Search Results

Search found 14910 results on 597 pages for 'programs and features'.

Page 22/597 | < Previous Page | 18 19 20 21 22 23 24 25 26 27 28 29  | Next Page >

  • How do I get an application to appear as a choice in update-alternatives?

    - by Jay
    I separately installed the Firefox Beta and Alpha channels, and have desktop configuration files pointing to them in ~/.local/share/applications. However, stable Firefox is being used as my default browser by the system. (Firefox Beta used to be used until I messed with the "Default Applications" in System Settings, where it is not listed.) I tried running sudo update-alternatives --config x-www-browser to manually change it, but it's only recognizing Chromium and Firefox (stable) and showing them as a choice. What can I do to get custom desktop configuration files in ~/.local/share/applications to be seen as default alternatives? I think I may have to fiddle with the desktop config files, or with mimeinfo.cache or mimeapps.list? Running Oneiric. Here is the content of the firefox-beta.desktop file I created: [Desktop Entry] Name=Firefox Beta Exec=firefox-beta -P Beta -no-remote Icon=firefox Terminal=false X-MultipleArgs=false Type=Application StartupNotify=true StartupWMClass=Firefox Categories=GNOME;GTK;Network;WebBrowser; Comment[en_US]=Firefox Beta Channel MimeType=text/html;text/xml;application/xhtml+xml;application/xml;application/vnd.mozilla.xul+xml;application/rss+xml;application/rdf+xml;image/gif;image/jpeg;image/png;x-scheme-handler/http;x-scheme-handler/https;x-scheme-handler/ftp;x-scheme-handler/chrome;video/webm; Name[en_US]=Firefox Beta [NewWindow Shortcut Group] Name=Open a New Window Exec=firefox-beta -new-window about:blank TargetEnvironment=Unity

    Read the article

  • What one feature available in other IDEs should be added to Xcode? [closed]

    - by Graham Lee
    This is inspired by Which features from other IDEs/editors you wish you have in Visual Studio? Xcode is a very different tool from Visual Studio, with a different feature set. While some of its capabilities are very mature (it had RAD UI layout in Interface Builder since before most other platforms), it lacks some features that e.g. Visual Studio or Eclipse provide. If you could request one feature to be added to Xcode, which would it be? How would that feature help you write better code, or write the same code faster?

    Read the article

  • The report belongs to a package that is not installed

    - by user71700
    I have gotten the error message "The report belongs to a package that is not installed." Generally I would just ignore that since seemingly there is no problem except I got a crash report and then I said to report the problem and then I get this. Now, how can a package or program crash that is not even installed? Sounds a little paradox, isn't it? What should I do with this? Why does this even come up?

    Read the article

  • Know Thy Operating System?

    - by AdityaGameProgrammer
    As developers how much time, or do you spend time, In learning the hidden features tricks of your operating system ? How important do you feel is this for productivity in day to day programming? tasks. What do you mean when you list knowledge of an OS in your resume? What are your favorite hidden -less known features For example: A common problem of How can i open the cmd window in a specific location a do it yourself solution in say xp and what to do if something breaks Are these something you look into as and when you find the need to do so?

    Read the article

  • How to clean and add options to the Open With list of apps

    - by Luis Alvarado
    After installing several PPAs (Wine, PoL) and opening several files with other apps (Like changing from Totem to VLC) I discovered that the Open With option had 2 problems: Many items on the list are duplicated (As seen on the image for "A Wine Program") Sometimes the app I want to use to open is not shown there (For example, Virtualbox or VLC) So how can I edit this list to clean the duplicates and add missing apps from the list.

    Read the article

  • NetBeans 7.2 RC1 is published

    - by Ondrej Brejla
    NetBeans 7.2 RC1 was today published. You can download it here. You could read about the PHP features added to the NetBeans 7.2 release here on the blog, but the main features added or improved are: Support for PHP 5.4 PHP editing: Fix Uses action, annotations support, editing of Neon and Apache Config files and more Support for Symfony2, Doctrine2 and ApiGen frameworks FTP remote synchronization Support for running PHP projects on Hudson For more information, just look at New and Noteworthy page for NetBeans 7.2. And as obvious you can help us to test the build. Just try it and if you find an issue / error, please report it. Thanks for your help.

    Read the article

  • NetBeans 7.3 Beta2 is Out!

    - by Ondrej Brejla
    NetBeans 7.3 Beta2 was published today. You can download it. You could read about the PHP features added to the NetBeans 7.3 release here on the blog, but the main features added or improved are: Parsers for Namespaced Annotations (Symfony 2, Doctrine 2, etc.), Basic Composer Integration (Dependency Manager for PHP), Twig Code Completion (with documentation), Smarty Braces Matching for Related Tags, Smarty Parser Errors of Unmatched Tags. As obvious you can help us to test the build. Just try it and if you find an issue / error, please report it. Thanks for your help.

    Read the article

  • Changing default application

    - by ragebunny
    Hey i have recently installed Sublime text 2 and i must say it's one of the best text editors i've ever used or even seen in use. So i managed to install everything and added to the unity menu and also change all the entries of gedit.desktop to sublime.desktop and it works just fine for most files but i realised that some files still open in gedit, for example php files, i checked the default list for php but there isn't anything in there. So how would i set the default for opening php files? Update:Also Sublime doesn't show up in them properties menu where i can usually select the default program. Thanks for the help. I'm using 12.10.

    Read the article

  • ubuntu 12.04 how do I select a custom program to open specific file type

    - by user68477
    I want to set a filetype to be opened by a specific program other than ubuntu's default. I usually right click properties open with and select the application I need. This however does not work in this case. The application I want is not shown in the "open with" dialog and there apparently is no way I can browse to it or type in a custom command like I used to in 10.04. Is this a bug that I should file or a feature that can be worked around?

    Read the article

  • Set default app for opening an extension [closed]

    - by Primož Kralj
    I am using Linux Mint 13 with Cinnamon. I have a problem setting an application to open a file with some extension when the application is actually only extracted, not really installed using apt-ger or .deb file. For example - Blender. You can download a .tar.gz from official site which you then extract and you can just run the app. However, it isn't listed in the default apps list (Right Clickon file - Open with other application - Show all installed apps). So now I have to always open the application first and open the file from there. How can I solve this?

    Read the article

  • Stardict config is not working

    - by terces907
    I set Stardict autostart when poweron with option -h (hidden mode) and set scanning key to Alt+Ctrl (translates word if select text and press Alt+Ctrl). My problem is sometimes Stardict's config is not working (Alt+Ctrl is not working seem like program didn't read config). I checked on "~/.stardict/stardict.cfg" every thing it look OK. A Problem seems like Stardict program had ran before read a config file.

    Read the article

  • How do I set a wine program (ex. Foxit Reader for Windows) as the default program?

    - by To Do
    I regularly annotate pdf files and unfortunately there is no good linux pdf reader that supports decent annotations. Evince has a very rudimentary and buggy annotation feature. So I'm stuck using a Windows viewer through wine. This works pretty well but, when I simply right-click a file (in this case a pdf), properties, open with and selected Foxit Reader, the Unity Launcher icon remained the wine icon instead of the application icon. Has anyone set a wine program as the default program for any file? Any ideas?

    Read the article

  • Downloading anything is hell and arduous without an application. Care to help out an Ubunt-noob?

    - by FilthyCasual
    Torrent from the Pirate Bay even though I have Tixati installed and working? Asks for an application which I don't have (apturl or something similar?) Trying to get the latest AMD Radeon drivers automatically installed? the main file downloads...but just shows up in a window full of files but idk where to drag them. And PlayOnLinux my one hope for gaming since Steam is buggin out without proper drivers? Can't figure out where to "mount" the DVD version it asks for. I'm so lost. I dove head first into the Linux thing thinking it was wildly superior, and maybe it is, but dang if the simplest things aren't hard to do. Any help appreciated. I've been reading guides and manuals but what I need is human experience most of all.

    Read the article

  • Forcing programs to be installed to another drive

    - by zyboxenterprises
    I have an SSD as my main Windows drive, with a 640GB 2.5" HDD, partitioned to store programs and user settings, and also to act as backup (it's the only thing I had lying around at the time of building my PC). The task was to make the PC as fast as possible, while having an increased storage capacity available to store normal user data, and to assist in my small data recovery business. The problem is that whenever I install a program, it installs to C:\Program Files [(x86 for the 32 bit programs]\, although I have changed the environment variables. This wouldn't normally be an issue, however every installation program points its shortcut to my 640GB HDD. The root layout of both drives: To clarify: Program files get installed to C:\ Program shortcuts are always pointed to Z:\, my 640GB HDD Modifying the relevant environment variables doesn't do anything, I looked at this, but however it only talks about modifying the registry and environment variables, which I have already done so. I install to the Z:\ drive if the installation program lets me change the installation path, but however the installation programs sometimes don't let me change this. Is there a way that I can force every program to install to the relevant location on Z:\? Perhaps I'm missing something here? Edit: Found this program; would it be appropriate to use in my case? I would be able to move the entire Program Files (and its x86 version) to Z:\, without impacting on the performance.

    Read the article

  • Starting programs from terminal then exiting terminal exits started programs?

    - by sherrellbc
    I really was unsure how to phrase the question title. What I mean is that when I use the terminal to start a program, most of the time when the terminal is closed it also exits the programs started from it. Now this makes sense if we look at it from a hierarchical standpoint of the terminal being the parent process which spawns child processes, and any halt of the parent causes subsequent halting of the children as well. However, I've noticed this to not always be the case. For example, I downloaded Sublime Text Editor and created a symlink in PATH. I can start this program by issuing a sublime command from the terminal, but subsequent closure of the terminal program does nothing to sublime. However, other times either the child process that was started it also closed or it hangs up and causes problems. tl;dr: Is it always the case that programs started from a closed parent process will be closed when the parent is exited? And if so, is there way to start a program from the terminal and then close the terminal without exiting the started process? The whole point here is to start programs from the terminal so I do not overly-populate my desktop with symlinks.

    Read the article

  • Learn To Adjust Contrast Like a Pro in Photoshop, GIMP, and Paint.NET

    - by Eric Z Goodnight
    Brightness and Contrast tools are for beginners! Ever wondered what graphics programs offer advanced users to ensure their photographs have a great value range? Read on to learn about Levels, Curves, and Histograms in three major programs. Curves and Levels are not as intuitive as the more basic Brightness and Contrast sliders Photoshop, GIMP, and Paint.NET all share. However, they offer a great deal more control over images that professionals and skilled image editors will demand. Combine these tools with a knowledge of how basic histograms work, and you’ll be well on your way to editing contrast like a pro! Latest Features How-To Geek ETC Learn To Adjust Contrast Like a Pro in Photoshop, GIMP, and Paint.NET Have You Ever Wondered How Your Operating System Got Its Name? Should You Delete Windows 7 Service Pack Backup Files to Save Space? What Can Super Mario Teach Us About Graphics Technology? Windows 7 Service Pack 1 is Released: But Should You Install It? How To Make Hundreds of Complex Photo Edits in Seconds With Photoshop Actions Add a “Textmate Style” Lightweight Text Editor with Dropbox Syncing to Chrome and Iron Is the Forcefield Really On or Not? [Star Wars Parody Video] Google Updates Picasa Web Albums; Emphasis on Sharing and Showcasing Uwall.tv Turns YouTube into a Video Jukebox Early Morning Sunrise at the Beach Wallpaper Data Networks Visualized via Light Paintings [Video]

    Read the article

  • Printer features don't work when printing to Canon Printers ir5185 and ir7095 in Snow Leopard

    - by Ken
    Recently updated to iMac running Snow Leopard. Connected Canon printers ir7095 and ir5185 via ethernet and downloaded latest drivers from Canon website. Can print to both from InDesign CS3, however, when I select printer features such as heavy paper and printing to stack bypass, it prints but just defaults to plain paper in drawer 1. Also, when printing to ir7095, 0001 prints five times on sheet in background. Is there any way to get the printer features that are available to work?

    Read the article

  • OneNote 2007 - recommendation(s) of a good place/tutorials to learn features

    - by studiohack23
    I'm pretty familiar with Office 2007, however, I have just recently acquired OneNote 2007. It is such a big and powerful tool, that I'm pretty much lost on the features and how to use it. I don't really know where to start. I'm looking for some recommendation(s) on a good place to learn more about OneNote and its features and what it does...for what it's worth, I'm a student, so student perspectives on how to use/learn OneNote would be awesome! Thanks!

    Read the article

  • Xen guest miration to host with missing features

    - by deploymonkey
    If I want to move a xen guet (domU) from one host (dom0) to another host on another hardware platform which misses some capabilities, say virtualisation features, especially directio or likewise, will my image be able to run despite missing capabilities of the new host? This is important because I need to know if I can prepare XEN images on my workstation with full virtualisation features and deploy same images on less capable Servers later on. Thanks for the help and input

    Read the article

  • How to Sync Any Folder With SkyDrive on Windows 8.1

    - by Chris Hoffman
    Before Windows 8.1, it was possible to sync any folder on your computer with SkyDrive using symbolic links. This method no longer works now that SkyDrive is baked into Windows 8.1, but there are other tricks you can use. Creating a symbolic link or directory junction inside your SkyDrive folder will give you an empty folder in your SkyDrive cloud storage. Confusingly, the files will appear inside the SkyDrive Modern app as if they were being synced, but they aren’t. The Solution With SkyDrive refusing to understand and accept symbolic links in its own folder, the best option is probably to use symbolic links anyway — but in reverse. For example, let’s say you have a program that automatically saves important data to a folder anywhere on your hard drive — whether it’s C:\Users\USER\Documents\, C:\Program\Data, or anywhere else. Rather than trying to trick SkyDrive into understanding a symbolic link, we could instead move the actual folder itself to SkyDrive and then use a symbolic link at the folder’s original location to trick the original program. This may not work for every single program out there. But it will likely work for most programs, which use standard Windows API calls to access folders and save files. We’re just flipping the old solution here — we can’t trick SkyDrive anymore, so let’s try to trick other programs instead. Moving a Folder and Creating a Symbolic Link First, ensure no program is using the external folder. For example, if it’s a program data or settings folder, close the program that’s using the folder. Next, simply move the folder to your SkyDrive folder. Right-click the external folder, select Cut, go to the SkyDrive folder, right-click and select Paste. The folder will now be located in the SkyDrive folder itself, so it will sync normally. Next, open a Command Prompt window as Administrator. Right-click the Start button on the taskbar or press Windows Key + X and select Command Prompt (Administrator) to open it. Run the following command to create a symbolic link at the original location of the folder: mklink /d “C:\Original\Folder\Location” “C:\Users\NAME\SkyDrive\FOLDERNAME\” Enter the correct paths for the exact location of the original folder and the current location of the folder in your SkyDrive. Windows will then create a symbolic link at the folder’s original location. Most programs should hopefully be tricked by this symbolic location, saving their files directly to SkyDrive. You can test this yourself. Put a file into the folder at its original location. It will be saved to SkyDrive and sync normally, appearing in your SkyDrive storage online. One downside here is that you won’t be able to save a file onto SkyDrive without it taking up space on the same hard drive SkyDrive is on. You won’t be able to scatter folders across multiple hard drives and sync them all. However, you could always change the location of the SkyDrive folder on Windows 8.1 and put it on a drive with a larger amount of free space. To do this, right-click the SkyDrive folder in File Explorer, select Properties, and use the options on the Location tab. You could even use Storage Spaces to combine the drives into one larger drive. Automatically Copy the Original Files to SkyDrive Another option would be to run a program that automatically copies files from another folder on your computer to your SkyDrive folder. For example, let’s say you want to sync copies of important log files that a program creates in a specific folder. You could use a program that allows you to schedule automatic folder-mirroring, configuring the program to regularly copy the contents of your log folder to your SkyDrive folder. This may be a useful alternative for some use cases, although it isn’t the same as standard syncing. You’ll end up with two copies of the files taking up space on your system, which won’t be ideal for large files. The files also won’t be instantly uploaded to your SkyDrive storage after they’re created, but only after the scheduled task runs. There are many options for this, including Microsoft’s own SyncToy, which continues to work on Windows 8. If you were using the symbolic link trick to automatically sync copies of PC game save files with SkyDrive, you could just install GameSave Manager. It can be configured to automatically create backup copies of your computer’s PC game save files on a schedule, saving them to SkyDrive where they’ll be synced and backed up online. SkyDrive support was completely rewritten for Windows 8.1, so it’s not surprising that this trick no longer works. The ability to use symbolic links in previous versions of SkyDrive was never officially supported, so it’s not surprising to see it break after a rewrite. None of the methods above are as convenient and quick as the old symbolic link method, but they’re the best we can do with the SkyDrive integration Microsoft has given us in Windows 8.1. It’s still possible to use symbolic links to easily sync other folders with competing cloud storage services like Dropbox and Google Drive, so you may want to consider switching away from SkyDrive if this feature is critical to you.     

    Read the article

  • Beginner Geek: How to Use Multiple Monitors to Be More Productive

    - by Chris Hoffman
    Many people swear by multiple monitors, whether they’re geeks or just people who need to be productive. Why use just one monitor when you can use two or more and see more at once? Additional monitors allow you to expand your desktop, getting more screen real estate for your open programs. Windows makes it very easy to set up additional monitors, and your computer probably has the necessary ports. Why Use Multiple Monitors? Multiple monitors give you more screen real estate. Hook up multiple monitors to a computer and you can move your mouse back and forth between them, dragging programs between monitors as if you had an extra-large desktop. People who swear by multiple monitors use them to display multiple things on-screen at a time. Rather than Alt+Tabbing and task switching to glance at another window, you can just look over with your eyes and then look back to the program you’re using. Some examples of use cases for multiple monitors include: Coders who want to view their code on one display with the other display reserved for documentation. They can just glance over at the documentation and look back at their primary workspace. Anyone who needs to view something while working. Viewing a web page while writing an email, viewing another document while writing an something, or working with two large spreadsheets and having both visible at once. People who need to keep an eye on information, whether it’s email or up-to-date statistics, while working. Gamers who want to see more of the game world, extending the game across multiple displays. Geeks who just want to watch a video on one screen while doing something else on the other screen. Hooking Up Multiple Monitors Hooking up an additional monitor to your computer should be very simple. Most new computers come with more than one port for a monitor — whether DVI, HDMI, the older VGA port, or a mix. Some computers may include splitter cables that allow you to connect multiple monitors to a single port. Most laptops also come with ports that allow you to hook up an external monitor. Plug a monitor into your laptop’s DVI or VGA port and Windows will allow you to use both your laptop’s integrated display and the external monitor at once. This all depends on the ports your computer has and how your monitor connects. If you have an old VGA monitor lying around and you have a modern laptop with only DVI or HDMI connectors, you may need an adapter that allows you to plug your monitor’s VGA cable into the new port. Be sure to take your computer’s ports into account before you get another monitor for it. Managing Multiple Monitors With Windows Windows makes using multiple monitors easy. Just plug the monitor into the appropriate port on your computer and Windows should automatically extend your desktop onto it. You can now just drag and drop windows between monitors. To control how this works, right-click your Windows desktop and select Screen resolution. Choose an option from the Multiple displays box. The Extend option extends your desktop onto an additional monitor, while the other options are mainly useful if you’re using an additional monitor for presentations — for example, you could mirror your laptop’s desktop onto a large monitor or blank your laptop’s screen while it’s connected to a larger display. Be sure to arrange your monitors properly so Windows understands how your monitors are physically positioned. Windows 8 allows you to extend your Windows taskbar across multiple monitors. You’ll find this option in the taskbar’s options window — right-click the taskbar and select Properties. You can also choose where you want Windows to display taskbar buttons for open programs — on any monitor’s taskbar or only on the taskbar on the associated monitor. Windows 7 doesn’t have these convenient features built-in — your second monitor won’t have a taskbar. To extend your taskbar onto an additional monitor, you’ll need a third-party utility like the free and open-source Dual Monitor Taskbar. If you just have a single monitor, you can also use the Aero Snap feature to quickly place multiple Windows applications side by side. On Windows 7 or 8, press Windows Key + Left or Windows Key + Right to make the current window take up the left or right half of your display. You could also drag any window’s title bar to the left or right edges of your screen and release the window. How useful this feature is depends on your monitor’s size and resolution. If you have a large, high-resolution monitor, it will allow you to see a lot. If you have a smaller laptop monitor with the seemingly standard 1366×768 resolution, you won’t be able to see much of each snapped window at once, so snapping windows may not be practical. Image Credit: Chance Reecher on Flickr, Camp Atterbury Joint Maneuver Training Center on Flickr, Xavier Caballe on Flickr     

    Read the article

  • Cleanest way to run/debug python programs in windows

    - by YGA
    Python for Windows by default comes with IDLE, which is the barest-bones IDE I've ever encountered. For editing files, I'll stick to emacs, thank you very much. However, I want to run programs in some other shell than the crappy windows command prompt, which can't be widened to more than 80 characters. IDLE lets me run programs in it if I open the file, then hit F5 (to go Run- Run Module). I would rather like to just "run" the command, rather than going through the rigmarole of closing the emacs file, loading the IDLE file, etc. A scan of google and the IDLE docs doesn't seem to give much help about using IDLE's shell but not it's IDE. Any advice from the stack overflow guys? Ideally I'd either like advice on running programs using IDLE's shell advice on other ways to run python programs in windows outside of IDLE or "cmd". Thanks, /YGA

    Read the article

  • PL/SQL execption and Java programs

    - by edwards
    Hi Business logic is coded in pl/sql paackages procedures and functions. Java programs call pl/sql packages procedures and functions to do database work. Issue now is pl/sql programs store excpetions into Oracle tables whenever a execption is raised. How would my java programs get the execptions since the exception instead of being propogated from pl/sql to java is getting persisted to a oracle table.

    Read the article

  • New Enhancements for InnoDB Memcached

    - by Calvin Sun
    In MySQL 5.6, we continued our development on InnoDB Memcached and completed a few widely desirable features that make InnoDB Memcached a competitive feature in more scenario. Notablely, they are 1) Support multiple table mapping 2) Added background thread to auto-commit long running transactions 3) Enhancement in binlog performance  Let’s go over each of these features one by one. And in the last section, we will go over a couple of internally performed performance tests. Support multiple table mapping In our earlier release, all InnoDB Memcached operations are mapped to a single InnoDB table. In the real life, user might want to use this InnoDB Memcached features on different tables. Thus being able to support access to different table at run time, and having different mapping for different connections becomes a very desirable feature. And in this GA release, we allow user just be able to do both. We will discuss the key concepts and key steps in using this feature. 1) "mapping name" in the "get" and "set" command In order to allow InnoDB Memcached map to a new table, the user (DBA) would still require to "pre-register" table(s) in InnoDB Memcached “containers” table (there is security consideration for this requirement). If you would like to know about “containers” table, please refer to my earlier blogs in blogs.innodb.com. Once registered, the InnoDB Memcached will then be able to look for such table when they are referred. Each of such registered table will have a unique "registration name" (or mapping_name) corresponding to the “name” field in the “containers” table.. To access these tables, user will include such "registration name" in their get or set commands, in the form of "get @@new_mapping_name.key", prefix "@@" is required for signaling a mapped table change. The key and the "mapping name" are separated by a configurable delimiter, by default, it is ".". So the syntax is: get [@@mapping_name.]key_name set [@@mapping_name.]key_name  or  get @@mapping_name set @@mapping_name Here is an example: Let's set up three tables in the "containers" table: The first is a map to InnoDB table "test/demo_test" table with mapping name "setup_1" INSERT INTO containers VALUES ("setup_1", "test", "demo_test", "c1", "c2", "c3", "c4", "c5", "PRIMARY");  Similarly, we set up table mappings for table "test/new_demo" with name "setup_2" and that to table "mydatabase/my_demo" with name "setup_3": INSERT INTO containers VALUES ("setup_2", "test", "new_demo", "c1", "c2", "c3", "c4", "c5", "secondary_index_x"); INSERT INTO containers VALUES ("setup_3", "my_database", "my_demo", "c1", "c2", "c3", "c4", "c5", "idx"); To switch to table "my_database/my_demo", and get the value corresponding to “key_a”, user will do: get @@setup_3.key_a (this will also output the value that corresponding to key "key_a" or simply get @@setup_3 Once this is done, this connection will switch to "my_database/my_demo" table until another table mapping switch is requested. so it can continue issue regular command like: get key_b  set key_c 0 0 7 These DMLs will all be directed to "my_database/my_demo" table. And this also implies that different connections can have different bindings (to different table). 2) Delimiter: For the delimiter "." that separates the "mapping name" and key value, we also added a configure option in the "config_options" system table with name of "table_map_delimiter": INSERT INTO config_options VALUES("table_map_delimiter", "."); So if user wants to change to a different delimiter, they can change it in the config_option table. 3) Default mapping: Once we have multiple table mapping, there should be always a "default" map setting. For this, we decided if there exists a mapping name of "default", then this will be chosen as default mapping. Otherwise, the first row of the containers table will chosen as default setting. Please note, user tables can be repeated in the "containers" table (for example, user wants to access different columns of the table in different settings), as long as they are using different mapping/configure names in the first column, which is enforced by a unique index. 4) bind command In addition, we also extend the protocol and added a bind command, its usage is fairly straightforward. To switch to "setup_3" mapping above, you simply issue: bind setup_3 This will switch this connection's InnoDB table to "my_database/my_demo" In summary, with this feature, you now can direct access to difference tables with difference session. And even a single connection, you can query into difference tables. Background thread to auto-commit long running transactions This is a feature related to the “batch” concept we discussed in earlier blogs. This “batch” feature allows us batch the read and write operations, and commit them only after certain calls. The “batch” size is controlled by the configure parameter “daemon_memcached_w_batch_size” and “daemon_memcached_r_batch_size”. This could significantly boost performance. However, it also comes with some disadvantages, for example, you will not be able to view “uncommitted” operations from SQL end unless you set transaction isolation level to read_uncommitted, and in addition, this will held certain row locks for extend period of time that might reduce the concurrency. To deal with this, we introduce a background thread that “auto-commits” the transaction if they are idle for certain amount of time (default is 5 seconds). The background thread will wake up every second and loop through every “connections” opened by Memcached, and check for idle transactions. And if such transaction is idle longer than certain limit and not being used, it will commit such transactions. This limit is configurable by change “innodb_api_bk_commit_interval”. Its default value is 5 seconds, and minimum is 1 second, and maximum is 1073741824 seconds. With the help of such background thread, you will not need to worry about long running uncommitted transactions when set daemon_memcached_w_batch_size and daemon_memcached_r_batch_size to a large number. This also reduces the number of locks that could be held due to long running transactions, and thus further increase the concurrency. Enhancement in binlog performance As you might all know, binlog operation is not done by InnoDB storage engine, rather it is handled in the MySQL layer. In order to support binlog operation through InnoDB Memcached, we would have to artificially create some MySQL constructs in order to access binlog handler APIs. In previous lab release, for simplicity consideration, we open and destroy these MySQL constructs (such as THD) for each operations. This required us to set the “batch” size always to 1 when binlog is on, no matter what “daemon_memcached_w_batch_size” and “daemon_memcached_r_batch_size” are configured to. This put a big restriction on our capability to scale, and also there are quite a bit overhead in creating destroying such constructs that bogs the performance down. With this release, we made necessary change that would keep MySQL constructs as long as they are valid for a particular connection. So there will not be repeated and redundant open and close (table) calls. And now even with binlog option is enabled (with innodb_api_enable_binlog,), we still can batch the transactions with daemon_memcached_w_batch_size and daemon_memcached_r_batch_size, thus scale the write/read performance. Although there are still overheads that makes InnoDB Memcached cannot perform as fast as when binlog is turned off. It is much better off comparing to previous release. And we are continuing optimize the solution is this area to improve the performance as much as possible. Performance Study: Amerandra of our System QA team have conducted some performance studies on queries through our InnoDB Memcached connection and plain SQL end. And it shows some interesting results. The test is conducted on a “Linux 2.6.32-300.7.1.el6uek.x86_64 ix86 (64)” machine with 16 GB Memory, Intel Xeon 2.0 GHz CPU X86_64 2 CPUs- 4 Core Each, 2 RAID DISKS (1027 GB,733.9GB). Results are described in following tables: Table 1: Performance comparison on Set operations Connections 5.6.7-RC-Memcached-plugin ( TPS / Qps) with memcached-threads=8*** 5.6.7-RC* X faster Set (QPS) Set** 8 30,000 5,600 5.36 32 59,000 13,000 4.54 128 68,000 8,000 8.50 512 63,000 6.800 9.23 * mysql-5.6.7-rc-linux2.6-x86_64 ** The “set” operation when implemented in InnoDB Memcached involves a couple of DMLs: it first query the table to see whether the “key” exists, if it does not, the new key/value pair will be inserted. If it does exist, the “value” field of matching row (by key) will be updated. So when used in above query, it is a precompiled store procedure, and query will just execute such procedures. *** added “–daemon_memcached_option=-t8” (default is 4 threads) So we can see with this “set” query, InnoDB Memcached can run 4.5 to 9 time faster than MySQL server. Table 2: Performance comparison on Get operations Connections 5.6.7-RC-Memcached-plugin ( TPS / Qps) with memcached-threads=8 5.6.7-RC* X faster Get (QPS) Get 8 42,000 27,000 1.56 32 101,000 55.000 1.83 128 117,000 52,000 2.25 512 109,000 52,000 2.10 With the “get” query (or the select query), memcached performs 1.5 to 2 times faster than normal SQL. Summary: In summary, we added several much-desired features to InnoDB Memcached in this release, allowing user to operate on different tables with this Memcached interface. We also now provide a background commit thread to commit long running idle transactions, thus allow user to configure large batch write/read without worrying about large number of rows held or not being able to see (uncommit) data. We also greatly enhanced the performance when Binlog is enabled. We will continue making efforts in both performance enhancement and functionality areas to make InnoDB Memcached a good demo case for our InnoDB APIs. Jimmy Yang, September 29, 2012

    Read the article

  • The next next C++ [closed]

    - by Roger Pate
    It's entirely too early for speculation on what C++ will be like after C++0x, but idle hands make for wild predictions. What features would you find useful and why? Is there anything in another language that would fit nicely into the state of C++ after 0x? What should be considered for the next TC and TR? (Mostly TR, as the TC would depend more on what actually becomes the next standard.) Export was removed, rather than merely deprecated, in 0x. (It remains a keyword.) What other features carry so much baggage to also be more harmful than helpful? ISO Standards' process I'm not involved in the C++ committee, but it's also a mystery, unfortunately, to most programmers using C++. A few things worth keeping in mind: There will be 10 years between standards, barring extremely exceptional circumstances. The standard can get "bug fixes" in the form of a Technical Corrigendum. This happened to C++98 with TC1, named C++03. It fixed "simple" issues such as making the explicit guarantee that std::vector stores items contiguously, which was always intended. The committee can issue reports which can add to the language. This happened to C++98/03 with TR1 in 2005, which introduced the std::tr1 namespace.

    Read the article

< Previous Page | 18 19 20 21 22 23 24 25 26 27 28 29  | Next Page >