|
Binary Badger posted:Translation: they went to Intel and got someone to fix the drivers. Pretty inexcusable for a major hardware vendor. Er, what? Graphics driver bugs happen and get fixed all the time, and more so when the hardware is new and the drivers are young. Not seeing where this outrage comes from. Sure, it'd be nice if all software shipped bug free 1.0 versions, but we live in the real world.
|
#
¿
Jul 12, 2013 00:24
|
|
|
|
| # ¿ Apr 27, 2024 13:35 |
|
|
Binary Badger posted:The Intel OS X drivers for the GMA 950 and X3100 implement only just enough of the OpenGL standard so that the OS X desktop and Quartz Extreme environment can run without issues. Ok, ok, I get it. You hate Intel GPUs because you've been burned by their "hey we have a few square millimeters free on the north bridge chip, guess we'll throw in a token GPU" era (no joke, that was the actual design philosophy behind GMA 950 and X3100 and many others). Understandable. Not sure how much blame to put on Intel for the lack of ML support, though, since those GPUs are still supported in Windows 8. Our favorite fruit company might have a bit more to do with that decision than you think. I was just surprised to see you that angry at Intel over a graphical glitch in early drivers for new hardware because it's hardly an Intel-specific thing. I've seen the same and worse (outright crashes) with ATI/AMD and NVidia so many times I've lost count.
|
|
#
¿
Jul 12, 2013 11:26
|
|
|
SRQ posted:I know but I wasn't expecting such a noticeable improvement on old hardware, it's really nice. Haswell's new features (very fast power state transitions and ultra low power active idle) make timer coalescing more effective, but old CPUs benefit too. More time spent in low power states is always better. Also, they kinda needed to get idle CPU percentage back up to where it used to be. I noticed a significant power regression in Safari when I first installed 10.8 last year. In 10.7 it was able to throttle non-visible tabs, but somehow that got lost in 10.8. (Filed a Radar about it and never got feedback, though that's pretty much par for the course.) 10.8 also seems to suffer from some daemons which maybe aren't as efficient as they could be. In 10.9 they're clearly looking to find and fix that kind of power bug, and they've added App Nap too. Using less CPU time is a great way to save power on any CPU.
|
|
#
¿
Aug 3, 2013 07:53
|
|
|
SRQ posted:I want to move my OSX install from an HDD to an SSD. An alternate to CCC which I've been using in recent times: If you've been using 10.7 / 10.8 / 10.9, you probably have a recovery partition by now. Boot to it. (I forget what the special boot-the-recovery-partition key combo is, I just use the option key to bring up the graphical chooser.) Now you can run Disk Utility instead of performing an install, and use it to "restore" your existing OS X partition to the SSD. No need to go through the intermediate step of creating an image file. You may also want to use Recovery Disk Assistant (free download from Apple) to set up a recovery partition on the SSD before starting the clone.
|
|
#
¿
Aug 13, 2013 11:40
|
|
|
Kilometers Davis posted:I tried that and have a wall of text. Lots of "failed to resolve library dependencies". After a few seconds the no symbol came up again in the middle of the screen for some reason. Stuck again with "Still waiting for root device" as the most recent string. FYI, the "failed to resolve" errors are the kernel trying (and failing) to load kernel extensions (drivers), and the "Still waiting for root device" is the kernel complaining that it needs to mount a root (aka boot) filesystem to continue but hasn't been able to. Probably because it can't due to all the KEXTs failing to load. The "no" symbol is the kernel telling you it's given up on making forward progress. Basically, something hosed up one or more files in your OS X install. Could've been software you installed (if so it would've been something that required an admin password to install), could've been filesystem damage due to bugs, could've been your SSD failing. I'm seconding Japtor's recommendation to use a SMART tool to try to rule that last possibility out.
|
|
#
¿
Aug 19, 2013 11:00
|
|
|
You Am I posted:Not happening till they fix my serious bug with 10.9 Yeah, I just had a hilariously awful one myself. Bad enough to make me glad I didn't install it as my main OS.
|
|
#
¿
Aug 26, 2013 06:59
|
|
|
NESguerilla posted:My macbook pro has locked out the internal speakers and won't output any sound. This has happened to me before (it always seems related to plugging iphone headphones into it for an extended amount of time) but it is usually something I can fix by restarting the computer. This time it's not working though. The computer doesn't even seem to be aware that it has speakers all of a sudden, so I can't fix it through system preferences or anything. What the gently caress? Could be corrupted firmware settings. Try a PRAM reset: http://support.apple.com/kb/HT1379 since PRAM is where speaker volume is stored, and if that doesn't work try a SMC reset: http://support.apple.com/kb/HT3964 Also when you say "isn't aware it has speakers", do they even show up in the list when you run the "Audio MIDI Setup" app?
|
|
#
¿
Aug 26, 2013 08:55
|
|
|
jototo posted:I'm looking through the documentation in Carbon Copy Cloner about making a bootable backup of my HDD to an external drive. It says in there not to buy Western Digital external drives for this as they're not always bootable on Mac. I'm thinking of getting this one. I use that exact drive as a Time Machine backup drive and it works fine. Not sure why CCC docs think WD drives are a bad idea; some drives arrive formatted the wrong way and the mfr bundles software to make the one touch backup button on the drive work or some such thing, but in my experience you can safely ignore bundled software, nuke and pave with Disk Utility, install an OS, and bam bootable external. Just make sure to set the partition table format to GPT.
|
|
#
¿
Aug 28, 2013 00:59
|
|
|
Binary Badger posted:At the beginning of the OS X era, some French author wrote something called Diablotin that allowed you to pick and choose what system extensions / prefpanes got loaded and which didn't but someone must've yelled hard at him or he realized playing fast and loose with system extensions was a losing proposition. Huh, I hadn't heard of that one before. Looked it up and apparently it lets you control most plugin-ish stuff that goes under /Library. Neat, but kinda pointless. Classic Mac system extensions were a nightmare because they worked by patching into the base OS at a low level using mostly undocumented techniques. Update the OS, patches break. Install too many, some of them will conflict. OS X pref panes, kernel extensions, contextual menu items, etc. all interface to OS X APIs formally defined and maintained by Apple, so there's less potential for explosions and excitement. See also Unsanity, the company devoted to bringing the "wonderful" experience of classic Mac hack & patch style extensions forward into the OS X era! They even invented a cutesy name for their products, "haxies", and tried to evangelize their framework for injecting haxies into the system to other developers. This did not work out very well for them because Apple gave no fucks about radically changing OS X internals at whim. They spent a lot of their time playing catch-up to get their stuff working on the latest OS X release. Also, their software was so notorious for destabilizing the system that many application developers (and Apple itself) refused to take bug reports from users who had Unsanity products installed ("go uninstall that poo poo and then tell me if the bug still happens"). They lasted a lot longer than Diablotin, but finally seem to have disappeared off the web after years of hanging around promising that yes they'd finally be catching up to modern OS X releases real soon now.
|
|
#
¿
Aug 28, 2013 20:55
|
|
|
Splinter posted:Are there commands I can run to turn the second hard drive in my MBP off and on? I just installed a data doubler with a 7200rpm drive to go along with the factory SSD. I'm currently only using the second drive for music production (sample storage & recordings). I want to turn the 2nd drive off when I'm doing non-music tasks to minimize battery usage, heat and noise. I already reduced the disk idle time to disk sleep from 10 minutes to 5 minutes, but I still hear the drive occasionally when it doesn't need to be on. Since I know exactly when I need the 2nd drive on, I figure hard on/off commands will be a better solution. Try opening Disk Utility and unmounting it. You can maybe even eject it, though I'm not sure offhand whether you'd have to reboot to get an internal disk mounted again after an eject.
|
|
#
¿
Sep 10, 2013 10:22
|
|
|
QuarkJets posted:I can, I'm just venting frustration at having to modify settings that I think should have been defaulting to something else. I can't figure out a reason to not broadcast pageup/pagedown to open terminal programs, but at least it was relatively easy to change the behavior once someone told me what to do. I was always told that OSX is incredibly easy to use and user-friendly, so I was expecting a lot, but so far the Linux to OSX transition has been a lot more frustrating than the Windows to Linux. It's just a little disheartening is all. I'll get it You will, and it happens both ways. My personal computers are Macs, but I use RHEL at work, and having to use shift-PgUp/PgDn to scroll Linux terminals always bugs me. I think the key is that I never do significant text editing in a terminal. I can see how you'd be in a world of hurt if you do, but when you're just using a shell that behavior is annoying. Sending cryptic escape sequences to your command line serves no useful purpose, while paging through the terminal's buffer does. This is probably why Apple made that choice -- they expect most users to use GUI apps to edit text. (Even hardcore vi users can do that with MacVim -- if they're allowed to download poo poo! That is going to be a huge annoyance for you.) Also, the truly big advantage for OS X lies in other areas. Namely, the lack of driver or dependency hell. I have these disfiguring scars from trying to install decent X11 drivers on RHEL 5, and the less said about trying to install non-RPM programs on RHEL the better. 
|
|
#
¿
Sep 26, 2013 11:27
|
|
|
evol262 posted:As a "hardcore vi user", I would never use MacVim. The entire advantage of vim is that it comes with the operating system. The entire advantage to running a *nix (from my perspective) is that it's easy to get out to other systems via SSH and that everything behaves mostly the same from terminal to terminal once you get stty or whatever hosed thing that UNIX uses configured. Different strokes I guess. That comment was based on a friend of mine who is a certified Linux and Vim fiend, and who was also a certified Apple hater right up until he bought an iPhone, which we all gave him crap for.  Then a couple years later he followed the iPhone up with a MBP because it was in his opinion the best Sandy Bridge laptop hardware... and he proceeded to not install Linux on it, which (predictably) we all gave him crap for. He uses MacVim.quote:OSX doesn't have driver dependency hell because it doesn't have drivers to speak of. It certainly does have drivers, but I understand what you're saying -- most people never need anything but the drivers Apple distributes with the OS to cover their own hardware. The thing is, the infrastructure is there. There is a stable driver API and even a stable driver ABI. Third parties can (and do) distribute OS X kernel extensions which don't break when Apple sneezes. Apple isn't as good about driver API/ABI stability as Microsoft, but they're both way better than Linux. That's no accident either. It's a conscious policy decision by several core Linux kernel developers, including Linus Torvalds. They have their reasons, and I understand them, but it does cause usability issues all the same. Like this: quote:You don't need "non-RPM programs" to get "decent X11 drivers" on RHEL5. AMD, nVidia, Intel, S3, and everyone else publishes RPMs. RHEL keeps the kernel ABI the same the whole way through, so it's really easy (complaints about getting nVidia binary blobs working on nightly versions of X.org on Fedora/Gentoo/whatever are relevant, just that RHEL isn't). Er, they don't all publish RPMs. From very recent experience, I know for a fact that NVidia distributes self-extracting archives which, when you run them, literally compile a shim kernel module to interface the Linux kernel to the NVidia binary driver blob, also contained and installed by the self-extracting archive. And I also know that when this house-of-cards "install" process goes wrong it can be insanely frustrating to get working. (Thankfully, NVidia does at least make sure that failures are graceful -- all that happens is that you don't get your drivers installed.) Also, I didn't write that totally clearly. The RPM comment was about application software, not drivers. Not even all open source applications have been packaged as RPM, and even when they have they may not have been packaged for the specific RHEL I'm using. Sometimes installing a non-native RPM works. Other times I end up needing to install a non-native SRPM and hack on the source and/or build scripts till it builds something which does work. This is not fun. quote:Every proprietary UNIX has its own weird poo poo, including NeXT, and Apple dragged the (awful) NetInfo poo poo from NeXT into OSX. It's gone now, fyi. They added LDAP, migrated things to it over one or two OS X releases, and finally removed NetInfo in 10.5.
|
|
#
¿
Sep 27, 2013 08:36
|
|
|
agarjogger posted:To get a dev acct and download it early cost $99. Who said you have to pay? They seed to plenty of people on the free tier dev accounts. Of course, you probably need a track record of using said free account to file bugs now and then. 
|
|
#
¿
Oct 5, 2013 00:53
|
|
|
strokevictim posted:I still have one open from 2010: Hah. I've got one open on Safari from 2004.  Of course now that I've checked it for the first time in years, it looks like it's been fixed even though the bug's still open. But I'm pretty sure it took them at least five years, because I used to check on it once in a blue moon and it just kept persisting. (Not that it ever would've been high priority -- it was a minor text wrapping bug. Trailing quote marks and some other punctuation could get wrapped to the next line rather than staying glued to the word they were next to.)
|
|
#
¿
Oct 18, 2013 10:23
|
|
|
japtor posted:Didn't they kill off the MacBU (or roll it into some other unit) a while back? The latter. Speculation was that they needed to shuffle the deck a little to make the other unit look better (it was losing money), and MacBU was a relatively independent group that was turning a profit.
|
|
#
¿
Oct 19, 2013 06:26
|
|
|
Bob Morales posted:Just saw a fix on MacRumors: I also had problems getting 603 but that MacRumors fix is overkill IMO. What worked for me: 1. Delete "Install OS X Mavericks.app" from /Applications (if present). 2. Download from the App Store, but don't use the purchased apps pane. Search for it like you weren't ever in the seed program and download it from there. My theory: Apple probably uses different App Store application IDs for seed builds and releases. Same name, but different so far as the store is concerned. As soon as they release an OS to the public, they disable seed downloads. So if you're like me and were lazy about moving to 603 because 598 worked, trying to download using 598 and your registered entitlement to the Mavs seed (which is what it knows about in "Purchases") began failing yesterday even though on the 21st it probably would have grabbed the 603 installer. The end of the ML seed program was smoother because they just gave us App Store redemption codes for the real thing. (yup, it turned out that being in the ML seed program got you free ML)
|
|
#
¿
Oct 24, 2013 02:51
|
|
|
twoot posted:Mavericks seems to have done something to the fan control on my 2013 Air. Probably not, because OS X doesn't control the fans directly. That's done by the SMC (System Management Controller), a chip on every Intel Mac's motherboard which runs all the vital heartbeat functions. It uses a feedback loop with the computer's temperature sensors to determine how fast to run the fans, so increased fan noise implies higher temps implies higher power use. Mavericks is a very power focused release, but some of the rules for applications have changed, and it's easy to imagine apps which were already trying to do the right thing getting faked out by some of the new behaviors. Civ 5 probably tries to cap its own frame rate, running faster in the foreground than in the background. If that code is now hitting a bug when you tab out, and the failure path behavior is no frame rate cap whatsoever, that would explain why your fan's going nuts. That's guesswork, but if something like that is responsible, you may actually need a Civ 5 patch to fix it. Could be either party's fault, though. You could also try a SMC reset, since it can't hurt. You can search for how to do this on Apple's website.
|
|
#
¿
Oct 24, 2013 14:01
|
|
|
lord funk posted:Ahaha, no, Pages. That's actually bullshit. Oh great, this is going to make migration of the Momputer to a modern machine real fun. (She's on an old 2006ish iMac running 10.6 and the contemporary iWork. Decided to slow down on updating it a few years ago because 10.7 would have required some retraining and it was working fine as is. Was hoping to move her to a new computer soonish; maybe I should wait till a new iWork release to see if they treat it like FCPX and add in some of the missing poo poo over time.)
|
|
#
¿
Oct 26, 2013 04:23
|
|
|
Dick Trauma posted:http://www.macworld.com/article/1163755/vmware_fusion_update_lets_users_virtualize_leopard_snow_leopard.html What the article doesn't say is that you are going to have to find a way to get precisely that version of VMWare Fusion, and no other. Permitting installation of Snow Leopard client was considered an unintentional bug and they removed it from the very next Fusion patch. I don't have any links to share offhand, but you may be able to google up some methods of hacking disk images of the SL installer and/or SL installations to fool Fusion into accepting it as SL Server even when it's not.
|
|
#
¿
Oct 30, 2013 20:57
|
|
|
Binary Badger posted:Hmm, in Mountain Lion/Lion, the purge command could be issued from the terminal with no special permissions. This is guesswork because I don't have a L/ML install to check, but: I bet it always required root (telling the kernel to do something that drastic with memory is a root kind of thing), but was a SUID root binary so it could elevate its own privilege. Apple's probably trying to cut down on the number of SUID root binaries shipped with the system. Every one is an extra attack surface for malware.
|
|
#
¿
Oct 30, 2013 21:09
|
|
|
Ninja Rope posted:When people mail me directly and an email list that I'm on in the same email I only get one copy in Mail but it alerts twice. It seems like Mail is smart enough to de-duplicate the message in storage but it still plays the new message alert/notification for each dupe. Is there a way to stop this? It's a little annoying to get so many alerts. This isn't going to sound real helpful but step 1 is "file a bug report" and step 2 is "wait, and hope".
|
|
#
¿
Nov 5, 2013 21:41
|
|
|
the_lion posted:Is there a good way to find and remove files with certain names? Finder becomes clunky on my external drive (it found over 21,000 files) and the find function isn't getting all the files etc. Highly dangerous don't-make-a-typo UNIX shell technique, offered in addition to the almost-certainly-better-for-you Automator solution already posted because we unix weenies love this poo poo: code:
|
|
#
¿
Nov 8, 2013 06:55
|
|
|
JamesOff posted:I run this without -delete to check it's matching the right things, and then run it properly. Heh. Piping find output to xargs is so habitual for me (for grepping and whatever else) I didn't even remember find can delete stuff by itself. (and I know find has -exec too, but xargs is better in some ways)
|
|
#
¿
Nov 8, 2013 09:30
|
|
|
fookolt posted:Is there a way to have two Safari instances with independent profiles? I'd love to have one Safari I can use with my work logins and then one Safari I can use for all my home logins to the same site without having to log in and out all the time. This won't get you profiles, but you can run two instances of any Cocoa app by starting the second from the command line, like this: code:
|
|
#
¿
Nov 9, 2013 08:51
|
|
|
Star War Sex Parrot posted:This is what I generally do because it's way faster, but is there a way to make it not capture the window's drop shadow? I thought I saw a .plist change posted at one point. There is. Don't know the name, but TinkerTool can do it for you.
|
|
#
¿
Nov 11, 2013 22:32
|
|
|
^^^^ Over time, Apple added a bunch of stuff to AFP to make networked TM backups more reliable, and in the process made TM dependent on those special AFP features. They documented it here: https://developer.apple.com/library...08951-CH100-SW1 I expect TM's AFP requirement to go away in a future OS X major release. Obviously as of Mavericks they feel their homegrown SMB implementation is now robust enough to use for more than just light-use Windows compatibility, so the next step is to rewrite Time Machine's AFP-dependent data safety features for SMB. If you want to live dangerously, close Preferences and run this in Terminal: code:
|
|
#
¿
Nov 19, 2013 11:43
|
|
|
Djimi posted:I'd be surprised if it was just about getting a 'new' protocol finalized that other devices could use, to be a viable solution. "Think Different" has pretty much become, "Do As We Say". My eyes are rolling so hard. Network file systems are way more complex and fragile than most realize. (Note that it has taken Apple three major OS releases to get their brand new SMB2 implementation up to speed, and I suspect the reason for shifting away from AFP has a lot to do with not wanting to maintain multiple high perf networked FSes.) And Time Machine hammers the living poo poo out of them. Ever notice that TM needs a HFS FS to accurately back up HFS file systems, and when you use it to back up over the network it stores this HFS FS in a sparse bundle disk image? Every time they make changes to this containerized file system, they must ensure the modifications to blocks in the sparse bundle band files are written out to physical media in proper journal order, as otherwise there would be a high probability of corrupting the FS when the server connection breaks for a moment. This is not easy, or even possible, in most network file systems. And they do a lot of these modifications, every hour, on the hour, thanks to versioning and so on. You think this is trivial and that their design which channels you down the path of using their co-designed network FS is arbitrary and controlling, I think about the edge cases and frankly it's total  territory.I understand your bitterness that something got taken away which seemed to work fine, but sometimes the truth is that making systems like this reasonably robust for nontechnical users requires narrowing the support matrix to no more than what you control and working your rear end off on finding and fixing as many bugs as possible. For better or worse that's pretty much how Apple defines itself.
|
|
#
¿
Nov 20, 2013 01:46
|
|
|
^^^ They learned to leave the old versions of the app still installed so you have them available while they re-implement features in their ground-up rewrite. Which, apparently, is what the new iWork is. Also, they learned to be more communicative about plans to re-add said features. But only slightly more communicative. (This may still leave you wondering why they didn't learn "just loving implement everything before releasing v1.0 jesus christ")
|
|
#
¿
Nov 26, 2013 13:02
|
|
|
Star War Sex Parrot posted:I used to use iChat screen sharing but I'm not even sure if it's still in Messages anywhere. What up, fellow screen sharing for elderly-relative tech-support guy? It's still there. You just have to add a Jabber or AIM account to Messages on each end. The UI looks exactly the same, so I'm guessing none of the code changed. Ironically, I think you can't initiate screen sharing via iMessage/iCloud, it has to be Jabber/AIM accounts.
|
|
#
¿
Dec 1, 2013 05:26
|
|
|
Warning, possibly far too detailed EE sperging about power supplies lies ahead!Binary Badger posted:MacBook Pro 15-inch adapters are 85W, MacBook Air is usually a 45W adapter. There should be no ill effects as the adapters are intelligent enough to detect what kind of machine they're being hooked into and actually step down the wattage to the appropriate level. I read some guy's dissection / reverse engineering of a MagSafe once. Turns out the adapter doesn't actually know what it's connected to; there's only two wires (+V and ground) in the cable. There is an ID system, but it works in the opposite direction: there's an ID chip embedded in the MagSafe connector which tells the computer what kind of MagSafe is plugged in. A supply doesn't need to know anything on the order of "user has plugged me into a 15 inch MBP" to vary its output power. In fact its output power is fluctuating quite a bit while it's plugged in to any computer, since no computer demands a fixed amount of power. The supply fixes only one parameter: its output voltage. Power is equal to current times voltage (P=IV), so with voltage constant, current and power must change in proportion to one another. The supply has a voltage feedback loop which adjusts current (and therefore power output) in whatever direction is needed to keep its output voltage constant. "Adjusting" power thus consists of the computer just trying to draw more (or less) current and blindly trusting the supply to do its best to hold voltage constant. quote:There CAN be ill effects if you plug a lower wattage adapter into a machine that requires a higher wattage than what the adapter can provide. Seems unlikely much will happen other than being unable to charge at full rate (or possibly unable to charge at all in some cases), and draining the battery when running software which needs more than the adapter can provide. I'd be interested to know if there's anything else. eames posted:Surely connecting a 45W charger to an idle rMBP would be the same as connecting a 85W charger to a machine running at 50% GPU/CPU TDP. Sort of. MagSafe chargers don't all produce the same DC voltage. I just checked my 2011 Air and its 45W supply puts out 14.5V, while this brand new rMBP 15 85W supply is 20V. Generally speaking, the smaller the MagSafe's wattage rating, the lower its output voltage. But even at the same wattage rating things can change: apparently the original MagSafe 85W adapters were 18.5V. (With, of course, a higher maximum current rating than the new 20V 85W bricks.) This doesn't cause problems when mixing and matching because most circuitry inside the computers isn't powered directly from the MagSafe port. Instead it's fed by DC-to-DC converters, which step the high voltage MagSafe output down to lower voltages at higher currents. DC-to-DC converters are readily designed to accept a wide range of input voltages. I once worked on one which could run on anything from about 10V to 90V, which was a wide enough range to require a little bit of extra design effort. ~14V to ~20V? Easy as pie. You get that kind of input range without trying very hard at all. TLDR summary: if both are supplying a 30W load, an 85W MagSafe will generate less current at a higher voltage than a 45W MagSafe, so it's technically not quite the same thing. 
|
|
#
¿
Dec 4, 2013 13:10
|
|
|
^^^ Seconding this, in fact I'd go further and say that VNC depends on the combination of client and server. A paired VNC client/server that were co-developed usually work well together. Once you get away from that it tends to be random, so try clients till you find one that works well with your server, even the one you thought was crap because it didn't work so well with a different server.
|
|
#
¿
Dec 10, 2013 20:52
|
|
|
The identical version numbers for HD 4000 and 5000 suggests the drivers are built from the same source, which is good news for HD4K long term support. Not unheard of for two generations of GPU to be similar enough to share most of the driver. HD3000 is left out in the cold, alas. I know of a significant bug and at this point doesn't look like it'll ever be fixed. Had my hopes up when Mavericks shipped with a bump to the HD3K driver version but nope, still there. (Avoidance strategy for said bug: reboot at least once a week. It's a hang which never shows up (for me anyways) before a week plus of uptime.)
|
|
#
¿
Dec 17, 2013 01:13
|
|
|
Japtor, if that was a question for me, the hd 3000 bug I know of is a hang. I never tracked down a reliable way to reproduce on demand so I don't know a specific trigger other than long uptime and heavy use. When it happens the machine will freeze for 10 to 15 seconds, and then the video driver re-inits and the system is semi usable again. However it's only semi usable, from then on the hang will retrigger on just about any little thing, such as opening or closing windows. If you're patient you can cleanly reboot, just takes a long time to quit apps since closing windows triggers so many hangs.
|
|
#
¿
Dec 17, 2013 01:38
|
|
|
PRADA SLUT posted:Is there a reason some YouTube videos won't load without flash? Workaround? I believe they require it on popular videos where they want to get ad money. No workaround that I know of. Well, sort of. One way to have your cake and eat it too is to use Chrome as a secondary browser just for viewing flash stuff. Chrome includes its own private copy of flash so that Google can auto update it for security, so it doesn't require flash to be installed as a system wide browser plugin.
|
|
#
¿
Jan 15, 2014 01:32
|
|
|
Laserface posted:Upgraded my Mid 2011 iMac from 4GB to 16GB since it was running a little sluggish in 10.9 and while the performance bump is nice, I noticed 8GB is being used up by file cache. This seems to just grab whatever it can and hang on to it. There is no setting to tweak, and as far as I know OS X doesn't even have a limit per se, so you're going to have to learn to stop worrying and love the bomb. You mention seeing memory use go up despite quitting apps to try to get it to go down. Well, one of the clever things the OS X virtual memory system does is to keep lots of a quit program's memory around. Whatever it decides to keep is reclassified as "inactive". The parts of the OS which fill memory allocation requests treat inactive memory as an extension of free memory. As soon as truly free runs out, inactive is used. Only if the system uses up all free and inactive memory will it even begin to consider swapping stuff out to make room. Inactive memory is an optimization because hey, it's still valid data. If nothing reclaims it, maybe some other program will try to read the file data which the inactive memory is a copy of, in which case the OS notices that it's still got that and doesn't have to waste time rereading it from disk. In this and other ways, OS X is lazy about throwing memory away. But in a good way, because with proper design there's really no downside to procrastination. All that said... open a Terminal window and do a "pmset -g". What does it show for hibernatemode? Non zero means your system may be hibernating instead of sleeping. I ask because (far as I know) a sleeping Mac should not take longer to come out of sleep no matter how much swap it used while awake. Also, if your Mac is hibernating rather than sleeping you might be in for a rude surprise, in that hibernation performance can easily get worse with more memory installed. Your worst-case hibernate restore is now up to 16GB read from the hibernation image instead of 4GB...
|
|
#
¿
Jan 17, 2014 14:08
|
|
|
My PIN is 4826 posted:Whyyyy? Because UNIX user accounts were designed to provide data security in the context of large institutions where dozens or hundreds or even thousands of people are sharing one computer. The assumption is that default behavior should avoid inadvertently exposing your data to other accounts. You don't want trashing a file to unintentionally share it with anyone else just because the trash folder is shared. The design of OS X inherits that philosophy, so trash folders are per user. Apple is trying to change security philosophy for modern needs. Today the main threat comes from inside the browser or the plugin or the downloaded app: software you run (intentionally or not) from inside your own account. That's what the whole sandboxing push for MAS apps and Apple's bundled apps like Safari is about, limiting or eliminating the damage and spying that can be done on your files as you. The old unix philosophy security model is still kicking around at the base of OS X, just not being paid as much attention.
|
|
#
¿
Jan 18, 2014 03:34
|
|
|
Amberskin posted:If I'm not wrong the "5 seconds press" is hardwired so it should not be affected by an operating system level remap. This is correct. It's a SMC function because it has to work even when OS X is hardlocked and incapable of processing keystrokes. Same reason why the SMC does not rely on OS X supervision to control fan speeds, it has all the relevant sensor data and makes its own decisions.
|
|
#
¿
Jan 24, 2014 04:04
|
|
|
jackpot posted:Say for example that giraffe is my old profile, and banana is my new one. Why, when I open terminal, does it read: Do you come from a Windows background by any chance? For what it's worth, in unix land the terminology is user accounts. There are also groups, to which users may belong. Unix shell prompts often show the computer name followed by what directory you're presently in, the former since a terminal might be SSH'd to another machine, and the latter because it's always useful to know where in the file system you're at. What you're seeing to the left of the colon is the computer name. It's only related to your first user account name because that's what Apple's first-boot setup program does. Until you change the computer name in System Prefs you'll see "giraffe" in the command prompt for any user account, and it has no implications for file permissions. Next in the prompt is the current working directory. Note that "~name" is unix shorthand for whatever the home directory of user "name" happens to be, which on OS X will be "/Users/name" unless you've put some effort into relocating that user's home directory. When you open a terminal it starts you out in your home dir, so that's where the "~banana" comes from. Now, finally, we get to why "chmod 755" might not be doing what you expect. Every file and folder in a unix system is owned by exactly one user and one group. The basic permissions system has three groups of three permission bits (R/W/X), one for the user who owns the file, one for the group, and the last for everyone else. 755 is an octal number; the left digit is user, the middle is group, and the right others. So 755 gives the owner permission to do anything, and the group or anyone else to read and execute. (If it's a directory, X means the ability to list contents rather than execute). Thing is, when you do a chmod 755 you aren't changing ownership. Your new user account is getting either the group or the "everyone" permissions, and they are still more restrictive than the owner permissions. You need to use the 'chown' command to change what account owns it, or set the permissions to something other than 755 (eg 777 to let anyone do anything, although that is not best practice). If you want it to be visible to only those two accounts you'd want to grant the group higher permissions and add your new user account to the owning group (which is a little more involved). That's the old school UNIX privilege model which dates back to 1969. OS X does implement the modern-UNIX POSIX Access Control List extensions, which do things more like the fine grained model you're expecting, where you can individually tag files and folders with who can do what to them. However I'm honestly not familiar with POSIX ACL manipulation commands since on a personal machine you don't usually need that level of flexibility. (If you didn't need a brief course in How Unix Works then my apologies in advance for misreading the question)
|
|
#
¿
Jan 25, 2014 02:10
|
|
|
BobHoward posted:However I'm honestly not familiar with POSIX ACL manipulation commands since on a personal machine you don't usually need that level of flexibility. So that effortpost got me curious about playing with ACLs. I think this is the right incantation to give user "gorilla" the ability to do anything you (that is, the user running this command) can do in directory "foo". The inheritance clauses make these rights extend to every file or folder that is inside foo, including nested files and folders. I might have erred on the side of adding too many permissions. chmod +a "gorilla allow file_inherit,directory_inherit,list,search,add_file,add_subdirectory,delete_child" foo Learn more than you wanted to about all this by typing "man chmod". Also, I slightly screwed up the description of what's in the OS X command prompt by default: sloth:~ gorilla$ This is actually computername:currentdirectory username$
|
|
#
¿
Jan 25, 2014 06:51
|
|
|
|
| # ¿ Apr 27, 2024 13:35 |
|
|
I don't know of any way to make a USB installer from Windows which doesn't involve  at some level.When you're booted into the recovery partition or internet recovery, you can run Disk Utility. It's under one of the menus after the installer / recovery OS boots. Run it and click on your disk; does it say "S.M.A.R.T. Status: Verified" at the bottom? What happens if you then click the "Verify Disk" button? It's not totally clear from your post, so: have you tried completely reformatting the partition? Before you kick off an installation, run Disk Utility, click the partition you're going to install to, select the Erase tab, leave it on Mac OS Extended (Journaled), give it whatever name you like, and click the Erase button. (Is it a spinny disk, or a SSD? If it's a hard drive and installation fails again after erasing the partition and starting over, you could try doing an erase again but this time use the "Erase free space..." button then configure it to write zeroes (one pass). HDDs having problems with a few bad sectors can sometimes be resurrected by overwriting every sector, which gives the drive a chance to remap spares good sectors in place of the bad. Be prepared for a full HDD erase to take many hours. Also this is by no means foolproof, if a drive has lots of bad sectors it's probably on the way to total failure and this will be at best a temporary measure.) Finally, if the installation gets to a point where it is doing stuff with a progress bar, you can hit Command-L to open a log window. Set it to display all logs and let it progress till it fails. There will be a lot of stuff that looks like errors even in a healthy install, so focus on what happens right before it gives up. Note that if you have the log window open and in the foreground it will not allow automatic reboots that are part of the installation process to take place, just close it to proceed.
|
|
#
¿
Feb 2, 2014 04:11
|
|