[jmincey]

Over the years people will periodically observe that Safari updates -- even minute dot-releases -- often require system restarts, and the reply by the Mac community is invariably to point out that Safari leverages an infrastructure of web kit objects on a subsystem level and this is why restarts are necessary.

What, then, is the possible explanation for the latest dot release of Safari on WINDOWS to require a restart? Is Apple modifying the Windows kernel now too?

It's untenable for the incremental Safari 3.1.2 to require a reboot when a major change of Firefox 2.x to 3.0 did not. The rationale about low-level underpinnings simply doesn't wash. What the hell are Apple's engineers doing that would require these incessant system reboots? It's ridiculous.






Jeff Mincey

[Tom_Diola]

You're not talking about 'that' Windows system that requires reboots on every little thing are you? Certainly you can't compare updates from Windows to Apples updates can you?

[jmincey]

Tom, I'm definitely not comparing Windows system updates to OS X system updates. What I am comparing, though, are Windows APP updates to OS X APP updates. Application updates should not generally require reboots. I realize there are exceptions; it's just that I see no plausible rationale that a Safari dot-release should require Windows to restart. Even if the latest version of Safari includes a new or modified DLL or two, these can be registered without requiring an entire reboot of the system.

[DMurray431]

Easy solution........ Don't update.

[jmincey]

To D Murray: Are you always happy with your software? Do you think no software can ever be improved? The next time you find fault with a new version of software or you voice your opinion on how it could be made better, I hope a self-satisfied ignoramus retorts, "Well, don't update then."

Jeff Mincey

[DMurray431]

Being the genius you think your are why don't you write the perfect browser that never needs updates or re-booting? "Those that can-do...... those who can't........ Bitch".

[jmincey]

Apparently you think technology forums should be confined only to praising software and never to pointing out its flaws and how it could be better, but I have a different view.

For the record, I don't object to software which calls for periodic updating nor even the occasional reboot, though, again, I think the latter should be rare for applications as opposed to updates of the OS itself. What I object to are gratuitous reboots for minor dot releases of software when clearly it need not be engineered that way.

What do the Mozilla engineers of Firefox know that Apple's engineers of Safari do not?

Jeff Mincey

[Dan Frakes]

Let's keep it civil.

[Typhoon14]

Jmincy: Read the description of the main security issue at http://liudieyu0.blo...og-entry-1.html

As you can see, the big part of it is the way Windows with IE 8 will auto-execute .DLL files residing on the desktop. Safari has to work around this in the update, and doing so might well require a system reboot for any changes to take effect. It's not so much about Safari needing a restart as it is about Windows needing a restart to change its DLL handling behavior.

[macademianut]

I know that you are a long time poster (even in the old forum days). I finally managed to a new account. I find it funny that whenever we don't understand something, we think it must be wrong. Wouldn't it be nice if you had just asked the question of why does Safari 3.1.2 on windows require a reboot? rather than coming to "ridiculous" conclusions?

[jmincey]

Typhoon14, thanks for the information.

To macademianut, I came to no conclusions about this specific case. I was making a general observation about the frequency with which I have to restart the whole system simply on account of an application update. Again, the ground-breaking Firefox 3 release required no reboot -- not even under Windows. And in my experience, Safari wants reboots with almost every update.

Thus this is not about explaining a specific isolated instance, though I still appreciate what Typhoon14 has to say. Instead I'm using this example to make a point about best practices in software development and how applications are engineered.

By the way, you inquire as to why I don't simply ask for information. Even though I'm making a general observation, I do indeed ask about this specific case. Perhaps you missed that. Here's the quote from my first post:

"What...is the possible explanation for the latest dot release of Safari on WINDOWS to require a restart?"

In my experience, most of the posters on MacWorld think only in terms of individual end users. Most of us are consumers. Seldom do I read posts in which people realize that organizations use Apple software too. And on a large scale it's very disruptive to an organization to have to restart systems every time they deploy something as modest as a dot-release update of an app.

Again, I don't mind the exception here and there, but I contend that Safari requires reboots more often than not and that might not be a bother for the "Ma and Pa Kettle" demographic, but for other markets it is an issue.

Jeff Mincey

Message was edited by Jeff Mincey.

[Martian]

jmincey said:

...most of the posters on MacWorld think only in terms of individual end users. Most of us are consumers. Seldom do I read posts in which people realize that organizations use Apple software too. And on a large scale it's very disruptive to an organization to have to restart systems every time they deploy something as modest as a dot-release update of an app...

This makes sense. I had been wondering why a simple reboot should be an issue because a single reboot is complete long before I return from getting a cup of coffee or taking a leak (I wonder if there is a connection between those two).


Maybe you IT guys should allow the end user more authority—let them do their own updates. :D

[BigFatDuck]

i love mac just as much as the next macworld user, but honestly i think apple needs to chill with the updates- especially those that require full on restarts. in the early days of OS X that was a major bragging right of owning mac, that you could keep your system running for MONTHS without needing a restart. now, even if my system could take it (and in my experience they no longer can) the damn software update will bug me at least once every other week needing a restart. it started in tiger i think but leopard seems to really take it to the next level with it's new system of downloading the packet and then installing during a restart. LAME. (sorry to hijack the thread, this just really struck a chord with something that's been bugging me lately)

[macademianut]

OK. Let's get this straight here. Safari allows you to download something to your desktop! It is the Internet Explorer that allows you to run. People make a big fuss about "carpet bomb"ing the system because of this. Apple releases an update, which might make some changes to the standard behavior of said DLL (which will take effect only on restart).

Which is worse? Leaving the "carpet-bomb" issue unresolved because some people will get offended when their system had to be restarted? or make sure there's a fix for the above issue?

1. IF the carpet-bomb issue is that big a deal, then a restart should no nothing compared to the impact of an unpatched hole.

2. If you think restart will screw your business and cause inconvenience, your business should be such that people in your business should not be visiting questionable sites in their workplace. So, the "carpet-bomb" issue is really not an issue for you. So, this update is not critical and can be pushed back.

In both the scenarios, it appears that the IT folks can make some good decision!