Opera’s new browser is here. Make people swim (not sure if that still applies).
They still fail all xml:id tests. Now I’m not sure anymore if someone on (Opera) IRC promised me they will be fixed, but after the change logs of the previous beta I would have expected them to do something about it. (Note that the change logs have been changed since then. See some links in my post on Opera 8 beta 3 for more information.)
Some things that caught my interest:
XMLHttpRequest.
xv. I’m not sure if they prefixed the new ‘unknown’ media type speech as well, but I guess not.
XMLHttpRequest when reading the abstract by the way. Anyone with more insight?
within the first four days of the launch
It was launched on the 19th, so today will probably be the last day. Does he have to get himself a wetsuit and start training?
I downloaded in for Windows and two Linux distros. Hope he swims :)
Anyhow, I don't actually use Opera, its only for testing, but there are a few things that irk me:
Its my understanding that most of these issues wouldn't exist if I used KDE, but I don't.
XSLT support would be nice. But still, Opera has better standards support than any browser has.
Jero, that's not completely true. Look it up.
All others: some people aren't so happy with Opera.
Rob, check out this chart. But you also have a point, though. The quality of Opera's implementations isn't always good.
Support charts are overrated. Unless they come with over 20000 solid test cases. (And that is just for testing CSS2.1.)
Jero, the support for CSS in Opera is terrible. They try to implement all the stuff but the quality of it is ... well, there's no quality. It just sucks. I mean, come on it even has text-decoration bugs, that's like really crazy.
Anne, you shouldn't even mention the support for XMLHttpRequest - it's buggy and it's like 1/4 of it. The problem is they implemented some of the basic stuff, but completely left out things like setRequestHeader which leaves the scripts broken and is very hard to work with...
I have been against Opera for a long time and Opera 8 is not going to change that. When will they realize that it's not about the quantity, but about the quality. Oh, well...
Anatoly: I can provoke the Gecko rendering engine into having total hissy-fits (display:table and friends is a prime candidate), so I could easily say that "there's no quality in Gecko. It just sucks".
But I'm not going to. I'm just going to point out that every browser out there has bugs. They all have different bugs, and when you are accustomed to designing for one browser, you tend to avoid these bugs unconciously, because you more or less automatically use techniques that doesn't trigger bugs in your favourite browser.
text-decoration bugs? It just doesn't support blink. I'm really missing seeing pages where text blinks in front of my face consistantly. I've programmed many webpages where the site looks exactly the same in Opera as it does in Firefox except my newest one where Firefox ignores line-height on a floated :first-letter. I'm unsure if that is a bug or not, though. As of right now I'm really unsure of the poor support for CSS you're speaking of since typically websites that I've programmed work exactly the same in Opera and Firefox. Minor bugs have popped up from each occasionally which the majority if not all have been fixed. I usually just have hell with IE.
I do remember a while back where my colleague, J. King, ran Opera through the W3C's own tests. Out of the ones he tested 28 failed. And many of them that I have tested on recently still fail in Opera and also in Firefox. There's a few that fail only in Firefox, and a few that fail only in Opera.
Opera mentioned themselves that Opera 8 only supports enough of XMLHttpRequest to get gmail working correctly. They're only supporting it because of gmail. And yes, Anne, DOM3 Load and Save is strikingly similar and is there only to allow you an alternate way to get stuff to work in Opera while they finish out support for XMLHttpRequest.
Also it's still surprising to see just how hatred people can have for competing applications. I personally use Opera because I can't stand the snail's pace of Firefox. It's not a reason to hate it.
Support for DOM3 load and save. That specification looks very similar to XMLHttpRequest when reading the abstract by the way. Anyone with more insight?
From my understanding of conversations with Opera employees, Opera's implementation of xmlhttprequest (what a tongue-twister...) is essentially aliases to DOM3-L&S methods. I read most of the DOM3-L&S spec myself a little while back when playing with user scripts, but I couldn't get it to work. I have enough trouble understanding the basics of DOM stuff.
I do remember a while back where my colleague, J. King, ran Opera through the W3C's own tests. Out of the ones he tested 28 failed.
I would have put Firefox through the tests for comparison, but like you I couldn't stand the half-second it takes to go back to the test list---that's a lot of time wasted for over five hundred tests in an agent I don't particularly care for.
Has anyone put Firefox through the tests in question?
Dustine: I guess you have never tried to work with more complex stuff, the floating in Opera is completely messed up and I won't go into how many bugs there are related to floats, it's just too many. If you ever try to code with minimal amount of divs and full-power usage of CSS you'll find that Opera just bugs and bugs, sure I can get it working for Opera... no problem, but I have to specify useless stuff which shouldn't have to be there...
Oh, and maybe try to set position relative to link and see what it does to text-decoration in a more complex code.
Arve: Well, that's nice of you, but I can provide tons of examples where Opera doesn't behave correctly - it doesn't get the whole floating stuff correctly, it doesn't get relative positioning correctly, it screws up on setting auto widths, it doesn't round correctly plus tons more stuff. (Oh and it has problems with styling list-items.)
Well, that's nice of you, but I can provide tons of examples where Opera doesn't behave correctly - it doesn't get the whole floating stuff correctly, it doesn't get relative positioning correctly, it screws up on setting auto widths, it doesn't round correctly plus tons more stuff.
Write some test cases and report some bugs. That would be much more useful than non-specific complaint on a Web log entry's comments. I'd be quite interested in seeing test cases for float bugs you're encountering.
Antoly: You are making a lot of claims, and do not care to back them up, and some of the things you mention are stuff that are
Firefox has it's own share of rounding errors, it has a history with floating/clearing problems.
As for the rest of your "claims" about Opera: I suggest that you do What J. King suggest: Write valid testcases that demonstrate the problem, because right now, it sounds like you're just whining to waste both your and other people's time.
I think Opera and Firefox have the same amount of problems, just in different areas, so whining about particular gaps in Opera doesn't make different ones in Firefox excusable.
People should just be happy that a fine alternative to Internet Explorer is getting new users and press coverage. Besides, Opera and Firefox are on the same side, the standards-friendly side, making life easier for users and developers.
This is the same reason I get upset that sites will support Firefox but not Opera (looking at you, Basecamp). The marketshare of standards-friendly browsers should be grouped together. I don't design pages for Internet Explorer and Firefox. I design for the standards-browsers and "that other one." Opera isn't a minority, it's just another choice in the collective whole of "good" browsers.
Opera 8 is pretty good. They always sucked in the Javascript department, but have fixed a lot now. Only problem I’ve encoutered so far is that it doesn’t seem to allow the dynamic addition of CSS (external files - it will probably work with inline CSS, I haven’t tried). And it still has quite some issues with corrupting graphics on dynamically changing pages (e.g. click on the bottom link in this test I made earlier, all it does is show/hide a couple of paragraphs in a box with overflow: auto;)
With a number of tweaks, I have been succesful in making the Javascript engine that’s created where I work (Backbase) work on Opera 8. It still has some issues, but nothing serious we can’t resolve, it seems.
So, thumbs up!
~Grauw
I do so agree with comment 8.
Arve is right, of course: all browsers have bugs -- they are just not the same bugs. If they were the same, then they might have become "standard", like some IE-bugs have.
Opera Software gets my bug reports and case-references. Cross-testing of potential bugs in all browsers goes on all the time, when one has the time. Would have been nice to have all bugs fixed yesterday, but after some frustration over bugs in each new release, the bottom line is: it's going in the right direction.
Now, if someone could point me to some good software that'll pick up some crashes I've had for Opera8 on win2K-pro, then I could file some proper bug reports on that too.
I finally got some time to test the latest Opera 8 beta for Mac today, and I was pretty happy with what I saw. Especially for block level elements, things have improved lots since 7.5x. It is not yet perfect. And the interface is much better. On the other side, I'm rather disappointed with the latest Safari (1.3), esp on the CSS front: some improvements (min-height!), but lots of small bugs remaining.
Dustin (comment 9), that line-height problem with :first-letter in Firefox is a bug, imho. I filed bug 290125 for that.
CSS blink seems to work in Opera 8. But as far as I remember, Opera can't handle percentage value correct, e.g. 10.1% is treated as 10%. I once saw a testcase drawing a slope with CSS. The testcase is rendered as a staircase in Opera. I didn't bookmark it and I can't find that page anymore. Does anyone know where that testcase is?
DOM Load and Save is a much more general Interface then XMLHttpRequest, imo, it allows e.g. to parse directly in a Node ( parseWithContext ). Also it is actually a standardised Method for innerHTML as it allows to parse Strings and to serialize a DOM Tree, but in both Directions in a much more controlled Way ( e.g. with defined Namespaces, if it is implemented complete ).
( On the other side it does not support e.g. getAllResponseHeaders )
Georg: Regarding crash bugs in Opera (on Win32)
Install Inspector IIXII: It will create crash logs for you. As far as I understand, this log file can then be sent to the bug feedback e-mail address you get when you have reported a bug through the bug wizard (these crash logs tend to get into the megabyte size, so you might want to compress them before doing so).
-DOM Load and Save
-( On the other side it does not support e.g. getAllResponseHeaders )
It's also xml only, isn't it, so not nearly as useful?
I really don't understand why they implemented it and not xmlhttprequest first like everyone else. Doesn't make any sense. They really should have taken a month and added post support in xmlhttprequest and fixed the bugs there before releasing 8.
Opera still has the annoying bug where it computes percentage dimensions of an absolutely positioned box relative to the parent block, rather than the containing block. Other than that, Opera 8 looks quite good.
There was some weird bug in the beta versions that caused problems with my blog, but they seem to have ironed it out in the final version.