Post Archive
› February 5, 2004
Apple Safari 1.2
If you're a mac user you have probably already heard it, the new version of Safari is out. Like version 1.1, this is an OS X 10.3.1 (Panther) specific update. Improvements include LiveConnect support (Java and plug ins are scriptable from JavaScript), personal certificate authentication, much improved site navigation interface with full keyboard navigation, improved event handling, improved HTML error correction, title attribute tooltips, support for the IE propriatary marquee tag, DHTML performance optimisations, better cache handling, smallcaps support, outline support, cursor support, improved tables rendering, saving images and downloading linked files from the context menu, and support for resumable downloads.
There are a few complaints and issues you can see on mac related forums, that I will talk about.
Why is it Panther only? (Mac OS X 10.3.1 to be exact, because of the Java update to 1.4.2)
Because of a number of reasons. One is that WebKit is a part of the operative system in Panther, but only an added layer in Jaguar. Another is that it relies on OS libraries (CoreGraphics, libxml2, SecureTransport among others) and APIs that are changed between the OS versions, or absent in the older one. A third would be that Apple would not be able to continue development at the pace they currently do if they had to produce what must essentially be two different versions of it, and they would have far harder to track and repair bugs.
You say caching is improved? But it seems to work even worse than before!
The caching system is revamped, so the old cache must be cleansed before things start working as they should.
The full keyboard navigation you claim is added doesn't seem to be especially full to me!
Well, the full keyboard navigation is OS powered. Go to System Preferences:Keyboard & Mouse:Keyboard Shortcuts and tick the "Turn on full keyboard access" checkbox. Then you can go to the Safari Preferences:Advanced and chose what behavior you want - should regular tabbing highlight links, or should it just highlight form controls? Option tabbing will be assigned the other behavior.
Safari's JavaScript support is really bad. It goes to 100% CPU usage when you go to JavaScript using pages, and almost nothing that works in IE5/Mac works in Safari.
Safari has a pretty good JavaScript support. Sure, it's less complete than Mozilla and slower than Opera, but it supports most stuff. The problem lies in that JavaScript objects and functions are not standardised, and that most sites are coded only to IE/Win, and possibly Mozilla. However, most things that work in Mozilla work in Safari. As for the CPU load, it does that even for pages without JavaScript, and it always goes down to 4-10% after loading.
Comments
1. February 5, 2004 09:14 PM
2. February 6, 2004 12:22 AM
George Posted…
This is great information. First and foremost it does a good job explaining the lack of Jaguar support in the recent updates to Safari, which has been a beef of mine shared with others. And the tip to turn full keyboard access on is "killer". I was wondering why tabing to a pulldown, such as for US state names, and typing a letter didn't focus on my desired state. Now it does.
As for the DHTML/Javascript improvements... keep them coming. As long as the Apple team sticks to the standard DOM much like the one Mozilla uses, I'll be happy. There's no point in following IE's scripting, that practice is what mostly got us in the morass we are finally starting to get out of.
Lastly, I'd love to see Apple provide an offline content feature that's innovative and simple to use (a la Apple). It would be great to click my tab sets on my PowerBook and be able to read down one level or two between my commute to and from work. Now I'm stuck just getting teaser article headlines. I think I'll go suggest this to them now.
3. February 6, 2004 04:21 AM
Ben Posted…
Not that I'd like to see this happen, but I have to worry that Apple choosing to integrate the browser into the operating system could have legal ramifications. I mean, didn't a certain other company do this and get into quite a bit of trouble (well, for a while at least) ?
One thing I wouldn't mind knowing however, is if you can disable the new tabbed link-highlighting for certain links using CSS. My current project uses the a:focus pseudo-selector to provide an effect for focused (and hovered) links so that its selection is apparent. For Mozilla I've used -moz-outline:none; to disable the dotted border for focused links... does such a thing exist for Safari 1.2 also?
4. February 6, 2004 04:48 AM
liorean Posted…
Outline Support is added, as I mentioned in the main post. It supports the CSS2 Outline properties as described in the spec, so it's not got a -vendor- flag on it.
5. February 6, 2004 05:05 AM
Aleksandar Vacić Posted…
Moose, your page http://www.literarymoose.info/=/css.xhtml is blank white page in Safari 1.2. View source reveals that page is there, but Safari display none of it.
6. February 6, 2004 05:12 AM
liorean Posted…
Was just going to mention that. The example pages seems to work (in the cases they do) however.
7. February 6, 2004 05:28 AM
Joshua Kaufman Posted…
Is it just me or does Safari 1.2 not render ol li bullets correctly?
Here's an entry from my site where this is happening. Any reports are very welcome.
8. February 6, 2004 05:49 AM
Moose Posted…
Oh, I have no idea why. The page is valid, and the CSS doesn't contain errors, despite what the validator claims. The only thing that comes to my mind as an explanation is that I linked stylesheets via the XML processing instruction, and not via LINKs. I don't actually remember why I did so for this particular page (must have been annoyed at someone I guess :), but I removed the instructions now. If the synopses work, there is no reason why the front page shouldn't.
One problem that I have is that since Safari in many CSS respects is closer to Opera than to Mozilla, I have a lousy choice as of now. Either I give the maximum to the aforementioned two and leave the latter to its fate, or add overrides for Mozilla. If I do, Safari eats them too, something I don't want to do. Perhaps I will remove all overrides...
I so wish I could afford a Mac. And hey, you Apple users will soon get Opera 7.5 for OSX... Will I sound like a broken record if I quietly mention that CSS3 generated content will then appear on a Mac platform for the first time? :)
M.
9. February 6, 2004 05:13 PM
George Posted…
Moose, I've been able to get content to render successfully in Safari 1.1 even. Check it out here. Hover over the DE VERA box in the top - left in Safari 1.1+ and "about" is prepended via CSS. Check the source to see what I did. The stylesheet is internal. Is there more to CSS generated content support that is lacking from Safari? I'm curious to find out.
It sounds as though you don't have Safari available to you, so you'll just have to take our word for it.
10. February 6, 2004 06:14 PM
Moose Posted…
Is the Destroy front page still blank in Safari?
The page you mentioned does use generated content, and all versions of Safari, from what I know, will display it. Yet if you try to add position or float declarations to the generated content, 1.1 and earlier versions will not obey.
I am not sure, either, whether the content type has anything to do with the rendering here. I send my documents as application/xhtml+xml to Safari. In Opera 7.5, it does matter a lot, in some cases. The rendering is stricter, if the document is sent to it as an application. I have no idea whether it matters in Safari...
One of my friends just told me that he now has access to Safari 1.2, so he will test all my experiments in detail, sending me screenshots, including the dynamic actions, which are unavailable in the online Safari tester, which by the way shows only the top of the page, and hence is not as useful as it might have been otherwise.
M.
11. February 21, 2004 04:48 AM
Sam Walker Posted…
Not that I’d like to see this happen, but I have to worry that Apple choosing to integrate the browser into the operating system could have legal ramifications. I mean, didn’t a certain other company do this and get into quite a bit of trouble (well, for a while at least) ?
The reason Microsoft got in trouble is because IE truly is integrated into the OS – as in, you can't uninstall it. Safari is not "integrated" in the same way IE is; you can drag it to the trash and have it gone forever, just like any other application. Webkit is an integrated part of the OS, but that's different than Safari itself – webkit is just a framework available to all applications for rendering html. Besides, Safari doesn't suck, so maybe the judges will turn a blind eye ;) .
Also, you say that Safari 1.2 has support for the IE-specific marquee feature. I could be wrong, but I think the marquee support in Safari is actually based on the w3c specs. IT's got nifty new features like being able to stop and re-start marquees. Not sure on the specifics, though.
12. February 24, 2004 02:49 PM
Jim Dabell Posted…
No, the reason Microsoft got into trouble was that they did that while they had a monopoly. Apple doesn't have a monopoly.
Anyone interested in tracking down bugs in Safari might find the KDE bugs database useful (search for Konqueror bugs).
Moose Posted…
I don't have access to it, but Safari 1.2 supposedly supports positioning and floating of generated content, first supported by Opera 7, partially since 7.0, and then fully in 7.2. I wish I could test my experiments in this browser...