counter properties on 'display: none' elements Fails Passes Passes Passes Passes Passes
171 Counter scope Fails Passes Passes Fails Passes Passes
172 Counter scope and nesting on elements Fails Passes Passes Fails Passes Passes
173 Shorthand Properties (border) Fails Passes Passes Passes Passes Passes NS 7.2, Amaya 10.0.1
174 margin-right Fails Passes Fails Passes Passes Passes
175 margin-left Fails Passes Passes Passes Passes ?Fails? NS 7.2, Konqueror 4.1.2
176 border-right Passes Passes Passes Passes Passes Passes NS 7.2
link to actual test page IE 7 IE 8 Firefox 3.0.3 Opera 9.60 Safari 3.1.2 Konq 4.1.3 Other browsers which
pass the test page
177 border-left Passes Passes Passes Passes Passes Passes NS 7.2
178 float Passes Passes Passes Passes Passes Passes NS 7.2, Amaya 10.0.1, Hv3 TKHTML alpha 16
179 The Height of Lines Fails Passes Passes Passes Passes Passes (almost) NS 7.2
180 line-height Passes Passes Passes Passes Passes Passes (almost) NS 7.2
181 background-repeat Passes Passes Passes Passes Passes Passes NS 7.2, Hv3 TKHTML alpha 16
182 background-repeat Passes Passes Passes Passes Passes Passes NS 7.2, Hv3 TKHTML alpha 16
183 background-repeat Passes Passes Passes Passes Passes Passes NS 7.2
184 background-repeat Passes Passes Passes Passes Passes Passes NS 7.2
185 background-repeat Passes Passes Passes Passes Passes Passes NS 7.2
186 background-position Passes Passes Passes Passes Passes Passes NS 7.2, Hv3 TKHTML alpha 16
187 background-position Passes Passes Passes Passes Passes Passes NS 7.2, Hv3 TKHTML alpha 16
188 word-spacing Invalid test according to IE Team
189- Escape of new line inside a string with backslash is not ignored Inside a string, a backslash followed by a newline should be ignored (i.e., the string is deemed not to contain either the backslash or the newline) as stated by CSS 2.1, section 4.1.3 Characters and case. Firefox 3, Shiretoko 3.1, NS 7.2, Seamonkey 1.x, Seamonkey 2.x, K-meleon 1.x, Galeon 2.0.7, Epiphany 2.22 all pass this test.

189- A /* comment */ between rgb numerical values should have no influence on parsing and rendering A comment may occur anywhere between tokens and its content should have no influence on the rendering as stated by CSS 2.1, section 4.1.9 Comments. Internet Explorer 7, Firefox 3.0.3, Opera 9.62, Safari 3.1.2, Safari 4.0, Konqueror 4.1.1, Google Chrome 0.3, NS 7.2, Hv3 TKHTML alpha 16, Amaya 10.0.1, Seamonkey 1.x, Seamonkey 2.x, K-meleon 1.x, Galeon 2.0.7, Epiphany 2.22 all pass both of these tests. This bug has been FIXED in Internet Explorer 8 pre-RC1 (build 6001.18343)

190- Negative margin not applied to generated content Credit must go to Mitch 74 for discovering, originally reporting this bug in IE blog. Firefox 3.0.3, Opera 9.62, Safari 3.1.2, Safari 4.0, Konqueror 4.1.1, Seamonkey 2.x all pass this test. This bug has been reported at connect's IE beta feedback as bug 380412. This bug has been FIXED in Internet Explorer 8 pre-RC1 (build 6001.18343)

192- A long sentence with white-space: nowrap does not generate an horizontal scrollbar One element necessary to trigger this incorrect behavior in IE 8 beta 2 is the presence of an image inside the tested long sentence. Removing that image restores the correct behavior in IE 8 beta 2. Firefox 3.0.4, Opera 9.62, Safari 3.2, Safari 4.0, Konqueror 4.1.1, NS 7.2, Hv3 TKHTML alpha 16, Seamonkey 2.x, K-meleon 1.x, Galeon 2.0.7, Epiphany 2.22.3 all pass this test. This bug has been reported at connect's IE beta feedback as bug 382232. This bug has been FIXED in Internet Explorer 8 pre-RC1 (build 6001.18343)

193- Content spilling out of a floated child does not make parent container styled with overflow: auto generate vertical scrollbar An additional testcase (furthermore reduced): Overflowing floated child does not make parent container styled with overflow: auto generate vertical scrollbar: testcase 2. If the behavior is undefined, if the rendering is not explicitly defined in CSS 2.1, then it usually means that user agents have a normal and reasonable latitude. Here, the logical and accessibility-preferred choice would be to render a vertical scrollbar. Internet Explorer 7, Firefox 1.5, Firefox 2.0.0.4, Firefox 3.0.4, Safari 3.2, Safari 4.0, Konqueror 3.5.10, Konqueror 4.1.1, NS 7.2, Seamonkey 2.x, K-meleon 1.x, Galeon 2.0.4, Epiphany 2.22 all pass both of these tests. This bug has been reported at connect's IE beta feedback as bug 382351. This bug has been FIXED in Internet Explorer 8 pre-RC1 (build 6001.18343)

194- Greek letters are not rendered when using Symbol font Firefox 3.0.4, Opera 9.62, Safari 3.2, Safari 4.0, Konqueror 4.1.1, NS 7.2, Hv3 TKHTML alpha 16, Amaya 10.0.1, Seamonkey 2.0a2pre all pass both of these tests. This bug has been FIXED in Internet Explorer 8 RC1 (build 6001.18372)

195- Hovered block-level elements do not change background-color when they should Firefox 1.5, Firefox 2.x, Firefox 3.0.4, Opera 9.62, Safari 3.2, Safari 4.0, Konqueror 3.5.10, Konqueror 4.1.1, Hv3 TKHTML alpha 16, NS 7.2, Seamonkey 2.x, K-meleon 1.x, Galeon 2.0.7, Epiphany 2.22 all pass both of these tests. This bug has been reported at connect's IE beta feedback as bug 382795. This bug has been FIXED in Internet Explorer 8 pre-RC1 (build 6001.18343)

197- Mousedown, click or text selection (with mouse or via F7 caret browsing) on text of absolutely positioned block unexpectedly moves scroll view to the top of document view in IE 8 beta 2 Firefox 1.5, Firefox 2.x, Firefox 3.0.4, Opera 9.62, Safari 3.2, Safari 4.0, Konqueror 3.5.10, Konqueror 4.1.1, Hv3 TKHTML alpha 16, Seamonkey 1.x, Seamonkey 2.x, NS 7.2, K-meleon 1.x, Galeon 2.0.7, Epiphany 2.22 all pass this test. This bug has been reported at connect's IE beta feedback as bug 383156. This bug has been FIXED in Internet Explorer 8 pre-RC1 (build 6001.18343)

198- Table with a cell with colspan="2" miscalculates, misrenders table cells' width in Internet Explorer 7 and Internet Explorer 8 beta 2: test #1 Firefox 3.0.4, Safari 3.2, Safari 4.0, Konqueror 4.1.1, Amaya 11 build 20081216, Seamonkey 2.x, K-meleon 1.x, NS 7.2, Hv3 TKHTML alpha 16, Google Chrome 0.4 all pass this test. This bug has been reported at connect's IE beta feedback as bug 384963. This bug has been closed by the IE team and will not be fixed for the release of IE 8 final.

This bug 198 has been closed because it turns out it is INVALID. There is no specific steps to execute (or algorithm) to render a table with colspan when table-layout: auto is implicitly defined.

199- Table with a cell with colspan="2" miscalculates, misrenders table cells' width in Internet Explorer 7 and Internet Explorer 8 beta 2: test #2 Firefox 3.0.4, Safari 3.2, Safari 4.0, Konqueror 4.1.1, Seamonkey 2.x, K-meleon 1.x, NS 7.2, Google Chrome 0.4 all pass this test. This bug has been reported at connect's IE beta feedback as bug 384963. This bug has been closed by the IE team and will not be fixed for the release of IE 8 final.

This bug 199 has been closed because it turns out it is INVALID. There is no specific steps to execute (or algorithm) to render a table with colspan when table-layout: auto is implicitly defined.

200- Percentage height in table-cell element not rendered correctly in Internet Explorer 8 beta 2. Firefox 3.0.4, Opera 10.00 alpha 1, Safari 3.2, Konqueror 4.1.1, NS 7.2, K-meleon 1.x, Seamonkey 2.x all pass both of these tests. This bug has been FIXED in Internet Explorer 8 RC1 (build 6001.18372)

205- Application CRASH reproducible 100% when dragging custom cursor up and down outside image border. Steps to reproduce: load this Dynamic magnification of an image DHTML demo, then depress the left mouse button and drag (up or down) first inside the image and then outside the green border. The application crash can be reproduced in backward-compatible rendering mode and in standards rendering mode. This crash bug affects IE 8 pre-RC1 Partner build 6001.18344. Firefox 1.x, Firefox 2.x, Firefox 3.0.5, Opera 9.63, Safari 2, Safari 3.2, Konqueror 4.1.2, Seamonkey 1.x, Seamonkey 2.x, NS 7.2, K-meleon 1.x, Galeon 2, Epiphany 2 all pass this test. This bug has been reported at connect's IE beta feedback as bug 389755. This bug has been FIXED in Internet Explorer 8 RC1 (build 6001.18372)

206- Outline border mispositioned with text-align Firefox 1.x, Firefox 2.x, Firefox 3.0.5, Opera 9.63, Safari 2, Safari 3.2, Konqueror 4.1.2, Seamonkey 1.x, Seamonkey 2.x, NS 7.2, K-meleon 1.x, Galeon 2, Epiphany 2 all pass this test. This bug has been reported at connect's IE beta feedback as bug 389799. This bug has been FIXED in Internet Explorer 8 RC1 (build 6001.18372)

208- Incorrect parsing on malformed declaration #FirstLine { color:green; color: } /* malformed declaration missing value */ is not correctly handled by IE 8 pre-RC1 Partner Build 18344 when/while it was correctly handled in IE 8 beta 1. So this is a regression. This bug has been reported at connect's IE beta feedback as bug 404185. This bug has been FIXED in Internet Explorer 8 RC1 (build 6001.18372)

209- Reflow (and repaint) required when increasing or decreasing text size When text-resizing a webpage with View/Text Size/Larger or Smaller, text lines are not reflowed and line boxes are sometimes not repainted correctly. This bug happens in IE 8 RC1 build 18372. This bug has been FIXED in Internet Explorer 8 final release (build 6001.18702)

211- list-style: none set on individual <li> does not cancel graphic-imaged bullets set on <ul> list. Firefox 2.x, Firefox 3, Safari 3.1.1, Opera 9.27, Opera 9.50, Konqueror 3.5.8, Seamonkey 1.x, Seamonkey 2.x, NS 7.2, K-meleon 1.x, Galeon 2.0.4, Epiphany 2.22, Amaya 10 (build 20080229) all pass this test. Credit must go to John Russel for originally discovering and reporting this bug. This bug has been reported at connect's IE beta feedback as bug 407932. This bug has been FIXED in Internet Explorer 8 final release (build 6001.18702)

212- scrollWidth less than clientWidth By definition, scrollWidth should always be equal or greater than clientWidth. IE 8 RC1 fails. Internet Explorer 6, Internet Explorer 7, Firefox 2.x, Firefox 3, Safari 3.1.1, Opera 9.27, Opera 9.50, Konqueror 3.5.8, Seamonkey 1.x, Seamonkey 2.x, NS 7.2, K-meleon 1.x honor this requirement. This bug has been reported at connect's IE beta feedback as bug 411707. This bug has been FIXED in Internet Explorer 8 final release (build 6001.18702)

213- Text underlining of font-family Courier (monospace) is too high and too thick. The underline is on the baseline and the underline is so thick that, in IE 8 RC1 build 18372, it masks, covers completely a period. The same buggy layout problem happens with other system fonts like "MS Serif" and "MS Sans Serif". Credit must go to Philip A. Hagen for originally discovering and reporting this bug; credit must go to a certain "iecustomizer" for searching, forwarding and validating this bug. This bug has been reported at connect's IE beta feedback as bug 408071. This bug has been FIXED in Internet Explorer 8 final release (build 6001.18702)

214- Accurate ex unit implementation. The CSS 2.1 offers a loose and flexible implementation of ex as a length unit offering 2 distinct, separate manners to implement it. So this bug is void for now. This bug has been reported at connect's IE beta feedback as bug 409460.

215- <col></col> can make a whole webpage become blank. 2 initial conditions for this bug: Tools/Internet Options/Advanced tab/Browsing category/Underline links must be set to Hover and the checkbox "Automatically recover from page layout errors with Compatibility View" must be unchecked. Steps to reproduce: load the page, then zoom in or zoom out the page (with the drop down list on the status bar or with the keyboard), navigate Back and Forward from the page and move the mouse over and out of the browser window viewport (in and out of <body> box is also good enough). Actual results in RC1 build 8372: the whole page becomes blank. There are many other ways to trigger this bug. This bug is a rather serious bug. This bug has been reported at connect's IE beta feedback as bug 409478. This bug has been FIXED in Internet Explorer 8 final release (build 6001.18702)

216- Absolutely positioned reactive-clickable images positioned on top of another image are not reactive in their padding area NS 7.0, NS 7.1, NS 7.2, Firefox 1.5.x, Firefox 2.0.0.3, Firefox 3.0.6, Opera 9.63, Safari 2.0.4, Safari 3.0.4, Konqueror 3.5.5, Konqueror 4.1.3, Seamonkey 1.x, Seamonkey 2.x, K-meleon 1.x all pass this test. This bug has been reported at connect's IE beta feedback as bug 411772. This bug has been FIXED in Internet Explorer 8 final release (build 6001.18702)

217- Height percentage of a zero height container is incorrectly rendered in IE 8 RC1 build 18372 Firefox 3.0.6, Opera 9.63, Konqueror 4.1.3, Hv3 TKHTML alpha 16, NS 7.2, K-meleon 1.5.2, Seamonkey 2.x all pass this test. This bug has been reported at connect's IE beta feedback as bug 411825. This bug has been FIXED in Internet Explorer 8 final release (build 6001.18702)

218- Only 1 absolutely positioned reactive-clickable image gets styled by pseudo-class :hover Firefox 3.0.6, Opera 9.63, Safari 3.2, Konqueror 4.1.3 all pass this test. This bug has been reported at connect's IE beta feedback as bug 412500. This bug has been FIXED in Internet Explorer 8 final release (build 6001.18702)

226- Image shrinks unexpectedly after hovering a link The image can be unshrinked by text-selecting/highlighting the images from left to right. According to developer tools, the image(s) shrink from 155px by 155px to 32px by 32px. Credit must go to Vasil Rangelov for originally discovering and reporting this bug. This bug has been reported at connect's IE beta feedback as bug 415317. This bug has been FIXED in Internet Explorer 8 final release (build 6001.18702)

Specific MSIE 8 bug collection sites (FIXED)

Individual testcases or webpages that have now been fixed in MSIE 8

Valid HTML 4.01! CSS compliant Web standards project HTML Tidy