Wikipedia:Reference desk/Archives/Computing/2012 May 27

Computing desk
< May 26	<< Apr | May | Jun >>	May 28 >

Welcome to the Wikipedia Computing Reference Desk Archives
The page you are currently viewing is an archive page. While you can leave answers for any questions shown below, please ask new questions on one of the current reference desk pages.

May 27

Linux, file created on?

Why do linux files do not have a creation date metainformation? OsmanRF34 (talk) 01:18, 27 May 2012 (UTC)

They can, but why should they? ¦ Reisio (talk) 03:51, 27 May 2012 (UTC)

All versions of ext have create, modify, and access date metadata. Shadowjams (talk) 03:45, 28 May 2012 (UTC)

Wrong. Traditional unix filesystems have atime, mtime, and ctime. Lots of people guess that ctime stands for "creation time", and then that guess turns in to a solid belief over a long period of failing to read any proper documentation, and that solid belief sometimes becomes an authoritative pronouncement reinforcing other people's bad guesses... Actually ctime stands for "change time" and it's updated whenever any attribute of the file changes. Among native Linux filesystems, ext4 has a creation time field, but ext3 and earlier don't. Just see the articles ext3 and ext4, where it says "Dates recorded".

As for why it took so long, I don't know, but it might have been a chicken/egg problem. Even now, with ext4 recording the creation times, it's hard to actually look at them, since 40 years of unix tools and programming languages have been built around the assumption that the original 3 timestamps are the only ones that exist. 98.226.12.79 (talk) 04:50, 28 May 2012 (UTC)

ctime used to stand for "creation time". Richie & Thompson's 1974 paper "The UNIX Time-Sharing System" says (in section IV) that an inode records "time of creation, last use, and last modification". Similarly the 1971 Unix system manual's section on the stat(2) call shows it yielding 4 bytes for "creation time" (ctime) and 4 for "modification time" (so we're before the age of atime). There's nothing in the brief descriptions of inode-altering calls like chown(2) and chmod(2) that says these change the ctime. So at some point after Unix defined ctime someone decided that meta-calls like chown should change the ctime, and thus they had to back-change what the abbreviation means to the rather mealymouthed "change time". This blog posting puts the change at around UNIXv7. I'll bet the change happened when someone was writing an incremental tape backup program and complained that they needed to capture permissions changes, but didn't want to have to write a whole file to the tape whenever it was merely chmod-ed. Apart from that use (which only makes sense on sequential media like tape; rsync's method make much more sense when one has random-access backup media) I'm rather lost to think of a useful application for ctime. Happily many modern unixen have taken a "better good than POSIX" approach to atime (cf Stat (Unix)#Solutions); perhaps a ctimecreate option for mount would make sense. -- Finlay McWalterჷTalk 14:09, 28 May 2012 (UTC)

I'm going to update Stat (system call) a little bit to reflect some of what Finlay has said. Shadowjams (talk) 18:00, 28 May 2012 (UTC)

To my mind there's a lot of stuff in Stat (system call) that should be in inode, including almost all of the extensive atime discussion (and inode is where I'd expect details about ctime worked and works). These data are held in the inode and changed by a variety of syscalls; stat is merely the means by which we see them. -- Finlay McWalterჷTalk 18:12, 28 May 2012 (UTC)

I was kind of thinking that too, but the "userspace" caveat gave me some pause because I suppose while one could find a way to alter those by tinkering with the disk directly, that might not technically be considered userspace. I apparently already made a technical "blunder" here already =). Shadowjams (talk) 07:12, 30 May 2012 (UTC)

Although the blog link provided by Finlay McWalter says that the earliest versions of unix source code were lost, most of them have since been found.

In V1, there is a "ctim" field, which appears to be a working creation timestamp. It's set in the kernel by the function called "maknod", and never updated anywhere else. C didn't exist yet, so this is all in assembly.

V2 kernel source is still lost, or at least I can't find it.

In V3, maknod is gone, the new function is maknode, written in C, and there is no creation timestamp set there.

It seems plausible (WP:OR) that in the transition from assembly to C, the ctim got accidentally left out, and nobody noticed for a long time because it wasn't frequently needed, and then the issue was resolved by removing it from the documentation rather than reinserting the code to make it work again (which by that point may have required an incompatible change to the on-disk inode format). 98.226.12.79 (talk) 04:46, 31 May 2012 (UTC)

Lyrics on phones

Can the Samsung SII display lyrics while a song plays? I wouldn't think so, but, just asking. --Sp33dyphil ©^hat_ontributions 09:35, 27 May 2012 (UTC)

Which version of Windows is Wine supposed to be?

OsmanRF34 (talk) 16:12, 27 May 2012 (UTC)

winecfg lets you set which version it tries to pretend to be. The docs say it defaults to Windows XP - but in practice it implements some subsystems from newer versions, and doesn't fully implement everything from older ones. -- Finlay McWalterჷTalk 16:20, 27 May 2012 (UTC)

Bluetooth on Win7 and Samsung Galaxy Phone

I am unable to connect to my phone using Bluetooth from my Win7 laptop. My phone connects ok, and the computer even notices that I have changed the name of the device, but when I try to transfer any files, it just tells me it can't find the device. I have been into Control Panel>Administrative Tools>Services, and everything appears to be running. Does anyone know what to do here? KägeTorä - (影虎) (TALK) 17:55, 27 May 2012 (UTC)

Not all Bluetooth stacks are equal, you might try http://www.broadcom.com/support/bluetooth/update.php or http://aps2.toshiba-tro.de/bluetooth/?page=download. ¦ Reisio (talk) 19:03, 27 May 2012 (UTC)

Thanks. My PC is a HP laptop with Win7. I have not been able to update the driver, and I think it just gets done automatically. So, basically, what you're saying is, my phone's Bluetooth and my PC's Bluetooth may not be compatible? That's annoying, because neither is my iPod's Bluetooth (which works only with iPhone and Mac computers). As I also don't have a USB cable for the phone, I have taken to sending all the files I want by email. Cheers, anyway. KägeTorä - (影虎) (TALK) 13:22, 28 May 2012 (UTC)

Not so much incompatible as much as your laptop may well not have even come with any stack (to speak of). If it doesn't have internal bluetooth this is a foregone conclusion, if it does it's probably something else. ¦ Reisio (talk) 20:59, 28 May 2012 (UTC)

Well, my laptop's Bluetooth worked with my old Sony Ericsson and works with my Android tablet. I'm surprised it won't work with my Galaxy phone (which is also Android)...... KägeTorä - (影虎) (TALK) 21:22, 29 May 2012 (UTC)

You should be able to send files over wifi without using email, Samsung Kies for example seems to support wifi. (It's even possible to enable ADB over wifi although I'm not sure if some phones prevent this.) Alternatively a microUSB cable should be at most a few dollars; under $5 shipped if bought from eBay, DealExtreme, KaiDomain, Buyincoins or similar. Or complain to Samsung, some/many? other manufacturers provide a power adapter with a USB socket and a USB to micro cable meaning as long as you have your power adapter you have your cable which IMO seems better then the Samsung practice of hardwiring the microUSB to the power adapter. If you don't have a wifi network whereever you are and are relying on mobile data, I presume your laptop must at least have a wifi NIC and it's possible to enable AP mode in various versions of Windows 7 [1] [2] presuming the NIC and drivers support it. (Alternatively you can often connect to an adhoc network but I think this usually relies on a rooted phone.) Nil Einne (talk) 04:19, 30 May 2012 (UTC)

Twitter following

Occasionally, I find that somehow a load of people have somehow made me follow them on Twitter without me wanting to. It's bloody annoying. How can I stop this happening? Thank you. DuncanHill (talk) 18:50, 27 May 2012 (UTC)

If that is happening I think your account might compromised. You might want to change your password and look through the list of applications that you've allowed to use Twitter on your behalf. More: http://support.twitter.com/articles/76036 --Tow Trucker talk 06:18, 31 May 2012 (UTC)

Hmm, two apps, klout.com and tweinds seem to be giving themselves approval, have just removed them for the second time. DuncanHill (talk)