Read the full changelog. The line is red. (What a horrible color for a pass condidition, read Writing Test Cases for Web Browsers.) We support xml:id. Including the weird case of appending it to a text/html document. Please find bugs and report them here, on the forums or in the bug database.
First browser with experimental support for Web Forms 2. Please, play with it! We also support html:canvas now. On that element we support toDataURL(), but not toDataURL(in DOMString type) as it was added to the specification at a later date.
And more. Download for Windows.
Hooooray for Opera!
It crashes when I press the first button in the Sample Order Form test case.
Anne, does it support POST using XMLHttpRequest now? I didn't find evidence of this in changelog
I am glad to see that Opera added the support for the HTTP Link header field. And passing Acid 2 will not last long anymore. You are on the right way!
Their Web Forms 2.0 implementation needs a lot of work.
zcorpan, yeah known bug. We should have fixed that though before the preview. Dean, I would welcome test cases. Like I said, it is experimental.
Alexey, you would have to test. At the moment I am underinformed about all the things we support. (It is hard to keep track of too.)
Rob Eberhardt, I do not really get your point. It is incredibly useful to attach a style sheet to thousands of documents without having to modify the markup of them. The Link header originates from HTTP 1.0. It was not included in HTTP 1.1 due to lack of implementations. Now there are two implementations, there is a chance of getting it back in, or standardizing it in some other way.
Seems like they've fixed some absolute positioning bugs, so your menu renders all the way across the screen now. Is this your doing, by chance? They also have fixes for things concerning HTML comments and @import rules. Things like your recent posts have been about. Coincidence? :-)
I need to get over opening new tabs with Ctrl+T instead of Ctrl+N, though; I just opened three different windows. @_@
I pushed some things, yes. There are other people too though, like Tim Altman and Tarquin. I'm very happy with the fact that we follow the initial containing block rules correctly now. People no longer have to bitch at me for making my site in a way that it does not look equivalent in Opera and Firefox.
Yeah, red sucks for pass conditions. Now I never get to work at Opera (or Google) ;(
Oh well, it makes people laugh and it's good for elitism.
Ah. As long as Opera allows you to redefine keyboard shortcuts I don't mind them changing opening a new page from CTRL+N to CTRL+T, it's easy to change back.
I'm still missing the last-child selector though
Which parts of the WebForms 2.0 specs is supported ? Is there a list of implemented specs ?
(Noticed that the demo of the DateTime input type works but the demos of repition model and prefilling the repetition model don't.)
Nice improvement to SVG support too! :)
What about NPRuntime stuff? Do you know if that will that be included?
Alexey: Opera has supported POST for XMLHttpRequest since 8.0, in a workable manner since 8.01 (with the introduction of support for setRequestHeader), and correctly since 8.02 (in 8.01 they always appended "text/xml; charset=utf-8" to any content-type that was sent).
Woei! 3 out of 4 bugs I once reported have (finally) been fixed in this release!
The fourth one is a bug that is subject to discussion because of vagueness in the specs: in Opera clipped content can obstruct underlaying elements (such as links) as shown on http://therealcrisp.xs4all.nl/meuk/opera_clip_bug.html
:lang() selector no longer applies to xml:lang attributes.
What? Why not?
Overall, big number of fixes, great step forward.
~Grauw
Laurens, I think that was a mistake. From what I have tested we only fail HTTP related ones on :lang().
Looks like there's some interesting stuff going on there.
fieldset legend { display: none; } seems to be broken :?
That was a known issue. It was fixed today. Your other bug is being worked on, we might have it in final. I can not promise anything though.
Alright Opera! Hopefully this continues to push Web Forms 2.0 forward. It's all so very simple when you think about it. How slowly these standards develop. Arghhh!!!
Laurens,
Do you mean for HTML (text/html) or XHTML (application/xhtml+xml)?
As far as I could test the :lang() selector never works in Opera 8 with xml:lang for real XHTML documents. I'd really like to find out I'm wrong about that and why.
On another subject I have the same problem with XML stylesheets that you pointed out on Tim Altman's weblog and for which he made a test case. On my site it's pretty radical.
What I think that the release notes meant to say was that as we do no longer support namespaces in HTML (as in text/html) xml:lang is no longer recognized in text/html as a "lang" kind of attribute and therefore is not used for :lang() based matching. Unless it really is in a namespace which can only be done through the DOM for text/html documents.
And also, I can assure you we fix the alternate style sheet processing instruction bit in time. As a matter of fact, it should be fixed in internal builds that have yet to come out.
Thanks for the precisions Anne
I was pretty sure it was more an overlook than a hard bug.
Also I noticed you said we
when talking about Opera..., hé hé !
Another piece of good news: Opera 9 recognizes xml:lang attribute values for CSS :lang() selectors in XHTML pages: test case.