Comments: Safari Newsflash: CSS Parser Integrated

Does it fix this bug, which I believe is a CSS parsing error but which has been identified as "lack of support for the Q element"?

--Kynn

Posted by Kynn at February 21, 2003 6:59 PM

Er...

This bug?

http://kynn.com/macosx/safari-q-bug.html

Posted by Kynn at February 21, 2003 7:00 PM

Yes, the Q element now works with the default rules in html4.css.

Posted by hyatt at February 21, 2003 7:14 PM

After looking up Fabrice's Inversion, it looks like it *should* be vulnerable to Fabrice's Inversion. After all, technically, the CSS is enclosed in comments, even though it looks really strange.

Posted by Kevin at February 21, 2003 7:29 PM

This is great news! I'm quite excited and happy. This was my one main item on my wishlist for Safari and it looks like it's been taken care of. Thanks Dave!

Posted by Brent at February 21, 2003 7:32 PM

How about the 1px dotted border?

Posted by Scottish at February 21, 2003 7:54 PM

Please excuse my ignorance, but will it be possible for Safari (or any other browser for that matter) to get to 100% compliance with all the approced W3C standards? If so, is Apple working towards that?

Posted by lunarboy at February 21, 2003 7:55 PM

Out of curiosity, does the International Herald Tribune site now render correctly? It was one of the most off sites in terms of how it looked / how it was supposed to look. I submitted a bug report at some point about it. To see what I'm talking about, point v60 to http://www.iht.com/ and click on any story. It'll come out looking... interesting.

Posted by eon at February 21, 2003 8:15 PM

I'm not sure 100% compliance is strictly desirable. I'd prefer that standards compliance be weighed along side other important attributes and that dumb or performance degrading or never-used standards (e.g., ) be consider for omittance. Standards compliance is very important but I think performance and reliability are even more so.

Posted by pb at February 21, 2003 8:19 PM

I was going to ask about the IHT site too... Their dhtml is very impressive when viewed with Gecko, but safari doesn't seem to get even close.

Posted by xml dreamer at February 21, 2003 8:26 PM

I too, hope that the 1px dotted border problem is fixed. I almost feel like I'm using IE 6 when I see all of them dashed borders... almost.

Posted by Neil at February 21, 2003 8:58 PM

I would like to add my voice to the clamor for corrected rendering of the 1px dotted border, and also add that text-variant: small-caps desperately needs to be fixed.

Posted by jeremy at February 21, 2003 11:22 PM

Great news, indeed. My only response to the respondees who question "is 100% w3c compliance a Good Thing" is a resounding "YES".

The HTML 4.0 and CSS 2.x specs are not new. They've been around for long enough for at least ONE browser to implement them correctly. OK, they have a few ambiguities, largely resolved by later versions and drafts, but there really is no excuse for browsers written in the last few years to not implement standards.

Posted by robert at February 22, 2003 12:29 AM

Yeah, I read IHT.com as well, and the formatting is really out of whack.. It uses some kind of funky column dhtml... Will it be fixed?

Posted by Onar Vikingstad at February 22, 2003 5:32 AM

Please excuse my ignorance, but will it be possible for Safari (or any other browser for that matter) to get to 100% compliance with all the approced W3C standards?


IMO, and that of many more knowledgable the answer is unfortunately no. Getting close is an appropriate goal and should be something vendors strive for, but technically speaking there are some area of the specs that are either too vague or contradictory for a browser to be able to follow the specs line by line 100% and have that 100% be the same from vendor to vendor. 95%? 99%? 99.9%? Perhaps. 100%, no.

Posted by Another Chris at February 22, 2003 4:36 PM

> for at least ONE browser to implement them
> correctly.

If they were implementable, sure. CSS2 is not implementable. CSS2.1 is not even implementable (though it's trying, but it's still full of ambiguities, contradictions, and implied understandings; the only way to even attempt to implement it is to have someone in the CSS WG advising you as to all the vital details left out of the spec).

> OK, they have a few ambiguities, largely
> resolved by later versions

You must never have tried to implement CSS2....

Posted by Boris at February 22, 2003 6:30 PM

There are vast chunks of the standards (take most of CSS2, for example) that aren't ambiguous at all.

It certainly can only help Apple's competitive position by having a very fast, standards compliant browser on the Mac--including standards that are missing on other platforms.

Lets not forget that Mac OS X has lots of built-in technologies (like Quartz) that would make implementing some of the W3C standards easier than on other platforms.

I would bet that Apple is aware of this and we'll see Safari (eventually) implement more of the W3C standards than we've seen before.

Posted by Al at February 22, 2003 10:22 PM

CSS2 is chock full of ambiguities. It is not even close to being well-specified. Boris is right. As things stand now, it is impossible for any UA to really achieve CSS2 compliance.

Posted by hyatt at February 22, 2003 10:26 PM

> take most of CSS2, for example

And your experience with implementing CSS2 is....

Seriously, I can't see anyone who's tried to implement the damn thing being able to say that with a straight face.

Posted by Boris at February 22, 2003 11:05 PM

What do yo mean it's "no longer vulnerable" to the Star HTML Bug? My copy of v60 doesn't render the text in this test page as green:

http://centricle.com/ref/css/filters/tests/star_html/

But, AFAICT, that's the correct behavior. The initial asterisk shouldn't match anything because the HTML element has no parent.

Likewise, Fabrice's Inversion should hide the declarations from a conforming UA, so Safari is correct in ignoring it.

Posted by ksmith at February 23, 2003 3:10 PM

The initial asterisk should be able ot match no parent, though, shouldn't it? I don't know enough about CSS2 sytax, but that's what I would assume.

However, I would like to note that v60 was released BEFORE this code change, so we won't see its effects until the next public beta.

Posted by Kevin at February 23, 2003 6:06 PM

"nor should we"

Amen.

Posted by codepoet at February 24, 2003 8:00 AM

As for standards compliance: UAs will be 100% compliant with standards that are realistic. Most stuff W3C is cranking out these days is nowhere near realistic. With standards like CSS2 or XSLT they're trying to solve all the world's problems in one standard, and the result is so amazingly bloated that no development team could possibly implement all of it.

Most standards that *really* work well are simple, limited in scope, and reasonably easy to implement. TCP, HTTP, XML are good examples. Stuff like CSS2 will always result in incomplete implementations, and each UA's incompleteness will be different, hence the dreaded "browser compatibility matrix."

Anyway, just rambling. I don't envy people writing browsers these days. BTW Dave: you're doing a fantastic job. Safari is great!

Best regards,
Josh

Posted by josh at February 25, 2003 12:23 PM
Post a comment