Working on it now? Isn't it Sunday? ;)

Take a break, man.

Why are all the publicized bug fixes aimed at correcting rendering errors in people's blogs?

Everyone knows blogs are the only thing anyone looks at anymore... ;)

Andy: International Herald Tribune is not a blog.

Ben: Depends on your definition of a blog.

Bloggers are spending all their time trying to be cool, so they waste way too much time reading specs trying to come up with things that'll either look cool or break something, preferably breaking something in a cool way... that way they'll be loved by all the wannabegeeks out there reading their blog, which will give their life a sort of pseudo-purpose.

As soon as I've finished rewriting my blogsoftware I will, of course, contact my lawyer and have him sue our favorite safarihacker unless he deletes this comment. ;-)

I don't think it matters where the problem is noted when it's something that crashes a browser? Admittedly, my blog is also a testing ground for things that I put into production for business/corporate sites, but this is regular, valid, CSS.

Being able to used advanced CSS without killing browsers is a good thing for all sites, not just blogs. IMO, of course.

"Why are all the publicized bug fixes aimed at correcting rendering errors in people's blogs?"

Uh, low blow. This bug is a crasher, which makes it extremely important to fix (it had been manifesting itself on multiple sites). You can assume that bugs like this that are discovered and reduced by bloggers are happening on many more sites.

The last thing I talked about prior to this was overflow, the implementation of which fixed briefcase.yahoo.com, iht.com, giantrobot.com, etc. Real Web sites.

Finally, I work on WebCore, with a focus on rendering and CSS issues. If I babble about what I'm working on, that is naturally going to primarily be rendering issues and CSS issues. I have gone out of my way to share information about other work in WebCore (such as the spelling/textarea fix).

I will of course only be talking about WebCore and JavascriptCore, since those are the open source frameworks. As I've said many times before, I will not be commenting on other areas of the code that are not open source, nor will I comment on GUI features or release dates.

Just yesterday, I managed to consistently crash Safari by using :before on dt elements, with display set to "marker" and an image as the content.

I didn't get around to reducing the problem to find whether it was from dt:before or display: marker. Is this the same bug, or is there a known bug, or should I boil it down and submit a proper report?

You know, I don't know if it's just multiple types. I have three other tags that change the A tag (initial style, hover stle, themed style) and then put in a A:after{ url(); } and that took Safari down hard. I added that code to the stylesheet at http://www.codepoetry.net/sage.css in various ways and that's the code that kills it.

Emptying the cache and restarting the program gives me one page load, though! =)

Our union website is broken, on Safari, when you highlight the menu system it is all a bit dodgy... a lot better than in previous versions of Safari though :)
Don't know if this is the place to tell you or what but I'd assume you'd like to know of it
Thanks,
Robert

I think it would probably become unbearable if every random bug were posted in the comments of whatever the latest blog entry is. :)

If your bug is about CSS generated content crashing Safari, then, yeah, this would be the place for it. Otherwise I bet things could get out of hand...

Jeremy, codepoet, could you guys produce some simple test pages and post the URLs here in the comments section when you have the pages up? I want to test my bug fix on your pages to make sure the crash is fixed.

Same goes for you, Suzanne, if you want to post a test page with your original CSS+HTML that was making your blog crash.

It would be nice to just test all three to make sure I've really fixed all the combinations. :)

Next time I purposely crash my browser, remind me that the window I'm writing about it in shouldn't be from the same application. :)

It's a bit weird -- it displays fine, the first time. Then if you navigate away (for example, click the "back" button) and then click a link to return to the page, presumably from the cache, the browser goes down hard. So I can make it happen every time, but only when the page is being read from the cache.

The simplest I could make it is here:

http://www.supernews.com/testing.html

Since you need a link to the page in order to demonstrate the bug, I put one here:

http://www.supernews.com/testing2.html

load up the testing2 page, click the link. Click back. Click the link again.

Now, I'm on the east coast, so at nearly 5am, I think it's time for me to give thought to going to bed. :)

(sorry about the bogus email address, I find that spam is proportional to the number of times my address is on the web...)

I got it down pretty far, I think. One CSS code and one tag good enough for you? =)

http://www.codepoetry.net/test.html

It only happens when the WunderLust script is included onto the page. All my other code is fine. Fun, fun, fun!

Enjoy.

I don't know if it's exactly the same issue but a crash occurs consistently when going to the Quality Assurance pages on the World Wide Web Consortium (W3C). Loading a second page causes a crash (unless you 'empty cache' before clicking a link).

http://www.w3.org/QA/

Wait, the link in your blog is a crasher, Dave? I've gone to the site three times now and no crashes in sight..

Thanks for the tests, guys. Your pages are both fixed now. Hugo, I'll look into that. It does sound similar.

I brought this up a couple posts ago but got shot down in the comments (rightly so) over using invalid code. It's all fixed up now and I'm still interested to know whether this bug is deemed fix-worthy. Thanks.

http://homepage.mac.com/walt/

Sorry for the delay -- http://www.synapticimpulse.com/safari_go_boom.html is the page with the crashing bits on it.

(For Jer -- I removed the code before posting the explanation of it so it _wouldn't_ crash Safari, that's why you were able to visit it. Try the page above?)

Walt, technically Safari could draw dancing panda bears and play the hamster dance jingle where an <hr> is supposed to be, and that would be standards-compliant.

So while what you've reported is not a bug from a standard-compliance perspective, it *is* probably a bug from a "We should match other browsers" perspective.

I have no idea what the rule is for sizing of hr tags though, so I'm not sure how it's supposed to be fixed. I guess I'll have to write a bunch of test cases. :(

I can't argue with that. =-) I certainly didn't mean to imply that HRs were being rendered out of spec with standards. I simply wanted to point out that other browsers force a bit of a verticle buffer between HRs and the content that surround them. And that I hoped Safari would one day behave similarly. Thanks so much for looking into it further.

I have to disagree with you there Walt. In your example you have specifically stated in your code that you want NO margins and NO padding in both your paragraphs and your HR.

The ONLY browser that gives you EXACTLY what you asked for is Safari. The others have overruled your code and inserted padding where they (as in the browser developers) see fit.

Personally I think that it's the other browsers that should take a lead from Safari and remove the padding that wasn't asked for in the first place.

That POV was called out in the previous discussion thread as well. This is definitely one of those cases where Safari can either follow the style designated by other browsers, or break with convention and render HRs its own way (with dancing panda bears or whathaveyou). One way to look at it is to consider what an HR consists of. Dave is right, the HTML spec is no help, "The amount of vertical space inserted between a rule and the content that surrounds it depends on the user agent." But the convention adopted by browsers is clear. HR's have a bit of unremovable vertical padding. My vote (if we get to vote) is to follow the convention but all I really want is to get the same output from each browser.

That being said, this is a relatively tiny issue and I (again) appreciate Dave spending his time to look into it and to everyone else for addressing the issue.

Hum

I have a page with some CSS <HR> tests on it, although they're more for sizing and coloring the line itself than for padding. I could add some padding tests...

http://www.natural-innovations.com/test/csshr.html

-Walter

Is it me or has hotmail suddenly stopped working?
The send button just won't do anything....
Yes I have sent a bug report but I thought I'd drop it in here too.

Otherwise I love safari, The tabs are really well implemented.

If you try to buy a ticket, perhaps SNA to SFO for WWDC, select outbound and inbound flights, and then click "buy now" to log in, Safari does not let you select the passengers.

(You are not charged until later, so this is safe to try at home...)

Clicking the 'continue with purchase' button just brings you back to the same page.

This is a show stopper for travelocity.

Walt - if you put borders on the paragraphs, you'll see that, with the exception of a single extra pixel below the HR before the paragraphs, your issue is not one of non-compliance with the spacing of the HR, but perhaps with the internal padding on paragraphs. (There is 'copious' padding above the top of the ascenders before the top border on the 2nd paragraph.)

I suspect that the following is a serious bug and a potential security vulnerability - automatic download of files to the download folder when visiting a web page, even with "Block Pop-Up Windows" turned-on...

Visiting the web page "www.headbone.com" which is supposed to be a site for kids, Safari AUTOMATICALLY downloads two files to the visiting computer's download folder (desktop, in my case). You can see them afterwards in the "Downloads" window. In this case they are HTML files, but a hacker could potentially download some other file with this mechanism, without the visitor's consent. I can envision average people finding a cleverly named file on their desktop and mistakenly using it without knowing what it is or where it came from. Spooky.

In this site's case the files are called "ad.empty.html" and "popup_code.html"... I submitted a bug report on Safari's bug button but thought of sharing here, so it can be dealt with quickly in case it's serious.

Speaking of misrendering blogs, http://talkingpointsmemo.com/ has an odd problem I've not seen elsewhere.

It renders fine. But ask Safari to redraw the page (switching tabs or changing the text size will do the trick) and, viola, no wrapping. Reload the page, and it's fixed, for now.

Not that the page validates. But Safari's behavior is strange, regardless.

There seems to be a problem with content sizing for <embed> tags - try looking at

http://patimg1.uspto.gov/.piw?Docid=06000000&idkey=NONE

and use the scrollbars to actually find the picture which should be sitting right on top...

Seems like the attributes (width="570" height="840") are somewhat ignored... Also, why is the image quality so much crappier in Safari then it is in IE even though both seem to use the QuickTime plugin?

Forget EMBED -- it is intrinsically evil. Make OBJECT work per HTML spec, and let EMBED die the death it deserves -- a quick, unmerciful death.

I would sooner commit hara-kiri than put EMBED on a Web page. It's even more evil than JavaScript, and that's pretty damn evil.

-Walter
one of the 3 remaining people on Earth who care about accessibility.

I have also just submitted a bug report complete with test case HTML where Safari v73 does not render a text input element if position is absolute, relative or fixed and width is specified as a percentage. No matter what you do, it just does not show up. Using any other length-unit works fine. The element still gets (invisible) focus and you can type and extract text from the input, but it does not appear at all. Is this a known bug yet?

As to why it's a good idea to focus on blog-oriented bugfixes...

Many of the more popular blogs melong to people using advanced and, at times, esoteric HTML and CSS skills. Most major Websites use things which are, comparatively speaking, far simpler. Often, when the advanced stuff is made to work right, the simpler stuff just falls into place, because advanced tricks tend to rely on features or aspects of the spec that more basic designers -and sometimes even programmers- just take for granted and may not even be consciously aware of.

In other words, advanced-markup-using blogs make sure you've dotted the i and crossed the t. Everyone benefits from that.

Arthur, can you post a URL here in comments with an example?

Xerox PARC has been working on some interesting and simple browser features lately. Think these ideas will influence Apple someday?

http://www2.parc.com/csl/projects/popoutprism/

A picture (or video) is worth a thousand words...

http://www2.parc.com/csl/projects/popoutprism/movies/popout_prism.mpg

I know that this is not directly on topic, and that Dave doesn't seem to be involved with Safari's gui, but I found this article about the deficiencies of Safari's interface as compared to iTunes kind of interesting. Maybe you could pass it along to the appropriate people Dave. It's here:

http://daringfireball.net/2003/05/interface_details_itunes_vs_safari.html

I apologize if this has already been mentioned elsewhere.

So I was downloading Unreal Tournament 2003 Demo at 150 k/sec using Safari and when I was halfway through Safari CRASHED! AAARRGH. And it doesn't support resuming downloads. It should at least do that. (Yeah I know it's beta).

Hyatt,

I have uploaded the test file here:

http://www.xfinitegames.com/~arthur/SafariBug.html

While I'm at it, here's an example that will crash Safari, yay! It's a JavaScript bug. Just start typing away in the edit box and pretty soon, Safari will crash. (Holding a single key does not work). See the comments in the html for more info.

http://www.xfinitegames.com/~arthur/safaricrash.html

(NOTE: this will CRASH safari, do not try if you have important windows open)

What it boils down to is that if you have a lot of events flying around and you clone the event object passed to the event handler function, then sooner or later Safari will crash. (The "safe" version included in the html only seems to delay this effect, not solve it, sadly.)

Also, Safari will also crash if you pass something other than a function to addEventListener(). Very interesting.

And finally, Safari's overall event handling support on most elements is in a rather sad state. (Again, see comments in the html)

Thanks for listening.

Arthur, I was able to reproduce the crash bug. I was not able to reproduce the input field bug (all of them show up just fine in the latest builds).

Ah, that's good news. I also just tested the addEventListener() crash bug again under v73 and it didn't crash, no matter what nonsense I passed to the function, so that has been fixed as well.

I just have some minor gripes with Safari now, them being: getElementsByTagName("*") return an empty array instead of all child element nodes.

Furthermore, Safari does not support the Function.caller property. I use this in combination the arguments property to create full stack traces of functions in case of exceptions and in Safari this is impossible.

Speaking of exceptions, Safari does not have a native Error object, does it? Related to that, will Safari support an option in the future to report JavaScript errors in any way?

Finally, like I said, Safari only supports a small set of events currently. In v73, basic things like keydown, keyup and keypress simply do not work if I add them to textareas or text inputs.

Heh, I'm sorry to inundate you with all this, but when the little bug report window just slides closed, it feels like my bug reports are conveniently disposed of, as I get no feedback on whether the reporting process succeeded.

Do the key events work if you use addEventListener to add a bubbler instead of a capturer (i.e., pass false instead of true)? Is it just "capturing" key listeners that don't work?

According to my test it doesn't work with either bubbling or capturing. I made a simple test page and uploaded it here:

http://www.xfinitegames.com/~arthur/safariael.html

You can specify the capturing flag and add a keypress event to a textarea. See the html source for more. Safari does not call the event handler at all. (Same goes for keydown and keyup)

Also, other components are not receiving focus and blur events, double click events, etc. In IE and Mozilla, ALL elements receive all user input events, Safari only handles a few. Adding a double click handler to a button input does not work in Safari, for example.

Again, many thanks for taking the time to look into all of this. It is very appreciated. I really want Safari to be the king of browsers and proper DOM / UIEvents support is part of that. You guys are doing a great job!

Regards, Arthur

Dave:
You might want to check out Zen Garden, as some of the CSS designs are not too Safari friendly!

http://www.mezzoblue.com/zengarden/

Hello, I have a question:
My website is build using Cascading Style Sheets, and works ok in IE and NS, but the css is ignored in Safari. Does anybody know why?
http://www.corega.be/welkom.php

bart,

What happens if you remove the bogus META tag that's somehow inserted ahead of the DOCTYPE?

Also, is the LINK to your stylesheet static or dynamic? If it's dynamic and you serve different stylesheets based on browser sniffing (I have a CGI which does that), is Safari getting the right one?

-boo

rob and ScottEllsworth both mention troubles that I've also encountered in other contexts.
It seems that sometimes (not always - this form works fine) Safari just plain fails to submit forms. The place I'm noticing it right now is in the SQL tab of the phpmyadmin tool (I've reproduced that on another machine too). I also noticed it on Nokia's site when (ironically) trying to submit a feedback report on how part of their site doesn't render legibly in ANY Mac browser!
The symptoms: fill in a form and click submit; nothing happens at all - no progress bar, no page submission or anything.
I've sometimes managed to make a form submit by hitting return in a single-line text field, but in a form that doesn't contain any, that doesn't help. It sounds like a problem with submit buttons rather than form submission itself.
I've seen this happening way back to the first version of Safari released, and in all cases, the forms I've had trouble with work OK in IE on the same machine.

By the way, in a follow-up to the security bug mentioned earlier on this comments page (www.headbone.com), the "Activity" window of Safari shows "Location change interrupted" in red for every file that gets downloaded to the download folder automatically. Today it's five files...

Apple just released an update to Safari, 1.0 Beta 2 (v74) - that appears to have fixed the above security issue. Visiting the www.headbone.com page with this newest version no longer causes that error.

PS: Good job, Safari team @ Apple!

there will be quite a selection of free dish network channels in the near future. This could dishsinologies also lead to a re-emerging of terrestrial free directv TV DXing.
It seems that a mini antenna free direct tv and DVB-T receiver is used for reception direcway of national and local programs. And direct tv since in most cases these same people directv want to know more about what is going free satellite tv on in the world, besides what is happening digital satellite tv in their own backyard, they also get a satellite receiver to bring this into their living rooms as well. One

If you want to use a doublezowie signal meter,
satellite dish I suggest you swing the dish until satellite tv you get the highest needle deflection
doublezowie or audible tone then maintaining that free satellite tv horizontal angle, raise or lower the get free satellite tv for free elevation of the dish to see if the Dish Network meter will go higher still. At the
Direct tv highest signal level, watch the screen dish network for the digital readout changing. If satellite tv it doesn't, then you're not pointing at the right satellite system.

New or reconditioned parts and products

get satellite tv for free

is clearly the best choice for satellite internet. 



free-satellite-dish-network-directv-direct-tv free satellite tv. Don't worry



free satellite tv will be available forever! 

satellite tv, satellite dish

instead of cable

immediately. after switching to the



Free Satellite dish tv Dish Network DirectTV tv sets,

Directv

is the next obvious step.

Those free satellite tv

customers who own Dolby Digital systems by including



dish network are probably okay doublezowie.