Internet Explorer 8 Beta 1 for Developers – Standards Highlights


Dean mentioned a bunch of things we are doing in Internet Explorer 8 Beta 1 for Developers. I want to point you to more details specifically about the developer focused changes to CSS, the DOM and the new version targeting.

Standards support (CSS/HTML)

IE8 improves rendering of content authored to various web standards in standards mode. As we have mentioned before, IE8 Beta 1 for Developers ships with standards mode as its default formatting engine. In order to maintain backwards compatibility, sites can opt-into IE7-like handling of content by inserting a special meta element into the web page, that triggers the “IE7 standards mode”. For more complete details regarding document compatibility, see Defining Document Compatibility.

While the behavior of the browser is unchanged from Internet Explorer 7 in “IE7 Standards Mode”, in standards mode (the default IE8 rendering mode), IE8 supports Data: URIs, the abbr tag, CSS generated content, the display: table CSS properties, in addition to fixing a lot of CSS and HTML parsing bugs.  For a complete list, see CSS Improvements in Internet Explorer 8.  

Standards support (DOM) and AJAX

CSS is not the only place where we tuned IE8, we tweaked the DOM as well.  IE8 features an enhanced and standardized DOM that brings it in line with implementations in other browsers. Highlights include changes in the behavior of the getAttribute, setAttribute and removeAttribute modifiers to make IE8 more interoperable with other browsers.    Additionally, IE8 has dramatically enhanced AJAX support with features like DOM: Storage, Cross Document Messaging (XDM) and the Selectors APIs. For a more complete list of AJAX and DOM improvements, see What’s New in Internet Explorer 8.  

Last but certainly not least, we have added support for the Accessible Rich Internet Applications (ARIA) specification. This specification describes how site authors can mark up their content semantically such that users who might use assistive technologies may more completely experience the content. See What’s New for Accessibility in Internet Explorer 8 for a comprehensive discussion of this topic.

It has been great to see all of the wonderful feedback on the beta thus far and remember that IE8 is still a work in progress. Stay tuned for more details here on our blog and on the IE Development Center on MSDN. 

Thanks and Happy Browsing,

Doug Stamper
Principal Program Management Lead
Internet Explorer Developer Experience Team

Comments (134)

  1. steve says:

    The BIGGEST issues with the DOM are still BROKEN in IE8.

    * Can’t Prototype on ANY DOM Element Level

    * document.getElementById( id ) Still Fails

    * Element.hasAttribute( name ) Still Fails

    I haven’t tested document.getElementsByName( name ) but I’m going to pressume that hasn’t been fixed either.

    Glad to see you had time to ADD NEW FEATURES, before FIXING THE BROKEN / MISSING IMPLEMENTATIONS!

    – Not impressed.

  2. EricLaw [MSFT] says:

    @Steve: You need to be using a standards/strict mode document to observe changes/fixes in DOM handling.  

  3. JeremyE [MSFT] says:

    @steve: Could you provide an example of how getElementById() and hasAttribute() fail?

  4. cwilso says:

    @Steve and Element.hasAttribute().  Test cases would be much appreciated.

  5. Gyrobo says:

    Will document.getElementsByClassName be supported? It seems like it should be trivial to do now that you’ve got the Selector API in.

  6. Jake says:

    When exiting IE8 B1 with multiple tabs, the IE7 feature "Open these next time I use Internet Explorer" is no longer an option…

  7. Michael C. says:

    Pardon my bluntness, but IE7 was a disaster for right-to-left pages. I’ve installed the IE8 beta 1, and it seems to be far, far worse. I know it’s just a beta, however, so this is merely to bring it to your attention that part of standards is proper support of RTL documents.

    …pleease? :-/

  8. I agree with Michael C. that, so far, in this first beta of IE 8, right-to-left webpages/tests are not better than in IE 7.

    Regards, Gérard

  9. EricLaw [MSFT] says:

    @Gerard, @Michael C: RTL is not yet implemented for standards mode pages.  Rest assured, we’re very well aware of the importance of this feature.

  10. Arieta says:

    Will GIF animation speed and PNG gamma support gets fixed by the final release? Also, progressive JPEGs don’t display fine mid-loading…

  11. Arieta says:

    yay for display: table support though, it’s gonna make a lot of things easier. But, the only reason I’d use table display divs is because IE6/7/8 eats a lot of cpu power displaying normal tables…

  12. Michael M. says:

    What i would really like to see added is styling for select boxes, so developers can change the look of the button, drop down, and border of select elements.  Color is most important to me, but it would be nice if we could stick images into select structures as replacement for the default arrow button and as entries for the option tags.

  13. Two things…

    1.) For a first Beta of course IE8’s new standards rendering engine is VERY solid. On my site when you resize the browser the pixel-perfect layout (1px of "margin" around always % width based form elements) is always maintained. I also agree with select elements set to block to be measured as block-level elements. A select element set to display: block, width: 343px; and border-width: 1px; should render as 345px;. I’m not sure about inline measurements, I presume that they adjust according to the content.

    2.) There are some MAJOR if not severe performance issues on my website and I’m having difficulty creating a minimal test case. I watched the video on channel 9 where Chris Wilson says that performance isn’t concentrated in the layout aspect though I’m running a static test case page without any JavaScript loaded in IE8 no-addons mode and IE8’s CPU usage is about 52% and the :hover and :focus selectors are very slow to respond. My CPU is an AMD Opteron 185 and my system is very fine tuned for performance. So is there a known issue with pages that have an abundant number of selectors or certain types of layouts that could help me concentrate on figuring out the specific issue to create a good minimal test case to file a good bug report?

    There seems to be a lot of broken JavaScript on my site in IE8B1. I haven’t yet taken the time to look closely at the scripting though my level of scripting isn’t too high so I’ll be looking through it this weekend.

    For a beta 1 this is pretty solid. I do have a non-layout request: spellchecker! Not so much for me but for other people please. 😉

    Good job so far and keep up the great work! 🙂

  14. Tino Zijdel says:

    text-overflow seems to be broken, or is this a general overflow issue?

    Data-URI’s seem to be too restricted to be of any use.

    XDM looks proprietary, why not follow the W3C’s Web Application Formats WG efforts?

    What’s up with the increase of the max connections per host of the (RFC compliant) 2 to 6? It that only for XHR or general? Is it to get a perceived performance advantage over other browsers which still default to 2?

    Why didn’t MS discuss or announce any of these changes (also wrt DOM storage, ARIA and changed parsing rules) to the various W3C WG’s MS is also participating in? And when is MS finally going to present it’s comments on the HTML5 WD (which I believe was due August of last year)?

  15. EricLaw [MSFT] says:

    @Tino: DataURLs are primarily useful for embedding small images into CSS for use as icons, etc; IE8 permits this use.  Unfortunately, for some content types, DataURLs can be dangerous, which is why IE8 scopes them to use where they are not a security threat.  

    Note that there are many cases where using a DataURL will actually be LESS performant than using a HTTP URL with appropriate caching and compression headers.

    XDM is part of the HTML5 specification.  Search for "onMessage" for more information.

    The change to the per-host connection limit applies to all HTTP/HTTPS traffic, not just that from XHR.  RFC2616 suggests that user-agents "SHOULD" (not MIST) limit themselves to 2 connections per server.  The RFC2616bis working group is currently deliberating over removing that suggestion, as the characteristics of the Intranet have changed significantly since this implementation guidance was provided.  Opera defaults to 8 connections per host, and Firefox will dynamically scale up to 8 as well.

  16. EricLaw [MSFT] says:

    Ahem… typo "MUST" not "MIST" above. 🙂

  17. Hello,

    (For bug numbers, please refer to my IE 7 bug website)

    Bugs (#51, #72, #79) regarding text increase/decrease or related to EM have not been fixed.

    Some testcases have border disappearing: bug #101 (which is fixed! thank you!) but now, we do not see the border… which must be another bug surfacing, appearing. bug #15 is also another example of border disappearing, reappearing.

    If the last character of a block-level element is a exclamation mark, question mark, period or comma, or some other typical ponctuation character/sign, then it can not be selected/highlighted with the mouse when dragging the text.

    Eg.: <h1>Hello World</h1> is entirely selectable/highlightable with a mousedown+mousedrag from left to right or from right to left.

    But

    <h1>Hello World!</h1> and the ! is not selectable/highlightable with mousedragging. It can be easy to provide lots of tests of this buggy behavior.

    Bug #102 is now worse: hovering the mouse cursor hover the padding-left area does *not* trigger the :hover class but hovering the mouse cursor over the  padding-right area does trigger the :hover class. This must be a regression. I have other tests in which the same regression is noticed.

    Bugs #49 and #67 are not fixed: when a link is styled with display: block, then its padding area should be reactive to :hover.

    There is something wrong going on, definitely a regression, with list markers and with images as list-style-image. a) previous list marker is duplicated when hovering the mouse over a list item in bug #67

    b) list-style-image is not visible, viewable in bug #68. Others elsewhere have also mentioned problems with list-style-image.

    Best regards, Gérard

  18. getElementById(), hasAttributes() and attributes as a NamedNodeMap are not correctly implemented. See bugs #13, #54 and #55 respectively at my IE 7 bug webpage for testcases.

    Many other DOM attributes and methods are still not correctly implemented: see bugs #23, #24, #26, #30, #42, etc.

    Regards, Gérard

  19. Still Waiting says:

    Still Waiting for one of those elusive invites to file bugs in the public bug tracking system.

    I must say that I am far from impressed about how the bug tracking system works this time around.

    At least last time I could file a bug.  Now I just get to sit here and hope that someone files the bugs I’ve seen.

    Where’s Al when you need him?… oh, never mind, he works for a company that has a FULLY PUBLIC BUG TRACKING SYSTEM now.

    Grumble, grumble, grumble.

  20. Gyrobo says:

    My last comment doesn’t seem to have gotten through: will you support document.getElementsByClassName? It should be trivial, now that the Selector API is in.

    Also, whatever happened to inline find? Is that still a go?

  21. Steve says:

    @Gerard,

    Glad to see I’m not the only one!

    Once again, we would file a bug report in the bug tracker, BUT WE DON’T HAVE ACCESS!

  22. Margin-bottom is no longer rendered in

    http://www.gtalbot.org/BrowserBugsSection/Opera9Bugs/MarginBottom.html

    This appears to be a regression.

    Padding-right is no longer rendered in

    http://www.gtalbot.org/BrowserBugsSection/Opera9Bugs/Opera9PaddingRightMisrendered.html

    This appears to be a regression.

    <object height="100%"> is rendered with only 75 pixels. See/just load

    http://www.w3.org/Style/CSS/Test/CSS1/20070302/sec44.htm

    This appears to be a regression.

    Regards, Gérard

  23. @ Steve

    I believe I have access to a Microsoft IE bug tracker system of some sort, somewhere … not sure… I was invited to – if I understood this correctly – … but anyway, I prefer my own website for all this. So, as soon as I have more time, I’ll list the bugs, upload testcases at my IE 8 bug webpage.

    hasAttribute() seems to be correctly implemented: at least, I see no problems or issue.

    But hasAttributes() is different and it has at least one bug.

    "Not yet fixed bugs" are one category of bugs; regression bugs are more important bugs, I would say, and represent a more urgent category of bugs to tackle, eradicate.

    So, in that sense, I fully understand why Microsoft IE team did not implement DOM 2 Events for this beta.

    When there are so many bugs to be fixed, tackled, investigated, it is understandable that you want to restrict bug fixing to delimited areas and to "advance" methodically, in a structured manner. IE 6 and IE 7 had well over 750 bugs (most likely around 1250 bugs of all sorts); so, a first beta can not aspire to fix everything… it’s not a realistic goal, it would be too wild, too complex, too demanding.

    Regards, Gérard

  24. Gyrobo, John A. Bilicki III: inline search and spell checker are included in the very useful add-on IE7pro. Furthermore this adds a whole lot of other features that are probably also in the list of most downloaded FF extensions 🙂

    And Eric Law: Could you point to where data: URIs have been a security threat in browsers that implement them? The only thing I found was an address spoofing flaw in Opera and I don’t mind the fact that data: URIs don’t work from the address bar, as this is probably not their primary use. But predicting that data URIs won’t have any use outside of image/ MIME types might be a bit bold, I think.

  25. M says:

    Hotmail / IE 8 compatibility issues known defect?

  26. Dynamically disableing a command button does not work: the command button still receives + fires onclick event.

    Dynamically defaultChecking a command button does not work: this worked in IE 7.

    For a demo for both problems with/in IE 8, see bug #21 at my webpage.

    Regards, Gérard

  27. According to HTML 4 spec. (section 11.3.2), the default vertical-alignment for table header cells is middle, not top.

    "middle: Cell data is centered vertically within the cell. This is the default value."

    Default vertical alignment value for all table header cells was middle in IE 7. Now, in IE 8, default vertical alignment value for table header cells is top.

    Regards, Gérard

  28. Gyrobo says:

    @Johannes Rössel

    Thanks, but I’m aware of the plug in. I remember hearing about inline find being something the IE team really wanted to include natively. Just wondering if that was on the menu.

    @IE Team

    And what’s with the grayed-out address bar? That "feature" was backed out of Firefox a year ago because it hinders readability. Plus, it doesn’t even highlight subdomains! Completely USELESS for blogs. I hope you’ll reconsider this functionality.

  29. David Naylor says:

    How come max- and min-widths/heights aren’t in the list of CSS fixes? They must be fixed since IE8 passes Acid2…

  30. Rowan says:

    @David, those properties are already part of IE7 as far as I know.

  31. Lancelot says:

    Hello IE8 team,

    This is probably not exactly the right place to post what I’m going to post, but at least I’m sure someone from the IE team will read it. So, let’s go.

    First, congratultions for the software. Great work. The tools for developers are really awesome! I’m going to try to add slices to my own site.

    It’s too bad that they made you back down on the issue of default rendering engine, though. IMHO, it’s really too early to make IE8’s rendering engine the default. Probably this has to do with Opera and its EU action.

    Now I guess you mostly have to fix bugs and make the whole think work much quicker. As far as I can say, IE8 as a software is slower than IE7 as a software, and, within IE8, IE8 rendering is slower than IE7 rendering. Contrarily to what people say on the Web, IE7 in my experience was indeed quicker than FireFox. I hope IE8 will keep the edge once in the release version.

    Apart from that I’ve noticed two things on which it’s probably worthy that I give you some feedback:

    1) negative text-indents are not correctly displayed by the IE8 rendering engine when viewing my site on my computer (scroll down the main column of my site to witness this). Sometimes they display, sometimes not, varies as I scroll; which makes me say it’s propably not a standards-compliance issue with my site. Negative text-indents are standard aren’t they? I guess they shouldn’t be cut part of the time like they are.

    2) Italic text is (at least sometimes) cut at the end of lines. See the links on the right column of my site for an example. I hope it does it also for you.

    I guess there is some weirdness with the way inline boxes are rendered, but hey, you know your job better than I.

    FYI, if ever it matters, I’m using XP sp2.

    Besides, I have a two questions that someone from Microsoft or someone better skilled than I am can probably answer:

    1) I don’t understand the View menu of the developer tools window. Does skipping to quirks mode automatically amount to viewing my site as rendered by IE5? I would say no myself, because some incorrect or non-existing DTD is required for a navigator to skip to quiks mode, and my site has a correct DTD; but I’m far for being an expert.

    2) I’ve read many times that IE8 enables me the viewing of my site as rendered by IE6. How do I achieve this?

  32. Webdesign says:

    What i would really like to see added is styling for select boxes, so developers can change the look of the button, drop down, and border of select elements.  Color is most important to me, but it would be nice if we could stick images into select structures as replacement for the default arrow button and as entries for the option tags.

  33. Markus says:

    Hey um, I registered on thet Connect site; but I can’t file a report.

    Also I can’t start a new thread on the Maillist, the new-thread-Button doesn’t work!

  34. Dave H says:

    Hi,

    In 2005, Chris Wilson said that Internet Explorer 7 wouldn’t have XHTML media type support[1].  This didn’t sound at all unreasonable to me, as the IE team had other priorities.

    Now the IE8 beta 1 is out, I note that support hasn’t been added, but work has been done to add XML-ish namespace support to regular HTML.  (This is a little unexpected, given that neither the HTML 4 Recommendation nor the current HTML 5 drafts have any such feature!)

    Can you please let us know what’s going on.  Is this part of a plan working to add XHTML support in a future beta?  Or does Microsoft no longer intend to support this standard?

    Thanks.

    [1] http://blogs.msdn.com/ie/archive/2005/09/15/467901.aspx

  35. mocax says:

    overflow:hidden and opacity isn’t working in IE8

  36. Hey Gérard, can you please contact me through my site? I can’t find any way to contact you through yours. I’d like to discuss IE8 and bugs. 🙂

  37. gabe says:

    id like to know how many ppl work on the ie team

  38. mocax says:

    More specifically a DIV with overflow:hidden ignores the property when the content in the DIV is a UL with position:relative.

    overflow:scroll and auto works fine though.

  39. Hi Doug Stamper,

    My suggestions to the next build of IE 8 beta

    when to click with shift key + left button mouse in to link:

    – to open new pages new tabs instead of new windows as presented currently.

    When browser is minized:

    – Hide it in an icon in the task bar.

  40. Frank: Ok, I have to admit, I resorted to a simple Google search, turns out that most things either aren’t indexed or not linked enough (and the experience of digging through Bugzilla myself was ugly enough for me one time I filed a bug :))

    Álisson P. F.: Especially that last feature will probably never see IE natively. The notification area is defined in the UX guidelines as reserved for … well … notifications. If you desperately want that functionality, there are 3rd party plug ins out there that do that. You can click links with the middle mouse button or use Ctrl-click to open them in a new tab. This is consistent with other browsers and works since IE 7.

  41. Steve says:

    @Gerard,

    Very good points… I’m well aware that several dozen bugs were fixed, and for that I do thank MS.

    However, there is 1 bug that if fixed would make us all very happy.

    Element.prototype.xxxx

    If we could prototype on any element type, we could fix the setAttribute, getAttribute, hasAttribute, hasAttributes, getElementsByName, add getElementsByClassName (for descendants), autoCompletionStorage, onchange for radio/checkboxes, etc.

    Adding the ability to prototype won’t break any existing content yet it would enable end developers to fix what IE won’t, or has decided not to yet.

  42. Tino Zijdel says:

    @EricLaw:

    "DataURLs are primarily useful for embedding small images into CSS for use as icons, etc; IE8 permits this use.  Unfortunately, for some content types, DataURLs can be dangerous, which is why IE8 scopes them to use where they are not a security threat."

    I was more talking about the 32KB limit which doesn’t really make it a viable option to compensate for the lack of <canvas> and native SVG. 32KB won’t let you generate things like complex graphs and such, even when you use a compressed imageformat like PNG or GIF (I actually do have a GIF implementation in javascript that uses data-url’s to show the image :P) instead of a simple format like BMP. I haven’t really compared IE8’s implementation to those in other browsers yet so I don’t really know what restrictions they impose, but at least they offer alternatives.

    "Note that there are many cases where using a DataURL will actually be LESS performant than using a HTTP URL with appropriate caching and compression headers."

    I know that, but I had other usecases in mind 😉

    "XDM is part of the HTML5 specification.  Search for "onMessage" for more information."

    But then why isn’t IE8’s implementation following the HTML5 specification? I noticed several differences and Anne van Kesteren confirms my initial findings ( http://annevankesteren.nl/2008/03/ie8-bad ). MS is acting again like the big gorilla in the room doing as they please without consulting the WG. Effectively they are again forcing their interpretation upon the world leaving the WG no choice but to alter the specification to IE’s implementation on several points. That is bypassing the WG to get what you want without having to reach concensus on possible issues and is actually hurting the standards process and interoperability (again).

    The same goes for MS’ decision to increase the per-host connection limit. You’ve effectively already decided over RFC2616bis WG deliberations, probably without having participated in that debat. Although I understand the reasoning behind this change it’s more the way that MS is again making this decision on it’s own as if they own the web. Afaik Firefox still ships with the default set to 2 (although that can be changed in the configuration or by using some extensions), I don’t know about Opera.

  43. David says:

    the emulate IE7 seen it dos not want to work

    i down lode the new IE8  and i did re start it when i got done down lodeing it  like it ask me to but it seen lie it dos not want to work

    do i need to down  lode any thing???

  44. Greg says:

    @Nicole,

    I too can verify that there are issues with RGB color values in IE8.

    I also noted that asking for an attribute via getAttribute() won’t necs. return the same thing you set.

    in the case of style, if you pass in "border: 1px solid #ff0000;" it will return MSIE’s broken UPPERCASE PROPERTY names exploded for each top, bottom, left, right. e.g. "BORDER-TOP: 1px solid #ff0000; BORDER-RIGHT: 1px solid #ff0000; BORDER-BOTTOM: 1px solid #ff0000; BORDER-LEFT: 1px solid #ff0000;"

    Its a minor thing, but something to keep in mind since IE already does strange stuff with form element’s value attributes, if the user has typed something in.

  45. EricLaw [MSFT] says:

    @Tino: The 2-connection per host suggestion in RFC2616 is a SHOULD, not a MUST requirement.

    I’ve already noted that both Firefox and Opera have surpassed the 2 connection limit, so implying that IE is somehow unique in this regard is incorrect.

    There’s no real dissent in the HTTP WG related to removing the connection limit.  See http://lists.w3.org/Archives/Public/ietf-http-wg/2008JanMar/0423.html for the discussion.  

    In http://lists.w3.org/Archives/Public/ietf-http-wg/2008JanMar/0429.html, Roy Fielding (whose name you’ll find atop RFC2616) says: "in regards to the topic, I see no reason for the connection limits."

    Please note, Anne’s post does not make ANY mention of IE8’s support for HTML5 Cross-Document-Messaging.  For any of you who might not be aware (and because it is not mentioned on his post), Anne is an employee of Opera Software ASA.

  46. Tino Zijdel says:

    @EricLaw: I’m not discussing whether or not it would be good to increase (or even abandon) the per-host connection limit but the more general issue of making implementations interoperable. That doesn’t mean cherry-picking and interpret the existing standards in a broader or strict manner where you see fit for your /own/ need but it means reaching some sort of consensus on certain issues in the field and then act in a simular way. I’ll see if I can find some more information on other browsers behaviour in that respect. Do you have any pointers?

    As for Cross-Document-Messaging, that’s my bad; I mistook XDM for XDomainRequest. And yes I’m well aware that Anne works for Opera, but he is also deeply involved in standards work so I have no reason to doubt his opinion.

  47. Rowan says:

    Is there any chance the IEBlog could introduce some new blog categories (tags)? For instance I might want to see all the posts relating to IE7, or newly supported CSS properties, or user interface changes, Add-ons, etc…

    Some of the existing tags are quite vague as well, what can one expect to find in "IE on the Web"? Everything except Intranet-related posts?

    Thanks

  48. JSMinch says:

    Nice beta.

    I have had a few pages not display properly,  but I was able to track those down to non-standard coding.

    The only pages that REFUSE to work properly are live.com and hotmail, LOL.

  49. Steve C says:

    Ok, I’m getting real impatient about this now.

    What is the deal with the invites to submit bugs?  (no offense to Gerald) but how come the people that want to submit bugs can’t submit bugs?  This seems totally counter intuitive!

    My invite status in Connect just sits there, with no activity, just status=active, while I wait for approval.

    I’m flabbergasted that you shut this thing down for over a year, and when you open it up, you don’t even make it public!?!?!?

    Did none of the ranting on this blog by just about everyone not indicate that they would like to participate in this process?

    For the record, I’m fine if you reject my request to enter tickets… but at least provide the professional courtesy to let us know that you are blocking us from entering them, and give an explanation as to why!

    Steve C.

  50. Alan Gresley says:

    @Gérard Talbot

    http://www.gtalbot.org/BrowserBugsSection/Opera9Bugs/MarginBottom.html

    http://www.gtalbot.org/BrowserBugsSection/Opera9Bugs/Opera9PaddingRightMisrendered.html

    "This appears to be a regression."

    That’s not a regression, that IE8 following the CSS standards for an overflow box (I like IE7 behavior better anyway). The extra space that Opera 9.26 and earlier showa in the second test case is actually an old Opera bug.

    http://css-class.com/test/css/overflow/float-container-margin-overflow.htm

    But your first case give me a good example of a bug with frozen backgrounds when scrolling. My test cases are worse since I have non scrollable overflow boxes.

    http://css-class.com/test/css/overflow/floated-overflow-length-with-inner-box-float.htm

    Go on try to scroll those boxes in IE8. I have many more new bugs to demonstrate and my test pages show existing bugs in IE7 still happening in IE8.

    http://css-class.com/test/bugs/ie/renderingbands.htm

    The fun begin. 🙂

  51. Manuel says:

    Correct me if I’m wrong, but in IE8’s version of Standards mode, the following are still messed up.

    .setAttribute() for "cellpadding", "cellspacing" (on tables), as well as "style"… which sometimes get confused as an object, esp. when calling object.getAttribute(‘style’) which should return a string, not IE’s deprecated quick access to the style object that you might expect when using object.style

    Where is the full list of what is fixed, and what is still broken in IE8?

    Where can I file these bugs with test cases so others can verify these bugs, and IE developers can verify that the fixes in Beta 2 address these?

    Manuel

  52. AC says:

    @Steve C

    I would venture a guess that if any invites are being given out, it’s to participants on here with an established, consistent, "Identity" with contact information of some type that’s contributed to this blog in a constructive way with given examples.

  53. Manuel says:

    Confirmed.

    You can’t set the style attribute on a TABLE tag in IE8’s version of Standards mode.

    You still need to use the IE6 hack,

    TableObject.style.setAttribute(‘cssText’, ‘Your CSS Property String Here’);

    You can use the DOM tool in IE8 to verify.

    Man I was really hoping bug 245 was going to be fixed in this release!

    http://webbugtrack.blogspot.com/2007/10/bug-245-setattribute-style-does-not.html

    IE8 does fix it for DIV, FORM, and TD tags from my testing, but TABLE is definitely broken still.

  54. Steve C says:

    @AC

    You would have figured that their top submitters of verified bugs in the IE7 Feedback Connect Site would have been given the first invites… but that didn’t happen.  (I was 1 or 2)

    I get the feeling they don’t want a lot of volume in the queue because they think it will look bad.

  55. RaFi says:

    What about finally making IE ICM-aware?

    Try and go to http://www.color.org/version4html.xalter

  56. Josh says:

    What HTML5 tags and attributes do you support? I’d like to use them on my website.

  57. MacDaddy says:

    @Tino

    [QUOTE1] MS is acting again like the big gorilla in the room doing as they please without consulting the WG. Effectively they are again forcing their interpretation upon the world leaving the WG no choice but to alter the specification to IE’s implementation on several points. That is bypassing the WG to get what you want without having to reach concensus on possible issues and is actually hurting the standards process and interoperability (again).

    [QUOTE2]EricLaw: I’m not discussing whether or not it would be good to increase (or even abandon) the per-host connection limit but the more general issue of making implementations interoperable. That doesn’t mean cherry-picking and interpret the existing standards in a broader or strict manner where you see fit for your /own/ need but it means reaching some sort of consensus on certain issues in the field and then act in a simular way. I’ll see if I can find some more information on other browsers behaviour in that respect. Do you have any pointers

    Most browsers already use more than 2 connections. The related RFC allows this. Where is the problem!?

    [QUOTE3]The same goes for MS’ decision to increase the per-host connection limit. You’ve effectively already decided over RFC2616bis WG deliberations, probably without having participated in that debat. Although I understand the reasoning behind this change it’s more the way that MS is again making this decision on it’s own as if they own the web. Afaik Firefox still ships with the default set to 2 (although that can be changed in the configuration or by using some extensions), I don’t know about Opera

    IE8 is not alone and the RFC doesn’t even prohibit this, at all!

    You’ve been wrong about several points you’ve made. Stop trolling/fan-boying and research first.

    It sounds like you’re Neelie Kroes’ advisor…

  58. MacDaddy says:

    @Steve C

    [QUOTE]but how come the people that want to submit bugs can’t submit bugs?

    They should definitely reconsider this.

    As someone else said, you have to wait and hope someone else posts the issue you want to report and then vote on it 😉

    IE8b1 has many problems and they won’t even be known unless (all) volunteers can freely report these issues on the Connect website.

    If they actively moderate/clean the issue reports then the volume increase would be manageable. Right now I can’t do anything but … "report a website problem" 😀

  59. Kamran says:

    Nice to see the IE8 beta.. I’m using it now.. seems nice and smooth..

    BUT can you please implement that "page title" functionality where the title of the page appears right to the URL when you pull the drop down menu of the address bar.

    This is implemented in some other browsers (FFox) and its really helpful when going through a long list of unfriendly URLs in the address bar’s drop down to have the title of the page. It simply fetches the <title> tag of that particular page and shows it.

    Hope I made enough sense 🙂

  60. Tino Zijdel says:

    @MacDaddy: I thought I was quite civilized in my posts and already admitted that I wasn’t aware of the fact(?) that other browsers already went beyond the 2 per-host limit, that’s exactly why I asked for some references.

    Fwiw, I still think that ‘SHOULD’ is quite a strong directive and that’s why I think that on points where the specifications ‘allows’ (as in: doesn’t prohibit by using ‘MUST’) you to be more loose implementations should be more equal. I can pinpoint several items in several RFC’s where IE’s implementation is either more slack or more strict than those in other browsers and that seems to me to be a bad thing.

    I know that there is more and more colobaration  between browservendors to resolve those kind of interoperability issues, but Microsoft just doesn’t seem willing to participate in such efforts, which to me is disappointing at least.

  61. Tino Zijdel says:

    And on a sidenote: I’m already doing quite some investigation on several items regarding IE8 in my spare time. It’s very time-consuming and I’m sorry that I haven’t really focussed on the good things. I’m just trying to get as much information I can before I can really make judgements on detailed points.

    At this point the only judgement I made is that MS is not really into colobaration with the other browservendors and the several W3C WG’s they’re participating in. We’ve seen what that leads to in the past, and even recently wrt the X-UA-Compatible flag were they had to revise that behaviour for several reasons. I hate to see MS making those same mistakes when implementing other new features or changing existing behaviour.

  62. David says:

    when will Internet Explorer 8 Beta 2 be out ???

  63. Web Developer says:

    Now that’s something good to hear.

    Now just make sure you fixed the damage you’ve already done and get everyone to update (when the public version is done).

  64. Ivan says:

    When there is a MASS posting like the past 2-3 days, you really need a TOC (Table Of Contents) for this blog, as major info gets lost in the blur of posts.

    If you had a recent posts section on the right, listing say the past 8-10 posts, readers would be able to see at a glance the last post they read, and what they’ve missed.

    Thus as of Sunday March 9th, the list would look like:

    ———————

    * Internet Explorer 8 Beta 1 for Developers – Standards Highlights

    * Improved Productivity Through Internet Explorer 8 Developer Tools

    * The Default Layout Mode

    * IE8 and CSS 2.1 Testing

    * IE8 and JScript

    * Activities and WebSlices in Internet Explorer 8

    * IE8 and IP Licensing

    * Why Isn’t IE8 Passing Acid2?

    * IE8 Beta Feedback

    * Internet Explorer 8 Beta 1 for Developers Now Available

    I was reading the feedback on the "IE8 Beta Feedback" post… by the time I was done (read over a few days)… 7 or 8 more posts had sailed through!

    Thanks, ivan

  65. @Steve

    You said "there is 1 bug that if fixed would make us all very happy.

    Element.prototype.xxxx

    If we could prototype on any element type, *_we could fix_* the setAttribute, getAttribute, hasAttribute, hasAttributes, getElementsByName, add getElementsByClassName (for descendants), autoCompletionStorage, onchange for radio/checkboxes, etc."

    prototype is just another bug. It should not be fixed so that *we* could work around lots of other bugs in IE. This hacking and workaround-IE-bugs mentality must die with IE 8. The correct, adequate, appropriate, responsible and proactive mentality (starting with IE 8) should be, must be:

    1- I find/discover a bug in IE

    2- I create a reduced testcase of it

    3- I report the bug to Microsoft so that **Microsoft IE Team** fixes it for everyone and forever, period.

    That mentality should have been dominating, prevailing in 1999, 2000, 2001, 2002, 2003, 2004, 2005, 2006, etc…

    Just one example coming from Eric Meyer himself. The incorrect CSS 1 box model was known to Microsoft as soon as IE 3 was shipped and RTM.

    "The classic example of this was the original implementation of height and width in Internet Explorer, which was wrong per the CSS specifications. The IE team at the time became aware of this *_fairly soon after they shipped it in IE3_* … and yet the problem wasn’t fixed until IE6, a delay that slowed the adoption of CSS and gave rise to a whole family of JavaScript sniffers and CSS hacks."

    In fact, it never was fixed everywhere and for everyone: it was just pushed into a new created rendering mode and then fix in that newly created rendering mode.

    So today, the web is full of hacks of all sorts, ugly fixes, non-forward-compatible workarounds, junk code, etc… which are working around IE bugs. Again, if Microsoft fixes bugs soon and fast after they are discovered, then the web developers will not have sufficient time to create and to publish non-webstandards-sensical hacks about them.

    Regards, Gérard

  66. Thompson says:

    Once again the Connect site has proven to be the worst public bug tracking system ever!

    Not only can we not log in to submit bugs, but even viewing them is messed up.

    Just how am I supposed to view a screenshot or a test case if there is no link to the file?

    What gives?!

    Are you seriously telling us that in the time since IE7 was announced as being "On the roadmap", that no one has spent *ANY* time putting something together to manage this whole process?

    I can’t believe it is 2008 and IE is the only major browser without a realistic public bug tracking system.

    Talk about your prime candidate for a feature on the dailywtf!

  67. andy says:

    Great.  Now work on the new acid 3 test

    http://acid3.acidtests.org/

    IE7 scores 14/100

    IE8 scores 17/100

    ffox 3.0bpre5 scores 69/100

    and let the live.com team know their site is completely broken.

  68. andy says:

    … I sent the live.com team a note yesterday to talk to Dean, Chris, Alex & Markus.  If Microsoft wants to hire me for being an inside man let me know:)  I just like learning about how you all do business.

  69. @Manuel

    "You can’t set the style attribute on a TABLE tag in IE8’s version of Standards mode."

    "the following are still messed up.

    .setAttribute() for "cellpadding", "cellspacing" (on tables), as well as "style"."

    Manuel, that is not what I see, notice at these webpages:

    http://www.gtalbot.org/DHTMLSection/CreatingTable.html

    http://www.gtalbot.org/DHTMLSection/DynamicTableFormatting.html

    Of course, some functions do not work perfectly but overall, I can set the style of a table object and table cell objects.

    Regards, Gérard

  70. html namespaces are dangerous says:

    It would be great, if the IE8 team should be working on XHTML support, but please don’t add HTML namespace support in IE8 final.

    If you honestly intend to have namespaces in HTML, then please coordinate this with the HTML5 WG.

    Personally, I even think that xml-like namespaces would be great to have in plain html, because this would hopefully keep the use of microformats at a tolerable level and instead add a reasonable to use html extensions, But such changes should be introduced in the proper way and it would be dangerous to introduce an unspecified xhtml/html hybrid mode of parsing.

  71. @Steve, Still Waiting and all those who want to file bugs in the public bug tracking system.

    Create a reduced testcase of the bug you see: shorter is usually better. Make sure there is a clear PASS and clear FAIL in such minimized testcase: the test should be easy to figure out for anyone. Make sure your markup code and CSS code are perfectly valid. Your reduced testcase should be using a doctype declaration (preferably) referring to a strict DTD. Upload it on your website.

    I’m sure now Microsoft is looking for such testcases. Again, the best ones are the ones that are short preferably minimalist, clear, easy to figure out, using valid code, identifying only 1 bug per testcase. Be as factual and as formal as you can in such testcases. In longer testcases, clearly identify steps to reproduce and expected results.

    CSS2.1 Test Case Authoring Guidelines

    http://www.w3.org/Style/CSS/Test/guidelines.html

    If you do that, I’m absolutely sure Microsoft will appreciate such testcases and fix the bugs because such testcases help not only figuring out bugs but to fix them and to verify that they are fixed.

    2 examples (from Nicole’s report):

    http://www.gtalbot.org/BrowserBugsSection/MSIE8Bugs/ShorthandBackgroundAndRGB-1.html

    411 bytes

    http://www.gtalbot.org/BrowserBugsSection/MSIE8Bugs/ShorthandBackgroundAndRGB-2.html

    328 bytes

    ——–

    Other idea. You can still report your good, clear testcases with the

    Report a webpage problem (in IE add-on)

    http://www.microsoft.com/windows/products/winfamily/ie/ie8/readiness/ResourcesAndLinks.htm

    Regards, Gérard

  72. @Thompson

    I agree with you that the Connect IE Bug database reporting/feedback website is not an ideal or even just a good bug database reporting website.

    But complaining at this point is not going to be useful, helpful to anyone. At the very least, identify what is wrong/incorrect with that Connect bug tracking system and what you would improve in it and how you would improve it (hints: make useful, constructive criticisms, be suggestive, preferably with concrete proposals, concrete proposals, examples, screenshots, etc) and do so preferably at the best place, at the opportune moment. A good place to do all this would be on your own website.

    If you want to report bugs, you still can do it in many ways (read the other IE blog messages pointing to IE8 documentation webpages).

    Download and install the Report a webpage problem (with IE add-on).

    File your bugs with that "Report a webpage problem" IE add-on.

    Create your own webpage/testcases on your site.

    Regards, Gérard

  73. grd says:

    I am trying to use the Emulate IE7 as some of the pages I was trying to visit were broken.

    Identified 3 important issues.

    1. When I use the vertical scroll bar it comes down but again it goes on to the top. It happens randomly

    2. When I use favorites it crashes my IE

    3. I think still it has some issues displaying CSS based sites. The width and the height are a bit abnormal to what I see with Ie7. The Emulate IE7 doesnt work exactly as its expected to work.

    Hope this feedback helps.

  74. Arieta says:

    A question… if the developer toolbar can switch between IE7 and IE8 rendering at will, why can’t the IE7 emulation button do this, only after restarting the browser?

  75. PatriotB says:

    @html namespaces are dangerous — IE has supported namespaces in HTML documents since IE 5 (1999).  The only thing that they are doing in IE8 is allowing these custom elements to function without IE-specific goo (an Object tag to pull in the behavior, and an Import directive to associate the behavior with the namespace).

    I’ve heard some HTML5 folk complaining about supporting namespaces in HTML but still haven’t heard a good argument against it.  Frankly, it would be better if HTML5 would focus on the basics (interoperable error handling, clarification to undefined behaviors, etc) and include an extensibility mechanism (namespaces) so that the remaining "goodies" (video tag, etc) could be implemented separately from the main spec.

    Incidentally, IE has supported a video tag (the SMIL video tag) via HTML+TIME since 1999.  Why this video tag isn’t "good enough" for the HTML5 folk, and why they feel the need to re-invent the wheel, is beyond me.

  76. Omar says:

    How can i hide the horizontal… scroll bar… it shows up… even if not needed… thx…

  77. G says:

    Standards? Where is your support for SVG?

    How are you going to pass Acid 3 without it?

  78. Omar says:

    How can I hide the horizontal scroll bar?…

  79. AC says:

    @gabe

    Bad press? About Microsoft? On Slashdot?

    Inconceivable!

  80. Anonymous says:

    Yea, this horizontal scroll thing is killing me. how did that make it to a beta?

  81. Roger Wilson says:

    So after watching several of the MIX 08 videos on IE 8, I took the plunge and installed IE 8 on my Vista (64-bit) laptop. After restarting the laptop, IE 8 had completed its installtion.

    Everything seemed to be working fine until I visited a page containing a Silverlight control. IE 8 just stopped, dead in its tracks. No CPU utilization, no mouse response, etc.

    I tried uninstalling Silverlight Beta 2 and installing Silverlight 1.0. The result was the same.

    Ended up uninstalling IE 8, which worked. Upon rebooting IE7 was back. When I visited a silverlight page everything worked.

    Disapointed.

    Roger Wilson

  82. Paul Allsopp says:

    Well, for a B1 I personally thought it was fine.

    My only real issue is that a page designed in Expression Web (XHTML/CSS2.1) looks different in IE7 and IE8. So I’m hoping its not soon going to be a case of having to design for IE7, IE8 and FX (what’s Opera?).

    Comformance with standards is great, but who decides on the standards? Surely it should be the first team to fully develop a procedure, not a bunch of over-educated tie-wearing monkeys with nothing better to do than agree with each other about nothing.

    I like the new features. Not something I would use as a user very much, but certainly one more item for the toolbox.

    If this was an alpha I could imagine all of the above bitching being warranted, but as it is just a beta maybe some people need to rein in their attitude. If you got asked to testdrive the new Lambi (still in development) you wouldn’t come back and bitch about it.

    Constructed criticism is a great developer tool…stressed and bitching developers are just tools. 😉

    Great work IE team.

  83. nonA says:

    "My only real issue is that a page designed in Expression Web (XHTML/CSS2.1) looks different in IE7 and IE8. So I’m hoping its not soon going to be a case of having to design for IE7, IE8 and FX (what’s Opera?)."

    One word: meta

    "Comformance with standards is great, but who decides on the standards? Surely it should be the first team to fully develop a procedure, not a bunch of over-educated tie-wearing monkeys with nothing better to do than agree with each other about nothing."

    That’s brilliant! Congratulationz, you win a cookie!

  84. Anonymouse says:

    RE > Gérard Talbot

    Very constructive work – a job well done.

  85. Alan Gresley says:

    @nonA

    —————–

    "My only real issue is that a page designed in Expression Web (XHTML/CSS2.1) looks different in IE7 and IE8. So I’m hoping its not soon going to be a case of having to design for IE7, IE8 and FX (what’s Opera?)."

    One word: meta

    "Comformance with standards is great, but who decides on the standards? Surely it should be the first team to fully develop a procedure, not a bunch of over-educated tie-wearing monkeys with nothing better to do than agree with each other about nothing."

    That’s brilliant! Congratulationz, you win a cookie!

    —————–

    Hello anonymousN. Do you have a name? You really think the solution is a meta tag. Ok, I have one test case to show you dealing with display table.

    http://css-class.com/test/css/table/table-property1.htm

    Indeed no meta, no hack, no nothing is going to correct this. All boxes should be 100px high. IE8 support of display table properties is badly broken. This will result in a large number of sites breaking if an author has used display table properties for the better browsers. This I see as the most critical CSS bug to be fixed.

    So instead of presenting cookies to anyone, considerer the real issues. Try navigating IE8 to this page.

    WARNING!

    http://www.ruzee.com/blog/2007/05/align-list-items-horizontally-with-css/

    WARNING: This page will cause IE8 to freeze for about 5 minutes before you can stop it. The blog entry is able using display table properties. Test for yourself anonymousN.

  86. Ken says:

    @PAul Allsopp: "what’s Opera?"

    Its the browser that renders much faster than IE, had Tabs before Firefox and Safari and IE and the one you should be making sure your pages/sites work in.

    Currently when designing a site, you want to make sure it will work in:

    Firefox 2.x-3.x, Safari 3.x, Opera 9.x (not the Alpha’s), IE 7 (possibly IE6 if you still support it) (IE8, when the RTM is out), Konqueror 3.5-4.x.

    There are other browsers that you’ll likely want to check in too, but this is the core set.

    "Comformance with standards is great, but who decides on the standards" – Easy, the W3C

    http://www.w3c.org/

    This is your "Bible" to web development.

    "If this was an alpha I could imagine all of the above bitching being warranted, but as it is just a beta maybe some people need to rein in their attitude"

    Well, if you had been around for the IE7 beta fiasco, then you’d know that RANTING early is REQUIRED, because before you know it, MS is going to announce that they are "Render Complete" indicating that they’ve stopped fixing UI bugs in the layout engine, stopped fixing broken/unimplemented JavaScript, and have given up on fixing the browser’s UI/Chrome.

    Microsoft haven’t even fully opened their bug tracking system **YET**, so the bugs have yet to flood in.  I’m scared because I know that there are a ton of bugs that will be "missed" because the IE Team didn’t hear about them until it was "too late".

    At the moment, this beta is more like an Alpha release, I sincerely hope that there are at least 2 more betas before the RC’s start.

  87. Stifu says:

    Ken: Privileging Opera 9 over IE6 is currently not realistic at all for most commercial/professional sites. IE6 still has over 1/3 of market share, compare that with Opera’s 1%…

    But yeah, for a personal site, it can be done. Dean Edward’s (updated) IE7 JavaScript is also a nice way to bypass many limitations of IE6 and 7.

    As for Opera, with its 1% of market share, it’s hard to justify tweaking your site just for it. Personally, at work, I only bother fixing the site if Opera is right to display it as broken. If it’s obviously an Opera bug, I don’t bother. And since Opera cares about standards, chances are they’ll end up fixing things on their side.

    However, on a side note, Opera’s closed bug-tracking system is a pain to work with.

  88. Sean says:

    @Stifu,  In general I find that if I code to standards, my content works in Firefox, Safari and Opera automatically! (hence the beauty of standards), the only catch is Opera not fully supporting overflow-x/-y properly.

    After coding to standards, comes the joy and pain of tweaking for IE6 and IE7 (read: horrible ugly hacks).

    IE8 now in the mix, has made things even more interesting, as it has its own set of fixes, bugs, regressions and other random oddities.

    currently, if IE8 were to go live tomorrow, I would recommend to all customers to use the emulate IE7, because IE8 is still way to immature to use for work purposes.

    Your comments on Opera are somewhat just… the bug tracking does need to be more open (although at least we can file bugs in Opera!!!!), but as for market share, it may be low, but it is really easy to support.

  89. Jack T. says:

    Bug Report: JT-001

    IE8B1:

    Scope: JavaScript

    Synopsis: Element.hasAttibute(‘style’); returns false when Element has a style attribute set.

  90. Jack T. says:

    Bug Report: JT-002

    IE8B1:

    Scope: JavaScript

    Synopsis: TableElement.setAttibute(‘style’,’…’); does not correctly set the style attribute on the table.

  91. Jack T. says:

    Bug Report: JT-003

    IE8B1:

    Scope: JavaScript

    Synopsis: Element.hasAttributes(); does not even execute (I believe it is not implemented), but also stops any further JavaScript execution.

  92. Jack T. says:

    Bug Report: JT-004

    IE8B1:

    Scope: JavaScript

    Synopsis: Element.innerHTML; returns UPPERCASE tagnames, not what was set (lowercase), this occurs regardless if it is the original HTML sent on the response or dynamically added via JavaScript

  93. Jack T. says:

    Bug Report: JT-005

    IE8B1:

    Scope: JavaScript

    Synopsis: TableElement.setAttibute(‘class’,’…’); does not correctly set the class attribute on the table.  Works on other elements.

  94. Jack T. says:

    Bug Report: JT-006

    IE8B1:

    Scope: JavaScript

    Synopsis: Element.setAttribute(‘on{eventName}’,’handler’);

    Sets the HTML attribute but does not register the event handler with the Element.

    Setting event handlers like this in HTML works fine, thus not supporting it fully via JavaScript is a bug, especially since it does store the attribute value.

  95. Jack T. says:

    Bug Report: JT-007

    IE8B1:

    Scope: Browser UI

    Synopsis: User can’t select a text fragment on screen and drag-n-drop into an input text box on the page.

  96. Jack T. says:

    Bug Report: JT-008

    IE8B1:

    Scope: Browser UI

    Synopsis: User can’t select a hyperlink on screen  and drag-n-drop to *anywhere* except the favorites bar (bar only), or another application outside IE8.

    Minor exception, user can drag and drop to the 1 pixel bottom border underneath the IE8 tab row.

  97. Jack T. says:

    Bug Report: JT-009

    IE8B1:

    Scope: Browser UI

    Synopsis: User can’t right click on an iframe (or frame) and choose an option to view that frame in another tab or window.  This is a major usability issue when trying to view specific contents, or trying to print a specific frame only.

  98. Mike says:

    @Jack T

    Some of the ‘bugs’ you are reporting aren’t actually bugs. Sure, they’re good usability points, but the IE team have never announced they were introducing drag and drop text, and the like.

  99. Jesse Pelton says:

    I see no mention of supporting the standard DOM 2 event model (Element.addEventListener(), Event.target, bubbling, capturing, that sort of thing).  I’m wondering if we can expect any movement in that direction.  Maybe a later beta, or IE 9?

  100. Jack T. says:

    @Mike

    They haven’t announced the ability to drag and drop text or hyperlinks… no, however these are regression issues that creeped* into IE7, and have not be fixed in IE8.

    When IE7 was being released MS indicated that they did lose UI functionality when they added tabs.  They were sorry that it didn’t make it into the rushed IE7 RTM release, but getting a new version out was the priority.

    Thus here we are with IE8 on the horizon, and these issues have not been addressed.

    Regardless, a bug by definition, is any non-expected behavior.  It may be resolved as a non-issue if the new/odd behavior has a rationale to support it, but in this case there is no respectable reason why drag-n-drop of text, and in particular hyperlinks is not appropriate when other browsers (including IE) have supported this for ages.

  101. Brent says:

    I thought hasLayout was finally dropped in IE8 Standards mode?

    If I do:

    javascript:alert(‘Has Layout? ‘ + window.document.body.currentStyle.hasLayout);

    IE8 in standards mode returns true.

    So what’s the story, is it dropped, or not?

  102. Jack T. says:

    Bug Report: JT-010

    IE8B1:

    Scope: JavaScript

    Synopsis: SelectElement.setAttibute(‘style’,’…’); does not correctly set the style attribute on the select element.

  103. Jack T. says:

    Bug Report: JT-011

    IE8B1:

    Scope: JavaScript

    Synopsis: Setting the style (via style.setAttribute(‘cssText’) due to bug JT-010) on a Select element changes the chrome (on XP pro) rendering of the drop arrow, if the background color, or border is changed.

    The chrome does not get altered if the font is changed (font face or color), width, or height, etc.

  104. Jack T. says:

    Bug Report: JT-012

    IE8B1:

    Scope: JavaScript

    Synopsis: SelectElement.setAttibute(‘class’,’…’); does not correctly set the class attribute on the select element.  Works on other elements (except Table as noted in bug JT-005).

  105. Jack T. says:

    Bug Report: JT-013

    IE8B1:

    Scope: JavaScript

    Synopsis: SelectElement.getAttribute(‘class’,’…’); does not correctly return the class attribute on the select element.  This bug occurs regardless how it was set, either via HTTPResponse HTML class="foo" or SelectElement.setAttribute(‘className’, ‘foo’);

    (using className in this example since bug JT-012 will not allow correctly setting the ‘class’ attribute)

  106. Jack T. says:

    Bug Report: JT-014

    IE8B1:

    Scope: JavaScript

    Synopsis: SelectElement.hasAttribute(‘class’); fails to execute and halts remaining JavaScript execution. (likely related to bugs JT-012, & JT-013)

  107. Jack T. says:

    Bug Report: JT-015

    IE8B1:

    Scope: JavaScript

    Synopsis: Potential security issue/spoofing ability.  On any page with an iframe, you can set the src to "about:Tabs" which renders the first time user screen for a new tab.  A malicious page could render "what appears to be" the new tab page, whilst JavaScript in the parent container page does undesired things. (e.g. launching random popup ads.) (Minor issue)

  108. Reid M. says:

    Thank you, IE team, for pushing through this decision and reverting your earlier decision. Strict/standards mode is much easier to opt into and is a fair compromise.

    It’s bad enough that web developers will have to be testing for IE6 for years to come, much like the slow death of IE5.

    It feels like this will finally start moving the web forward.

  109. Pundit says:

    @@JackT #15: How is that even ~remotely~ a security or spoofing issue?  The bad guy could just as easily serve the same HTML instead of going to about:tabs.  And the popup blocker would block the popups from the outer page anyway… there’s no need to show the new tabs page, as it doesn’t get the attacker anything…

  110. Francis says:

    It’s great to see improved accessibility in IE, but keep in mind that accessibility is not only about semantics. It’s also about increasing usability for disabled persons (e.g., big print–or ALT tags–may help the disabled, but if they can’t turn–or navigate–the pages, they’re no better off.) And in terms of usability for those who lack the characteristics required for the use of a point-and-click medium like the web (high dexterity, visual acuity, etc.), IE comes up short.

    The IE team should consider developing a holistic alternative to the reigning point-and-click paradigm. By that, I do not mean a TAB here, an ALT+LEFT there. Document semantics and structure should not only accessible but also actionable to them, regardless of whether they point, type, or dictate. Though IE comes up short in this regard, other browsers excel. Take Opera. It provides users with the ability to, among others:

    – spatially navigate elements within a document (even with frames) with the arrow keys: SHIFT+arrow keys

    – select (and follow) links/text by typing with inline link/text search:  ,+SEARCHLINK (ENTER) and .+SEARCHTEXT

    – select the next/previous link, heading, and document element within a document with a single key: a/q, s/w, d/e

    – predictively follow links to a succeeding/preceding page (e.g. in search results) with a single shortcut: CTRL+SHIFT+ or CTRL+ALT+RIGHT/LEFT

    – fast forward/rewind to a succeeding/preceding site (instead of a page) with a single shortcut

    – follow embedded, site-specific navigation elements (next, previous, home, up, search) with a single shortcut or navigation bar

    – move focus from widgets (chrome and controls) to page: F9

    – move focus to next/previous frame: 3/SHIFT+3

    – navigate by mouse gesture

    – navigate and control the browser by dictation

    Provisions such as these make a difference. They’ve personally saved me a great deal of pain (and time.) If IE team is serious about improving accessibility, they should take a long and hard look at usability. Trying to use IE sans mouse would be a good starting point.

  111. Rowan says:

    @Alan G

    I made pretty simple test case for IE8 with CSS tables, check this out: http://www.rowanw.com/bugs/ie8b1-display-table-cell.htm

    Definitely broken.

  112. @Jack T.

    1- Have you tested

    Element.hasAttibute(‘style’);

    or

    Element.hasAttribute(‘style’);

    ?

    What’s the URL of your testcase? Can you upload it somewhere?

    2- Have you tested

    TableElement.setAttibute(‘style’,’…’);

    or

    TableElement.setAttribute(‘style’,’…’);

    ?

    setAttribute() works as expected and as specified by the spec on style at

    http://www.quirksmode.org/dom/tests/attributesBasics.html#setAttribute()

    3- Have you tested TableElement.setAttibute(‘class’,’…’);

    or

    TableElement.setAttribute(‘class’,’…’);

    ?

    etc.

    Can you provide URLs for those tests of yours?

    We know that Element.hasAttributes(); does not work as expected: see bug #54 at my webpage

    ——-

    Regarding "User can’t right click on an iframe (or frame) and (…)": that would be a feature request; not a bug like a spec violation or incorrect implementation.

    Regards, Gérard

  113. @Rowan

    border-spacing and border-collapse properties only apply to "table" elements and "inline-table" elements. They do not apply to table-cell elements.

    Also, the default border-collapse value is separate. So, your whole CSS code can be reduced to

    table {border-spacing: 0;}

    which is the initial, default value of border-spacing.

    I’m afraid your testcase needs work.

    Regards, Gérard

  114. Rowan says:

    @Gérard

    Thanks, I forgot to remove the extra CSS, I was experimenting. 😉

  115. PatriotB says:

    I find this one kinda funny:

    "Bug Report: JT-004

    IE8B1:

    Scope: JavaScript

    Synopsis: Element.innerHTML; returns UPPERCASE tagnames, not what was set (lowercase), this occurs regardless if it is the original HTML sent on the response or dynamically added via JavaScript"

    Not only is innerHTML a proprietary property invented by Microsoft (meaning it can return whatever Microsoft wants), HTML 4.01 has no case sensitivity in tag names.  The browser can return eVeRy oThEr letter uppercase if they’d want to and it would be perfectly valid.

    It’s a nice request, but why should the parser keep track of the case that each element was parsed in at?  I imagine that the element uses an enumeration (e.g. ELEMENT_TAG_ID, http://msdn2.microsoft.com/en-us/library/aa769842(vs.85).aspx) to internally store what type of tag it is; for each element to have to keep track of the case-sensitive string of the tag name would be a waste of memory.

  116. Jack T. says:

    @Gerard

    Yes I have tested all the .setAttribute bugs outlined above! (JT-001, JT-002, JT-003, JT-004, JT-005, JT-006, JT-010, JT-011, JT-012, JT-013, JT-014) Thats how I found them!  Not to mention that making my applications work in IE8 standards mode required me to tweak my code.

    I had a wrapper for setAttribute, that fixed all bugs outlined in bug 242 (here)

    http://webbugtrack.blogspot.com/2007/08/bug-242-setattribute-doesnt-always-work.html

    Which works great for IE6, IE7.  I foolishly removed the ENTIRE wrapper for IE8, and found things broke right left and center.

    Anyway, I *do* have test cases, but I’m not going to upload any of them until I get an invite to IE Feedback on Connect (or it goes public).  The WHOLE purpose of a centralized bug tracking system is that there is ONE location where EVERYONE can view/add/edit bugs testcases and workarounds.

  117. Francis, moving focus from/to chrome is done in IE with F6 and works since at least IE6. Navigating and controlling the browser by dictation is works with IE 7 on Vista.

  118. Lee says:

    How about the last-child pseudo class?  I find this useful in cases where i have a div with multiple p tags (margin-bottom: 15px) within and I wish to remove the bottom margin for the last element in the div so i don’t get extra padding.

  119. Gunther says:

    Hi!

    I’m having trouble with the installation of IE8 beta.

    Everything seems to work well (including "Updates are being configured …"), but after the Reboot nothing has chenged – still IE7 in place.

    Any hints or ideas?

    Greetings

    Gunther

  120. AC says:

    @Francis

    I’ve been using IE sans mouse since IE 6. I thought the keys were more or less known, but maybe it’s just my being used to them.

    I always found Opera kind of awkward that way because of it.

  121. Richard says:

    It would be extremely helpful if we could specify what box model, inclusive or exclusive of borders, to use while retaining standards compliance in other respects.

    This is absolutely a MUST HAVE.

  122. PatriotB says:

    @Richard — I’ve noticed IE8 supports the CSS3 box-sizing attribute (-ms-box-sizing:border-box).

  123. Adriaan says:

    You claim standard compliance with IE8 beta 1, is that only html compliance or what?  Most of my sites is xhtml transitional 1.0 but they hardly work in IE8 beta1… :hover effects take forever, layout is screwed completely…ha!

    why can’t IE just die and everybody switch to firefox?  have you seen the firefox 3 beta 4…damn man, that’s how a browser should behave…

  124. I know there have already been some test pages posted for this problem, but I’ve made my own that I hope makes the problem really clear.

    http://patrick1978.com/ie8test.html

    Unfortunately I have yet to see this bug posted to the MS Connect Site so I can’t vote on it or enter it myself, otherwise I would.

  125. Rag says:

    Adriaan,

    Instead of just complaining, why not actually post what the issue is, or better yet test cases (assuming you have the skill for that)? Do you remember how lousy FF 3 beta 1 was?

  126. @ Lee

    You said:

    "

    How about the last-child pseudo class?  I find this useful in cases where i have a div with multiple p tags (margin-bottom: 15px) within and I wish to remove the bottom margin for the last element in the div so i don’t get extra padding.

    "

    CSS 2.1, margin collapsing is now fully supported and *correctly* implemented in IE 8. So, what you are asking for no longer needs to be worked around or coded for.

    "The bottom margin of an in-flow block-level element (…) is adjoining to its last in-flow block-level child’s bottom margin if the element has no bottom padding or border."

    CSS 2.1, section 8.3.1

    Visit my webpage and bug #36 to verify for yourself

    Regards, Gérard

  127. Rory says:

    Looking at IE8 for the first time, I work as a web developer and am trying to ensure that our web applications will work with the new browser. Thinking that I would have to create yet another style sheet to ensure my sites are rendered correctly you can imagine my joy when I heard about the X-UA-Compatible method. That Joy was however shortlived. Contrary to the documenation changing this mode to IE7 does not render the site as the shipped version of IE7 does… could you explain why this is, if it doesn’t render the same then I realy don’t see the use of it and its extremely frustrating to be led down the garden path like this…

    Has anyone else noticed this and is there any plan to resolve this before IE 8 hits the market????

    Things like this cost small business like mine thousands in lost man hours!

  128. After some further testing I’ve updated my test page to be much more descriptive of the CSS background/RGB color problem, and to test more combinations of background declarations.

    http://patrick1978.com/ie8test.html

  129. Our friends over in the Internet Explorer building recently released a developer preview version of IE8.

  130. IEBlog says:

    With the release of IE8 Beta 1, I’m pleased to be able to talk about the first round of improved standards

  131. IEBlog says:

    This blog post frames our approach in IE8 for delivering trustworthy browsing. The topic is complicated