Wikipedia:Reference desk/Archives/Computing/2011 May 9

= May 9 =

Open Office Writer trouble
Hello. I use OO Writer and recently whenever I go to "Format" (top of screen) -> "Page...", and then "Page" tab, it freezes. I have tried this several times and always with the same result. I have also restarted my machine, which did not help. THis is the most recent version. What can I do to fix this? Thanks. 72.128.95.0 (talk) 00:45, 9 May 2011 (UTC)


 * Always try a reboot when this type of thing happens. If that doesn't fix it, it may be that the software has become corrupted, in which case you need to uninstall it and then reinstall.  If that still doesn't fix it, then perhaps there's an incompatibility with something else on your PC.  Trying running OOW alone, after a reboot.  You can also try a system restore to before the problem occurred, or else perhaps an older version of OOW.   StuRat (talk) 18:23, 10 May 2011 (UTC)


 * This happens quite frequently to me as well. The Page tab seems to be pretty buggy, it always crashes the program when I try it. Perhaps it's a known issue? 74.190.117.8 (talk) 18:39, 12 May 2011 (UTC)

Windows Live Mail
There was an unexpected error and Windows Live Mail can't sync the messages in your Hotmail account. To sync your messages, in Windows Live Mail, remove your Hotmail account and then add it again. To view messages in your Hotmail account, go to http://hotmail.live.com on the Web.

Subject '' Server Error: 4102 Windows Live Mail Error ID: 0x8DE20003

The solution given yields no result, help? Plasmic Physics (talk) 03:41, 9 May 2011 (UTC)


 * Have you previously synced successfully, or is this the first time you have tried? Are you able to access your mail via the website? If you have an older version of Windows Live Mail, you may want to update it. Removing and re-adding the problem account seems to be the most commonly suggested solution, but if that hasn't worked for you, it could be a temporary issue with the Hotmail servers, which would hopefully be resolved in a day or two. (See ) -- Kateshortforbob talk  15:23, 9 May 2011 (UTC)

Yes, no. Yes. Newest version. OK. Plasmic Physics (talk) 02:49, 10 May 2011 (UTC)


 * Problem hasn't gone away yet. Plasmic Physics (talk) 01:26, 11 May 2011 (UTC)

Comcast speeds? / Bridging two internet connections?
I am moving into a new apartment complex. The complex is wired for ethernet and provides a basic internet service as part of the rent. My understanding is that approximately 40 units share a symmetric 20 Mbps connection (not sure what the backend technology is). This is good when the network is not congested, but that's relatively little bandwidth for the number of people it represents and I've measured effective connection speeds less than 1 Mbps on a number of occasions.

With that in mind, I am strongly considering purchasing a separate broadband service. Based on the choices in my area (in the US), and a variety of other factors, I'm leaning towards Comcast at the 50/10 Mbps tier. However, I know that Comcast also shares bandwidth across multiple residences. I'm wondering how big a problem that is in practice, and whether I would just be trading one network congestion problem for another? (I'm led to believe that there are a few apartments that already subscribe to Comcast internet, but not very many, so it might not be a big deal.) Secondly, is there any practical way to bridge the two networks so that I could exploit both the Comcast service and the included-in-the-rent-but-often-congested service? Ideally, one might split the connection across both channels and utilize whatever was available, but even having a simple fail-over in the event of a disruption might still be something.

Assuming I will have two internet connections, are there practical (and not too costly) ways of exploiting that? Dragons flight (talk) 09:32, 9 May 2011 (UTC)


 * The formal term for the sharing of bandwidth like this is contention ratio; the maximum contention you'll experience on a given connection is often a characteristic of the contract. ISPs in general sell a "business" grade connection with a lower contention ratio than the "consumer" grade - all that differs is the worst-case contention - I don't know about Comcast. Beyond whatever minima they guarantee in the contract, there's no way really of determining what you'll get beyond measurement, as all the ISP will tell you is the contractually mandated CR, not the details of the actual connection you have (which will change, anyway, as new subscribers are added and subscribers change their usage patterns).    To use two connections you need a router with dual-WAN ports (or two routers with a smart switch between you and them). A friend of mine uses a Draytek Vigor 3200 to do something like you want (but with ADSL connections). -- Finlay McWalter ☻ Talk 11:09, 9 May 2011 (UTC)


 * Strictly when I say "you need a router with dual-WAN port" I mean a router which also has fail-over and load-balancing capabilities. I'd expect that most dual WLAN routers would, but check to make sure. -- Finlay McWalter ☻ Talk 13:21, 9 May 2011 (UTC)

Centering Images in CSS without Floating?
Is there a way to center images in CSS without floating them or using display:block? --CGPGrey (talk) 09:40, 9 May 2011 (UTC)


 * The sensible way is to use display:block (and centre it with margin-left:auto;margin-right:auto); in what circumstances are you prevented from doing that? In this case using display:table would work too (in combination with the same margin settings). -- Finlay McWalter ☻ Talk 10:54, 9 May 2011 (UTC)


 * I'm trying to avoid images splitting over pages when the HTML document is printed. The don't split if I take out display:block, but they are no longer centered.  --CGPGrey (talk) 11:43, 9 May 2011 (UTC)


 * The "correct" thing to do is to use the page-break-before and page-break-after properties, but in my experience effective browser handling of these is pitiful (see "paged media" in Comparison of layout engines (Cascading Style Sheets)). Absent that, where browsers choose to break the page (and how they handle things like images) seems terribly arbitrary: the trouble you will find, I fear, is that even if you achieve the printout you want on one browser, others will behave in distressingly different ways. All I can suggest is:
 * Try tags around the image (they do centre it; I don't know how they handle pages)
 * Mercilessly exploit @media print styles to change margins, insert shims, resize objects, and position things absolutely (in mm)
 * But I fear you'll find printing for web pages, CSS or no, is variable and mostly rubbish. -- Finlay McWalter ☻ Talk 11:58, 9 May 2011 (UTC)


 * Another alternative: no CSS for the image, but wrap it in a div which has text-align:center. That's logically the same as wrapping it in a <tt> </tt> tag. -- Finlay McWalter ☻ Talk 13:19, 9 May 2011 (UTC)


 * Only the most ignorant computer users are going to be printing websites for the images. Most people would save the images directly, I'd say, and print for the text. ¦ Reisio (talk) 14:01, 9 May 2011 (UTC)
 * It depends on what the images are, obviously. If they are scans of document, for example, they will probably be printed. --Mr.98 (talk) 23:22, 9 May 2011 (UTC)
 * Yes, images of text online is a whole 'nother ball of fail. ¦ Reisio (talk) 17:14, 10 May 2011 (UTC)
 * If you're worried about how the page will print, providing a print-ready PDF as an alternative is always acceptable. I've done that with web materials that were intended to also be easy to print. --Mr.98 (talk) 23:22, 9 May 2011 (UTC)

What does "\033[1;1H\033[2J" mean?
I have a C program that runs on a Windows machine that simply clears the command line screen. It does so by writing "\033[1;1H\033[2J" to stdout. What's the significance of the characters of this string to Windows? Or, in other words, could somebody who knows break down, for my curiosity's sake, how this set of characters coming into the standard output stream gets the operating system to clear the terminal? Thanks. 20.137.18.50 (talk) 15:56, 9 May 2011 (UTC)


 * 033 is octal for the ASCII char <tt>ESC</tt>. So that translates into <tt>ESC [ 1 ; 1 H ESC [ 2 J</tt>.  It's an ANSI escape code (which I'll decode momentarily...) -- Finlay McWalter ☻ Talk 16:08, 9 May 2011 (UTC)


 * <tt>ESC [</tt> is the Control Sequence Introducer (or CSI). So that string further translates to <tt>CSI 1 ; 1 H CSI 2 J</tt>. The first sequence <tt>CSI 1 ; 1 H</tt> means "set the cursor to position (1,1) on the screen". The second, <tt>CSI 2 J</tt> means "clear the whole screen".  -- Finlay McWalter ☻ Talk 16:16, 9 May 2011 (UTC)


 * As to how this works in compiler land: the C string literal "\033[1;1H\033[2J" is compiled into a byte sequence in the .text segment of the final executable that looks like <tt>1b 5b 31 3b 31 48 1b 5b 32 4a 00</tt> (where 1b is the hex for ESC). When that's passed to printf it scans it and doesn't find anything to process, so it passes it to puts which squirts it down the socket to the terminal verbatim. The terminal has a simple state machine which tracks its input looking for CSIs and which intercepts and handles supported ANSI codes. -- Finlay McWalter ☻ Talk 16:44, 9 May 2011 (UTC)
 * I haven't seen those kinds of things for a long while. They are control codes for a VT100 character cell terminal (and its many emulators) and should work independently of the OS - it is not the OS which is clearing the screen, but these codes are supposed to be intercepted by the terminal to clear the screen.  In this particular case, as as Finlay says, they mean move to the top-left of the screen and clear the screen.  This site gives a comprehensive list of the codes.  Astronaut (talk) 11:18, 10 May 2011 (UTC)


 * Thanks all. It's funny, I wrote "\033[31;4mHello!\033[0m" to a Cygwin xterm and it appeared red and underlined as expected, but when I ran it with a DOS cmd terminal, it made the "Hello!" turquoise and not underlined. 20.137.18.50 (talk) 13:39, 10 May 2011 (UTC)

Javascript virus ?
I saw this on a page and it made my browser report an error. What is it?


 * JavaScript code removed following a post at the Help Desk. Apparently it makes the page trigger anti-virus alerts -- John of Reading (talk) 11:05, 10 May 2011 (UTC)


 * That is broken JavaScript; it appears that the intended purpose was to use JavaScript to load a program (specifically, a Java plugin application) from the website hosted at 178.92.90.81 (which does not even appear to be a valid web server). Unfortunately, it appears that the programmer of that website made at least one or more typos in his JavaScript, and the program did not load properly.  Nimur (talk) 18:27, 9 May 2011 (UTC)


 * (For those interested in more detailed diagnosis: it appears that the programmers misspelled "append"; mistyped the URL for his JAR file, and does not actually host an HTTP server at that IP address). That's all the errors I caught via a quick glance.  Nimur (talk) 18:30, 9 May 2011 (UTC)


 * Also browser sniffing like this is unnecessary (read: lame and awful). ¦ Reisio (talk) 19:25, 9 May 2011 (UTC)


 * Looks like the piece of code is trying to run malicious codes on your computer. Lucky it contains a typo and didn't run. 118.96.160.78 (talk) 03:34, 10 May 2011 (UTC)


 * That's sort of FUD`ing. There is absolutely no reason to assume anything about intent just because the program is from the internet.  In fact, a Java JAR is probably one of the most safe things to run from the internet, because it is sandboxed from the get-go.  Here's an article from JavaWorld Understanding the keys to Java security -- the sandbox and authentication, and here's the official documentation on the Java Security Environment.  If you know anything about WebStart, you'll know that a sandboxed, web-started JAR can not open sockets, can not access local files, can not access most multimedia hardware, and in general is "impossible" to use for most malicious purposes.  Is there any reason to suspect that the linked JAR is actually malicious, other than its hosting in the Ukraine and its obfuscated classname?  It could equally-well be a legitimate Java program (in fact, it is more likely the program is legitimate, rather than "malicious," given the security restrictions I noted above).  The "exploit database" linked above seems to be a post by an amateur who has "discovered" that Java Web Start is a thing that exists, but has not bothered to learn what it actually is.  Nimur (talk) 16:47, 10 May 2011 (UTC)


 * I'm pretty sure it's a real vulnerability. It's Secunia advisory SA39260, and Oracle appears to have acknowledged and patched it. -- BenRG (talk) 23:00, 10 May 2011 (UTC)
 * Ah. But did the original code posted above actually have an injected argument?  I didn't see one.  Somebody removed the sample-code from the page.  Nimur (talk) 23:59, 10 May 2011 (UTC)
 * On further review, the intent is to run the JAR outside the WebStart sandbox, which is accomplished via the command argument injection; (and therefore might have access to the local file-system and network). The unknown JAR will still run in an ordinary Java sandbox, though.  My error, for not reading the original alert more closely.  Technically, the security flaw is in the Sun-distributed NPAPI plugin and the IE ActiveX control, not in Sun Java 6.  Nimur (talk) 00:07, 11 May 2011 (UTC)
 * I added a more useful (sub)title. StuRat (talk) 23:14, 10 May 2011 (UTC)

Fedora 12 device automounting comes and goes randomly
It appears that the device automounting on my Fedora 12 system switches on and off randomly. I haven't touched any configuration files or turned any services or daemons on or off, but one day it just starts working and then one day it just stops working. I can always mount the devices manually when it's not working, but I'd like to have control over it myself. What could possibly be causing this? J I P &#124; Talk 18:14, 9 May 2011 (UTC)
 * Are you using <tt>/etc/fstab</tt> or some other method to automount? Nimur (talk) 18:32, 9 May 2011 (UTC)
 * I'm certainly not using <tt>/etc/fstab</tt> to automount, I use it strictly for manual mounting. I'm not actually sure what is doing the automounting, but based on the questions I've asked here, I guess it's HAL. J I P  &#124; Talk 18:52, 9 May 2011 (UTC)
 * The most probable culprit, then, is Gnome (your graphical login manager). The <tt>gnome-volume-manager</tt> daemon is on by default, and I have had strange interactions with various drives mounting and unmounting on various types of linux.  I usually disable the Gnome auto-mount completely and let Linux perform automounts at the kernel level (with udev or autofs on Ubuntu).  See the Fedora Project page for gnome-volume-manager, or see this tutorial on Using Gnome Volume Manager to learn about setting user policy.  Nimur (talk) 23:05, 9 May 2011 (UTC)
 * My phone's charger is a usb connector. When I'm charging my phone but don't have it set to mount USB, I cannot get Fedora to automount anything via usb. If I disconnect the phone or mount the phone manually, everything works great. So, is there something you are plugging in that may be causing issues like that? -- k a i n a w &trade; 18:41, 9 May 2011 (UTC)
 * Well, the last time it magically started working was when I plugged in a CF memory card to my card reader instead of plugging my camera in with the card in place, and the last time it magically stopped working was when I last rebooted the computer. J I P  &#124; Talk 18:52, 9 May 2011 (UTC)

HP DV 100 Pavilion internal mic does not work after loading Win 7 Home premium.
Hi anyone,

I have a HP Pavilion DV1000 laptop, that is several years old, but has been successfully upgraded from the riginal XP operating system to Vista and now subsequently to Win 7. Home Premium.

The machine has also been upgraded to 2G of RAM and runs fine, but whilst the sound card works to provide (output) sound, the internal mic "refuses" to work with Win 7. The mic itself does not seem faulty and interestingly an external mic plugged into the mic input will work, but I would like to be able to find a driver/solution that will make the internal mic work........any ideas?

Brian —Preceding unsigned comment added by 116.212.193.250 (talk) 23:21, 9 May 2011 (UTC)
 * I did a quick search for the audio chipset of your laptop which appears to be Conexant CX20468-31 AC97. There are no windows 7 drivers for this chipset on the HP website. However i did a search for windows 7 drivers for this chipset and came across a few sites you might try, like this (note there's a 2nd link in that page) and this. They don't specifically mention "microphone" as far as I can see, but worth a try. Vespine (talk) 00:51, 10 May 2011 (UTC)