Thanks John =) It's probably getting old... but keep up the good work. You guys have a lot of fans out here. My company is busy working at eliminating dependance on IE and is doing more and more testing in Safari for its projects. Great stuff. Jacked about a new Xserve we're installing/configuring/maintaining for a client project too. Say hi to my buddy Wisher if you see him (or are related to him ;) )

Hmm, persistent spellcheck.

Does that also mean the option for persistent *no* spellcheck?
I can't stand those red lines. Want them gone!

I need to find the hack someone helped me get... ah:
defaults write -g "CheckSpellingWhileTyping" '0'
That gets rid of the nuisance. :-)

I had to search for it because I lost a bunch of preferences recently when my disk ran out of space due to an overabundance of swapfiles (75!!!). This swapfile problem only appeared after installing v73 of Safari, but I haven't filed a bug report yet because I only "suspect" a correlation. I whipped up a "Swapfile Watcher" applet to notify me (and write to a log file) whenever a new swapfile is created, so hopefully I'll discover a pattern. (Maybe I should publicly release the applet after tweaking it a bit more...)

-Walter

Thanks John!

"Does that also mean the option for persistent *no* spellcheck?"

Spell-checking SETTINGS for textareas will now be persistent

"settings"

Uhh... oh! OK, I get it now. Thanks. :)

Yaay! Now it's only the GIF bug and the "scrolling while dock icon animates" to fix and we have the perfect browser.

Yeah! ONLY those two bugs and we have a final product :) NOT!

Very cool. That indeed makes my day.

No, the mad jumping about while you use the scroll wheel and move the mouse at the same time is still kinda inexcusable. :-)

Nice - how about Undo working properly {Ducks}

I have been using the Sneaky Peek of Omniweb for about two weeks now. It's really getting good. It already had persistent spell check. It also has a type-ahead-find for links. I've had SP13 running all night and not one crash yet.

Is Omni using a fork of the Webcore? Because there are some things that work correctly in Omni that don't in Safari.

I don't really see this mouse wheel bug (I'm using a Microsoft Intellimouse). Can someone provide steps to reproduce)?

Also, what is "the GIF bug"?

Yay!

I think the GIF bug refers to a problem with GIF anims. I think Zeldman mentioned it recently on his site. I haven't seen the bug myself - just recall reading about it.

yep - just checked - it's on zeldman 23rd April.

OK I just posted this via the bug button, but I figured I could spark some debate as to why this is rendering wrong here...

http://dhcpp1402.resnet.oberlin.edu/secondfilmstrip.html

For some reason two film strips are shown instead of one, like in every other browser. There are only two dependencies I believe. 1. That table has to be there and 2. the img tag has to have align=right in it. Any ideas? BTW I didn't design the page ;)

"Yaay! Now it's only the GIF bug and the "scrolling while dock icon animates" to fix and we have the perfect browser."

Hmm, what about the Frames bug that causes target frame to go blank? That is to me the most annoying bug, since it has persisted since public beta v58.

Great thing about spellchecking, but what about this frames bug?

Keep up the great work. I still love this browser.

Scrolling bug easy to reproduce

I use the Microsoft Intellimouse also. Anyway...

Find a web page that is at least a few screens of scrolling in length now spin your mouse
wheel to scroll down. (for me I roll wheel towards me) AND AT THE SAME TIME move
your mouse towards you also. The faster the better doing both at same time. Scrolling
stutters or scrolls some and jumps back or doesn't scroll at all. Only Safari does this
no other app has ever done this for me.

Same thing happens when i scroll up and move mouse at the same time away from me.

Go here to see a bunch of people talking about it who use all types of mouse and drivers

http://bbs.xlr8yourmac.com/ubb/Forum23/HTML/001428.html

this is my #1 complaint with Safari and yes I submitted the bug twice. First beta release
and then again on v73

"Is Omni using a fork of the Webcore? Because there are some things that work correctly in Omni that don't in Safari."

Perhaps those differences aren't WebCore related.

The scroll wheel bug has been there since Jan 7.

I've noticed it when using a Logitech Wheel Mouse Optical, and a Logitech MX 700. I don't have any drivers installed (meaning I'm just relying on OS X's built-in support for scroll wheels).

It happens to me when I move the mouse and scroll at the same time, but it's most noticeable when moving the mouse in the opposite direction that I'm scrolling (i.e. moving the mouse up while I scroll down). It's like it's doing some kind of accelerated scrolling, where just a small movement of the scroll wheel scrolls through the whole document.

At first I thought it was a hidden feature for really, really lazy people, but noticed it was too erratic to really be intentional.

Hoi John, dank je wel!

there is a god... john be thy name.

:)

Hmm, frames bug... that must be what makes the pages go blank when I try to configure my MacSense XRouter.
I have to view source on the left frame and then copy the links to the address box.

The secondfilmstrip bug (mentioned above) has been fixed. I just fixed it yesterday actually.

The wheel bug is bizarre. I see it now. I was trying to move the mouse in the opposite direction, but it's actually worse for me when I move the mouse in the same direction I'm wheeling.

So close to becoming the perfect browser, but please PLEASE let me set the browser cache to zero. I design and build a lot of websites and testing them in Safari is a pain in the ass because I keep seeing a previous version unless I hard refresh...

Dave, I really appreciate the insight into how a major professional app is developed that I get by reading your blog.

Can you provide any info on how defects are evaluated when they're turned in via the bug button? (names are not necessary, of course, just interested in process).

Proxy, proxy, proxy. The only version that will work properly with the proxy server in my district is v48. Please tell me who to get this infomration to. I am sure it is not a real priorty. I will be happy to test anyway I can. I really hate to go back to IE.

Also, it will not work with Border Manger by Novell. The only browser that will work with Border Manager is IE and Netscape 7.02 at with OSX.

I also have one scrolling bug to report. I have reported it to Apple using the bug button, but... If I click on the form field for searching on VersionTracker for example and then hit the down arrow key a menu with a list of the previous searches I made. The problem is when I use the scroll weel instead of scrolling the contents of the menu it scrolls the hole page and the menu gets detached from the form field.

Overall Safari is getting a really good browser... Your doing a good job...

I'm most bugged by the "Find..." command not working on pages with frames. I assume this is a known bug.

Visit the following to test it out:
http://www-db.stanford.edu/~widom/cs145/

Russ, you can probably install a local copy of squid or privoxy and use that as your proxy, and have it use your district's proxy. You can install squid or privoxy using fink:

http://fink.sf.net/

Excellent... I'm glad to see textareas are getting some attention. :)

Now, is there any hope of addressing these three long-standing issues?

1) Safari sends textarea end-of-line characters as , not (as every other browser does it, and the RFCs dictate).

2) Safari still chops off the final end-of-line character from text entered into a textarea; if I press Return twice at the end of this post, it will be converted into a single return character in the next pane.

3) "Smart quotes" and other special characters (like em dashes), if entered verbatim, get converted to "?" by Safari when you submit the form. Every other browser seems able to encode them to usable variants.

Pretty please? :)

Whoops-- in my post above, (1) should be:

1) Safari sends textarea end-of-line characters as <LF>, not <CR/LF> (as every other browser does it, and the RFCs dictate).

Why does frames not work properly and why is it never commented?

I can't reproduce the scrolling bug with my Logitech MX300 mouse.

Bless you both!

If you can do undo and fix whatever it is that makes text editing in html forms g e t r e e e e l y s l o o o w sometimes, for no apparent reason - we'll be hunting big game.

Another "characteristic" has been nagging at me too: Why is it that when I select text from a web page in Safari and paste it into a Mail message I get all sorts of extra newlines between paragraphs? On the rendered page there appears to be only one.

And while I'm picking text wrangling nits, when you are drag selecting text on a web page and move to the bottom of the window to get it to scroll, it works like "burst mode" on an M-16 - it scrolls then stops then you have to twiddle the mouse again to get the scrolling to burp up another burst. The gesture reminds me of Alladin rubbing the magic lamp to get the genie to come out - you have to "rub" the status bar at the bottom of the window to get the scrolling to keep scrolling. This action is a major carpal-tunnel maker.

All that's left on my list of text-wrangling peeves is to get the shift-click to extend a selection working and you'll have the most perfect browser on the planet.

I'm an editor, so these nits are probably more important to me than most people, but I'd be even more grateful than I currently am for Safari if these were addressed! I use it every day all day in my job doing text manipulation.

The M16 thing irritates me too. It's a known bug that I actually filed myself. :)

About the best advice I can give to everyone complaining is to be patient. Realize that we make forward progress with each release. There are (quite obviously) a lot of bugs that we're working on.

I think we're doing pretty well for a browser that isn't even at a 1.0 and that contains a staggering amount of brand new code. :)

What about the 60 second timeout force stop? that prevents me from using safari on one or two sites that can sometimes get very slow, like citibank.

The 60-second timeout complaints really confuse me. Gecko has a 30 second timeout, and in all my years working on it, I never heard anyone complain about that.

Regarding the time out, when I log in to my web server to view site statistics, the server database takes a couple of minutes to prepare them, by which time Safari has timed out. No problem in IE.

Safari works OK with the time-out haxie that's available, but I don't like using Safari add-ons - they caused all sorts of problems when upgrading to the new version, so I don't use them now, except for the time-out one.

Titles on images and mouse-over messages in the status bar on links are two more features I'd like to see (and miss compared with IE).

And a page archiving feature, retaining all the images and the original URL, and zipped - like iCab does it is perfect.

I guess they're on the way.

Otherwise, works great, and I look forward to the advantages of the code base being available to all apps in the future... Help, in particular, and maybe for Mail and Sherlock.

Many thanks to you and your colleagues.

No worries, Dave. You and the team are doing a terrific job, and v73 is silky-smooth. I hate it when I have to use another browser for the bits where it's necessary. The distance Safari has come is phenomenal.

It's just a communication issue, really; we can submit bug reports via the inline reporting in Safari, or we can post comments here in your blog-- but unless you respond to them directly (which you obviously can't do for more than a choice few such bugs), we have no way of knowing whether the bug is being fixed, has been punted off into not-worth-fixing land, has been misinterpreted by someone for lack of necessary details and deemed not reproducible, or hasn't been filed or noticed at all. That's the trouble with these public feedback mechanisms-- they're only one-way communication, and for all we know, Safari's "Bug" button is a pipe to /dev/null. ;)

I know it isn't, but in the absence of two-way feedback, the price we all have to pay is for us who are reporting the bugs to occasionally sound petulant or repetitive. We just want to make sure we get our thoughts into the system. :)

Thanks again!

Some of y'all may want to try Unsanity's Safari No Timeout, which sets the timeout to 600 seconds (or anything else you'd like).
http://www.unsanity.com/

The 60 second timeout is rather annoying. I use camino for all my statistic pages (it never complains about time-out.. even if it takes 5 minutes to load). If this could be fixed I would certainly appreciate it greatly!!

Kenoyer, about Omniweb, says:

>It also has a type-ahead-find for links.

Wow, that would be nice. I'm not sure this could be shoe-horned into the human interface guidelines for Safari, though. There will eventually be a hack to make this work
(via personal XSLT stylesheets; I can dream, can't I?), but I think most people don't know that this is more useful than tabs. :-)

Spellcheck is for people who can't spell.

I personally dislike this feature.

A js error console is more important than the spellcheck feature. Perhaps if and when Safari has a js console, web app/presentation tier developers will start taking it more seriosly.

Great work! Beta 2 is great to work with - I'm looking forward to the release version.

Nicely Done.

- DAT

Safari's fantastic - handles complicated iframes and JS well enough for our commercial site to let v73 users in, which means now I use it as my standard browser I get to see our web design co's own site as it should be :)

The one thing that bugs me though is that the download window always comes to the front - which means to continue surfing I have to click on my original window or close the D/L window - a pointless UI 'feature' (especially as one of the window's functions primarily only gets used once a D/L is finished, and the other only when a mistake's been made)... perhaps a preference to allow the user to control the default behaviour would be good?

And I may be the only one waiting for the lower-case-'i'-in-tabs bug to be fixed, but it's definitely an interesting one ;)

Great to see it all looking so slick, thanks guys

I'm most grateful for Dave putting time into better text area handling. Spell check will be handy (though I have to admit I think undo would be a more useful function.)

This is one area where Camino shines over other browsers -- it has a solid multiple-undo function, and this is SUCH an important feature for people who do a lot of online content management via web forms. (Microsoft appears to completely ignore this, because as far as I can remember, there is no text undo in any of their browsers.)

Anyway, Dave, well done, Safari is now my browser of choice on OS X -- and I was pretty damn loyal to Camino. (But then you were largely behind that too -grin-)

P.S. It's rather annoying that my net banking appears for a second then the window blanks. The damn bank (National Australia Bank) has worked it so that you can ONLY open the login window via the javascript link on the bank's website -- so there's no way I can find to log in.

With regards to the timeout issue, one thing that safari seems to do is cache data before it renders it. Thus long running cgi's, that incrementally print data back to the browser, fail in Safari, because Safari stills seems to report that it timed out, even though information would have been sent from the server. Camino does not have this problem. For example go to:

http://genome-www5.stanford.edu/cgi-bin/SMD/publication/viewPublication.pl?pub_no=204

click on "Data Retrieval and Analysis"

Click "Proceed to Data Filtering"

click "Retrieve Data"

At this point, Safari fails, whereas Camino does the right thing. If this were fixed (or if the web page is doing something wrong, let me know and I'll fix it) then I could use Safari for just about everything.

I?m not a programmer, so I have no idea how much work this would require.. but is there any chance for the contextual menus to be expanded a bit to include: back, forward, reload, copy, select all, etc..

Thnx in advance.

Hmm, as long as we're talking CGIs at Stanford, I decided to test my own. It's a Perl script that uses sysread and syswrite to send the data in "chunks" (but it's not a 'chunk' protocol, just a loop of print statements). I tried it on a 34.5MB PDF; it took seven and a half minutes to download, and did not time out.

Could it be that only certain MIME types time out?

http://www.slac.stanford.edu/cgi-wrap/pubpage?slac-r-556

(I used the "whole thing" PDF for testing Safari with my 'getdoc' CGI.)

-Walter
(now, don't everybody go pounding the server, please)

Hmm... Content-Length header?

IIRC, my CGI sends the length (total file size) before doing the sysread/syswrite loop.

-boo

This is good news indeed. With all the other comments and bug reports appearing on this page, I'd like to emphasize that Safari, in its present beta 2 incarnation, is already a very fine browser, that is also fun to use. It's fast and reliable on my barely supported PowerBook G3 (333 MHz, 'bronze keyboard'). It shows remarkable and consistent progress, in terms of UI and ease of use, scripting support, and, most importantly, standards compliance.

Personal Certificates à la Netscape. Will that ever be supported? I can't use my banks site with Safari. Have to use Netscape/Mozilla (not Camino though for some reason). Hyatt, care to elaborate??

First of all, thanks Dave for the great work you are putting into this browser.
Can you check, please, why http://www.gazzetta.it/ is displaying that way? The vertical lines are misplaced, is it a css bug?
Another question: what about a better download manager, with the possibility to resume downloads (or at least an option to not delete partially downloaded files), and the possibility to put the url of the downloaded files in the file's finder comments?

ps. you are making a great work keeping us well informed on webcore development progress; is there anyone who keeps us informed on javascriptcore progress?

Great news... now if only you guys can fix the quicktime bug, where an embedded movie file wont load unless it's in the active tab, I'd be really happy...

F

The Timeout bug is a pain in the ass, it forbid me to use a lott of in hose statiscal php scripts.

I've never add any problems with mozilla.

So kill this timeout.

Even if for vanlla user it could be seen as a great feature (I don't know why since it actually prevent them to use some sites) the fact that the timeout puts the faulty window on top is just plain ridiculous.

How about more intelligence on the URL autocomplete?

It only works if I type the beginning of the URL, unlike IE's which gives me a choice as long as what I type is *anywhere* in the URL, or the name of the site for that matter. It's let me use my bookmarks as kind of a free form reference tool. (I've been keeping this file since 1993, it still has a huge OS/2 section ;-)

Fixing autocomplete would be *good*

Maybe I'm a bit stupid - but this WebCore isn't available for download yet, is it?

Below is some code that opens a new window and shows you a preview of text or html, in an online mass email module my production team has built, it works in all browsers except Safari. In Safari, it shows the preview while it loads and then goes to a blank white page? Code is below... anything that can be done to make Safari more compatible is greatly appreciated. I love the browser aside from things that are just not there yet... Great work so far!

The code didn't post for some reason... if it doesnt post this time, shoot me an email I will send it on over... Thanks!

~SCRIPT LANGUAGE="JavaScript" TYPE="Text/JavaScript">

I like Gavin Sherlock's comment:

"Spellcheck is for people who can't spell. I personally dislike this feature."

And then he goes and misspells seriously in the last sentence. :)

Great work guys. As you can see, everyone really appreciates your effort, and you get paid!

I have been posting crash logs with th ebug tool in Safari. Is this a good idea? I have been getting memory-leak like crashes and didn't really know what else to do with teh crash logs.

Daniel

"Spellcheck is for people who can't spell."

Which would be 99% of the internet-using population.

"I personally dislike this feature."

Then don't use it.

"A js error console is more important than the spellcheck feature. Perhaps if and when Safari has a js console, web app/presentation tier developers will start taking it more seriosly."

Or, perhaps when it becomes the most widely-used browser on the Macintosh platform because the developers are wisely putting usability, UI and rendering issues on the top of their priority stack, then developers will take it more seriously.

And, John, whoever you are, you rule. Is there any chance we could get a beta bump with JUST this bug fixed? :)

What also would be nice is this shortcut like in IE on Mac "control arrow right" gets you "http://www." Just a thought :-)

Now just bug Apple Norway to implement Norwegian spell checking and I'll be a happy camper.

;-)

Nice job all, Safari is really shaping up. For me the only main issues are:

(a) HTTPS doesn't work behind a HTTP proxy, since this bug exists in most Apple Applications I guess it is somewhere in CoreFoundation (still a fix is essential).

(b) Bugzilla and other sites that do weird stuff with cgi don't work. Bugzilla (at least the version we use) sends Please Wait followed by the result of the form.. IE/Mozilla/Camino refresh correctly, OmniWeb shows both, Safari saves the o/p as 'query.cgi to disk!
(c) Copy+Paste of Styled text from a Table is broken, it pastes top-bottom, left-right rather than the correct left-right, top-bottom. Again a Bugzilla Bug List shows this off quite well.

And minor issues:

(d) (A problem with all mac-browsers) JNLP (Java WebStart) files are downloaded to the download directory before launch. These should be downloaded to a temp dir.

(e) No feedback on the 'Report Bugs' to Apple sheet!

One of the most common "spelling" errors will never be caught by a spellcheck program, because the typo results in a real word (just the wrong word).
You know, where people type "loose" instead of "lose".

You need a *grammar* checker to catch those errors. Why is there no grammar checker? That's much more important than a spelling checker.

-boo

Walter wrote: "Why is there no grammar checker? That's much more important than a spelling checker."

Probably, IMO, because that's a far, far harder endeavor than a spell checker.

Which is not to say that simply because it's harder that it should not be reviewed as a possible feature. Often the harder programming tasks attempted in order to help the end user (spare the programmer, spoil the UI :) are the most differentiating and beneficial. It just a far fuzzier category not nearly as well suited to computer implementation.

If you really want to report Safari or other OS X bugs and be able to track whether they're being fixed or not, get an ADC account. The free account will allow you to enter bugs on the web, and you will get feedback as the bug is moved through the hoops at Apple. You won't get to see all the details necessarily but at least you will know for sure that your comments were received, and the engineers will also have a way to contact you for more information if/when they need it.

Anyone know if WebCore/Safari is running the new iTunes store client?

It sure looks like it, It's even got back/forward buttons like a browser. Dave?

I believe Jobs said it used the same XML engine as Safari.

It'd be pointless to make a grammar checker just for Safari. It would have to a system-wide thing, liek the spell checker, if anything, and that means that the OS X developers would be responsible for it.

Of course, I think that the FTFF bug takes precedence over grammar checking and just about everything else.

And yes, it looks like iTunes is using WebCore. Very cool. I can't wait for Mail to start using it.

Dammit. I misspelled "like," and all because I don't have the version of Safari that keeps spell checking on. :)

In prior versions (69+) of Safari this comments window would retain its location but now in PB 2 it defaults to an odd position each time. This occurs with all blog comment windows on different sites (such as Unsanity.org). This happen to anyone else?

Desperate bug report "204 No Content" Handling:

I've reported this bug about a month ago, both via the "bug button", and Apple's BugReporter, but I haven't seen any progress on it. (I also sent a desperate e-mail to Dave about it, but understandably, he hasn't responded to me).

Where is the proper place to report Safari bugs?

The problem is that when a user clicks on a link, and the server responds with "204 No Content", Safari replaces the current page, with a blank page. All other browsers that I've ever tested leave the page as-is. We use links like this to control server functions, and it would be a real pain to work around this.

I don't know if this is a "non-compliant" behavior with any particular standard, but it is not the same as every other browser, and it sure is a good way to initiate a server action.

An example of this behavior can be found at:
http://www.stevek.com/ForApple/Safari-1/script1.cgi

Just click on the link, and nothing should happen, but in Safari, you'll get a blank page.

Stevek,

What happens if you send a content-length of 0?

Dave, One of our developers suggested that I contact you concerning a Safari issue I'm experiencing. I'm the webmaster for SAS' JMP division.

We have a java menu on our website (www.jmp.com) that I'm told is now visible to Safari users that have implemented the Java 1.4.1 plug-in. Three different Safari users have told me that they can see it. However, I can't get it to show up on my own computer (G4, 10.2.5). I've installed and re-installed Beta 2 v73 and the Java 1.4.1 plug in, but still can't get it to show up. I'm stumped as to why other users can see it, but I can't. I'd really appreciate any suggestions, since this hampers my ability to test my own website. Thanks!
Roger

I have a suggestion for a great Safari feature and I can't seem to find a place to land this on the apple site..here goes. There are many people out there who try to minimize mouse use as much as possible (due to injury, disability, etc.). Web surfing is one application where it's almost impossible not to use the mouse (esp. if there are pages with hundreds of links, active gifs, etc.). It would be an amazingly great feature to number each link on a web page (maybe a little # next to each link or button, basically anything that submits forms or urls) and a keyboard shortcut to activate that link, like "CRTL-L-23" to follow the link 23. This would be such a HUGE productivity jump for those of us who cannot use the mouse.

And I quote from Dave's personal blog:

"The comments in my Safari blog have become totally useless."

:D

"I have a suggestion for a great Safari feature"

Well, it would have to be a setting you could turn on, of course.

You would have better luck getting this feature into OmniWeb, me thinks.

I tend to agree with Dave. Most of these comments really are useless here. Why don't the above 80 something people stop spamming his blogg with mostly useless stuff ?

You got a bug? Use the spider icon in safari to submit it. This isn't Hyatt's Safari Bug Submission Blogg.

I think everyone is so antsy for another update in this blog that they're really getting rather useless. Calm down.

Not that this post is incredibly helpful.

Apparently the iTunes 4 GUI is generated on the server side with WebObjects and delivered to iTunes as XML.

If iTunes 4 used WebCore wouldn't running "otool -L" on its Mach-O executable reveal an entry for WebCore shared library?

>but is there any chance for the contextual menus to
>be expanded a bit to include: back, forward,
>reload, copy, select all, etc

Unlikely. Apple likes to make things easy and simple and this removes the possibility of doing anything sophisticated.