Search Results

Search found 11135 results on 446 pages for 'thread safe'.

Page 125/446 | < Previous Page | 121 122 123 124 125 126 127 128 129 130 131 132  | Next Page >

  • Dispatcher.CheckAccess() isn't working from my console application, is there a better way.

    - by zimmer62
    I wrote an application in WPF / VB and separated the business logic and UI into different projects. The business layer uses a serial port which runs on a different thread, Now that I'm trying to write a command line interface for the same business layer, it seems to fail when .Invoke() is called. (no error, just doesn't work) I'm pretty sure the reason I had to add in checkaccess and .invoke was because I have collections that would be changed during processing the serial port data and wanted the NotifyCollectionChanged to be handled by WPF data binding. (The reason I'm not 100% sure is because it was months ago I wrote that part and it all worked great from the GUI, now adding the console app has made me rethink some of this) I would like my business layer to run these processes on the thread they were created, I need this to work from both my GUI version and the command line version. Am I misusing the Dispatcher in my business layer? Is there a better way to handle an event from the serial port, and then return to the main thread to processes the data?

    Read the article

  • MySQL Syslog Audit Plugin

    - by jonathonc
    This post shows the construction process of the Syslog Audit plugin that was presented at MySQL Connect 2012. It is based on an environment that has the appropriate development tools enabled including gcc,g++ and cmake. It also assumes you have downloaded the MySQL source code (5.5.16 or higher) and have compiled and installed the system into the /usr/local/mysql directory ready for use.  The information provided below is designed to show the different components that make up a plugin, and specifically an audit type plugin, and how it comes together to be used within the MySQL service. The MySQL Reference Manual contains information regarding the plugin API and how it can be used, so please refer there for more detailed information. The code in this post is designed to give the simplest information necessary, so handling every return code, managing race conditions etc is not part of this example code. Let's start by looking at the most basic implementation of our plugin code as seen below: /*    Copyright (c) 2012, Oracle and/or its affiliates. All rights reserved.    Author:  Jonathon Coombes    Licence: GPL    Description: An auditing plugin that logs to syslog and                 can adjust the loglevel via the system variables. */ #include <stdio.h> #include <string.h> #include <mysql/plugin_audit.h> #include <syslog.h> There is a commented header detailing copyright/licencing and meta-data information and then the include headers. The two important include statements for our plugin are the syslog.h plugin, which gives us the structures for syslog, and the plugin_audit.h include which has details regarding the audit specific plugin api. Note that we do not need to include the general plugin header plugin.h, as this is done within the plugin_audit.h file already. To implement our plugin within the current implementation we need to add it into our source code and compile. > cd /usr/local/src/mysql-5.5.28/plugin > mkdir audit_syslog > cd audit_syslog A simple CMakeLists.txt file is created to manage the plugin compilation: MYSQL_ADD_PLUGIN(audit_syslog audit_syslog.cc MODULE_ONLY) Run the cmake  command at the top level of the source and then you can compile the plugin using the 'make' command. This results in a compiled audit_syslog.so library, but currently it is not much use to MySQL as there is no level of api defined to communicate with the MySQL service. Now we need to define the general plugin structure that enables MySQL to recognise the library as a plugin and be able to install/uninstall it and have it show up in the system. The structure is defined in the plugin.h file in the MySQL source code.  /*   Plugin library descriptor */ mysql_declare_plugin(audit_syslog) {   MYSQL_AUDIT_PLUGIN,           /* plugin type                    */   &audit_syslog_descriptor,     /* descriptor handle               */   "audit_syslog",               /* plugin name                     */   "Author Name",                /* author                          */   "Simple Syslog Audit",        /* description                     */   PLUGIN_LICENSE_GPL,           /* licence                         */   audit_syslog_init,            /* init function     */   audit_syslog_deinit,          /* deinit function */   0x0001,                       /* plugin version                  */   NULL,                         /* status variables        */   NULL,                         /* system variables                */   NULL,                         /* no reserves                     */   0,                            /* no flags                        */ } mysql_declare_plugin_end; The general plugin descriptor above is standard for all plugin types in MySQL. The plugin type is defined along with the init/deinit functions and interface methods into the system for sharing information, and various other metadata information. The descriptors have an internally recognised version number so that plugins can be matched against the api on the running server. The other details are usually related to the type-specific methods and structures to implement the plugin. Each plugin has a type-specific descriptor as well which details how the plugin is implemented for the specific purpose of that plugin type. /*   Plugin type-specific descriptor */ static struct st_mysql_audit audit_syslog_descriptor= {   MYSQL_AUDIT_INTERFACE_VERSION,                        /* interface version    */   NULL,                                                 /* release_thd function */   audit_syslog_notify,                                  /* notify function      */   { (unsigned long) MYSQL_AUDIT_GENERAL_CLASSMASK |                     MYSQL_AUDIT_CONNECTION_CLASSMASK }  /* class mask           */ }; In this particular case, the release_thd function has not been defined as it is not required. The important method for auditing is the notify function which is activated when an event occurs on the system. The notify function is designed to activate on an event and the implementation will determine how it is handled. For the audit_syslog plugin, the use of the syslog feature sends all events to the syslog for recording. The class mask allows us to determine what type of events are being seen by the notify function. There are currently two major types of event: 1. General Events: This includes general logging, errors, status and result type events. This is the main one for tracking the queries and operations on the database. 2. Connection Events: This group is based around user logins. It monitors connections and disconnections, but also if somebody changes user while connected. With most audit plugins, the principle behind the plugin is to track changes to the system over time and counters can be an important part of this process. The next step is to define and initialise the counters that are used to track the events in the service. There are 3 counters defined in total for our plugin - the # of general events, the # of connection events and the total number of events.  static volatile int total_number_of_calls; /* Count MYSQL_AUDIT_GENERAL_CLASS event instances */ static volatile int number_of_calls_general; /* Count MYSQL_AUDIT_CONNECTION_CLASS event instances */ static volatile int number_of_calls_connection; The init and deinit functions for the plugin are there to be called when the plugin is activated and when it is terminated. These offer the best option to initialise the counters for our plugin: /*  Initialize the plugin at server start or plugin installation. */ static int audit_syslog_init(void *arg __attribute__((unused))) {     openlog("mysql_audit:",LOG_PID|LOG_PERROR|LOG_CONS,LOG_USER);     total_number_of_calls= 0;     number_of_calls_general= 0;     number_of_calls_connection= 0;     return(0); } The init function does a call to openlog to initialise the syslog functionality. The parameters are the service to log under ("mysql_audit" in this case), the syslog flags and the facility for the logging. Then each of the counters are initialised to zero and a success is returned. If the init function is not defined, it will return success by default. /*  Terminate the plugin at server shutdown or plugin deinstallation. */ static int audit_syslog_deinit(void *arg __attribute__((unused))) {     closelog();     return(0); } The deinit function will simply close our syslog connection and return success. Note that the syslog functionality is part of the glibc libraries and does not require any external factors.  The function names are what we define in the general plugin structure, so these have to match otherwise there will be errors. The next step is to implement the event notifier function that was defined in the type specific descriptor (audit_syslog_descriptor) which is audit_syslog_notify. /* Event notifier function */ static void audit_syslog_notify(MYSQL_THD thd __attribute__((unused)), unsigned int event_class, const void *event) { total_number_of_calls++; if (event_class == MYSQL_AUDIT_GENERAL_CLASS) { const struct mysql_event_general *event_general= (const struct mysql_event_general *) event; number_of_calls_general++; syslog(audit_loglevel,"%lu: User: %s Command: %s Query: %s\n", event_general->general_thread_id, event_general->general_user, event_general->general_command, event_general->general_query ); } else if (event_class == MYSQL_AUDIT_CONNECTION_CLASS) { const struct mysql_event_connection *event_connection= (const struct mysql_event_connection *) event; number_of_calls_connection++; syslog(audit_loglevel,"%lu: User: %s@%s[%s] Event: %d Status: %d\n", event_connection->thread_id, event_connection->user, event_connection->host, event_connection->ip, event_connection->event_subclass, event_connection->status ); } }   In the case of an event, the notifier function is called. The first step is to increment the total number of events that have occurred in our database.The event argument is then cast into the appropriate event structure depending on the class type, of general event or connection event. The event type counters are incremented and details are sent via the syslog() function out to the system log. There are going to be different line formats and information returned since the general events have different data compared to the connection events, even though some of the details overlap, for example, user, thread id, host etc. On compiling the code now, there should be no errors and the resulting audit_syslog.so can be loaded into the server and ready to use. Log into the server and type: mysql> INSTALL PLUGIN audit_syslog SONAME 'audit_syslog.so'; This will install the plugin and will start updating the syslog immediately. Note that the audit plugin attaches to the immediate thread and cannot be uninstalled while that thread is active. This means that you cannot run the UNISTALL command until you log into a different connection (thread) on the server. Once the plugin is loaded, the system log will show output such as the following: Oct  8 15:33:21 machine mysql_audit:[8337]: 87: User: root[root] @ localhost []  Command: (null)  Query: INSTALL PLUGIN audit_syslog SONAME 'audit_syslog.so' Oct  8 15:33:21 machine mysql_audit:[8337]: 87: User: root[root] @ localhost []  Command: Query  Query: INSTALL PLUGIN audit_syslog SONAME 'audit_syslog.so' Oct  8 15:33:40 machine mysql_audit:[8337]: 87: User: root[root] @ localhost []  Command: (null)  Query: show tables Oct  8 15:33:40 machine mysql_audit:[8337]: 87: User: root[root] @ localhost []  Command: Query  Query: show tables Oct  8 15:33:43 machine mysql_audit:[8337]: 87: User: root[root] @ localhost []  Command: (null)  Query: select * from t1 Oct  8 15:33:43 machine mysql_audit:[8337]: 87: User: root[root] @ localhost []  Command: Query  Query: select * from t1 It appears that two of each event is being shown, but in actuality, these are two separate event types - the result event and the status event. This could be refined further by changing the audit_syslog_notify function to handle the different event sub-types in a different manner.  So far, it seems that the logging is working with events showing up in the syslog output. The issue now is that the counters created earlier to track the number of events by type are not accessible when the plugin is being run. Instead there needs to be a way to expose the plugin specific information to the service and vice versa. This could be done via the information_schema plugin api, but for something as simple as counters, the obvious choice is the system status variables. This is done using the standard structure and the declaration: /*  Plugin status variables for SHOW STATUS */ static struct st_mysql_show_var audit_syslog_status[]= {   { "Audit_syslog_total_calls",     (char *) &total_number_of_calls,     SHOW_INT },   { "Audit_syslog_general_events",     (char *) &number_of_calls_general,     SHOW_INT },   { "Audit_syslog_connection_events",     (char *) &number_of_calls_connection,     SHOW_INT },   { 0, 0, SHOW_INT } };   The structure is simply the name that will be displaying in the mysql service, the address of the associated variables, and the data type being used for the counter. It is finished with a blank structure to show that there are no more variables. Remember that status variables may have the same name for variables from other plugin, so it is considered appropriate to add the plugin name at the start of the status variable name to avoid confusion. Looking at the status variables in the mysql client shows something like the following: mysql> show global status like "audit%"; +--------------------------------+-------+ | Variable_name                  | Value | +--------------------------------+-------+ | Audit_syslog_connection_events | 1     | | Audit_syslog_general_events    | 2     | | Audit_syslog_total_calls       | 3     | +--------------------------------+-------+ 3 rows in set (0.00 sec) The final connectivity piece for the plugin is to allow the interactive change of the logging level between the plugin and the system. This requires the ability to send changes via the mysql service through to the plugin. This is done using the system variables interface and defining a single variable to keep track of the active logging level for the facility. /* Plugin system variables for SHOW VARIABLES */ static MYSQL_SYSVAR_STR(loglevel, audit_loglevel,                         PLUGIN_VAR_RQCMDARG,                         "User can specify the log level for auditing",                         audit_loglevel_check, audit_loglevel_update, "LOG_NOTICE"); static struct st_mysql_sys_var* audit_syslog_sysvars[] = {     MYSQL_SYSVAR(loglevel),     NULL }; So now the system variable 'loglevel' is defined for the plugin and associated to the global variable 'audit_loglevel'. The check or validation function is defined to make sure that no garbage values are attempted in the update of the variable. The update function is used to save the new value to the variable. Note that the audit_syslog_sysvars structure is defined in the general plugin descriptor to associate the link between the plugin and the system and how much they interact. Next comes the implementation of the validation function and the update function for the system variable. It is worth noting that if you have a simple numeric such as integers for the variable types, the validate function is often not required as MySQL will handle the automatic check and validation of simple types. /* longest valid value */ #define MAX_LOGLEVEL_SIZE 100 /* hold the valid values */ static const char *possible_modes[]= { "LOG_ERROR", "LOG_WARNING", "LOG_NOTICE", NULL };  static int audit_loglevel_check(     THD*                        thd,    /*!< in: thread handle */     struct st_mysql_sys_var*    var,    /*!< in: pointer to system                                         variable */     void*                       save,   /*!< out: immediate result                                         for update function */     struct st_mysql_value*      value)  /*!< in: incoming string */ {     char buff[MAX_LOGLEVEL_SIZE];     const char *str;     const char **found;     int length;     length= sizeof(buff);     if (!(str= value->val_str(value, buff, &length)))         return 1;     /*         We need to return a pointer to a locally allocated value in "save".         Here we pick to search for the supplied value in an global array of         constant strings and return a pointer to one of them.         The other possiblity is to use the thd_alloc() function to allocate         a thread local buffer instead of the global constants.     */     for (found= possible_modes; *found; found++)     {         if (!strcmp(*found, str))         {             *(const char**)save= *found;             return 0;         }     }     return 1; } The validation function is simply to take the value being passed in via the SET GLOBAL VARIABLE command and check if it is one of the pre-defined values allowed  in our possible_values array. If it is found to be valid, then the value is assigned to the save variable ready for passing through to the update function. static void audit_loglevel_update(     THD*                        thd,        /*!< in: thread handle */     struct st_mysql_sys_var*    var,        /*!< in: system variable                                             being altered */     void*                       var_ptr,    /*!< out: pointer to                                             dynamic variable */     const void*                 save)       /*!< in: pointer to                                             temporary storage */ {     /* assign the new value so that the server can read it */     *(char **) var_ptr= *(char **) save;     /* assign the new value to the internal variable */     audit_loglevel= *(char **) save; } Since all the validation has been done already, the update function is quite simple for this plugin. The first part is to update the system variable pointer so that the server can read the value. The second part is to update our own global plugin variable for tracking the value. Notice that the save variable is passed in as a void type to allow handling of various data types, so it must be cast to the appropriate data type when assigning it to the variables. Looking at how the latest changes affect the usage of the plugin and the interaction within the server shows: mysql> show global variables like "audit%"; +-----------------------+------------+ | Variable_name         | Value      | +-----------------------+------------+ | audit_syslog_loglevel | LOG_NOTICE | +-----------------------+------------+ 1 row in set (0.00 sec) mysql> set global audit_syslog_loglevel="LOG_ERROR"; Query OK, 0 rows affected (0.00 sec) mysql> show global status like "audit%"; +--------------------------------+-------+ | Variable_name                  | Value | +--------------------------------+-------+ | Audit_syslog_connection_events | 1     | | Audit_syslog_general_events    | 11    | | Audit_syslog_total_calls       | 12    | +--------------------------------+-------+ 3 rows in set (0.00 sec) mysql> show global variables like "audit%"; +-----------------------+-----------+ | Variable_name         | Value     | +-----------------------+-----------+ | audit_syslog_loglevel | LOG_ERROR | +-----------------------+-----------+ 1 row in set (0.00 sec)   So now we have a plugin that will audit the events on the system and log the details to the system log. It allows for interaction to see the number of different events within the server details and provides a mechanism to change the logging level interactively via the standard system methods of the SET command. A more complex auditing plugin may have more detailed code, but each of the above areas is what will be involved and simply expanded on to add more functionality. With the above skeleton code, it is now possible to create your own audit plugins to implement your own auditing requirements. If, however, you are not of the coding persuasion, then you could always consider the option of the MySQL Enterprise Audit plugin that is available to purchase.

    Read the article

  • Node.js Adventure - Node.js on Windows

    - by Shaun
    Two weeks ago I had had a talk with Wang Tao, a C# MVP in China who is currently running his startup company and product named worktile. He asked me to figure out a synchronization solution which helps his product in the future. And he preferred me implementing the service in Node.js, since his worktile is written in Node.js. Even though I have some experience in ASP.NET MVC, HTML, CSS and JavaScript, I don’t think I’m an expert of JavaScript. In fact I’m very new to it. So it scared me a bit when he asked me to use Node.js. But after about one week investigate I have to say Node.js is very easy to learn, use and deploy, even if you have very limited JavaScript skill. And I think I became love Node.js. Hence I decided to have a series named “Node.js Adventure”, where I will demonstrate my story of learning and using Node.js in Windows and Windows Azure. And this is the first one.   (Brief) Introduction of Node.js I don’t want to have a fully detailed introduction of Node.js. There are many resource on the internet we can find. But the best one is its homepage. Node.js was created by Ryan Dahl, sponsored by Joyent. It’s consist of about 80% C/C++ for core and 20% JavaScript for API. It utilizes CommonJS as the module system which we will explain later. The official definition of Node.js is Node.js is a platform built on Chrome's JavaScript runtime for easily building fast, scalable network applications. Node.js uses an event-driven, non-blocking I/O model that makes it lightweight and efficient, perfect for data-intensive real-time applications that run across distributed devices. First of all, Node.js utilizes JavaScript as its development language and runs on top of V8 engine, which is being used by Chrome. It brings JavaScript, a client-side language into the backend service world. So many people said, even though not that actually, “Node.js is a server side JavaScript”. Additionally, Node.js uses an event-driven, non-blocking IO model. This means in Node.js there’s no way to block currently working thread. Every operation in Node.js executed asynchronously. This is a huge benefit especially if our code needs IO operations such as reading disks, connect to database, consuming web service, etc.. Unlike IIS or Apache, Node.js doesn’t utilize the multi-thread model. In Node.js there’s only one working thread serves all users requests and resources response, as the ST star in the figure below. And there is a POSIX async threads pool in Node.js which contains many async threads (AT stars) for IO operations. When a user have an IO request, the ST serves it but it will not do the IO operation. Instead the ST will go to the POSIX async threads pool to pick up an AT, pass this operation to it, and then back to serve any other requests. The AT will actually do the IO operation asynchronously. Assuming before the AT complete the IO operation there is another user comes. The ST will serve this new user request, pick up another AT from the POSIX and then back. If the previous AT finished the IO operation it will take the result back and wait for the ST to serve. ST will take the response and return the AT to POSIX, and then response to the user. And if the second AT finished its job, the ST will response back to the second user in the same way. As you can see, in Node.js there’s only one thread serve clients’ requests and POSIX results. This thread looping between the users and POSIX and pass the data back and forth. The async jobs will be handled by POSIX. This is the event-driven non-blocking IO model. The performance of is model is much better than the multi-threaded blocking model. For example, Apache is built in multi-threaded blocking model while Nginx is in event-driven non-blocking mode. Below is the performance comparison between them. And below is the memory usage comparison between them. These charts are captured from the video NodeJS Basics: An Introductory Training, which presented at Cloud Foundry Developer Advocate.   Node.js on Windows To execute Node.js application on windows is very simple. First of you we need to download the latest Node.js platform from its website. After installed, it will register its folder into system path variant so that we can execute Node.js at anywhere. To confirm the Node.js installation, just open up a command windows and type “node”, then it will show the Node.js console. As you can see this is a JavaScript interactive console. We can type some simple JavaScript code and command here. To run a Node.js JavaScript application, just specify the source code file name as the argument of the “node” command. For example, let’s create a Node.js source code file named “helloworld.js”. Then copy a sample code from Node.js website. 1: var http = require("http"); 2:  3: http.createServer(function (req, res) { 4: res.writeHead(200, {"Content-Type": "text/plain"}); 5: res.end("Hello World\n"); 6: }).listen(1337, "127.0.0.1"); 7:  8: console.log("Server running at http://127.0.0.1:1337/"); This code will create a web server, listening on 1337 port and return “Hello World” when any requests come. Run it in the command windows. Then open a browser and navigate to http://localhost:1337/. As you can see, when using Node.js we are not creating a web application. In fact we are likely creating a web server. We need to deal with request, response and the related headers, status code, etc.. And this is one of the benefit of using Node.js, lightweight and straightforward. But creating a website from scratch again and again is not acceptable. The good news is that, Node.js utilizes CommonJS as its module system, so that we can leverage some modules to simplify our job. And furthermore, there are about ten thousand of modules available n the internet, which covers almost all areas in server side application development.   NPM and Node.js Modules Node.js utilizes CommonJS as its module system. A module is a set of JavaScript files. In Node.js if we have an entry file named “index.js”, then all modules it needs will be located at the “node_modules” folder. And in the “index.js” we can import modules by specifying the module name. For example, in the code we’ve just created, we imported a module named “http”, which is a build-in module installed alone with Node.js. So that we can use the code in this “http” module. Besides the build-in modules there are many modules available at the NPM website. Thousands of developers are contributing and downloading modules at this website. Hence this is another benefit of using Node.js. There are many modules we can use, and the numbers of modules increased very fast, and also we can publish our modules to the community. When I wrote this post, there are totally 14,608 modules at NPN and about 10 thousand downloads per day. Install a module is very simple. Let’s back to our command windows and input the command “npm install express”. This command will install a module named “express”, which is a MVC framework on top of Node.js. And let’s create another JavaScript file named “helloweb.js” and copy the code below in it. I imported the “express” module. And then when the user browse the home page it will response a text. If the incoming URL matches “/Echo/:value” which the “value” is what the user specified, it will pass it back with the current date time in JSON format. And finally my website was listening at 12345 port. 1: var express = require("express"); 2: var app = express(); 3:  4: app.get("/", function(req, res) { 5: res.send("Hello Node.js and Express."); 6: }); 7:  8: app.get("/Echo/:value", function(req, res) { 9: var value = req.params.value; 10: res.json({ 11: "Value" : value, 12: "Time" : new Date() 13: }); 14: }); 15:  16: console.log("Web application opened."); 17: app.listen(12345); For more information and API about the “express”, please have a look here. Start our application from the command window by command “node helloweb.js”, and then navigate to the home page we can see the response in the browser. And if we go to, for example http://localhost:12345/Echo/Hello Shaun, we can see the JSON result. The “express” module is very populate in NPM. It makes the job simple when we need to build a MVC website. There are many modules very useful in NPM. - underscore: A utility module covers many common functionalities such as for each, map, reduce, select, etc.. - request: A very simple HTT request client. - async: Library for coordinate async operations. - wind: Library which enable us to control flow with plain JavaScript for asynchronous programming (and more) without additional pre-compiling steps.   Node.js and IIS I demonstrated how to run the Node.js application from console. Since we are in Windows another common requirement would be, “can I host Node.js in IIS?” The answer is “Yes”. Tomasz Janczuk created a project IISNode at his GitHub space we can find here. And Scott Hanselman had published a blog post introduced about it.   Summary In this post I provided a very brief introduction of Node.js, includes it official definition, architecture and how it implement the event-driven non-blocking model. And then I described how to install and run a Node.js application on windows console. I also described the Node.js module system and NPM command. At the end I referred some links about IISNode, an IIS extension that allows Node.js application runs on IIS. Node.js became a very popular server side application platform especially in this year. By leveraging its non-blocking IO model and async feature it’s very useful for us to build a highly scalable, asynchronously service. I think Node.js will be used widely in the cloud application development in the near future.   In the next post I will explain how to use SQL Server from Node.js.   Hope this helps, Shaun All documents and related graphics, codes are provided "AS IS" without warranty of any kind. Copyright © Shaun Ziyan Xu. This work is licensed under the Creative Commons License.

    Read the article

  • The embarrassingly obvious about SQL Server CE

    - by Edward Boyle
    I have been working with SQL servers in one form or another for almost two decades now. But I am new to SQL Server Compact Edition. In the past weeks I have been working with SQL Serve CE a lot. The SQL, not a problem, but the engine itself is very new to me. One of the issues I ran into was a simple SQL statement taking excusive amounts of time; by excessive, I mean over one second. I wrote a little code to time the method. Sometimes it took under one second, other times as long as three seconds. –But it was a simple update statement! As embarrassing as it is, why it was slow eluded me. I posted my issue to MSDN and I got a reply from ErikEJ (MS MVP) who runs the blog “Everything SQL Server Compact” . I know little to nothing about SQL Server Compact. This guy is completely obsessed very well versed in CE. If you spend any time in MSDN forums, it seems that this guy single handedly has the answer for every CE question that comes up. Anyway, he said: “Opening a connection to a SQL Server Compact database file is a costly operation, keep one connection open per thread (incl. your UI thread) in your app, the one on the UI thread should live for the duration of your app.” It hit me, all databases have some connection overhead and SQL Server CE is not a database engine running as a service drinking Jolt Cola waiting for someone to talk to him so he can spring into action and show off his quarter-mile sprint capabilities. Imagine if you had to start the SQL Server process every time you needed to make a database connection. Principally, that is what you are doing with SQL Server CE. For someone who has worked with Enterprise Level SQL Servers a lot, I had to come to the mental image that my Open connection to SQL Server CE is basically starting a service, my own private service, and by closing the connection, I am shutting down my little private service. After making the changes in my code, I lost any reservations I had with using CE. At present, my Data Access Layer class has a constructor; in that constructor I open my connection, I also have OpenConnection and CloseConnection methods, I also implemented IDisposable and clean up any connections in Dispose(). I am still finalizing how this assembly will function. – That’s beside the point. All I’m trying to say is: “Opening a connection to a SQL Server Compact database file is a costly operation”

    Read the article

  • Gnome Shell segfault in libglib-2.0

    - by slohui
    I have been using Ubuntu 11.10 + Gnome Shell with a Nvidia card, but now I've moved it to my new PC which has an ATI card, at first it wasn't booting but I installed the driver from amd.com and then it worked. Anyway my problem is that gnome-shell is crashing, mostly when I try to start a VirtualBox machine (it happened in other times but I don't remember what I was doing). Sometimes gnome-shell respawns and it continue working but sometimes it doesn't so I have to restart lightdm and lose all the windows I was using. Here's some of the syslog when the crash occurs: Apr 9 12:20:08 desktop-1 NetworkManager[1032]: SCPlugin-Ifupdown: devices added (path: /sys/devices/virtual/net/vboxnet0, iface: vboxnet0) Apr 9 12:20:08 desktop-1 NetworkManager[1032]: SCPlugin-Ifupdown: device added (path: /sys/devices/virtual/net/vboxnet0, iface: vboxnet0): no ifupdown configuration found. Apr 9 12:20:08 desktop-1 NetworkManager[1032]: <warn> /sys/devices/virtual/net/vboxnet0: couldn't determine device driver; ignoring... Apr 9 12:20:08 desktop-1 kernel: [ 4498.689561] warning: `VirtualBox' uses 32-bit capabilities (legacy support in use) Apr 9 12:24:29 desktop-1 gnome-session[1617]: WARNING: Application 'gnome-shell.desktop' killed by signal Apr 9 12:24:45 desktop-1 gnome-session[1617]: WARNING: App 'gnome-shell.desktop' respawning too quickly Apr 9 12:24:45 desktop-1 gnome-session[1617]: CRITICAL: We failed, but the fail whale is dead. Sorry.... Apr 9 12:25:20 desktop-1 kernel: [ 4810.769775] show_signal_msg: 30 callbacks suppressed |----- > Apr 9 12:25:20 desktop-1 kernel: [ 4810.769785] gnome-shell[3427]: segfault at b0 ip b6bd09cd sp bfc9b650 error 4 in libglib-2.0.so.0.3000.0[b6b71000+f7000]** Apr 9 12:25:20 desktop-1 gnome-session[1617]: WARNING: Application 'gnome-shell.desktop' killed by signal Apr 9 12:25:23 desktop-1 kernel: [ 4814.055705] EXT4-fs (sda1): Unaligned AIO/DIO on inode 133295 by VirtualBox; performance will be poor. Apr 9 12:26:55 desktop-1 gnome-session[1617]: Gdk-WARNING: gnome-session: Fatal IO error 11 (Resource temporarily unavailable) on X server :0.#012 Apr 9 12:26:55 desktop-1 kernel: [ 4905.373256] [fglrx] IRQ 56 Disabled Apr 9 12:26:59 desktop-1 acpid: client 1124[0:0] has disconnected Apr 9 12:26:59 desktop-1 acpid: client connected from 3864[0:0] Apr 9 12:26:59 desktop-1 acpid: 1 client rule loaded Apr 9 12:26:59 desktop-1 kernel: [ 4909.700095] fglrx_pci 0000:02:00.0: irq 56 for MSI/MSI-X Apr 9 12:26:59 desktop-1 kernel: [ 4909.701466] [fglrx] Firegl kernel thread PID: 3867 Apr 9 12:26:59 desktop-1 kernel: [ 4909.701625] [fglrx] Firegl kernel thread PID: 3868 Apr 9 12:26:59 desktop-1 kernel: [ 4909.701852] [fglrx] Firegl kernel thread PID: 3869 Apr 9 12:26:59 desktop-1 kernel: [ 4909.702021] [fglrx] IRQ 56 Enabled Apr 9 12:26:59 desktop-1 kernel: [ 4909.861815] [fglrx] Gart USWC size:1280 M. Apr 9 12:26:59 desktop-1 kernel: [ 4909.861817] [fglrx] Gart cacheable size:508 M. Apr 9 12:26:59 desktop-1 kernel: [ 4909.861820] [fglrx] Reserved FB block: Shared offset:0, size:1000000 Apr 9 12:26:59 desktop-1 kernel: [ 4909.861821] [fglrx] Reserved FB block: Unshared offset:f8fd000, size:403000 Apr 9 12:26:59 desktop-1 kernel: [ 4909.861823] [fglrx] Reserved FB block: Unshared offset:3fff4000, size:c000 Does anyone could guide me on how to fix this? Or the proper place where to ask for help.

    Read the article

  • JRockit R28/JRockit Mission Control 4.0 is out!

    - by Marcus Hirt
    The next major release of JRockit is finally out! Here are some highlights: Includes the all new JRockit Flight Recorder – supersedes the old JRockit Runtime Analyser. The new flight recorder is inspired by the “black box” in airplanes. It uses a highly efficient recording engine and thread local buffers to capture data about the runtime and the application running in the JVM. It can be configured to always be on, so that whenever anything “interesting” happens, data can be dumped for some time back. Think of it as your own personal profiling time machine. Automatic shortest path calculation in Memleak – no longer any need for running around in circles when trying to find your way back to a thread root from an instance. Memleak can now show class loader related information and split graphs on a per class loader basis. More easily configured JMX agent – default port for both RMI Registry and RMI Server can be configured, and is by default the same, allowing easier configuration of firewalls. Up to 64 GB (was 4GB) compressed references. Per thread allocation profiling in the Management Console. Native Memory Tracking – it is now possible to track native memory allocations with very high resolution. The information can either be accessed using JRCMD, or the dedicated Native Memory Tracking experimental plug-in for the Management Console (alas only available for the upcoming 4.0.1 release). JRockit can now produce heap dumps in HPROF format. Cooperative suspension – JRockit is no longer using system signals for stopping threads, which could lead to hangs if signals were lost or blocked (for example bad NFS shares). Now threads check periodically to see if they are suspended. VPAT/Section 508 compliant JRMC – greatly improved keyboard navigation and screen reader support. See New and Noteworthy for more information. JRockit Mission Control 4.0.0 can be downloaded from here: http://www.oracle.com/technology/software/products/jrockit/index.html <shameless ad> There is even a book to go with JRMC 4.0.0/JRockit R28! http://www.packtpub.com/oracle-jrockit-the-definitive-guide/book/ </shameless ad>

    Read the article

  • The Top Ten Security Top Ten Lists

    - by Troy Kitch
    As a marketer, we're always putting together the top 3, or 5 best, or an assortment of top ten lists. So instead of going that route, I've put together my top ten security top ten lists. These are not only for security practitioners, but also for the average Joe/Jane; because who isn't concerned about security these days? Now, there might not be ten for each one of these lists, but the title works best that way. Starting with my number ten (in no particular order): 10. Top 10 Most Influential Security-Related Movies Amrit Williams pulls together a great collection of security-related movies. He asks for comments on which one made you want to get into the business. I would have to say that my most influential movie(s), that made me want to get into the business of "stopping the bad guys" would have to be the James Bond series. I grew up on James Bond movies: thwarting the bad guy and saving the world. I recall being both ecstatic and worried when Silicon Valley-themed "A View to A Kill" hit theaters: "An investigation of a horse-racing scam leads 007 to a mad industrialist who plans to create a worldwide microchip monopoly by destroying California's Silicon Valley." Yikes! 9. Top Ten Security Careers From movies that got you into the career, here’s a top 10 list of security-related careers. It starts with number then, Information Security Analyst and ends with number one, Malware Analyst. They point out the significant growth in security careers and indicate that "according to the Bureau of Labor Statistics, the field is expected to experience growth rates of 22% between 2010-2020. If you are interested in getting into the field, Oracle has many great opportunities all around the world.  8. Top 125 Network Security Tools A bit outside of the range of 10, the top 125 Network Security Tools is an important list because it includes a prioritized list of key security tools practitioners are using in the hacking community, regardless of whether they are vendor supplied or open source. The exhaustive list provides ratings, reviews, searching, and sorting. 7. Top 10 Security Practices I have to give a shout out to my alma mater, Cal Poly, SLO: Go Mustangs! They have compiled their list of top 10 practices for students and faculty to follow. Educational institutions are a common target of web based attacks and miscellaneous errors according to the 2014 Verizon Data Breach Investigations Report.    6. (ISC)2 Top 10 Safe and Secure Online Tips for Parents This list is arguably the most important list on my list. The tips were "gathered from (ISC)2 member volunteers who participate in the organization’s Safe and Secure Online program, a worldwide initiative that brings top cyber security experts into schools to teach children ages 11-14 how to protect themselves in a cyber-connected world…If you are a parent, educator or organization that would like the Safe and Secure Online presentation delivered at your local school, or would like more information about the program, please visit here.” 5. Top Ten Data Breaches of the Past 12 Months This type of list is always changing, so it's nice to have a current one here from Techrader.com. They've compiled and commented on the top breaches. It is likely that most readers here were effected in some way or another. 4. Top Ten Security Comic Books Although mostly physical security controls, I threw this one in for fun. My vote for #1 (not on the list) would be Professor X. The guy can breach confidentiality, integrity, and availability just by messing with your thoughts. 3. The IOUG Data Security Survey's Top 10+ Threats to Organizations The Independent Oracle Users Group annual survey on enterprise data security, Leaders Vs. Laggards, highlights what Oracle Database users deem as the top 12 threats to their organization. You can find a nice graph on page 9; Figure 7: Greatest Threats to Data Security. 2. The Ten Most Common Database Security Vulnerabilities Though I don't necessarily agree with all of the vulnerabilities in this order...I like a list that focuses on where two-thirds of your sensitive and regulated data resides (Source: IDC).  1. OWASP Top Ten Project The Online Web Application Security Project puts together their annual list of the 10 most critical web application security risks that organizations should be including in their overall security, business risk and compliance plans. In particular, SQL injection risks continues to rear its ugly head each year. Oracle Audit Vault and Database Firewall can help prevent SQL injection attacks and monitor database and system activity as a detective security control. Did I miss any?

    Read the article

  • Fix: SqlDeploy Task Fails with NullReferenceException at ExtractPassword

    Still working on getting a TeamCity build working (see my last post).  Latest exception is: C:\Program Files\MSBuild\Microsoft\VisualStudio\v9.0\TeamData\Microsoft.Data.Schema.SqlTasks.targets(120, 5): error MSB4018: The "SqlDeployTask" task failed unexpectedly. System.NullReferenceException: Object reference not set to an instance of an object. at Microsoft.Data.Schema.Common.ConnectionStringPersistence.ExtractPassword(String partialConnection, String dbProvider) at Microsoft.Data.Schema.Common.ConnectionStringPersistence.RetrieveFullConnection(String partialConnection, String provider, Boolean presentUI, String password) at Microsoft.Data.Schema.Sql.Build.SqlDeployment.ConfigureConnectionString(String connectionString, String databaseName) at Microsoft.Data.Schema.Sql.Build.SqlDeployment.OnBuildConnectionString(String partialConnectionString, String databaseName) at Microsoft.Data.Schema.Build.Deployment.FinishInitialize(String targetConnectionString) at Microsoft.Data.Schema.Build.Deployment.Initialize(FileInfo sourceDbSchemaFile, ErrorManager errors, String targetConnectionString) at Microsoft.Data.Schema.Build.DeploymentConstructor.ConstructServiceImplementation() at Microsoft.Data.Schema.Extensibility.ServiceConstructor'1.ConstructService() at Microsoft.Data.Schema.Tasks.DBDeployTask.Execute() at Microsoft.Build.BuildEngine.TaskEngine.ExecuteInstantiatedTask(EngineProxy engineProxy, ItemBucket bucket, TaskExecutionMode howToExecuteTask, ITask task, Boolean& taskResult)   This time searching yielded some good stuff, including this thread that talks about how to resolve this via permissions.  The short answer is that the account that your build server runs under needs to have the necessary permissions in SQL Server.  Youll need to create a Login and then ensure at least the minimum rights are configured as described here: Required Permissions in Database Edition Alternately, you can just make your buildserver account an admin on the database (which is probably running on the same machine anyway) and at that point it should be able to do whatever it needs to. If youre certain the account has the necessary permissions, but youre still getting the error, the problem may be that the account has never logged into the build server.  In this case, there wont be any entry in the HKCU hive in the registry, which the system is checking for permissions (see this thread).  The solution in this case is quite simple: log into the machine (once is enough) with the build server account.  Then, open Visual Studio (thanks Brendan for the answer in this thread). Summary Make sure the build service account has the necessary database permissions Make sure the account has logged into the server so it has the necessary registry hive info Make sure the account has run Visual Studio at least once so its settings are established In my case I went through all 3 of these steps before I resolved the problem. Did you know that DotNetSlackers also publishes .net articles written by top known .net Authors? We already have over 80 articles in several categories including Silverlight. Take a look: here.

    Read the article

  • SQL SERVER – CXPACKET – Parallelism – Usual Solution – Wait Type – Day 6 of 28

    - by pinaldave
    CXPACKET has to be most popular one of all wait stats. I have commonly seen this wait stat as one of the top 5 wait stats in most of the systems with more than one CPU. Books On-Line: Occurs when trying to synchronize the query processor exchange iterator. You may consider lowering the degree of parallelism if contention on this wait type becomes a problem. CXPACKET Explanation: When a parallel operation is created for SQL Query, there are multiple threads for a single query. Each query deals with a different set of the data (or rows). Due to some reasons, one or more of the threads lag behind, creating the CXPACKET Wait Stat. There is an organizer/coordinator thread (thread 0), which takes waits for all the threads to complete and gathers result together to present on the client’s side. The organizer thread has to wait for the all the threads to finish before it can move ahead. The Wait by this organizer thread for slow threads to complete is called CXPACKET wait. Note that not all the CXPACKET wait types are bad. You might experience a case when it totally makes sense. There might also be cases when this is unavoidable. If you remove this particular wait type for any query, then that query may run slower because the parallel operations are disabled for the query. Reducing CXPACKET wait: We cannot discuss about reducing the CXPACKET wait without talking about the server workload type. OLTP: On Pure OLTP system, where the transactions are smaller and queries are not long but very quick usually, set the “Maximum Degree of Parallelism” to 1 (one). This way it makes sure that the query never goes for parallelism and does not incur more engine overhead. EXEC sys.sp_configure N'cost threshold for parallelism', N'1' GO RECONFIGURE WITH OVERRIDE GO Data-warehousing / Reporting server: As queries will be running for long time, it is advised to set the “Maximum Degree of Parallelism” to 0 (zero). This way most of the queries will utilize the parallel processor, and long running queries get a boost in their performance due to multiple processors. EXEC sys.sp_configure N'cost threshold for parallelism', N'0' GO RECONFIGURE WITH OVERRIDE GO Mixed System (OLTP & OLAP): Here is the challenge. The right balance has to be found. I have taken a very simple approach. I set the “Maximum Degree of Parallelism” to 2, which means the query still uses parallelism but only on 2 CPUs. However, I keep the “Cost Threshold for Parallelism” very high. This way, not all the queries will qualify for parallelism but only the query with higher cost will go for parallelism. I have found this to work best for a system that has OLTP queries and also where the reporting server is set up. Here, I am setting ‘Cost Threshold for Parallelism’ to 25 values (which is just for illustration); you can choose any value, and you can find it out by experimenting with the system only. In the following script, I am setting the ‘Max Degree of Parallelism’ to 2, which indicates that the query that will have a higher cost (here, more than 25) will qualify for parallel query to run on 2 CPUs. This implies that regardless of the number of CPUs, the query will select any two CPUs to execute itself. EXEC sys.sp_configure N'cost threshold for parallelism', N'25' GO EXEC sys.sp_configure N'max degree of parallelism', N'2' GO RECONFIGURE WITH OVERRIDE GO Read all the post in the Wait Types and Queue series. Additionally a must read comment of Jonathan Kehayias. Note: The information presented here is from my experience and I no way claim it to be accurate. I suggest you all to read the online book for further clarification. All the discussion of Wait Stats over here is generic and it varies from system to system. It is recommended that you test this on the development server before implementing on the production server. Reference: Pinal Dave (http://blog.SQLAuthority.com) Filed under: DMV, Pinal Dave, PostADay, SQL, SQL Authority, SQL Query, SQL Scripts, SQL Server, SQL Tips and Tricks, SQL Wait Stats, SQL Wait Types, T SQL, Technology

    Read the article

  • Android -> Ruby Server Interface -> Mongodb

    - by MRabRabbit
    I've been wrecking my brain about this for a few days. I'll run my scenario by you and hopefully you can help me. In my head this is how it goes: I have an Android App. I want my Android App to make (function) calls to a MongoDB database via a Ruby Interface on the Server. e.g. Android app sends a HTTP GET ? with the function name, let's say getFriends for this user Ruby Interface receives this request from the app, grabs a thread from a thread pool and calls the appropriate function call implemented in Ruby, to the Mongodb. Ruby Interface gets results from Mongodb and sends a HTML POST to the Android app. So that's how I think it works. I know about the ruby driver for mongo db, and interacting with the mongodb from ruby but, how do I make a ruby back end listen for incoming messages and should these messages be done through sockets or a http interface ala Net::http in ruby?

    Read the article

  • HPET for x86 BSP (how to build it for WCE8)

    - by Werner Willemsens
    Originally posted on: http://geekswithblogs.net/WernerWillemsens/archive/2014/08/02/157895.aspx"I needed a timer". That is how we started a few blogs ago our series about APIC and ACPI. Well, here it is. HPET (High Precision Event Timer) was introduced by Intel in early 2000 to: Replace old style Intel 8253 (1981!) and 8254 timers Support more accurate timers that could be used for multimedia purposes. Hence Microsoft and Intel sometimes refers to HPET as Multimedia timers. An HPET chip consists of a 64-bit up-counter (main counter) counting at a frequency of at least 10 MHz, and a set of (at least three, up to 256) comparators. These comparators are 32- or 64-bit wide. The HPET is discoverable via ACPI. The HPET circuit in recent Intel platforms is integrated into the SouthBridge chip (e.g. 82801) All HPET timers should support one-shot interrupt programming, while optionally they can support periodic interrupts. In most Intel SouthBridges I worked with, there are three HPET timers. TIMER0 supports both one-shot and periodic mode, while TIMER1 and TIMER2 are one-shot only. Each HPET timer can generate interrupts, both in old-style PIC mode and in APIC mode. However in PIC mode, interrupts cannot freely be chosen. Typically IRQ11 is available and cannot be shared with any other interrupt! Which makes the HPET in PIC mode virtually unusable. In APIC mode however more IRQs are available and can be shared with other interrupt generating devices. (Check the datasheet of your SouthBridge) Because of this higher level of freedom, I created the APIC BSP (see previous posts). The HPET driver code that I present you here uses this APIC mode. Hpet.reg [HKEY_LOCAL_MACHINE\Drivers\BuiltIn\Hpet] "Dll"="Hpet.dll" "Prefix"="HPT" "Order"=dword:10 "IsrDll"="giisr.dll" "IsrHandler"="ISRHandler" "Priority256"=dword:50 Because HPET does not reside on the PCI bus, but can be found through ACPI as a memory mapped device, you don't need to specify the "Class", "SubClass", "ProgIF" and other PCI related registry keys that you typically find for PCI devices. If a driver needs to run its internal thread(s) at a certain priority level, by convention in Windows CE you add the "Priority256" registry key. Through this key you can easily play with the driver's thread priority for better response and timer accuracy. See later. Hpet.cpp (Hpet.dll) This cpp file contains the complete HPET driver code. The file is part of a folder that you typically integrate in your BSP (\src\drivers\Hpet). It is written as sample (example) code, you most likely want to change this code to your specific needs. There are two sets of #define's that I use to control how the driver works. _TRIGGER_EVENT or _TRIGGER_SEMAPHORE: _TRIGGER_EVENT will let your driver trigger a Windows CE Event when the timer expires, _TRIGGER_SEMAPHORE will trigger a Windows CE counting Semaphore. The latter guarantees that no events get lost in case your application cannot always process the triggers fast enough. _TIMER0 or _TIMER2: both timers will trigger an event or semaphore periodically. _TIMER0 will use a periodic HPET timer interrupt, while _TIMER2 will reprogram a one-shot HPET timer after each interrupt. The one-shot approach is interesting if the frequency you wish to generate is not an even multiple of the HPET main counter frequency. The sample code uses an algorithm to generate a more correct frequency over a longer period (by reducing rounding errors). _TIMER1 is not used in the sample source code. HPT_Init() will locate the HPET I/O memory space, setup the HPET counter (_TIMER0 or _TIMER2) and install the Interrupt Service Thread (IST). Upon timer expiration, the IST will run and on its turn will generate a Windows CE Event or Semaphore. In case of _TIMER2 a new one-shot comparator value is calculated and set for the timer. The IRQ of the HPET timers are programmed to IRQ22, but you can choose typically from 20-23. The TIMERn_INT_ROUT_CAP bits in the TIMn_CONF register will tell you what IRQs you can choose from. HPT_IOControl() can be used to set a new HPET counter frequency (actually you configure the counter timeout value in microseconds), start and stop the timer, and request the current HPET counter value. The latter is interesting because the Windows CE QueryPerformanceCounter() and QueryPerformanceFrequency() APIs implement the same functionality, albeit based on other counter implementations. HpetDrvIst() contains the IST code. DWORD WINAPI HpetDrvIst(LPVOID lpArg) { psHpetDeviceContext pHwContext = (psHpetDeviceContext)lpArg; DWORD mainCount = READDWORD(pHwContext->g_hpet_va, GenCapIDReg + 4); // Main Counter Tick period (fempto sec 10E-15) DWORD i = 0; while (1) { WaitForSingleObject(pHwContext->g_isrEvent, INFINITE); #if defined(_TRIGGER_SEMAPHORE) LONG p = 0; BOOL b = ReleaseSemaphore(pHwContext->g_triggerEvent, 1, &p); #elif defined(_TRIGGER_EVENT) BOOL b = SetEvent(pHwContext->g_triggerEvent); #else #pragma error("Unknown TRIGGER") #endif #if defined(_TIMER0) DWORD currentCount = READDWORD(pHwContext->g_hpet_va, MainCounterReg); DWORD comparator = READDWORD(pHwContext->g_hpet_va, Tim0_ComparatorReg + 0); SETBIT(pHwContext->g_hpet_va, GenIntStaReg, 0); // clear interrupt on HPET level InterruptDone(pHwContext->g_sysIntr); // clear interrupt on OS level _LOGMSG(ZONE_INTERRUPT, (L"%s: HpetDrvIst 0 %06d %08X %08X", pHwContext->g_id, i++, currentCount, comparator)); #elif defined(_TIMER2) DWORD currentCount = READDWORD(pHwContext->g_hpet_va, MainCounterReg); DWORD previousComparator = READDWORD(pHwContext->g_hpet_va, Tim2_ComparatorReg + 0); pHwContext->g_counter2.QuadPart += pHwContext->g_comparator.QuadPart; // increment virtual counter (higher accuracy) DWORD comparator = (DWORD)(pHwContext->g_counter2.QuadPart >> 8); // "round" to real value WRITEDWORD(pHwContext->g_hpet_va, Tim2_ComparatorReg + 0, comparator); SETBIT(pHwContext->g_hpet_va, GenIntStaReg, 2); // clear interrupt on HPET level InterruptDone(pHwContext->g_sysIntr); // clear interrupt on OS level _LOGMSG(ZONE_INTERRUPT, (L"%s: HpetDrvIst 2 %06d %08X %08X (%08X)", pHwContext->g_id, i++, currentCount, comparator, comparator - previousComparator)); #else #pragma error("Unknown TIMER") #endif } return 1; } The following figure shows how the HPET hardware interrupt via ISR -> IST is translated in a Windows CE Event or Semaphore by the HPET driver. The Event or Semaphore can be used to trigger a Windows CE application. HpetTest.cpp (HpetTest.exe)This cpp file contains sample source how to use the HPET driver from an application. The file is part of a separate (smart device) VS2013 solution. It contains code to measure the generated Event/Semaphore times by means of GetSystemTime() and QueryPerformanceCounter() and QueryPerformanceFrequency() APIs. HPET evaluation If you scan the internet about HPET, you'll find many remarks about buggy HPET implementations and bad performance. Unfortunately that is true. I tested the HPET driver on an Intel ICH7M SBC (release date 2008). When a HPET timer expires on the ICH7M, an interrupt indeed is generated, but right after you clear the interrupt, a few more unwanted interrupts (too soon!) occur as well. I tested and debugged it for a loooong time, but I couldn't get it to work. I concluded ICH7M's HPET is buggy Intel hardware. I tested the HPET driver successfully on a more recent NM10 SBC (release date 2013). With the NM10 chipset however, I am not fully convinced about the timer's frequency accuracy. In the long run - on average - all is fine, but occasionally I experienced upto 20 microseconds delays (which were immediately compensated on the next interrupt). Of course, this was all measured by software, but I still experienced the occasional delay when both the HPET driver IST thread as the application thread ran at CeSetThreadPriority(1). If it is not the hardware, only the kernel can cause this delay. But Windows CE is an RTOS and I have never experienced such long delays with previous versions of Windows CE. I tested and developed this on WCE8, I am not heavily experienced with it yet. Internet forum threads however mention inaccurate HPET timer implementations as well. At this moment I haven't figured out what is going on here. Useful references: http://www.intel.com/content/dam/www/public/us/en/documents/technical-specifications/software-developers-hpet-spec-1-0a.pdf http://en.wikipedia.org/wiki/High_Precision_Event_Timer http://wiki.osdev.org/HPET Windows CE BSP source file package for HPET in MyBsp Note that this source code is "As Is". It is still under development and I cannot (and never will) guarantee the correctness of the code. Use it as a guide for your own HPET integration.

    Read the article

  • Error in code of basic game using multiple sprites and surfaceView [on hold]

    - by Khagendra Nath Mahato
    I am a beginner to android and i was trying to make a basic game with the help of an online video tutorial. I am having problem with the multi-sprites and how to use with surfaceview.The application fails launching. Here is the code of the game.please help me. package com.example.killthemall; import java.util.ArrayList; import java.util.List; import java.util.Random; import android.app.Activity; import android.content.Context; import android.graphics.Bitmap; import android.graphics.BitmapFactory; import android.graphics.Canvas; import android.graphics.Rect; import android.os.Bundle; import android.view.SurfaceHolder; import android.view.SurfaceView; import android.widget.Toast; public class Game extends Activity { KhogenView View1; @Override protected void onPause() { // TODO Auto-generated method stub super.onPause(); while(true){ try { OurThread.join(); } catch (InterruptedException e) { // TODO Auto-generated catch block e.printStackTrace(); }} } Thread OurThread; int herorows = 4; int herocolumns = 3; int xpos, ypos; int xspeed; int yspeed; int herowidth; int widthnumber = 0; int heroheight; Rect src; Rect dst; int round; Bitmap bmp1; // private Bitmap bmp1;//change name public List<Sprite> sprites = new ArrayList<Sprite>() { }; @Override protected void onCreate(Bundle savedInstanceState) { // TODO Auto-generated method stub super.onCreate(savedInstanceState); View1 = new KhogenView(this); setContentView(View1); sprites.add(createSprite(R.drawable.image)); sprites.add(createSprite(R.drawable.bad1)); sprites.add(createSprite(R.drawable.bad2)); sprites.add(createSprite(R.drawable.bad3)); sprites.add(createSprite(R.drawable.bad4)); sprites.add(createSprite(R.drawable.bad5)); sprites.add(createSprite(R.drawable.bad6)); sprites.add(createSprite(R.drawable.good1)); sprites.add(createSprite(R.drawable.good2)); sprites.add(createSprite(R.drawable.good3)); sprites.add(createSprite(R.drawable.good4)); sprites.add(createSprite(R.drawable.good5)); sprites.add(createSprite(R.drawable.good6)); } private Sprite createSprite(int image) { // TODO Auto-generated method stub bmp1 = BitmapFactory.decodeResource(getResources(), image); return new Sprite(this, bmp1); } public class KhogenView extends SurfaceView implements Runnable { SurfaceHolder OurHolder; Canvas canvas = null; Random rnd = new Random(); { xpos = rnd.nextInt(canvas.getWidth() - herowidth)+herowidth; ypos = rnd.nextInt(canvas.getHeight() - heroheight)+heroheight; xspeed = rnd.nextInt(10 - 5) + 5; yspeed = rnd.nextInt(10 - 5) + 5; } public KhogenView(Context context) { super(context); // TODO Auto-generated constructor stub OurHolder = getHolder(); OurThread = new Thread(this); OurThread.start(); } @Override public void run() { // TODO Auto-generated method stub herowidth = bmp1.getWidth() / 3; heroheight = bmp1.getHeight() / 4; boolean isRunning = true; while (isRunning) { if (!OurHolder.getSurface().isValid()) continue; canvas = OurHolder.lockCanvas(); canvas.drawRGB(02, 02, 50); for (Sprite sprite : sprites) { if (widthnumber == 3) widthnumber = 0; update(); getdirection(); src = new Rect(widthnumber * herowidth, round * heroheight, (widthnumber + 1) * herowidth, (round + 1)* heroheight); dst = new Rect(xpos, ypos, xpos + herowidth, ypos+ heroheight); canvas.drawBitmap(bmp1, src, dst, null); } widthnumber++; OurHolder.unlockCanvasAndPost(canvas); } } public void update() { try { Thread.sleep(1000); } catch (InterruptedException e) { // TODO Auto-generated catch block e.printStackTrace(); } if (xpos + xspeed <= 0) xspeed = 40; if (xpos >= canvas.getWidth() - herowidth) xspeed = -50; if (ypos + yspeed <= 0) yspeed = 45; if (ypos >= canvas.getHeight() - heroheight) yspeed = -55; xpos = xpos + xspeed; ypos = ypos + yspeed; } public void getdirection() { double angleinteger = (Math.atan2(yspeed, xspeed)) / (Math.PI / 2); round = (int) (Math.round(angleinteger) + 2) % herorows; // Toast.makeText(this, String.valueOf(round), // Toast.LENGTH_LONG).show(); } } public class Sprite { Game game; private Bitmap bmp; public Sprite(Game game, Bitmap bmp) { // TODO Auto-generated constructor stub this.game = game; this.bmp = bmp; } } } Here is the LogCat if it helps.... 08-22 23:18:06.980: D/AndroidRuntime(28151): Shutting down VM 08-22 23:18:06.980: W/dalvikvm(28151): threadid=1: thread exiting with uncaught exception (group=0xb3f6f4f0) 08-22 23:18:06.980: D/AndroidRuntime(28151): procName from cmdline: com.example.killthemall 08-22 23:18:06.980: E/AndroidRuntime(28151): in writeCrashedAppName, pkgName :com.example.killthemall 08-22 23:18:06.980: D/AndroidRuntime(28151): file written successfully with content: com.example.killthemall StringBuffer : ;com.example.killthemall 08-22 23:18:06.990: I/Process(28151): Sending signal. PID: 28151 SIG: 9 08-22 23:18:06.990: E/AndroidRuntime(28151): FATAL EXCEPTION: main 08-22 23:18:06.990: E/AndroidRuntime(28151): java.lang.RuntimeException: Unable to start activity ComponentInfo{com.example.killthemall/com.example.killthemall.Game}: java.lang.NullPointerException 08-22 23:18:06.990: E/AndroidRuntime(28151): at android.app.ActivityThread.performLaunchActivity(ActivityThread.java:1647) 08-22 23:18:06.990: E/AndroidRuntime(28151): at android.app.ActivityThread.handleLaunchActivity(ActivityThread.java:1663) 08-22 23:18:06.990: E/AndroidRuntime(28151): at android.app.ActivityThread.access$1500(ActivityThread.java:117) 08-22 23:18:06.990: E/AndroidRuntime(28151): at android.app.ActivityThread$H.handleMessage(ActivityThread.java:931) 08-22 23:18:06.990: E/AndroidRuntime(28151): at android.os.Handler.dispatchMessage(Handler.java:99) 08-22 23:18:06.990: E/AndroidRuntime(28151): at android.os.Looper.loop(Looper.java:130) 08-22 23:18:06.990: E/AndroidRuntime(28151): at android.app.ActivityThread.main(ActivityThread.java:3683) 08-22 23:18:06.990: E/AndroidRuntime(28151): at java.lang.reflect.Method.invokeNative(Native Method) 08-22 23:18:06.990: E/AndroidRuntime(28151): at java.lang.reflect.Method.invoke(Method.java:507) 08-22 23:18:06.990: E/AndroidRuntime(28151): at com.android.internal.os.ZygoteInit$MethodAndArgsCaller.run(ZygoteInit.java:880) 08-22 23:18:06.990: E/AndroidRuntime(28151): at com.android.internal.os.ZygoteInit.main(ZygoteInit.java:638) 08-22 23:18:06.990: E/AndroidRuntime(28151): at dalvik.system.NativeStart.main(Native Method) 08-22 23:18:06.990: E/AndroidRuntime(28151): Caused by: java.lang.NullPointerException 08-22 23:18:06.990: E/AndroidRuntime(28151): at com.example.killthemall.Game$KhogenView.<init>(Game.java:96) 08-22 23:18:06.990: E/AndroidRuntime(28151): at com.example.killthemall.Game.onCreate(Game.java:58) 08-22 23:18:06.990: E/AndroidRuntime(28151): at android.app.Instrumentation.callActivityOnCreate(Instrumentation.java:1049) 08-22 23:18:06.990: E/AndroidRuntime(28151): at android.app.ActivityThread.performLaunchActivity(ActivityThread.java:1611) 08-22 23:18:06.990: E/AndroidRuntime(28151): ... 11 more 08-22 23:18:18.050: D/AndroidRuntime(28191): Shutting down VM 08-22 23:18:18.050: W/dalvikvm(28191): threadid=1: thread exiting with uncaught exception (group=0xb3f6f4f0) 08-22 23:18:18.050: I/Process(28191): Sending signal. PID: 28191 SIG: 9 08-22 23:18:18.050: D/AndroidRuntime(28191): procName from cmdline: com.example.killthemall 08-22 23:18:18.050: E/AndroidRuntime(28191): in writeCrashedAppName, pkgName :com.example.killthemall 08-22 23:18:18.050: D/AndroidRuntime(28191): file written successfully with content: com.example.killthemall StringBuffer : ;com.example.killthemall 08-22 23:18:18.050: E/AndroidRuntime(28191): FATAL EXCEPTION: main 08-22 23:18:18.050: E/AndroidRuntime(28191): java.lang.RuntimeException: Unable to start activity ComponentInfo{com.example.killthemall/com.example.killthemall.Game}: java.lang.NullPointerException 08-22 23:18:18.050: E/AndroidRuntime(28191): at android.app.ActivityThread.performLaunchActivity(ActivityThread.java:1647) 08-22 23:18:18.050: E/AndroidRuntime(28191): at android.app.ActivityThread.handleLaunchActivity(ActivityThread.java:1663) 08-22 23:18:18.050: E/AndroidRuntime(28191): at android.app.ActivityThread.access$1500(ActivityThread.java:117) 08-22 23:18:18.050: E/AndroidRuntime(28191): at android.app.ActivityThread$H.handleMessage(ActivityThread.java:931) 08-22 23:18:18.050: E/AndroidRuntime(28191): at android.os.Handler.dispatchMessage(Handler.java:99) 08-22 23:18:18.050: E/AndroidRuntime(28191): at android.os.Looper.loop(Looper.java:130) 08-22 23:18:18.050: E/AndroidRuntime(28191): at android.app.ActivityThread.main(ActivityThread.java:3683) 08-22 23:18:18.050: E/AndroidRuntime(28191): at java.lang.reflect.Method.invokeNative(Native Method) 08-22 23:18:18.050: E/AndroidRuntime(28191): at java.lang.reflect.Method.invoke(Method.java:507) 08-22 23:18:18.050: E/AndroidRuntime(28191): at com.android.internal.os.ZygoteInit$MethodAndArgsCaller.run(ZygoteInit.java:880) 08-22 23:18:18.050: E/AndroidRuntime(28191): at com.android.internal.os.ZygoteInit.main(ZygoteInit.java:638) 08-22 23:18:18.050: E/AndroidRuntime(28191): at dalvik.system.NativeStart.main(Native Method) 08-22 23:18:18.050: E/AndroidRuntime(28191): Caused by: java.lang.NullPointerException 08-22 23:18:18.050: E/AndroidRuntime(28191): at com.example.killthemall.Game$KhogenView.<init>(Game.java:96) 08-22 23:18:18.050: E/AndroidRuntime(28191): at com.example.killthemall.Game.onCreate(Game.java:58) 08-22 23:18:18.050: E/AndroidRuntime(28191): at android.app.Instrumentation.callActivityOnCreate(Instrumentation.java:1049) 08-22 23:18:18.050: E/AndroidRuntime(28191): at android.app.ActivityThread.performLaunchActivity(ActivityThread.java:1611) 08-22 23:18:18.050: E/AndroidRuntime(28191): ... 11 more

    Read the article

  • Why not Green Threads?

    - by redjamjar
    Whilst I know questions on this have been covered already (e.g. http://stackoverflow.com/questions/5713142/green-threads-vs-non-green-threads), I don't feel like I've got a satisfactory answer. The question is: why don't JVM's support green threads anymore? It says this on the code-style Java FAQ: A green thread refers to a mode of operation for the Java Virtual Machine (JVM) in which all code is executed in a single operating system thread. And this over on java.sun.com: The downside is that using green threads means system threads on Linux are not taken advantage of and so the Java virtual machine is not scalable when additional CPUs are added. It seems to me that the JVM could have a pool of system processes equal to the number of cores, and then run green threads on top of that. This could offer some big advantages when you have a very number large of threads which block often (mostly because current JVM's cap the number of threads). Thoughts?

    Read the article

  • Minecraft shows black screen on watt-os 64 after logon

    - by uffe hellum
    Minecraft appears to launch with oracle java 7, but crashes after logon. $ java -Xmx1024M -Xms512M -cp ./minecraft.jar net.minecraft.LauncherFrame asdf Exception in thread "Thread-3" java.lang.UnsatisfiedLinkError: /home/uffeh/.minecraft/bin/natives/liblwjgl.so: /home/uffeh/.minecraft/bin/natives/liblwjgl.so: wrong ELF class: ELFCLASS32 (Possible cause: architecture word width mismatch) at java.lang.ClassLoader$NativeLibrary.load(Native Method) at java.lang.ClassLoader.loadLibrary1(ClassLoader.java:1939) at java.lang.ClassLoader.loadLibrary0(ClassLoader.java:1864) at java.lang.ClassLoader.loadLibrary(ClassLoader.java:1825) at java.lang.Runtime.load0(Runtime.java:792) at java.lang.System.load(System.java:1059) at org.lwjgl.Sys$1.run(Sys.java:69) at java.security.AccessController.doPrivileged(Native Method) at org.lwjgl.Sys.doLoadLibrary(Sys.java:65) at org.lwjgl.Sys.loadLibrary(Sys.java:81) at org.lwjgl.Sys.(Sys.java:98) at net.minecraft.client.Minecraft.F(SourceFile:1857) at aof.(SourceFile:20) at net.minecraft.client.Minecraft.(SourceFile:77) at anw.(SourceFile:36) at net.minecraft.client.MinecraftApplet.init(SourceFile:36) at net.minecraft.Launcher.replace(Launcher.java:136) at net.minecraft.Launcher$1.run(Launcher.java:79)

    Read the article

  • Token based Authentication for WCF HTTP/REST Services: Authentication

    - by Your DisplayName here!
    This post shows some of the implementation techniques for adding token and claims based security to HTTP/REST services written with WCF. For the theoretical background, see my previous post. Disclaimer The framework I am using/building here is not the only possible approach to tackle the problem. Based on customer feedback and requirements the code has gone through several iterations to a point where we think it is ready to handle most of the situations. Goals and requirements The framework should be able to handle typical scenarios like username/password based authentication, as well as token based authentication The framework should allow adding new supported token types Should work with WCF web programming model either self-host or IIS hosted Service code can rely on an IClaimsPrincipal on Thread.CurrentPrincipal that describes the client using claims-based identity Implementation overview In WCF the main extensibility point for this kind of security work is the ServiceAuthorizationManager. It gets invoked early enough in the pipeline, has access to the HTTP protocol details of the incoming request and can set Thread.CurrentPrincipal. The job of the SAM is simple: Check the Authorization header of the incoming HTTP request Check if a “registered” token (more on that later) is present If yes, validate the token using a security token handler, create the claims principal (including claims transformation) and set Thread.CurrentPrincipal If no, set an anonymous principal on Thread.CurrentPrincipal. By default, anonymous principals are denied access – so the request ends here with a 401 (more on that later). To wire up the custom authorization manager you need a custom service host – which in turn needs a custom service host factory. The full object model looks like this: Token handling A nice piece of existing WIF infrastructure are security token handlers. Their job is to serialize a received security token into a CLR representation, validate the token and turn the token into claims. The way this works with WS-Security based services is that WIF passes the name/namespace of the incoming token to WIF’s security token handler collection. This in turn finds out which token handler can deal with the token and returns the right instances. For HTTP based services we can do something very similar. The scheme on the Authorization header gives the service a hint how to deal with an incoming token. So the only missing link is a way to associate a token handler (or multiple token handlers) with a scheme and we are (almost) done. WIF already includes token handler for a variety of tokens like username/password or SAML 1.1/2.0. The accompanying sample has a implementation for a Simple Web Token (SWT) token handler, and as soon as JSON Web Token are ready, simply adding a corresponding token handler will add support for this token type, too. All supported schemes/token types are organized in a WebSecurityTokenHandlerCollectionManager and passed into the host factory/host/authorization manager. Adding support for basic authentication against a membership provider would e.g. look like this (in global.asax): var manager = new WebSecurityTokenHandlerCollectionManager(); manager.AddBasicAuthenticationHandler((username, password) => Membership.ValidateUser(username, password));   Adding support for Simple Web Tokens with a scheme of Bearer (the current OAuth2 scheme) requires passing in a issuer, audience and signature verification key: manager.AddSimpleWebTokenHandler(     "Bearer",     "http://identityserver.thinktecture.com/trust/initial",     "https://roadie/webservicesecurity/rest/",     "WFD7i8XRHsrUPEdwSisdHoHy08W3lM16Bk6SCT8ht6A="); In some situations, SAML token may be used as well. The following configures SAML support for a token coming from ADFS2: var registry = new ConfigurationBasedIssuerNameRegistry(); registry.AddTrustedIssuer( "d1 c5 b1 25 97 d0 36 94 65 1c e2 64 fe 48 06 01 35 f7 bd db", "ADFS"); var adfsConfig = new SecurityTokenHandlerConfiguration(); adfsConfig.AudienceRestriction.AllowedAudienceUris.Add( new Uri("https://roadie/webservicesecurity/rest/")); adfsConfig.IssuerNameRegistry = registry; adfsConfig.CertificateValidator = X509CertificateValidator.None; // token decryption (read from config) adfsConfig.ServiceTokenResolver = IdentityModelConfiguration.ServiceConfiguration.CreateAggregateTokenResolver();             manager.AddSaml11SecurityTokenHandler("SAML", adfsConfig);   Transformation The custom authorization manager will also try to invoke a configured claims authentication manager. This means that the standard WIF claims transformation logic can be used here as well. And even better, can be also shared with e.g. a “surrounding” web application. Error handling A WCF error handler takes care of turning “access denied” faults into 401 status codes and a message inspector adds the registered authentication schemes to the outgoing WWW-Authenticate header when a 401 occurs. The next post will conclude with authorization as well as the source code download.   (Wanna learn more about federation, WIF, claims, tokens etc.? Click here.)

    Read the article

  • ASP.Net 4.5 Garbage Collection Improvement

    - by Aligned
    Originally posted on: http://geekswithblogs.net/Aligned/archive/2013/06/24/asp.net-4.5-garbage-collection-improvement.aspxI just read Five Great .NET Framework 4.5 Features on CodeProject by Shivprasad koirala. Feature 5 in his article mentions the GC background cleanup and has a good explanation of the work the GC has to do for ASP.Net on the server. “Garbage collector is one real heavy task in a .NET application. And it becomes heavier when it is an ASP.NET application. ASP.NET applications run on the server and a lot of clients send requests to the server thus creating loads of objects, making the GC really work hard for cleaning up unwanted objects.” “To overcome the above problem, server GC was introduced. In server GC there is one more thread created which runs in the background. This thread works in the background and keeps cleaning…objects thus minimizing the load on the main GC thread. Due to double GC threads running, the main application threads are less suspended, thus increasing application throughput. To enable server GC, we need to use the gcServer XML tag and enable it to true.” <configuration> <runtime> <gcServer enabled="true"/> </runtime> </configuration> This is not done by default. The MSDN information page says “There are only two garbage collection options, workstation or server. For single-processor computers, the default workstation garbage collection should be the fastest option. Either workstation or server can be used for two-processor computers. Server garbage collection should be the fastest option for more than two processors. Use the GCSettingsIsServerGC property to determine if server garbage collection is enabled.” “In the .NET Framework 4 and earlier versions, concurrent garbage collection is not available when server garbage collection is enabled. Starting with the .NET Framework 4.5, server garbage collection is concurrent. To use non-concurrent server garbage collection, set the <gcServer> element to true and the <gcConcurrent> element to false. “ So if you’re using ASP.Net 4.5 and have a multi-core server, you should try turning on the Server Garbage Collection and do some profiling to see if it improves the performance of your site.

    Read the article

  • Making your WCF Web Apis to speak in multiple languages

    - by cibrax
    One of the key aspects of how the web works today is content negotiation. The idea of content negotiation is based on the fact that a single resource can have multiple representations, so user agents (or clients) and servers can work together to chose one of them. The http specification defines several “Accept” headers that a client can use to negotiate content with a server, and among all those, there is one for restricting the set of natural languages that are preferred as a response to a request, “Accept-Language”. For example, a client can specify “es” in this header for specifying that he prefers to receive the content in spanish or “en” in english. However, there are certain scenarios where the “Accept-Language” header is just not enough, and you might want to have a way to pass the “accepted” language as part of the resource url as an extension. For example, http://localhost/ProductCatalog/Products/1.es” returns all the descriptions for the product with id “1” in spanish. This is useful for scenarios in which you want to embed the link somewhere, such a document, an email or a page.  Supporting both scenarios, the header and the url extension, is really simple in the new WCF programming model. You only need to provide a processor implementation for any of them. Let’s say I have a resource implementation as part of a product catalog I want to expose with the WCF web apis. [ServiceContract][Export]public class ProductResource{ IProductRepository repository;  [ImportingConstructor] public ProductResource(IProductRepository repository) { this.repository = repository; }  [WebGet(UriTemplate = "{id}")] public Product Get(string id, HttpResponseMessage response) { var product = repository.GetById(int.Parse(id)); if (product == null) { response.StatusCode = HttpStatusCode.NotFound; response.Content = new StringContent(Messages.OrderNotFound); }  return product; }} The Get method implementation in this resource assumes the desired culture will be attached to the current thread (Thread.CurrentThread.Culture). Another option is to pass the desired culture as an additional argument in the method, so my processor implementation will handle both options. This method is also using an auto-generated class for handling string resources, Messages, which is available in the different cultures that the service implementation supports. For example, Messages.resx contains “OrderNotFound”: “Order Not Found” Messages.es.resx contains “OrderNotFound”: “No se encontro orden” The processor implementation bellow tackles the first scenario, in which the desired language is passed as part of the “Accept-Language” header. public class CultureProcessor : Processor<HttpRequestMessage, CultureInfo>{ string defaultLanguage = null;  public CultureProcessor(string defaultLanguage = "en") { this.defaultLanguage = defaultLanguage; this.InArguments[0].Name = HttpPipelineFormatter.ArgumentHttpRequestMessage; this.OutArguments[0].Name = "culture"; }  public override ProcessorResult<CultureInfo> OnExecute(HttpRequestMessage request) { CultureInfo culture = null; if (request.Headers.AcceptLanguage.Count > 0) { var language = request.Headers.AcceptLanguage.First().Value; culture = new CultureInfo(language); } else { culture = new CultureInfo(defaultLanguage); }  Thread.CurrentThread.CurrentCulture = culture; Messages.Culture = culture;  return new ProcessorResult<CultureInfo> { Output = culture }; }}   As you can see, the processor initializes a new CultureInfo instance with the value provided in the “Accept-Language” header, and set that instance to the current thread and the auto-generated resource class with all the messages. In addition, the CultureInfo instance is returned as an output argument called “culture”, making possible to receive that argument in any method implementation   The following code shows the implementation of the processor for handling languages as url extensions.   public class CultureExtensionProcessor : Processor<HttpRequestMessage, Uri>{ public CultureExtensionProcessor() { this.OutArguments[0].Name = HttpPipelineFormatter.ArgumentUri; }  public override ProcessorResult<Uri> OnExecute(HttpRequestMessage httpRequestMessage) { var requestUri = httpRequestMessage.RequestUri.OriginalString;  var extensionPosition = requestUri.LastIndexOf(".");  if (extensionPosition > -1) { var extension = requestUri.Substring(extensionPosition + 1);  var query = httpRequestMessage.RequestUri.Query;  requestUri = string.Format("{0}?{1}", requestUri.Substring(0, extensionPosition), query); ;  var uri = new Uri(requestUri);  httpRequestMessage.Headers.AcceptLanguage.Clear();  httpRequestMessage.Headers.AcceptLanguage.Add(new StringWithQualityHeaderValue(extension));  var result = new ProcessorResult<Uri>();  result.Output = uri;  return result; }  return new ProcessorResult<Uri>(); }} The last step is to inject both processors as part of the service configuration as it is shown bellow, public void RegisterRequestProcessorsForOperation(HttpOperationDescription operation, IList<Processor> processors, MediaTypeProcessorMode mode){ processors.Insert(0, new CultureExtensionProcessor()); processors.Add(new CultureProcessor());} Once you configured the two processors in the pipeline, your service will start speaking different languages :). Note: Url extensions don’t seem to be working in the current bits when you are using Url extensions in a base address. As far as I could see, ASP.NET intercepts the request first and tries to route the request to a registered ASP.NET Http Handler with that extension. For example, “http://localhost/ProductCatalog/products.es” does not work, but “http://localhost/ProductCatalog/products/1.es” does.

    Read the article

  • Online Password Security Tactics

    - by BuckWoody
    Recently two more large databases were attacked and compromised, one at the popular Gawker Media sites and the other at McDonald’s. Every time this kind of thing happens (which is FAR too often) it should remind the technical professional to ensure that they secure their systems correctly. If you write software that stores passwords, it should be heavily encrypted, and not human-readable in any storage. I advocate a different store for the login and password, so that if one is compromised, the other is not. I also advocate that you set a bit flag when a user changes their password, and send out a reminder to change passwords if that bit isn’t changed every three or six months.    But this post is about the *other* side – what to do to secure your own passwords, especially those you use online, either in a cloud service or at a provider. While you’re not in control of these breaches, there are some things you can do to help protect yourself. Most of these are obvious, but they contain a few little twists that make the process easier.   Use Complex Passwords This is easily stated, and probably one of the most un-heeded piece of advice. There are three main concepts here: ·         Don’t use a dictionary-based word ·         Use mixed case ·         Use punctuation, special characters and so on   So this: password Isn’t nearly as safe as this: P@ssw03d   Of course, this only helps if the site that stores your password encrypts it. Gawker does, so theoretically if you had the second password you’re in better shape, at least, than the first. Dictionary words are quickly broken, regardless of the encryption, so the more unusual characters you use, and the farther away from the dictionary words you get, the better.   Of course, this doesn’t help, not even a little, if the site stores the passwords in clear text, or the key to their encryption is broken. In that case…   Use a Different Password at Every Site What? I have hundreds of sites! Are you kidding me? Nope – I’m not. If you use the same password at every site, when a site gets attacked, the attacker will store your name and password value for attacks at other sites. So the only safe thing to do is to use different names or passwords (or both) at each site. Of course, most sites use your e-mail as a username, so you’re kind of hosed there. So even though you have hundreds of sites you visit, you need to have at least a different password at each site.   But it’s easier than you think – if you use an algorithm.   What I’m describing is to pick a “root” password, and then modify that based on the site or purpose. That way, if the site is compromised, you can still use that root password for the other sites.   Let’s take that second password: P@ssw03d   And now you can append, prepend or intersperse that password with other characters to make it unique to the site. That way you can easily remember the root password, but make it unique to the site. For instance, perhaps you read a lot of information on Gawker – how about these:   P@ssw03dRead ReadP@ssw03d PR@esasdw03d   If you have lots of sites, tracking even this can be difficult, so I recommend you use password software such as Password Safe or some other tool to have a secure database of your passwords at each site. DO NOT store this on the web. DO NOT use an Office document (Microsoft or otherwise) that is “encrypted” – the encryption office automation packages use is very trivial, and easily broken. A quick web search for tools to do that should show you how bad a choice this is.   Change Your Password on a Schedule I know. It’s a real pain. And it doesn’t seem worth it…until your account gets hacked. A quick note here – whenever a site gets hacked (and I find out about it) I change the password at that site immediately (or quit doing business with them) and then change the root password on every site, as quickly as I can.   If you follow the tip above, it’s not as hard. Just add another number, year, month, day, something like that into the mix. It’s not unlike making a Primary Key in an RDBMS.   P@ssw03dRead10242010   Change the site, and then update your password database. I do this about once a month, on the first or last day, during staff meetings. (J)   If you have other tips, post them here. We can all learn from each other on this.

    Read the article

  • Voice echo in UDP based voice transmission [closed]

    - by Meherzad
    I have coded a java application for voice transmission between to ip in LAN. Here the code. public static Boolean flag= true; public static Boolean recFlag=true; DatagramSocket UDPSocket=null; AudioFormat format = null; TargetDataLine microphone=null; byte[] buffer=null; DatagramPacket UDPPacket=null; public void startChat(String ipAddress){ try{ buffer = new byte[1000]; UDPSocket=new DatagramSocket(1987); Thread th=new Thread(new Listener()); th.start(); microphone = AudioSystem.getTargetDataLine(format); format= new AudioFormat(8000.0f, 16, 1, true, true); UDPPacket = new DatagramPacket(buffer, buffer.length, InetAddress.getByName(ipAddress), 1988); microphone.open(format); microphone.start(); while (flag) { microphone.read(buffer, 0, buffer.length); UDPSocket.send(UDPPacket); } } catch(Exception e){ System.out.println(" ssss "+e.getMessage()); } } public class Listener extends Thread{ byte[] buff=new byte[1000]; DatagramSocket UDPSocket1=null; DatagramPacket recPacket=null; DataLine.Info info = new DataLine.Info(SourceDataLine.class, format); SourceDataLine line=null; @Override public void run(){ try{ UDPSocket1=new DatagramSocket(1988); format= new AudioFormat(8000.0f, 16, 1, true, true); line = (SourceDataLine) AudioSystem.getLine(info); line.open(format); line.start(); } catch(Exception e){ System.out.println("list "+ e.getMessage()); } recPacket=new DatagramPacket(buff, buff.length); while(recFlag){ try{ UDPSocket1.receive(recPacket); buff = (byte[])recPacket.getData(); line.write(buff, 0, buff.length); } catch(Exception e){ System.out.println("errr "+e.getMessage()); } } line.drain(); line.close(); } } Main problem which I am facing that I am getting only echo of my own voice. I am unable to hear voice from the other end only I am hearing is my own voice. Please suggest any solution.

    Read the article

  • 12.04 Wacom Volito tablet pressed without touching tablet surface

    - by Nox Deleo
    My tablet has worked okay in the past with older Ubuntu versions, and I think even with 12.04 a while back. Perhaps something has changed with the driver. Anyway my problem is the same as listed in this Ubuntu Forums thread, only the fixes in there don't seem to work for me. xsetwacom changes the value, but seems to have no effect on the tablet, and there is no entry in dconf-editor. Does anyone have any ideas? The tablet is basically unusable to me at the moment. It's an original Volito rather than the Volito 2 listed in the thread, but the problem is identical.

    Read the article

  • WPF: Timers

    - by Ilya Verbitskiy
    I believe, once your WPF application will need to execute something periodically, and today I would like to discuss how to do that. There are two possible solutions. You can use classical System.Threading.Timer class or System.Windows.Threading.DispatcherTimer class, which is the part of WPF. I have created an application to show you how to use the API.     Let’s take a look how you can implement timer using System.Threading.Timer class. First of all, it has to be initialized.   1: private Timer timer; 2:   3: public MainWindow() 4: { 5: // Form initialization code 6: 7: timer = new Timer(OnTimer, null, Timeout.InfiniteTimeSpan, Timeout.InfiniteTimeSpan); 8: }   Timer’s constructor accepts four parameters. The first one is the callback method which is executed when timer ticks. I will show it to you soon. The second parameter is a state which is passed to the callback. It is null because there is nothing to pass this time. The third parameter is the amount of time to delay before the callback parameter invokes its methods. I use System.Threading.Timeout helper class to represent infinite timeout which simply means the timer is not going to start at the moment. And the final fourth parameter represents the time interval between invocations of the methods referenced by callback. Infinite timeout timespan means the callback method will be executed just once. Well, the timer has been created. Let’s take a look how you can start the timer.   1: private void StartTimer(object sender, RoutedEventArgs e) 2: { 3: timer.Change(TimeSpan.Zero, new TimeSpan(0, 0, 1)); 4:   5: // Disable the start buttons and enable the reset button. 6: }   The timer is started by calling its Change method. It accepts two arguments: the amount of time to delay before the invoking the callback method and the time interval between invocations of the callback. TimeSpan.Zero means we start the timer immediately and TimeSpan(0, 0, 1) tells the timer to tick every second. There is one method hasn’t been shown yet. This is the callback method OnTimer which does a simple task: it shows current time in the center of the screen. Unfortunately you cannot simple write something like this:   1: clock.Content = DateTime.Now.ToString("hh:mm:ss");   The reason is Timer runs callback method on a separate thread, and it is not possible to access GUI controls from a non-GUI thread. You can avoid the problem using System.Windows.Threading.Dispatcher class.   1: private void OnTimer(object state) 2: { 3: Dispatcher.Invoke(() => ShowTime()); 4: } 5:   6: private void ShowTime() 7: { 8: clock.Content = DateTime.Now.ToString("hh:mm:ss"); 9: }   You can build similar application using System.Windows.Threading.DispatcherTimer class. The class represents a timer which is integrated into the Dispatcher queue. It means that your callback method is executed on GUI thread and you can write a code which updates your GUI components directly.   1: private DispatcherTimer dispatcherTimer; 2:   3: public MainWindow() 4: { 5: // Form initialization code 6:   7: dispatcherTimer = new DispatcherTimer { Interval = new TimeSpan(0, 0, 1) }; 8: dispatcherTimer.Tick += OnDispatcherTimer; 9: } Dispatcher timer has nicer and cleaner API. All you need is to specify tick interval and Tick event handler. The you just call Start method to start the timer.   private void StartDispatcher(object sender, RoutedEventArgs e) { dispatcherTimer.Start(); // Disable the start buttons and enable the reset button. } And, since the Tick event handler is executed on GUI thread, the code which sets the actual time is straightforward.   1: private void OnDispatcherTimer(object sender, EventArgs e) 2: { 3: ShowTime(); 4: } We’re almost done. Let’s take a look how to stop the timers. It is easy with the Dispatcher Timer.   1: dispatcherTimer.Stop(); And slightly more complicated with the Timer. You should use Change method again.   1: timer.Change(Timeout.InfiniteTimeSpan, Timeout.InfiniteTimeSpan); What is the best way to add timer into an application? The Dispatcher Timer has simple interface, but its advantages are disadvantages at the same time. You should not use it if your Tick event handler executes time-consuming operations. It freezes your window which it is executing the event handler method. You should think about using System.Threading.Timer in this case. The code is available on GitHub.

    Read the article

  • Dual Monitor results in 'greyed' windows

    - by paula
    This occurs in Maverick and Natty. Single screen is fine, mirror of single screen is fine. If the mirror box is unchecked and the second monitor is turned on to extend the desktop then all windows are greyed out (like they do when a process has timed out and is unresponsive) and pop up menus are greyed out but icons, panels and background are fine and the windows do operate (just can't see them well enough to use) I have a D620 with intel graphics. This machine did work with dual monitors at some time in the past, however I have been using another machine, a D630 with nvidia and it works fine. Yes, there have been any number of updates. I also upgraded from Maverick to Natty to see if it would go away. No joy. Also, the D620 has a dual boot windows system and the windows xp system works fine with daul monitors There is a forum thread that goes into more detail and there are a number of users experiencing this problem. Thread: greyed out windows Thanks for reading paula_ke

    Read the article

  • Using idle time in turn-based (RPG) games for updating

    - by The Communist Duck
    If you take any turn based RPG game there will be large periods of time when nothing is happening because the game is looping over 'wait_for_player_input'. Naturally it seems sensible to use this time to update things. However, this immediately seems to suggest that it would need to be threaded. Is this sort of design possible in a single thread? loop: if not check_something_pressed: update_a_very_small_amount else keep going But if we says 'a_very_small_amount' is only updating a single object each loop, it's going to be very slow at updating. How would you go about this, preferably in a single thread? EDIT: I've tagged this language-agnostic as that seems the sensible thing, though anything more specific to Python would be great. ;-)

    Read the article

  • Reusing WCF Proxy to reduce Memory Usage

    - by Sudheer Kumar
    I am working on a program that uploads BLOB from DB to a Document Management System. I have a WCF service to interact with the DMS. I have a multi-threaded client program that uploads the BLOBs to DMS and every thread used to create and dispose a proxy instance for every record to update. When I have a large no of records to convert, I found that the tool’s memory foot print keeps increasing. After a little debugging I found that the WCF proxies are the culprits for excessive memory usage. I changed the program to re-use the proxies to the service, having one proxy per thread. So in some scenarios, it might be beneficial to re-use WCF proxies.

    Read the article

< Previous Page | 121 122 123 124 125 126 127 128 129 130 131 132  | Next Page >