One side effect of the iTunes 11.2 update on Thursday, May 15th 2014 has been that some but not all Macs were seeing the /Users and /Users/Shared folders disappear.
The permissions on the /Users folder were also changed to be world-writable, so that anyone could read and write to the /Users folder.
As far as bugs go, this is a very fascinating one. Initially, people thought the OS X 10.9.3 update was the culprit, but as it turns out – it’s the iTunes 11.2 update. I’m interested to (eventually) hear the root cause of all this, but for now, the linked article contains a temporary workaround.
Is it my imagination, or has Apple’s QA/QC degraded seriously over the last four years or so, basically starting with the Lion release?
Well, it’s like they say …
http://www.patentlyapple.com/patently-apple/2013/09/apple-makes-it-…
I concur, Apple used to be the top notch in quality, now, not so much, and is not only a problem of OSX but also iOS.
Didn’t classic Mac OS lack preemtive multi tasking and memory protection? Event Windows 98 had that. I wouldn’t call that high quality. Only when they re-branded NeXTSTEP (Mach based UNIX with lots of BSD code) as Mac OS X that got fixed. I don’t use Mac myself, but that looks to me like it wasn’t Apple that ensured the quality. But it seems there is a correlation between these things and a certain Apple employee.
No CLI environment, other than via the MPW developer tools.
Since it was originally developed in Object Pascal and later ported to C and C++, the APIs were a bit Frankenstein with Pascal and C semantics.
I would dare to say, that Win16 APIs looked quite good in comparison.
Oh and the old Apple had zero support for open standards both in hardware and software, everything was Apple specific.
No Windows 98 did *not* have preemptive multitasking!
Yes it did.
No it didn’t. Just stick a scratched CD into a Win98 machine and see what happens to the so called pre-emptiveness of Win.
I guess we have complete different definition of what pre-emptive means in the end ..
Yes, we do.
One is the correct definition, in which running user mode processes are preempted by the kernel without intervention (or affecting) the process being preempted.
I’m not sure what you mean – but if it differs from the above definition, it’s wrong.
Perhaps you’re referring to kernel preemption, in which a user process can preempt certain kernel functions? Well, that’s preemption, too, but not at all a requirement for a system to be considered a preemptive multitasking system.
This has nothing to do with pre-emptiveness. That exact same behavior can be achieved in any OS one way or the other.
Yes. we haver yours, which is wrong, and then we have the commonly accepted one.
Windows 9X had pre-emptive multi-tasking. The fact that it had bugs doesn’t mean it wasn’t pre-emptive.
Edited 2014-05-18 05:52 UTC
There’s an academic definition of pre-emptiveness and there’s user experience, you see. If it makes you feel happy that you can stick ‘it has pre-emptive multi-tasking’ on your OS (though buggy enough that you end up having an unstable system), then go on. At the end, this is pure marketing (Win2000 had POSIX compatibility. Really? MSOffice supports the “open standard” (yeah, MSOOXML which is implemted by nobody), etc. etc). Fact is that Win9* was as unstable as MacOS and had to be re-booted at least once a day.
Pre-emptive multi-tasking is a non-value if it does not work properly. At the end you want to achieve system stability. You can have that with any sort of multi-tasking – even co-operative multi-tasking – if the software behaves properly. And you can have system-instability with pre-emptive systems if the implementation is buggy. There’s absolutely no difference for the user as the whole Win9* user experience shows.
At the end of the day, this is about the question if you can call a system pre-emptive if the implementation is buggy enough that it doesn’t show the expected qualities of pre-emptiveness.*
*) My experience in the software industry tells me that there is a big difference between sticker properties (what marketing says) and reality (what really works at the end of the day). I only care for the latter.
Edited 2014-05-18 08:12 UTC
It worked.
…yes, that’s what I said.
Well, you know, it has happened that my Linux system totally locks up due to bugs in the video drivers and similar. That just sucks for me as a user. I guess Linux doesn’t have pre-emptive multi-tasking.
As professional, there is only the CS term, regardless how you want to bend reality to suit your purposes.
I have used quite a few GNU/Linux unstable systems, so they also don’t fulfil your definition of reality.
Or do FOSS bugs not count in your world?
(Why should I care for FOSS bugs? Personally I use Mac OSX and what I said is not specially related to FOSS or Windows but to the ever so often seen pure marketing driven features lists in the software industry that do not live up to the hype once you start using those. Win9* pre-emptive multitasking falls into that category.)
Let me give you another non-OS* related example: we use Application Servers from a well known company. At some point in time, they had a version that supported (according to the feature list) OSGI class loading isolation. Unfortunately, the implementation was not good enough to be production ready. Would you really go and say this product supports OSGI class loading just because they tried to implement that (but failed)? Would you start to use OSGI class loading in your production environment because the sticker says it should work?
That’s why I see a huge difference between the academic definition and reality.
—
*) Unfortunately, people are abundant on OS News that are on some sort of “my OS is better than your’s” warfare.
Sure it did. Win32 apps are scheduled preemptively, whether on NT or 9x (including Windows 95). Of course, it wasn’t flawless.
http://support.microsoft.com/kb/117567
The KB article describes what has to happen for preemption to break in Windows 95:
1) A 16-bit application hangs.
2) A Win32 app calls an API that is implemented in 16-bit code.
Both conditions need to be met.
Only if you use Linus Torvalds definition* – not shared by anyone else with a clue.
The reason he called Amiga OS a non-preemtive system was because tasks could misuse the lack of protection to hinder the preemptive scheduler from doing its job.
The same is true of Windows 95 (and IIRC even using the Win32S subsystem on Windows 3.x).
However that isn’t the definition the rest of the technical world uses, the lack of protection have nothing to do with the scheduler policies.
(* https://groups.google.com/forum/?hl=en#!original/sfnet.atk.sodat/qBJ… – it’s in Finnish though)
Win32s processes are actually cooperatively multitasked.
Also, they run in a shared address space with other Win16 programs.
But, yeah, Torvalds saying that the only reason AmigaOS wasn’t a preemptive system was because it was possible to break it seems rather narrow.
Edited 2014-05-17 21:42 UTC
Apple’s “quality” was always an optical illusion. In fact, I’d say it’s misleading to even put the two words in the same sentence.
Nah. The chsh command in OS X 10.0 through 10.3.x (or .4.x?) would make /bin/ (IIRC) your user shell if used correctly, meaning you’d lose all shell access. That’s not about lack of quality control, it’s about not caring about quality at all. But hey, it’s certified unix, so it doesn’t have to work.
We have iPhones and iPads to make…
Updated iTunes to 11.2 (115) 64 bit version, still on OS X 10.9.2, opened terminal and did ‘ls -la’ on /, /Users, and /Users/me…
All folders showing up normally in finder, and ls shows permissions are what they should be. I may hold off on updating to 10.9.3. I’m surprised it’s even out yet… it’s not part of Apple’s new Public Beta, is it?
Now that I think of it, maybe that’s the issue… maybe we’re all beta-testers now… to Apple… and we just didn’t realize it.
You obviously didn’t read the linked article, it required to have “find my mac” enabled.
While I still think this is an edge case for testing, how can an itunes install can trigger such behavior ?
I really believe this is a start of convergence between iOS and OSX. Which is not a good thing, I’ve owned a iPad and an iPhone and I’m pretty sure I will never again. The way Apple tyrannical controls iOS is borderline obscene, where something as benign as choosing a different default web browser is not allowed. The EU went after Microsoft for just packaging IE with Windows, could you have imagine if Microsoft didn’t allow their users to set a different browser as their default, they would have crucified MS. However for some reason people seem to be okay with this practice as long as it’s Apple. It still alludes me today as to why I can’t see the entire directory tree of an iDevice when plugged into a computer or as to why Apple only allows their apps and a handful of third party ones to run in the background. Are they that hung up on cash that they need the extra margin from this limitation by only offering 1GB of RAM. As far as I’m concerned, what is the use of having UNIX underpinnings in iOS if you can never use that power. It’s one thing to make a device easy to use but it’s another to cripple the system.
There are three simple causes:
1) Form over function development
They have so much artsy fartsy garbage going on that I’m willing to bet “designers” outnumber people who actually know what they are doing 10:1
2) Too many cooks
A program this old likely has way too many hands in the pot, with none of them having any clue what anyone else is doing. This is actually becoming MORE of a problem as developers become more and more reliant on versioning tools and less reliant on actually TALKING TO EACH-OTHER or having a project manager who actually… well… manages things.
3) I’m going to use the Windows version as this example, BUT…
105 megabyte download with a 9 megabyte executable and 25 megabyte matching DLL at load time (not to mention the 200 or so megabytes of crap DLL’s it loads once running) FOR A LOUSY STINKING MUSIC PLAYER that to be frank, isn’t as useful as DOSAMP was twenty-five years ago. With all the stupid asshat bull it installs alongside it like “bonjour”… it’s such a fat bloated mess…
Lemme put it this way, the installer ALONE is ten times the Filesize of the ‘bloated’ WimAmp full (as opposed to the actually useful ‘lite’), Filezilla and uTorrent installers COMBINED. Once it’s up and running it has a memory footprint TWENTY times those three COMBINED.
For a really crappy music player/download client with a dreadfully useless and annoying UI. (no matter how ‘pretty’ it is)
Developer ineptitude? Could be… I don’t even want to THINK about how ridiculously massive the codebase behind it must be — and for WHAT?
Edited 2014-05-17 11:54 UTC
That is so true these days and, unfortunately, not only on Mac offerings.
And the reductio ad minima absurds with “Joe Sixpack has no use for this so no one has too!” and “Can not be implemented on a meaningful or easily presented form and as so must get canned!” excuses probably don’t help either.
I, for sure, wish this “rush to the bottom” “fashion” movement ends soon, the sooner, the better.
Point 1 and 2 are true for more or less all current Linux Distros / Desktops these days (and even more so for many previous iterations of those)
It’s an epidemic across the entire industry. See web development — where we have artsy fartsy types who know nothing about HTML, CSS, emissive colourspace, accessibility or practical limitations of the medium pissing out PSD’s that are impossible to ever turn into an accessible, semi-fluid, elastic and responsive design, then have the giant set of brass to call themselves “designers”… when to be frank most of them aren’t qualified to design a blasted thing for anyone.
No matter how pretty it is.
The whole industry is falling into this “false simplicity” trap… Remember the phrase “Make things as simple as they need to be, and no simpler”?
For those of you unfamiliar with false simplicity:
http://baymard.com/blog/false-simplicity
Simple fact is, if you simplify a complex task too much, it becomes impossible to do the task because it’s complex. A lesson lost on a LOT of people designing user interfaces these days, be it OS, Software, websites, ATM’s, credit card interfaces at checkout…
When user interfaces are simple enough that people can use them now, trying to simplify further is at best going to be diminishing returns on the effort, and at worst just piss off people who have a clue what they are doing.
OS UI’s reek of this, where IMHO everything since Windows 98 has served no real legitimate useful purpose other than code bloat and making things harder to use.
What a mess of idiotic comment for a stupid bug that’s already been fixed.
No software is perfect, if you want to use only perfect software you have to stop using any computer, any phone or vehicle built in the last 30 years and take all your money from the bank.
And all OSS fans should stick their heads in the sand for the next decade after that Hearthbleed thing.
It’s not so much that, more that it’s yet another sign that Apple’s QA systems are not effective enough.
Apples own “goto fail” SSL bug negates any criticism from Apple fans over Heartbleed, and is another example of failed QA processes at Apple.
Oh, and who’s QA system is? I’d like to hear some examples of software that that you think has superior QA, please.
Not Apples?
Sorry I didn’t realise this was a competition? “Yeah well (X) is just as bad [in my mind]” is an excuse now is it?
Gotofail wasn’t out as long as Heartbleed, wasn’t being exploited by the NSA and is 100% fixed.
Heartbleed is still an issue, people can still download versions of SSL and add Heartbleed onto machines that don’t have the issue.
Everyone has software QA problems. Normally like this Apples don’t lead to millions of dollars stolen! (Like Heartbleed is still causing)
Oh well that’s O.K then…
That we know of. Also, that’s O.K then.
You left out the “it’s already fixed” fact…but I suppose you think that omission is ‘ok.’
Edited 2014-05-18 09:46 UTC
So is Heartbleed. What’s your point?
Depending on where you browse, it may or may not be, and you would never know for sure.
When you compare the two, yes, I would rather use Apple then those other products. Apple found it fast and fixed it. And it was in the web browser, use a different web browser and you were fine.
Heartbleed is still a disaster.
Goto fail was in the web browser, us
I meant to say that bugs aren’t limited to commercial OSs, but in the editing the meaning was lost, my apologies.
Anyway I’m not an “Apple fan”. I just love my MBP and I’m quite happy with OSX, shortly followed by Win7.
“Love”? ;p
Edited 2014-05-20 15:43 UTC
I just updated OSX and iTunes today on my iMac and my User folder is not hidden, so I’m not exactly sure what everyone’s all up in arms about. Maybe it’s just a fluke and only happens to some Macs(Hackintoshes don’t count).
My Users folder is showing up just fine. As I said, a fluke. I know this website is full of Mac haters and never hesitate to jump on the chance to point fingers at a flaw that might only be happening to some users of the operating system. But….like….I have a brand new iMac. I guess if your Mac is 3 years old this might have happened, but I couldn’t be happier with my Mac. I left the PC(Windows) platform behind as soon as I bought this thing. That’s not to say I don’t dual boot OS X/Ubuntu but Windows 8 is just a horrible mess and I’m glad to be rid of it without having to downgrade my OS.
So you think that’s OK if somebody has older?…
There was a patch yesterday that fixed it. check the app store and if it says you’ve got iTunes 11.2.1, that’s why you’re not seeing the problem.
Exactly. Mac users are no more exempt from opening the app store to check for updates than Windows users are to open Windows Update. If you want to keep your computer working correctly, you have to make sure you keep it up to date. I have a Mac and a PC(which are really the same thing apart from the OS), and I check for updates every day. If your system is screwing up, how about you try searching for updates? We’re all supposed to be computer savvy here after all.
For many of us, it was not a problem before the 11.2.1 patch.
Again it would require to have “find my mac” enabled to trigger. Which apparently is not a commonly enabled option among computer savvy people (whose who own a mac anyway). As nobody here seemed to have the issue.
Actually, I think “find my mac” is getting disabled some where along the way with updates. It should of been enabled on my computer and at some point it turned itself off. This is the second time I noticed it was turned off without warning. I am curious if others have seen that glitch on them too.
Edited 2014-05-19 14:39 UTC
So you ar epointing out an interesting problem where an anti-theft feature is turning itself off.
This sound kind of worrying, because it defeats its purpose and usability, while I think it is slightly better than having none (like in linux or windows), it doesn’t look reliable enough to be trusted. Or worse there is actually a 3rd party software in the wild disabling it without user knowledge.
Heh, I only knew there were updates because of OSNews (the automatic updates sure take their damned time). I installed them right after I read about them.
Some time afterwards, I find this article, checked, and indeed, /Users was no longer visible in the Finder. Checked in Terminal and the permissions were all wrong.
Not too much after there was a patch to fix it. Automatic updates worked for once.
Although this was a local machine issue, I do miss the ability to encrypt one’s home folder instead of just the entire disk. It wouldn’t stop anyone messing with files in there with those bad permissions, but they’d not have been able to see anything in the clear.