Search Results

Search found 16940 results on 678 pages for 'disk drive'.

Page 372/678 | < Previous Page | 368 369 370 371 372 373 374 375 376 377 378 379  | Next Page >

  • SQL Server IO handling mechanism can be severely affected by high CPU usage

    - by sqlworkshops
    Are you using SSD or SAN / NAS based storage solution and sporadically observe SQL Server experiencing high IO wait times or from time to time your DAS / HDD becomes very slow according to SQL Server statistics? Read on… I need your help to up vote my connect item – https://connect.microsoft.com/SQLServer/feedback/details/744650/sql-server-io-handling-mechanism-can-be-severely-affected-by-high-cpu-usage. Instead of taking few seconds, queries could take minutes/hours to complete when CPU is busy.In SQL Server when a query / request needs to read data that is not in data cache or when the request has to write to disk, like transaction log records, the request / task will queue up the IO operation and wait for it to complete (task in suspended state, this wait time is the resource wait time). When the IO operation is complete, the task will be queued to run on the CPU. If the CPU is busy executing other tasks, this task will wait (task in runnable state) until other tasks in the queue either complete or get suspended due to waits or exhaust their quantum of 4ms (this is the signal wait time, which along with resource wait time will increase the overall wait time). When the CPU becomes free, the task will finally be run on the CPU (task in running state).The signal wait time can be up to 4ms per runnable task, this is by design. So if a CPU has 5 runnable tasks in the queue, then this query after the resource becomes available might wait up to a maximum of 5 X 4ms = 20ms in the runnable state (normally less as other tasks might not use the full quantum).In case the CPU usage is high, let’s say many CPU intensive queries are running on the instance, there is a possibility that the IO operations that are completed at the Hardware and Operating System level are not yet processed by SQL Server, keeping the task in the resource wait state for longer than necessary. In case of an SSD, the IO operation might even complete in less than a millisecond, but it might take SQL Server 100s of milliseconds, for instance, to process the completed IO operation. For example, let’s say you have a user inserting 500 rows in individual transactions. When the transaction log is on an SSD or battery backed up controller that has write cache enabled, all of these inserts will complete in 100 to 200ms. With a CPU intensive parallel query executing across all CPU cores, the same inserts might take minutes to complete. WRITELOG wait time will be very high in this case (both under sys.dm_io_virtual_file_stats and sys.dm_os_wait_stats). In addition you will notice a large number of WAITELOG waits since log records are written by LOG WRITER and hence very high signal_wait_time_ms leading to more query delays. However, Performance Monitor Counter, PhysicalDisk, Avg. Disk sec/Write will report very low latency times.Such delayed IO handling also occurs to read operations with artificially very high PAGEIOLATCH_SH wait time (with number of PAGEIOLATCH_SH waits remaining the same). This problem will manifest more and more as customers start using SSD based storage for SQL Server, since they drive the CPU usage to the limits with faster IOs. We have a few workarounds for specific scenarios, but we think Microsoft should resolve this issue at the product level. We have a connect item open – https://connect.microsoft.com/SQLServer/feedback/details/744650/sql-server-io-handling-mechanism-can-be-severely-affected-by-high-cpu-usage - (with example scripts) to reproduce this behavior, please up vote the item so the issue will be addressed by the SQL Server product team soon.Thanks for your help and best regards,Ramesh MeyyappanHome: www.sqlworkshops.comLinkedIn: http://at.linkedin.com/in/rmeyyappan

    Read the article

  • Ask HTG: How Can I Check the Age of My Windows Installation?

    - by Jason Fitzpatrick
    Curious about when you installed Windows and how long you’ve been chugging along without a system refresh? Read on as we show you a simple way to see how long-in-the-tooth your Windows installation is. Dear How-To Geek, It feels like it has been forever since I installed Windows 7 and I’m starting to wonder if some of the performance issues I’m experiencing have something to do with how long ago it was installed. It isn’t crashing or anything horrible, mind you, it just feels slower than it used to and I’m wondering if I should reinstall it to wipe the slate clean. Is there a simple way to determine the original installation date of Windows on its host machine? Sincerely, Worried in Windows Although you only intended to ask one question, you actually asked two. Your direct question is an easy one to answer (how to check the Windows installation date). The indirect question is, however, a little trickier (if you need to reinstall Windows to get a performance boost). Let’s start off with the easy one: how to check your installation date. Windows includes a handy little application just for the purposes of pulling up system information like the installation date, among other things. Open the Start Menu and type cmd in the run box (or, alternatively, press WinKey+R to pull up the run dialog and enter the same command). At the command prompt, type systeminfo.exe Give the application a moment to run; it takes around 15-20 seconds to gather all the data. You’ll most likely need to scroll back up in the console window to find the section at the top that lists operating system stats. What you care about is Original Install Date: We’ve been running the machine we tested the command on since August 23 2009. For the curious, that’s one month and a day after the initial public release of Windows 7 (after we were done playing with early test releases and spent a month mucking around in the guts of Windows 7 to report on features and flaws, we ran a new clean installation and kept on trucking). Now, you might be asking yourself: Why haven’t they reinstalled Windows in all that time? Haven’t things slowed down? Haven’t they upgraded hardware? The truth of the matter is, in most cases there’s no need to completely wipe your computer and start from scratch to resolve issues with Windows and, if you don’t bog your system down with unnecessary and poorly written software, things keep humming along. In fact, we even migrated this machine from a traditional mechanical hard drive to a newer solid-state drive back in 2011. Even though we’ve tested piles of software since then, the machine is still rather clean because 99% of that testing happened in a virtual machine. That’s not just a trick for technology bloggers, either, virtualizing is a handy trick for anyone who wants to run a rock solid base OS and avoid the bog-down-and-then-refresh cycle that can plague a heavily used machine. So while it might be the case that you’ve been running Windows 7 for years and heavy software installation and use has bogged your system down to the point a refresh is in order, we’d strongly suggest reading over the following How-To Geek guides to see if you can’t wrangle the machine into shape without a total wipe (and, if you can’t, at least you’ll be in a better position to keep the refreshed machine light and zippy): HTG Explains: Do You Really Need to Regularly Reinstall Windows? PC Cleaning Apps are a Scam: Here’s Why (and How to Speed Up Your PC) The Best Tips for Speeding Up Your Windows PC Beginner Geek: How to Reinstall Windows on Your Computer Everything You Need to Know About Refreshing and Resetting Your Windows 8 PC Armed with a little knowledge, you too can keep a computer humming along until the next iteration of Windows comes along (and beyond) without the hassle of reinstalling Windows and all your apps.         

    Read the article

  • VirtualBox Clone Root HD / Ubuntu / Network issue

    - by john.graves(at)oracle.com
    When you clone a root Ubuntu disk in VirtualBox, one thing that gets messed up is the network card definition.  This is because Ubuntu (as it should) uses UDEV IDs for the network device.  When you boot your new disk, the network device ID has changed, so it creates a new eth1 device.  Unfortunately, this conflicts with the VirtualBox network setup.  What to do? Boot the box (no network) Edit the /etc/udev/rules.d/70-persistent-net.rules Delete the eth0 line and modify the eth1 line to be eth0 --------- Example OLD ----------- # This file was automatically generated by the /lib/udev/write_net_rules # program, run by the persistent-net-generator.rules rules file. # # You can modify it, as long as you keep each rule on a single # line, and change only the value of the NAME= key. # PCI device 0x8086:0x100e (e1000) <-------------------- Delete these two lines SUBSYSTEM=="net", ACTION=="add", DRIVERS=="?*", ATTR{address}=="08:00:27:d8:8d:15", ATTR{dev_id}=="0x0", ATTR{type}=="1", KERNEL=="eth*", NAME="eth0" # PCI device 0x8086:0x100e (e1000) ---Modify the next line and change eth1 to be eth0 SUBSYSTEM=="net", ACTION=="add", DRIVERS=="?*", ATTR{address}=="08:00:27:89:84:98", ATTR{dev_id}=="0x0", ATTR{type}=="1", KERNEL=="eth*", NAME="eth1" .csharpcode, .csharpcode pre { font-size: small; color: black; font-family: consolas, "Courier New", courier, monospace; background-color: #ffffff; /*white-space: pre;*/ } .csharpcode pre { margin: 0em; } .csharpcode .rem { color: #008000; } .csharpcode .kwrd { color: #0000ff; } .csharpcode .str { color: #006080; } .csharpcode .op { color: #0000c0; } .csharpcode .preproc { color: #cc6633; } .csharpcode .asp { background-color: #ffff00; } .csharpcode .html { color: #800000; } .csharpcode .attr { color: #ff0000; } .csharpcode .alt { background-color: #f4f4f4; width: 100%; margin: 0em; } .csharpcode .lnum { color: #606060; } ---------------------------------------- --------- Example NEW ----------- # This file was automatically generated by the /lib/udev/write_net_rules # program, run by the persistent-net-generator.rules rules file. # # You can modify it, as long as you keep each rule on a single # line, and change only the value of the NAME= key. # PCI device 0x8086:0x100e (e1000) SUBSYSTEM=="net", ACTION=="add", DRIVERS=="?*", ATTR{address}=="08:00:27:89:84:98", ATTR{dev_id}=="0x0", ATTR{type}=="1", KERNEL=="eth*", NAME="eth0" .csharpcode, .csharpcode pre { font-size: small; color: black; font-family: consolas, "Courier New", courier, monospace; background-color: #ffffff; /*white-space: pre;*/ } .csharpcode pre { margin: 0em; } .csharpcode .rem { color: #008000; } .csharpcode .kwrd { color: #0000ff; } .csharpcode .str { color: #006080; } .csharpcode .op { color: #0000c0; } .csharpcode .preproc { color: #cc6633; } .csharpcode .asp { background-color: #ffff00; } .csharpcode .html { color: #800000; } .csharpcode .attr { color: #ff0000; } .csharpcode .alt { background-color: #f4f4f4; width: 100%; margin: 0em; } .csharpcode .lnum { color: #606060; } ----------------------------------------

    Read the article

  • Failed to install GRUB on a separate '/boot' partition on a fake RAID 0 (12.04LTS)

    - by gerben
    I'm having some problems getting GRUB configured for Ubuntu 12.04LTS on a fake RAID 0. I can either get the GRUB rescue prompt at startup, or just a GRUB prompt but I cannot boot to Ubuntu manually. How can I configure the GRUB to actually use the Ubuntu install? The steps taken: Installing Ubuntu on fake raid The Ubuntu installer cannot install Ubuntu on the drive. After defining the partitions to use it fails with "Error: ???", pressing OK terminates the installer. Therefore, I used GParted to configure the partitions: /dev/mapper/sil_agadaccfacbg : (the RAID configuration, created partition): /dev/mapper/sil_agadaccfacbg1:ext2, 200MiB, (with 'boot' flag) /dev/mapper/sil_agadaccfacbg3:ext2, 67.75GiB, (which will contain Ubuntu) /dev/mapper/sil_agadaccfacbg2:extended, 1.00GiB, (for swap) Contains: /dev/mapper/sil_agadaccfacbg5: unknown Because of the fake-RAID, I already mounted the destination partitions before running the Ubuntu installer: > mkdir /mnt/boot > sudo mount /dev/mapper/sil_agadaccfacbg1 /mnt/boot > mkdir /mnt/ubuntu > sudo mount /dev/mapper/sil_agadaccfacbg3 /mnt/ubuntu In the installer I chose the following partition usage: /dev/mapper/sil_agadaccfacbg1 ext2, mount at /boot (209MB) /dev/mapper/sil_agadaccfacbg3 ext2, mount at / (72751MB) /dev/mapper/sil_agadaccfacbg5 swap Device for boot loader installation: /dev/mapper/sil_agadaccfacbg, linux device-mapper (striped) (74.0GB) This will install Ubuntu, but will fail to install GRUB (it seems to use /dev/sda no matter which one I choose) Installing GRUB with dpkg-reconfigure I followed this guide, but adapted it for two partitions: sudo mount /dev/mapper/sil_agadaccfacbg3 /mnt/ubuntu sudo mount --bind /dev /mnt/ubuntu/dev sudo mount --bind /proc /mnt/ubuntu/proc sudo mount --bind /sys /mnt/ubuntu/sys sudo mount /dev/mapper/sil_agadaccfacbg1 /mnt/boot sudo mount --bind /boot /mnt/boot sudo chroot /mnt/ubuntu dpkg-reconfigure grub-pc However, it does not ask where to install GRUB (I should choose /dev/mapper/sil_agadaccfacbg somewhere..) After reboot I get the GRUB rescue prompt with message no such device Installing GRUB with grub-install After the same mount commands as above, I continued with: > sudo grub-install --root-directory=/mnt/boot /dev/mapper/sil_agadaccfacbg This gives the following message: /usr/sbin/grub-probe: error: cannot find a device for /mnt/boot/boot/grub (is /dev mounted?) It does succeed when mounting just the boot partition : sudo mount /dev/mapper/sil_agadaccfacbg1 /mnt sudo grub-install --root-directory=/mnt/ /dev/mapper/sil_agadaccfacbg This finishes with: Installation finished. No error reported. After reboot I get the GRUB console, with welcome text. Attempting to manually start Ubuntu: ls (hd0) (hd0,msdos3) : (Ubuntu install partition) (hd0,msdos1) : (Ubuntu boot partition) (hd1) (hd1,msdos1) : (Ubuntu live USB) ls (hd0,msdos3)/ contains: - vmlinuz - lib/ - tmp/ - initrd.img - mnt/ - var/ - proc/ - boot/ - root/ - etc/ - run/ - media/ - sbin/ - bin/ - selinux/ - dev/ - srv/ - home/ - sys/ ls (hd0,msdos1)/ contains: -grub/ -boot/ -initrd.img-3.8.0-29-generic -vmlinuz-3.8.0.29-generic -config-3.8 linux (hd0,msdos3)/vmlinuz This returns "error: out of disk" Installing GRUB on Ubuntu partition with grub-install > sudo mount /dev/mapper/sil_agadaccfacbg3 /mnt > sudo grub-install --root-directory=/mnt/ /dev/mapper/sil_agadaccfacbg This finishes with message: > Installation finished. No error reported. After reboot get the message "error: out of disk" and the GRUB rescue prompt. Configuring GRUB with grub-mkconfig Attempting to run grub-mkconfig with different destinations results in the same message: /usr/sbin/grub-probe: error: cannot find a device for / (is /dev mounted?). Remarks: Initially I didn't use a separate /boot partition, but the GRUB install then also failed. Because some mention that a small partition at the beginning of the drive is necessary on old machines, I retried with a /boot partition This is a single boot (no other OS's installed/used)

    Read the article

  • SQL Saturday 27 (Portland, Oregon)

    - by BuckWoody
    I’m sitting in the Seattle airport, waiting for my flight to Silicon Valley California for the SQL Server 2008 R2 Launch Event. By some quirk of nature, they are asking me to Emcee the event – but that’s another post entirely.   I’m reflecting on the SQL Saturday 27 event that was just held in Portland, Oregon this last Saturday. These are not Microsoft-sponsored events – it’s truly the community at work. Think of a big user-group meeting – I mean REALLY big – held in a central location, like at a college (as ours was) or some larger, inexpensive venue like that. Everyone there is volunteering – it’s my own money and time to drive several hours to a hotel for the night, feed myself and present. It’s their own time and money for the folks that organize the event – unless a vendor or two steps in to help. It’s their own time and money for the attendees to drive a long way, spend the night and their Saturday to listen to the speakers. Why do all this?   Because everybody benefits. Every speaker learns something new, meets new people, and reaches a new audience. Every volunteer does the same. And the attendees? Well, it’s pretty obvious what they get. A 7Am to 10PM extravaganza of knowledge from every corner of the product. In fact, this year the Portland group hooked up with the CodeCamp folks and held a combined event. We had over 850 people, and I had everyone from data professionals to developers in my sessions.   So I’ll take this opportunity to do two things: to say “thank you” to all of the folks who attended, from those who spoke to those who worked and those who came to listen, and to challenge you to attend the next SQL Saturday anywhere near you. You can find the list here: http://www.sqlsaturday.com/. Don’t see anything in your area? Start one! The PASS folks have a package that will show you how. Sure, it’s a big job, but the key is to get as many people helping you as possible. Even if you have only a few dozen folks show up the first time, no worries. The first events I presented at had about 20 in the room. But not this week.   See you at the Launch Event if you’re near the San Francisco area tomorrow, and see you at the Redmond SQL Saturday and TechEd if not.   Share this post: email it! | bookmark it! | digg it! | reddit! | kick it! | live it!

    Read the article

  • Caching factory design

    - by max
    I have a factory class XFactory that creates objects of class X. Instances of X are very large, so the main purpose of the factory is to cache them, as transparently to the client code as possible. Objects of class X are immutable, so the following code seems reasonable: # module xfactory.py import x class XFactory: _registry = {} def get_x(self, arg1, arg2, use_cache = True): if use_cache: hash_id = hash((arg1, arg2)) if hash_id in _registry: return _registry[hash_id] obj = x.X(arg1, arg2) _registry[hash_id] = obj return obj # module x.py class X: # ... Is it a good pattern? (I know it's not the actual Factory Pattern.) Is there anything I should change? Now, I find that sometimes I want to cache X objects to disk. I'll use pickle for that purpose, and store as values in the _registry the filenames of the pickled objects instead of references to the objects. Of course, _registry itself would have to be stored persistently (perhaps in a pickle file of its own, in a text file, in a database, or simply by giving pickle files the filenames that contain hash_id). Except now the validity of the cached object depends not only on the parameters passed to get_x(), but also on the version of the code that created these objects. Strictly speaking, even a memory-cached object could become invalid if someone modifies x.py or any of its dependencies, and reloads it while the program is running. So far I ignored this danger since it seems unlikely for my application. But I certainly cannot ignore it when my objects are cached to persistent storage. What can I do? I suppose I could make the hash_id more robust by calculating hash of a tuple that contains arguments arg1 and arg2, as well as the filename and last modified date for x.py and every module and data file that it (recursively) depends on. To help delete cache files that won't ever be useful again, I'd add to the _registry the unhashed representation of the modified dates for each record. But even this solution isn't 100% safe since theoretically someone might load a module dynamically, and I wouldn't know about it from statically analyzing the source code. If I go all out and assume every file in the project is a dependency, the mechanism will still break if some module grabs data from an external website, etc.). In addition, the frequency of changes in x.py and its dependencies is quite high, leading to heavy cache invalidation. Thus, I figured I might as well give up some safety, and only invalidate the cache only when there is an obvious mismatch. This means that class X would have a class-level cache validation identifier that should be changed whenever the developer believes a change happened that should invalidate the cache. (With multiple developers, a separate invalidation identifier is required for each.) This identifier is hashed along with arg1 and arg2 and becomes part of the hash keys stored in _registry. Since developers may forget to update the validation identifier or not realize that they invalidated existing cache, it would seem better to add another validation mechanism: class X can have a method that returns all the known "traits" of X. For instance, if X is a table, I might add the names of all the columns. The hash calculation will include the traits as well. I can write this code, but I am afraid that I'm missing something important; and I'm also wondering if perhaps there's a framework or package that can do all of this stuff already. Ideally, I'd like to combine in-memory and disk-based caching.

    Read the article

  • Perfect End to a Bad Day

    - by TehGrumpyCoder
    Yesterday's post about A Bad Day at Work actually had an addendum to it. There were apparently a bunch of guys on ice skates last night competing in some sport way the hell and gone over on the other side of the valley, and enough people couldn't live without seeing them that they had all major arteries heading west honked. I mean honked... the traffic guy reported the 101 had 16 miles of backup... yikes. Since I worked downtown for a number of years, my fallback is to cut across the city on surface streets to get to one of my old 'haunts' and just drive it home from there. Of course with the 101 backed up, then I17 would logically be as well, so I kept the news on rather than my Zune and heard where the bad stuff was going North. I popped out on the freeway about 7 miles south of my exit. Got to the exit which is about a mile from the house without killing or maiming me or anyone else. Waited patiently at the light in the inside lane to make a left and go under the freeway proceeding West. The light changed, I had full green, I started through and whoa... I've got someone in a little rat car crossing my bow! A little explanation... I drive a 3/4 ton pickup with a V-10, extended cab and shell on the back. It's not jacked up, but it sits up pretty good and is longer than any parking place I've ever tried to put it into. I consider this truck to be the consolation prize for paying uninsured motorist coverage for 45 years and having Pilar Martinez totally destroy a 3/4 ton Silverado on March 1, 2007 by plowing into me at traffic speed while I was stopped at a light. If you pay for uninsured motorist coverage, ask your insurance agent *exactly* what that means... I bet it's different than what you think it means. But I digress, sorry... So here I am with a car that is shorter from top to road than the hood on my truck, and the driver thought it would be safe to run a red light and see if they could get past me before I got into the lane. The right side of my front bumper was almost into the driver's window when I hit the brakes and wheeled it left. Fortunately for all involved, I saw it soon enough, and pulled into the 2nd lane for making a left to go back South. I looked in my mirror, signalled a move, then moved over behind the yuck in the rat car. I then punched it, and the future hood ornament and I both made it through the next light. I pulled alongside to let her know that she was DEFINITELY Number 1 in my book, and it's a middle-age woman looking at me with a "sorry, it was an accident" show of pouty face and arms held up. Tough $hit lady... that may have worked when you were 18, but it's not working anymore, and it wasn't an accident... you ran a freakin' red light and almost got yourself killed. That just about put a bow on the day... I was home later than usual, pissed off about work stuff, pissed off at traffic, and now that. I ate dinner, watched a little TV, and was asleep about 9:30 exhausted. Hope today is better.

    Read the article

  • Dynamically load and call delegates based on source data

    - by makerofthings7
    Assume I have a stream of records that need to have some computation. Records will have a combination of these functions run Sum, Aggregate, Sum over the last 90 seconds, or ignore. A data record looks like this: Date;Data;ID Question Assuming that ID is an int of some kind, and that int corresponds to a matrix of some delegates to run, how should I use C# to dynamically build that launch map? I'm sure this idea exists... it is used in Windows Forms which has many delegates/events, most of which will never actually be invoked in a real application. The sample below includes a few delegates I want to run (sum, count, and print) but I don't know how to make the quantity of delegates fire based on the source data. (say print the evens, and sum the odds in this sample) using System; using System.Threading; using System.Collections.Generic; internal static class TestThreadpool { delegate int TestDelegate(int parameter); private static void Main() { try { // this approach works is void is returned. //ThreadPool.QueueUserWorkItem(new WaitCallback(PrintOut), "Hello"); int c = 0; int w = 0; ThreadPool.GetMaxThreads(out w, out c); bool rrr =ThreadPool.SetMinThreads(w, c); Console.WriteLine(rrr); // perhaps the above needs time to set up6 Thread.Sleep(1000); DateTime ttt = DateTime.UtcNow; TestDelegate d = new TestDelegate(PrintOut); List<IAsyncResult> arDict = new List<IAsyncResult>(); int count = 1000000; for (int i = 0; i < count; i++) { IAsyncResult ar = d.BeginInvoke(i, new AsyncCallback(Callback), d); arDict.Add(ar); } for (int i = 0; i < count; i++) { int result = d.EndInvoke(arDict[i]); } // Give the callback time to execute - otherwise the app // may terminate before it is called //Thread.Sleep(1000); var res = DateTime.UtcNow - ttt; Console.WriteLine("Main program done----- Total time --> " + res.TotalMilliseconds); } catch (Exception e) { Console.WriteLine(e); } Console.ReadKey(true); } static int PrintOut(int parameter) { // Console.WriteLine(Thread.CurrentThread.ManagedThreadId + " Delegate PRINTOUT waited and printed this:"+parameter); var tmp = parameter * parameter; return tmp; } static int Sum(int parameter) { Thread.Sleep(5000); // Pretend to do some math... maybe save a summary to disk on a separate thread return parameter; } static int Count(int parameter) { Thread.Sleep(5000); // Pretend to do some math... maybe save a summary to disk on a separate thread return parameter; } static void Callback(IAsyncResult ar) { TestDelegate d = (TestDelegate)ar.AsyncState; //Console.WriteLine("Callback is delayed and returned") ;//d.EndInvoke(ar)); } }

    Read the article

  • The best computer ever

    - by Jeff
    (This is a repost from my personal blog… wow… I need to write more technical stuff!) About three years and three months ago, I bought a 17" MacBook Pro, and it turned out to be the best computer I've ever owned. You might think that every computer with better specs is automatically better than the last, but that hasn't been my experience. My first one was a Sony, back in the Pentium III days, and it cost an astonishing $2,500. That was even more ridiculous in 1999 dollars. It had a dial-up modem, and a CD-ROM, built-in! It may have even played DVD's. A few years later I bought an HP, and it ended up being a pile of shit. The power connector inside came loose from the board, and on occasion would even short. In 2005, I bought a Dell, and it wasn't bad. It had a really high resolution screen (complete with dead pixels, a problem in those days), and it was the first laptop I felt I could do real work on. When 2006 rolled around, Apple started making computers with Intel CPU's, and I bought the very first one the week it came out. I used Boot Camp to run Windows. I still have it in its box somewhere, and I used it for three years. The current 17" was new in 2009. The goodness was largely rooted in having a big screen with lots of dots. This computer has been the source of hundreds of blog posts, tens of thousands of lines of code, video and photo editing, and of course, a whole lot of Web surfing. It connected to corpnet at Microsoft, WiFi in Hawaii and has presented many a deck. It has traveled with me tens of thousands of miles. Last year, I put a solid state drive in it, and it was like getting a new computer. I can boot up a Windows 7 VM in about 19 seconds. Having 8 gigs of RAM has always been fantastic. Everything about it has been fast and fun. When new, the battery (when not using VM's) could get as much as 10 hours. I can still do 7 without much trouble. After 460 charge cycles, the battery health is still between 85 and 90%. The only real negative has been the size and weight. It's only an inch thick, but naturally it's pretty big with a 17" screen. You don't get battery life like that without a huge battery, either, so it's heavy. It was never a deal breaker, but sometimes a long haul across a large airport, you know you're carrying it. Today, Apple announced a new, thinner and lighter 15" laptop, with twice the RAM and CPU cores, and four times the screen resolution. It basically handles my size and weight issues while retaining the resolution, and it still costs less than my 17" did. So I ordered one. Three years is an excellent run, but I kind of budgeted for a new workhorse this year anyway. So if you're interested in a 17" MacBook Pro with a Core 2 Duo 2.66 GHz CPU, 8 gigs of RAM and a 320 gig hard drive (sorry, I'm keeping the SSD), I have one to sell. They've apparently discontinued the 17", which is going to piss off the video community. It's in excellent condition, with a few minor scratches, but I take care of my stuff.

    Read the article

  • Video works with 'Try me' but not after install. What is the difference? U12.04LTS,

    - by HarveyP
    My hard drive got corrupted so I did a reinstall. Tested Youtube in FF during 'try me' and it worked - jerky, but it worked. Instal without all the updates (576 outstanding now) in order to get ff installed as per the demo - to no avail. In 'try me' mode ff NEVER crashed! After install ff crashed whilst I was typing in 'youtube' in the address field. When I finally got to youtube - no video. What is the difference between ff in try me and ff after install? Off to try some selected updates now to see if I can see it for myself. In previous installation I had several profiles and aliased ff with -safe-mode switch to simplify startup of most stable ff. Also found that ff startup in graphic mode worked better (but still without video) with all of the extensions disabled and all of the plugins set to "ask" and always denied ... I have SiS graphic card in SiS Motherboard for XP and ancient Hyundai ImageQuest QV770 monitor. I have Ubuntu 12.04.01 LTS 1 day after install with only the immediate upgrades requested to language pack (English UK). Using FR Alternative keyboard. Connected with domestic wifi network from Orange (FT) I really want to use Skype, but won't bother installing it (again) without video as I can do my sms on FB - whilst ff is not crashed ... Update ... Is something overflowing? I have just had to reboot in order to get ff to restart in any way shape or form - restart on crash form generates new crash form, etc. It was however a good half hour before it crashed so some improvement over conditions before disk corruption. I have now installed all of the critical updates (332 recommended updates still outstanding) which included some relating to ff. Still no video. Still crashing - especially when on Grepolis website. Since the re-install I have had a lovely 1024x768 screen, but after last ff crash and reboot I got a message about 'low graphics mode' and 'setting things myself'. I was not sufficiently tuned in at the time to take proper note - I have no doubt I shall see it again and shall report accordingly. I still have only laptop options for my screen and do not know how to rectify this. Spent a few days with ubuntu on a different, newer machine which has now suffered a graphics breakdown. Returned to this old one again, but with new flat screen Monitor. Found SIS drivers for my graphics BUT it is intended for Red Hat 7.2. I chose this over the version for 7.0 because I thought what the hell, I might not be able to do anything with either of them but this is the later one ... The file will not open with software manager - found a similar problem on Overclock but it has not helped me to install this driver. File name is sis_drv.o-410 and it is currently idling away in my Downloads folder ... I have tried the solution offered on another sis problem, but this shows that my xserver-xorg-video-sis driver is up to date. I am now at a loss as to how to proceed if I can't install the latest sis driver from sis ... Does nobody know how FF changes from "try-me" to "installed"? Any time I MUST have video I reboot from the disk again, but this is tedious! Also one of the things I mock most about MS is the constant rebooting ... UPDATE 10/6/2014 I have installed chromium-browser - worse, crashes even more often than ff.I have installed epiphany - better; Video works but not the associated soundtrack.FireFox is version 14.01 in 'try me' and version 29.0 from my install. Would it be useful to try to downgrade FireFox in order to get video?

    Read the article

  • Why is x=x++ undefined?

    - by ugoren
    It's undefined because the it modifies x twice between sequence points. The standard says it's undefined, therefore it's undefined. That much I know. But why? My understanding is that forbidding this allows compilers to optimize better. This could have made sense when C was invented, but now seems like a weak argument. If we were to reinvent C today, would we do it this way, or can it be done better? Or maybe there's a deeper problem, that makes it hard to define consistent rules for such expressions, so it's best to forbid them? So suppose we were to reinvent C today. I'd like to suggest simple rules for expressions such as x=x++, which seem to me to work better than the existing rules. I'd like to get your opinion on the suggested rules compared to the existing ones, or other suggestions. Suggested Rules: Between sequence points, order of evaluation is unspecified. Side effects take place immediately. There's no undefined behavior involved. Expressions evaluate to this value or that, but surely won't format your hard disk (strangely, I've never seen an implementation where x=x++ formats the hard disk). Example Expressions x=x++ - Well defined, doesn't change x. First, x is incremented (immediately when x++ is evaluated), then it's old value is stored in x. x++ + ++x - Increments x twice, evaluates to 2*x+2. Though either side may be evaluated first, the result is either x + (x+2) (left side first) or (x+1) + (x+1) (right side first). x = x + (x=3) - Unspecified, x set to either x+3 or 6. If the right side is evaluated first, it's x+3. It's also possible that x=3 is evaluated first, so it's 3+3. In either case, the x=3 assignment happens immediately when x=3 is evaluated, so the value stored is overwritten by the other assignment. x+=(x=3) - Well defined, sets x to 6. You could argue that this is just shorthand for the expression above. But I'd say that += must be executed after x=3, and not in two parts (read x, evaluate x=3, add and store new value). What's the Advantage? Some comments raised this good point. It's not that I'm after the pleasure of using x=x++ in my code. It's a strange and misleading expression. What I want is to be able to understand complicated expressions. Normally, a complicated expression is no more than the sum of its parts. If you understand the parts and the operators combining them, you can understand the whole. C's current behavior seems to deviate from this principle. One assignment plus another assignment suddenly doesn't make two assignments. Today, when I look at x=x++, I can't say what it does. With my suggested rules, I can, by simply examining its components and their relations.

    Read the article

  • College Ratings via the Federal Government

    - by user9147039
    A few weeks back you might remember news about a higher education rating system proposal from the Obama administration. As I've discussed previously, political and stakeholder pressures to improve outcomes and increase transparency are stronger than ever before. The executive branch proposal is intended to make progress in this area. Quoting from the proposal itself, "The ratings will be based upon such measures as: Access, such as percentage of students receiving Pell grants; Affordability, such as average tuition, scholarships, and loan debt; and Outcomes, such as graduation and transfer rates, graduate earnings, and advanced degrees of college graduates.” This is going to be quite complex, to say the least. Most notably, higher ed is not monolithic. From community and other 2-year colleges, to small private 4-year, to professional schools, to large public research institutions…the many walks of higher ed life are, well, many. Designing a ratings system that doesn't wind up with lots of unintended consequences and collateral damage will be difficult. At best you would end up potentially tarnishing the reputation of certain institutions that were actually performing well against the metrics and outcome measures that make sense in their "context" of education. At worst you could spend a lot of time and resources designing a system that would lose credibility with its "customers". A lot of institutions I work with already have in place systems like the one described above. They are tracking completion rates, completion timeframes, transfers to other institutions, job placement, and salary information. As I talk to these institutions there are several constants worth noting: • Deciding on which metrics to measure is complicated. While employment and salary data are relatively easy to track, qualitative measures are more difficult. How do you quantify the benefit to someone who studies in one field that may not compensate him or her as well as another field but that provides huge personal fulfillment and reward is a difficult measure to quantify? • The data is available but the systems to transform the data into actual information that can be used in meaningful ways are not. Too often in higher ed information is siloed. As such, much of the data that need to be a part of a comprehensive system sit in multiple organizations, oftentimes outside the reach of core IT. • Politics and culture are big barriers. One of the areas that my team and I spend a lot of time talking about with higher ed institutions all over the world is the imperative to optimize for student success. This, like the tracking of the students’ achievement after graduation, requires a level or organizational capacity that does not currently exist. The primary barrier is the culture of "data islands" in higher ed, and the need for leadership to drive out the divisions between departments, schools, colleges, etc. and institute academy-wide analytics and data stewardship initiatives that will enable student success. • Data quality is a very big issue. So many disparate systems exist (some on premise, some "in the cloud") that keep data about "persons" using different means to identify them. Establishing a single source of truth about an individual and his or her data is difficult without some type of data quality policy and tools. Good tools actually exist but are seldom leveraged. Don't misunderstand - I think it's a great idea to drive additional transparency and accountability into the system of higher education. And not just at home, but globally. Students and parents need access to key data to make informed, responsible choices. The tools exist to not only enable this kind of information to be shared but to capture the very metrics stakeholders care most about and in a way that makes sense in the context of a given institution's "place" in the overall higher ed panoply.

    Read the article

  • On a BPM Mission with Process Accelerators. Part 1: BPM as an ATV

    - by Cesare Rotundo
    Normal 0 false false false EN-US X-NONE X-NONE /* Style Definitions */ table.MsoNormalTable {mso-style-name:"Table Normal"; mso-tstyle-rowband-size:0; mso-tstyle-colband-size:0; mso-style-noshow:yes; mso-style-priority:99; mso-style-qformat:yes; mso-style-parent:""; mso-padding-alt:0in 5.4pt 0in 5.4pt; mso-para-margin-top:0in; mso-para-margin-right:0in; mso-para-margin-bottom:10.0pt; mso-para-margin-left:0in; line-height:115%; mso-pagination:widow-orphan; font-size:11.0pt; font-family:"Calibri","sans-serif"; mso-ascii-font-family:Calibri; mso-ascii-theme-font:minor-latin; mso-fareast-font-family:"Times New Roman"; mso-fareast-theme-font:minor-fareast; mso-hansi-font-family:Calibri; mso-hansi-theme-font:minor-latin; mso-bidi-font-family:"Times New Roman"; mso-bidi-theme-font:minor-bidi;} Part 1: BPM as an ATV It’s always exciting to talk to customers that are in the middle of a BPM transformational journey. Their thirst for new processes to improve with BPM makes them explorers in a landscape of opportunities. They have discovered that with BPM the can “go places” they couldn’t reach before. In a way, learning how to generate value with BPM is like adopting a new mean of transportation. Apps are like regular cars: very efficient, but to be used on paved roads: the road/process has been traced, and there are fixed paths to follow to get from “opportunity to quote” or from “quote to cash”. Getting off the road is risky, and laying down new asphalt is slow and expensive. Custom development is like running: you can go virtually anywhere, following any path you like, yet it’s slow, and a lot of sweat. BPM allows you to go “off the beaten path” laid out by packaged apps, yet make fast progress compared to custom development. BPM is therefore more like an All-Terrain Vehicle (ATV): less efficient than a car, but much faster than running, with a powerful enough engine that can get you places. The similarities between BPM and ATVs don’t stop here: you must learn to ride it even if you already know how to drive a car; you can reach places but figuring out the path to your destination is harder. Ultimately, with BPM as with an ATV, you reach places that you thought you could never reach, and you discover new destinations that provide great benefit to you … and that you didn’t even know existed! That’s where the sense of accomplishment that we heard from our BPM customers comes from, as well as the desire to share their experience, or even, as in the case of a County, the willingness to contribute their BPM solutions to help other agencies that face the same challenges. The question we wanted to answer is how can we teach organizations to drive ATV/BPM, thus leading them to deeper success with BPM, while increasing their awareness of the potential for reaching new targets, and finally equip them with the right tools. Like with ATVs, getting from point A to point B is more of a work of art than cruising on the highway by car. There is a lot we can do: after all many sought after destinations are common: someone else has been on the same path before. If only you could learn from their experience …

    Read the article

  • Eclipse Hell . . . Failed to read the project description file (.project)

    - by Kobojunkie
    I think Eclipse is trying to make me miserable. A couple of hours ago, my project was working and compiling well. Suddenly that all changed. Eclipse somehow wipes out all changes I have made to my files(activity, manifest etc.) I make sure to save often but when I go to run the project, I get the error that I have a build error. I checked and there was none, so I go to close Eclipse, so I can reopen and see if the errors will go away. Instead what happens is Eclipse wipes clean all my files and I end up with a project on disk with lots of blank code files. I try to run anyway, and I get the error message below. Failed to read the project description file (.project) for 'com.example.android.nfc.simulator.FakeTagsActivity.FakeTagsActivity'. The file has been changed on disk, and it now contains invalid information. The project will not function properly until the description file is restored to a valid state. Anyone have an idea what in the world this is about and how I can rectify this?

    Read the article

  • Caching with AVPlayer and AVAssetExportSession

    - by tba
    I would like to cache progressive-download videos using AVPlayer. How can I save an AVPlayer's item to disk? I'm trying to use AVAssetExportSession on the player's currentItem (which is fully loaded). This code is giving me "AVAssetExportSessionStatusFailed (The operation could not be completed)" : AVAsset *mediaAsset = self.player.currentItem.asset; AVAssetExportSession *es = [[AVAssetExportSession alloc] initWithAsset:mediaAsset presetName:AVAssetExportPresetLowQuality]; NSString *outPath = [NSTemporaryDirectory() stringByAppendingPathComponent:@"out.mp4"]; NSFileManager *fileManager = [NSFileManager defaultManager]; [fileManager removeItemAtPath:outPath error:NULL]; es.outputFileType = @"com.apple.quicktime-movie"; es.outputURL = [[[NSURL alloc] initFileURLWithPath:outPath] autorelease]; NSLog(@"exporting to %@",outPath); [es exportAsynchronouslyWithCompletionHandler:^{ NSString *status = @""; if( es.status == AVAssetExportSessionStatusUnknown ) status = @"AVAssetExportSessionStatusUnknown"; else if( es.status == AVAssetExportSessionStatusWaiting ) status = @"AVAssetExportSessionStatusWaiting"; else if( es.status == AVAssetExportSessionStatusExporting ) status = @"AVAssetExportSessionStatusExporting"; else if( es.status == AVAssetExportSessionStatusCompleted ) status = @"AVAssetExportSessionStatusCompleted"; else if( es.status == AVAssetExportSessionStatusFailed ) status = @"AVAssetExportSessionStatusFailed"; else if( es.status == AVAssetExportSessionStatusCancelled ) status = @"AVAssetExportSessionStatusCancelled"; NSLog(@"done exporting to %@ status %d = %@ (%@)",outPath,es.status, status,[[es error] localizedDescription]); }]; How can I export successfully? I'm looking into copying mediaAsset into an AVMutableComposition, but haven't had much luck with that either. Thanks! PS: Here are some questions from people trying to accomplish the same thing (but with MPMoviePlayerController): Cache Progressive downloaded content in MPMoviePlayerController Simultaneously stream and save a video? Caching videos to disk after successful preload by MPMoviePlayerController

    Read the article

  • Problems with jQuery getJSON using local files in Chrome

    - by Tauren
    I have a very simple test page that uses XHR requests with jQuery's $.getJSON and $.ajax methods. The same page works in some situations and not in others. Specificially, it doesn't work in Chrome on Ubuntu. I'm testing on Ubuntu 9.10 with Chrome 5.0.342.7 beta and Mac OSX 10.6.2 with Chrome 5.0.307.9 beta. It works correctly when files are installed on a web server from both Ubuntu/Chrome and Mac/Chrome (try it out here). It works correctly when files are installed on local hard drive in Mac/Chrome (accessed with file:///...). It FAILS when files are installed on local hard drive in Ubuntu/Chrome (access with file:///...). The small set of 3 files can be downloaded in a tar/gzip file from here: http://issues.tauren.com/testjson/testjson.tgz When it works, the Chrome console will say: XHR finished loading: "http://issues.tauren.com/testjson/data.json". index.html:16Using getJSON index.html:21 Object result: "success" __proto__: Object index.html:22success XHR finished loading: "http://issues.tauren.com/testjson/data.json". index.html:29Using ajax with json dataType index.html:34 Object result: "success" __proto__: Object index.html:35success XHR finished loading: "http://issues.tauren.com/testjson/data.json". index.html:46Using ajax with text dataType index.html:51{"result":"success"} index.html:52undefined When it doesn't work, the Chrome console will show this: index.html:16Using getJSON index.html:21null index.html:22Uncaught TypeError: Cannot read property 'result' of null index.html:29Using ajax with json dataType index.html:34null index.html:35Uncaught TypeError: Cannot read property 'result' of null index.html:46Using ajax with text dataType index.html:51 index.html:52undefined Notice that it doesn't even show the XHR requests, although the success handler is run. I swear this was working previously in Ubuntu/Chrome, and am worried something got messed up. I already uninstalled and reinstalled Chrome, but that didn't help. Can someone try it out locally on your Ubuntu system and tell me if you have any troubles? Note that it seems to be working fine in Firefox.

    Read the article

  • Hive Based Registry in Flash

    - by Psychic
    To start with I'll say I've read the post here and I'm still having trouble. I'm trying to create a CE6 image with a hive-based registry that actually stores results through a reboot. I've ticked the hive settings in the catalog items. In common.reg, I've set the location of the hive ([HKEY_LOCAL_MACHINE\init\BootVars] "SystemHive") to "Hard Drive\Registry" (Note: the flash shows up as a device called "Hard Drive") In common.reg, I've set "Flags"=dword:3 in the same place to get the device manager loaded along with the storage manager I've verified that these settings are wrapped in "; HIVE BOOT SECTION" This is where it starts to fall over. It all compiles fine, but on the target system, when it boots, I get: A directory, called "Hard Disk" where a registry is put A device, name called "Hard Disk2" where the permanent flash is Any changes made to the registry are lost on a reboot What am I still missing? Why is the registry not being stored on the flash? Strangly, if I create a random file/directory in the registry directory, it is still there after a reboot, so even though this directory isn't on the other partition (where I tried to put it), it does appear to be permanent. If it is permanent, why don't registry settings save (ie Ethernet adapter IP addresses?) I'm not using any specific profiles, so I'm at a loss as to what the last step is to make this hive registry a permanent store.

    Read the article

  • Android SQLite database gets corrupted

    - by Seu
    There are about 100 people using my Android App right now and every once and while I get a crash report to the server with this stack trace: android.database.sqlite.SQLiteDatabaseCorruptException: database disk image is malformed at android.app.ActivityThread.performLaunchActivity(ActivityThread.java:2596) at android.app.ActivityThread.handleLaunchActivity(ActivityThread.java:2621) at android.app.ActivityThread.access$2200(ActivityThread.java:126) at android.app.ActivityThread$H.handleMessage(ActivityThread.java:1932) at android.os.Handler.dispatchMessage(Handler.java:99) at android.os.Looper.loop(Looper.java:123) at android.app.ActivityThread.main(ActivityThread.java:4595) at java.lang.reflect.Method.invokeNative(Native Method) at java.lang.reflect.Method.invoke(Method.java:521) at com.android.internal.os.ZygoteInit$MethodAndArgsCaller.run(ZygoteInit.java:860) at com.android.internal.os.ZygoteInit.main(ZygoteInit.java:618) at dalvik.system.NativeStart.main(Native Method) Caused by: android.database.sqlite.SQLiteDatabaseCorruptException: database disk image is malformed at android.database.sqlite.SQLiteQuery.native_fill_window(Native Method) at android.database.sqlite.SQLiteQuery.fillWindow(SQLiteQuery.java:75) at android.database.sqlite.SQLiteCursor.fillWindow(SQLiteCursor.java:295) at android.database.sqlite.SQLiteCursor.getCount(SQLiteCursor.java:276) at android.database.AbstractCursor.moveToPosition(AbstractCursor.java:171) at android.database.AbstractCursor.moveToFirst(AbstractCursor.java:248) The result is the app crashing and all the data in the DB being lost. One thing to note is that every time I read or write to the database I get a new SQLiteDatabase and close it as soon as I'm done. I thought this would simplify things, but perhaps that's causing the problem? Is it possible this is just a SQLite bug?

    Read the article

  • System.InvalidOperationException: Failed to map the path '/sharedDrive/Public'

    - by d03boy
    I'm trying to set up a page that will allow users to download files from a shared drive (where the file is actually sent via the page). So far this is what I have. public partial class TestPage : System.Web.UI.Page { protected DirectoryInfo dir; protected FileInfo[] files; protected void Page_Load(object sender, EventArgs e) { dir = new DirectoryInfo(Server.MapPath(@"\\sharedDrive\Public")); files = dir.GetFiles(); } } The aspx page looks kind of like this: <% Response.Write(System.Security.Principal.WindowsIdentity.GetCurrent().Name); %> <ul> <% foreach (System.IO.FileInfo f in files) { Response.Write("<li>" + f.FullName + "</li>"); } %> </ul> When I remove the erroneous parts of the code, the page tells me that the Windows Identity I'm using is my user (which has access to the drive). I don't understand what the problem could be or what it's even complaining about.

    Read the article

  • Recover files from corrupt filesystem

    - by Emile 81
    My situation: I have an older 80GB IDE internal hdd, with a few files on that I would like very much to recover: some word documents some latex documents (text files) and pictures (png, jpg, eps files) some other text documents and visual studio project files I had backed them (not the latex ones though) up using svn, but have not committed lately, and would loose a lot of work if I cant recover. the hdd seems to have lost its filesystem, i have no idea how it came about. I know it has/had 3 NTFS partitions, i know the files i want are on the second or third partition. I read http://superuser.com/questions/81877/recover-hard-disk-data Partition Find and Mount did not see all the partitions using intelligent scan TestDisk does (i think), I followed the step by step instructions here, but when I try to list the files it says: "Can't open filesystem, filesystem seems damaged." I'm not sure how to proceed here, as TestDisks wiki does not contain this error message afaik. I don't know if the hdd is gonna fail, or some prog has caused the filesystem to be corrupt, the hdd doesnt make a sound, so i guess that's good. I would like some guidance so I don't accidentally cause more damage. (eg. is it ok to let testdisk write the filesystem to disk? I'm pretty the partitions are listed ok, but not 100%)

    Read the article

  • Multiple dex files define Lcom/google/api/client/auth/oauth/AbstractOAuthGetToken;

    - by Elad Benda
    I have just followed this tutorial: https://developers.google.com/drive/quickstart-android so I don't see a reason for duplicated libs in my project. I have added the drive Client lib via Google plugin for eclipse When I build my android app with this manifest <uses-sdk android:minSdkVersion="15" android:targetSdkVersion="16" /> <uses-permission android:name="android.permission.READ_CALENDAR" /> <uses-permission android:name="android.permission.WRITE_CALENDAR" /> <uses-permission android:name="android.permission.READ_EXTERNAL_STORAGE"/> <uses-permission android:name="android.permission.GET_ACCOUNTS"/> <uses-permission android:name="android.permission.INTERNET" /> <application android:icon="@drawable/todo" android:label="@string/app_name" > <activity android:name=".TodosOverviewActivity" android:label="@string/app_name" > <intent-filter> <action android:name="android.intent.action.MAIN" /> <category android:name="android.intent.category.LAUNCHER" /> </intent-filter> </activity> <activity android:name=".TodoDetailActivity" android:windowSoftInputMode="stateVisible|adjustResize" > <intent-filter> <action android:name="android.intent.action.SEND" /> <category android:name="android.intent.category.DEFAULT" /> <data android:mimeType="image/*" /> </intent-filter> </activity> <provider android:name=".contentprovider.MyTodoContentProvider" android:authorities="de.vogella.android.todos.contentprovider" > </provider> </application> I get the following error: [2013-10-27 00:43:58 - Dex Loader] Unable to execute dex: Multiple dex files define Lcom/google/api/client/auth/oauth/AbstractOAuthGetToken; [2013-10-27 00:43:58 - de.vogella.android.todos] Conversion to Dalvik format failed: Unable to execute dex: Multiple dex files define Lcom/google/api/client/auth/oauth/AbstractOAuthGetToken; how can I fix this?

    Read the article

  • Deployed Qt5 Application Doesn't Print or Show Print Dialog

    - by MustacheMcLimey
    I'm experiencing Qt4 to Qt5 troubles. In my application when the user clicks the print button two things should happen, one is that a PDF gets written to disk (which still works fine in the new version, so I know that some of the printing functions are working properly) and the other is that a QPrintDialog should exec() and then send to a connected printer. I see the dialog when I launch from my development machine. The application launches on the deployed machine, but the QPrintDialog never shows and the document never prints. I am including print support. QT += core gui network webkitwidgets widgets printsupport I have been using Process Explorer to see what DLLs the application uses on my development machine, and I believe that everything is present. My application bundle includes: {myAppPath}\MyApp[MyApp.exe, Qt5PrintSupport.dll, ...] {myAppPath}\plugins\printsupport\windowsprintersupport.dll {myAppPath}\plugins\imageformats[ qgif.dll, qico.dll,qjpeg.dll, qmng.dll, qtga.dll, qtiff.dll, qwbmp.dll ] The following is the relevant code snippet: void PrintableForm::printFile() { //Writes the PDF to disk in every environment pdfCopy(); //Paper Copy only works on my dev machine QPrinter paperPrinter; QPrintDialog printDialog(&paperPrinter,this); if( printDialog.exec() == QDialog::Accepted ) { view->print(&paperPrinter); } this->accept(); } My first thought is that the relevant DLLs are not being found come print time, and that means that my application file system is incorrect, but I have not found anything that shows me a different file structure. Am I on the right track or is there something else wrong with this setup?

    Read the article

  • XenServer 5.5 local storage problem

    - by Jason Nerer
    Hi community, I have the following problem with a Citrix XenServer 5.5. I had to physically move the host, so I shut down all machines via console: xe vm-shutdown force=true vm=my-machine-uuis-s After that I shut down the machine itself by issuing: halt After the reboot today the local storage repository is unplugged. I was trying to repair it via XenCenter, but I don't trust this one. So I tried: [root@xenserver ~]# xe pbd-list uuid ( RO) : ef6e2f3b-5825-393c-23e1-391d105c87ec host-uuid ( RO): c4bcf09c-2e52-448f-8210-df5d13bd33a9 sr-uuid ( RO): 2fb3be9c-075c-53ed-acb6-42f0c4ad0614 device-config (MRO): device: /dev/disk/by-id/scsi-SATA_WDC_WD5001ABYS-_WD-WCAS83698154,/dev/disk/by-id/scsi-SATA_WDC_WD5001ABYS-_WD-WCAS83694262 currently-attached ( RO): false To reattach the storage I issued: xe pbd-plug uuid=ef6e2f3b-5825-393c-23e1-391d105c87ec That one is running now for a while but not talking to me. The local repo has around 1TB. Should I wait, or are there any other options to reattach the local repo? What could have caused this problem? Any ideas? Thx. J

    Read the article

  • Unable to fetch initial output of "defrag" commad in Windows Server 2008 R2 in WOW64 environment.

    - by Ganesh
    Hi All, [Application & code back ground] I have an MFC application which is executing on Windows Server 2008 R2 in WOW64 environment. In which on user input it defragments the selected drive on the disk. I initiated the process(.i.e. cmd /c defrag –v c:) of defragmentation using the CreateProcess() API, along with this to display output of the process on the main screen I created the pipe using CreatePipe() API. I used PeekNamedPipe() & ReadFile() API to get the output and display. [Problem Area] When the process is launched I am not getting the initial output as below: Microsoft Disk fragmenter Copyright © 2007 Microsoft Crop. Invoking defragmentation on (C:)…. I constantly monitor the output of the process while it is under progress but not able to get any thing as output in the pipe. It seem the process is not doing any thing and appears as if the application is not responding. But after certain period of time, once the process is about to completed I get result along with the initial data. [Sample code] //Pipe created if(0 == ::CreatePipe(&l_hStdOutRead, &l_hStdOutWrite, &l_SecurityAttribute, (DWORD)NULL)) { //Error code } //Process created/launched if (0 == ::CreateProcess(NULL, (LPTSTR)f_csProcessName, &l_stSecurityAttributes, NULL, TRUE, CREATE_NO_WINDOW, NULL,NULL, &l_StartupInfo, &l_CmdPI)) { //Error code } //Read output if (0 == ::PeekNamedPipe(m_hStdOutRead, l_cArrPeekBuffer, (DWORD)NULL, (LPDWORD)NULL, &l_dwAvailable, (LPDWORD)NULL)) { //Return to read again } if (MPLUSFALSE == ::ReadFile(m_hStdOutRead, l_cArrOutput, min(BUFFER_SIZE - 2, l_dwAvailable), &l_dwRead, NULL) || !l_dwRead) { //error code } //Display data. If any one is aware of similar problem or worked on the similar issue please let me know the solution. Thanks in Advance, Ganesh

    Read the article

  • C# File IO with Streams - Best Memory Buffer Size

    - by AJ
    Hi, I am writing a small IO library to assist with a larger (hobby) project. A part of this library performs various functions on a file, which is read / written via the FileStream object. On each StreamReader.Read(...) pass, I fire off an event which will be used in the main app to display progress information. The processing that goes on in the loop is vaired, but is not too time consuming (it could just be a simple file copy, for example, or may involve encryption...). My main question is: What is the best memory buffer size to use? Thinking about physical disk layouts, I could pick 2k, which would cover a CD sector size and is a nice multiple of a 512 byte hard disk sector. Higher up the abstraction tree, you could go for a larger buffer which could read an entire FAT cluster at a time. I realise with today's PC's, I could go for a more memory hungry option (a couple of MiB, for example), but then I increase the time between UI updates and the user perceives a less responsive app. As an aside, I'm eventually hoping to provide a similar interface to files hosted on FTP / HTTP servers (over a local network / fastish DSL). What would be the best memory buffer size for those (again, a "best-case" tradeoff between perceived responsiveness vs. performance). Thanks in advance for any ideas, Adam

    Read the article

< Previous Page | 368 369 370 371 372 373 374 375 376 377 378 379  | Next Page >