The macosxhints Forums - OS X 10.3.9 is out...

The macosxhints Forums (http://hintsforums.macworld.com/index.php)

- OS X Products -- News and Reviews (http://hintsforums.macworld.com/forumdisplay.php?f=26)

- - OS X 10.3.9 is out... (http://hintsforums.macworld.com/showthread.php?t=37935)

OS X 10.3.9 is out...

The OS X 10.3.9 update is out.

The delta is 51.3 MB, and the combo is 113 MB.

Downloading now...

Here is some information on what is in the update and here is a link to the download page

[The information link on the download page linked above takes you to the information page about the 10.3.8 update. Follow the information link in this post instead.]

ooooh Safari 1.3!

Been waiting a while for this! I'm envisioned Apple releasing this on the 29th, though.

JavaScript tests are MUCH improved, now i can finally play Yahoo! Pool in Safari with no chopping!

:)

I have installed 10.3.9 on my 1.2GHz iBook G4 and have seen no problems so far.

Safari 1.3 doesn't open

After updating to 10.3.9 from 10.3.8 and rebooting, Safari quit working. Everytime I start it up the "Unexpectedly Quit" dialog box pops up.

Quote:

Originally Posted by bugmenot

After updating to 10.3.9 from 10.3.8 and rebooting, Safari quit working. Everytime I start it up the "Unexpectedly Quit" dialog box pops up.

Remove all of your 3rd-party Safari enhancements.

Hey! Thanks very much Hayne.

I've been favorably impressed both with Safari 1.3 and the improvements made to network directory lookup performance.

Macfixit has a couple of comments today about some breakage caused by the update:

Quote:

Safari: Crashes on launch, broken JavaScript, Java

Crashes on launch A number of users report that Safari is crashing on launch after updating to Mac OS X 10.3.9.

This problem, in most cases, is caused by an apparent problem with the new version of Safari's ability to handle additions to the Library/Application Support/SIMBL/Plugins folder.

These include popular shareware utilities like AcidSearch and Pith Helmet.

Removing these items from the above folder (e.g. acidsearch.bundle) will resolve the on-launch crashes in most cases.

Broken JavaScript Some popular Web sites that make use of JavaScript are no longer working properly in Safari for some users since installing Mac OS X 10.3.9. These include a number of banking Web sites, and several of Google's tools.

MacFixIt reader Amuro Ray writes:

"Access to Google's Gmail is no longer possible after the Mac OS X 10.3.9 upgrade. The login page will still show normally. The problem, however, is that once you've 'successfully' logged in, Safari will crash (with the spinning ball) immediately. I have waited for a few minutes (with the spinning ball), and have finally forced to force quit Safari. Turning off JavaScript will allow me to login to my G-mail successfully, although Gmail will now tell me that my browser is not enabling all of its features (obviously, JavaScript)."

As noted in Knowledge Base article #301327, the version of Safari included with Mac OS X 10.3.9 includes a significant security update affecting JavaScript. According to Apple:

"This update closes a vulnerability that allowed remote websites to load javascript to execute in the local domain."

As such, turning off JavaScript (in the Security pane of Safari's preferences) will allow some problematic sites to be accessed, albeit in limited form.

Another alternative is to set Safari's user agent to a different browser. This is accomplished by first activating the Safari debug menu (most easily with a tool like Safari Enhancer) then selecting "Windows IE" or another agent.

MacFixIt reader David Benesch writes:

"Unfortunately, the first website I visited after updating to OS X Mac OS X 10.3.9 was maps.google.com. The new version of Safari (1.3) displays many things incorrectly. If another User Agent (like Windows IE) is selected via the debug menu, the site renders fine. I suspect that Google had to work around a rendering bug in Safari that is now fixed, but the fix no longer works correctly in the new Safari. I've tested this on two machines. Otherwise, 10.3.9 works fine for me."

Broken Java A number of readers are reporting that Java applets cause Safari to crash after updating to Mac OS X 10.3.9. This problem can generally be worked around by re-installing the most recent Java updater (Java 1.4.2 Update 2).

However, for many users, the Java update installer refuses to do its job because -- for some reason -- it does not recognize that there is an installed version of Java to be updated.

As noted by one MacFixIt reader:

"Java programs don't work (under Mac OS X 10.3.9). Typing 'java -version' in Terminal gives a segmentation fault. Trying to reinstall Java 1.4.2 v2 doesn't work, as it doesn't recognize that the installed version of Java can be upgraded."

Luckily, there is a workaround for this issue. Though the Java update installer refuses to properly place components, you can use a utility like unpkg to extract the various parts of the Java 1.4.2 update 2 package, and drag and drop the components to their appropriate locations.

MacFixIt reader Josh Camp writes:

"I was having the same problem as everyone else, any time safari try's to load a java applet it would crash Safari. The way I got around it was to reinstall the java142update2. Of course the installer would not let me so I used a free app called unpkg which unpacked it to my desktop. I then just did the drag and drop thing, and just like that Safari works with Java applets again. "

In order to perform this workaround, first download unpkg. Next download the Java 1.4.2 Update 2. Drag the file Java142Update2.pkg -- which now should be located on the mounted Java 1.4.2 Update 2 disk image -- onto the unpkg icon.

The files contained in the package will now be located on your desktop in a folder named Java142Update2. The files in this folder will be organized according to where they should be placed on your startup volume. For instance, you will see a Library folder, within which you will find folders named "Internet Plug-Ins" and "Java." Drag the individual files located in these folders to the to the folders /Library/Internet Plug-Ins and /Library/Java respectively.

When performing these procedure, copy at the file level only. In other words, do not attempt to copy an entire folder (e.g. the "Library" folder located in the Java142Update2 folder) or vital components may be deleted. You can drag multiple individual files simultaneously, but do not attempt to replace entire folders.

Positive reports MacFixIt reader Scott Leddy, though experiencing the aforementioned Java crashes, welcomes a fix for excessive fan noise that was occurring after Mac OS X 10.3.8.

Leddy writes:

"Just to let you know that Mac OS X 10.3.9 thankfully fixed the well described fan issue that 3.8 created with many dual G5's. Also, issues 3.8 created on my machine such as locking up upon awaking from sleep seem resolved with 3.9 as well. However, Safari does crash the instant I load a Java applet, as some other users have reported.

Especially the Java segfaulting looks bad.

I can now use Safari to connect to the secure campus 'Advisor PIN' system. [This system provides access to information about my advisees and lets me remove 'holds' on their ability to enroll in classes after they've met with me about scheduling.] In the past I had to keep Camino around to access that system because Safari would say: "Can't connect to ...... because Safari can't make a secure connection." or something like that.

I was in one of those moods, so I cleaned the goldfish tank, ran DiskWarrior prior to updating with this updater, and ran permissions repair afterwards. Permissions repair was more active than I had suspected. Among those changes were:

Code:

We are using special permissions for the file or directory ./Library/ColorSync/Profiles/Displays.  New permissions are 16893

We are using special permissions for the file or directory ./Library/ColorSync/Profiles.  New permissions are 16893

We are using special permissions for the file or directory ./System/Library/Filesystems/cd9660.fs/cd9660.util.  New permissions are 33261

Permissions differ on ./System/Library/Frameworks/CoreServices.framework/Versions/A/Frameworks/CFNetwork.framework/Versions/A/Resources/PACSupport.js, should be -rw-r--r-- , they are -rwxr-xr-x 

We are using a special uid for the file or directory ./private/var/at/jobs.  New uid is 1

We are using a special uid for the file or directory ./private/var/at/spool.  New uid is 1

Permissions differ on ./private/var/log/install.log, should be -rw-r--r-- , they are -rw-r-----

So far everything seems to be working nicely. I don't have a Google account.

Some folks who have experienced the Java segfault issue have claimed reapplying the Security Update 2005-002 patch has repaired it. Might be worth trying.

By the way, I had zero issues with either Java or the various Google services when testing them after installing 10.3.9 on any of three machines, when I read these reports.

Quote:

Originally Posted by Craig R. Arko

By the way, I had zero issues with either Java or the various Google services when testing them after installing 10.3.9 on any of three machines, when I read these reports.

Same thing here - zero problems. One notable about Safari: I recently had a problem with Safari failing on startup so I had downloaded a fresh copy, trashed absolutely every Safari file except my bookmarks, and reinstalled. When this upgraded, there were no problems of the kind reported.

Safari suffers from random crashes, wether surfing a Java enabled site or not. Applying the 2005-002 fix didn't work for me. It installed fine but has had no effect. Wish I hadn't bothered downloading 10.3.9 now.

The 10.3.9 Combo update is here:

http://www.apple.com/support/downloa...pdate1039.html

It's over 100MB! Man, we are definitely living in the 21st century now! It took around 3 minutes to download with my Comcast cable internet.

Quote:

Originally Posted by mclbruce

The 10.3.9 Combo update is over 100MB! Man, we are definitely living in the 21st century now! It took around 3 minutes to download with my Comcast cable internet.

As a guy who started with System 1 in a 512 K Mac which came on two 400K floppies that took more than 3 minutes to read in, having a 500k pipe to the world is still awesome. My System Folder is 1.15 GB now, so 100 MB is less than 10%.

I also can report no issues with 10.3.9/Safari 1.3 on any of the many web sites that I use including gmail.com, banking and financial sites, and sites that use java.
Sounds like it may be a machine-specific issue and not general issues with 10.3.9.

From the 10.3.9 documentation:
http://docs.info.apple.com/article.html?artnum=301327

Quote:

"This update removes the ability of Mac OS X to run SUID/SGID scripts."

While I decided based on some limited reading that I didn't know enough to risk attempting to use this method to do anything, it still feels like the loss of a feature. I'm just curious what the hardcore scripters think - are we better off without SUID scripts?

a java program i use did not work after the upgrade, but did work after reapplying the 2005-02 security patch

PithHelmet is BACK for OS 10.3.9's Safari v 1.2.3!

The only thing that OS 10.3.9 and its new Safari 1.3 broke on my G4 and G3 was PithHelmet, the must-have complement to Safari for browsing and one of the things that gives Safari a powerful advantage over some other popular browsers.

http://www.versiontracker.com/dyn/moreinfo/macosx/18158

Respectfully, Norm

Quote:

Originally Posted by biovizier

are we better off without SUID scripts?

I was slightly surprised to read about that security fix - I had always assumed that SUID scripts were not allowed in OS X like in most versions of Unix. The reason they are not allowed is that they open a subtle security hole related to the fact that the script is executed by an interpreter (the shell) and hence that there is a small time between when the shell starts and when the script starts during which time a skillful mal-doer might be able to switch to a different script.

Anyone who understands the security issues well enough to be trusted to write a SUID script is likely also a knowledgeable enough programmer to be able to do what they want in a compiled application instead of a script - or at least to write a wrapper application for the script.
So no-one is likely to be missing SUID scripts.

So if I understand you correctly, you're saying there is a danger in the underlying mechanism in addition to the pitfalls to avoid in writing the script, and the vulnerability will be present regardless of how carefully the script is written… ok, in that case, good riddance.
Thanks.