Search Results

Search found 1261 results on 51 pages for 'trivial'.

Page 46/51 | < Previous Page | 42 43 44 45 46 47 48 49 50 51  | Next Page >

• SQL Monitor’s data repository

  - by Chris Lambrou

  As one of the developers of SQL Monitor, I often get requests passed on by our support people from customers who are looking to dip into SQL Monitor’s own data repository, in order to pull out bits of information that they’re interested in. Since there’s clearly interest out there in playing around directly with the data repository, I thought I’d write some blog posts to start to describe how it all works. The hardest part for me is knowing where to begin, since the schema of the data repository is pretty big. Hmmm… I guess it’s tricky for anyone to write anything but the most trivial of queries against the data repository without understanding the hierarchy of monitored objects, so perhaps my first post should start there. I always imagine that whenever a customer fires up SSMS and starts to explore their SQL Monitor data repository database, they become immediately bewildered by the schema – that was certainly my experience when I did so for the first time. The following query shows the number of different object types in the data repository schema: SELECT type_desc, COUNT(*) AS [count] FROM sys.objects GROUP BY type_desc ORDER BY type_desc;  type_desccount 1DEFAULT_CONSTRAINT63 2FOREIGN_KEY_CONSTRAINT181 3INTERNAL_TABLE3 4PRIMARY_KEY_CONSTRAINT190 5SERVICE_QUEUE3 6SQL_INLINE_TABLE_VALUED_FUNCTION381 7SQL_SCALAR_FUNCTION2 8SQL_STORED_PROCEDURE100 9SYSTEM_TABLE41 10UNIQUE_CONSTRAINT54 11USER_TABLE193 12VIEW124 With 193 tables, 124 views, 100 stored procedures and 381 table valued functions, that’s quite a hefty schema, and when you browse through it using SSMS, it can be a bit daunting at first. So, where to begin? Well, let’s narrow things down a bit and only look at the tables belonging to the data schema. That’s where all of the collected monitoring data is stored by SQL Monitor. The following query gives us the names of those tables: SELECT sch.name + '.' + obj.name AS [name] FROM sys.objects obj JOIN sys.schemas sch ON sch.schema_id = obj.schema_id WHERE obj.type_desc = 'USER_TABLE' AND sch.name = 'data' ORDER BY sch.name, obj.name; This query still returns 110 tables. I won’t show them all here, but let’s have a look at the first few of them:  name 1data.Cluster_Keys 2data.Cluster_Machine_ClockSkew_UnstableSamples 3data.Cluster_Machine_Cluster_StableSamples 4data.Cluster_Machine_Keys 5data.Cluster_Machine_LogicalDisk_Capacity_StableSamples 6data.Cluster_Machine_LogicalDisk_Keys 7data.Cluster_Machine_LogicalDisk_Sightings 8data.Cluster_Machine_LogicalDisk_UnstableSamples 9data.Cluster_Machine_LogicalDisk_Volume_StableSamples 10data.Cluster_Machine_Memory_Capacity_StableSamples 11data.Cluster_Machine_Memory_UnstableSamples 12data.Cluster_Machine_Network_Capacity_StableSamples 13data.Cluster_Machine_Network_Keys 14data.Cluster_Machine_Network_Sightings 15data.Cluster_Machine_Network_UnstableSamples 16data.Cluster_Machine_OperatingSystem_StableSamples 17data.Cluster_Machine_Ping_UnstableSamples 18data.Cluster_Machine_Process_Instances 19data.Cluster_Machine_Process_Keys 20data.Cluster_Machine_Process_Owner_Instances 21data.Cluster_Machine_Process_Sightings 22data.Cluster_Machine_Process_UnstableSamples 23… There are two things I want to draw your attention to: The table names describe a hierarchy of the different types of object that are monitored by SQL Monitor (e.g. clusters, machines and disks). For each object type in the hierarchy, there are multiple tables, ending in the suffixes _Keys, _Sightings, _StableSamples and _UnstableSamples. Not every object type has a table for every suffix, but the _Keys suffix is especially important and a _Keys table does indeed exist for every object type. In fact, if we limit the query to return only those tables ending in _Keys, we reveal the full object hierarchy: SELECT sch.name + '.' + obj.name AS [name] FROM sys.objects obj JOIN sys.schemas sch ON sch.schema_id = obj.schema_id WHERE obj.type_desc = 'USER_TABLE' AND sch.name = 'data' AND obj.name LIKE '%_Keys' ORDER BY sch.name, obj.name;  name 1data.Cluster_Keys 2data.Cluster_Machine_Keys 3data.Cluster_Machine_LogicalDisk_Keys 4data.Cluster_Machine_Network_Keys 5data.Cluster_Machine_Process_Keys 6data.Cluster_Machine_Services_Keys 7data.Cluster_ResourceGroup_Keys 8data.Cluster_ResourceGroup_Resource_Keys 9data.Cluster_SqlServer_Agent_Job_History_Keys 10data.Cluster_SqlServer_Agent_Job_Keys 11data.Cluster_SqlServer_Database_BackupType_Backup_Keys 12data.Cluster_SqlServer_Database_BackupType_Keys 13data.Cluster_SqlServer_Database_CustomMetric_Keys 14data.Cluster_SqlServer_Database_File_Keys 15data.Cluster_SqlServer_Database_Keys 16data.Cluster_SqlServer_Database_Table_Index_Keys 17data.Cluster_SqlServer_Database_Table_Keys 18data.Cluster_SqlServer_Error_Keys 19data.Cluster_SqlServer_Keys 20data.Cluster_SqlServer_Services_Keys 21data.Cluster_SqlServer_SqlProcess_Keys 22data.Cluster_SqlServer_TopQueries_Keys 23data.Cluster_SqlServer_Trace_Keys 24data.Group_Keys The full object type hierarchy looks like this: Cluster Machine LogicalDisk Network Process Services ResourceGroup Resource SqlServer Agent Job History Database BackupType Backup CustomMetric File Table Index Error Services SqlProcess TopQueries Trace Group Okay, but what about the individual objects themselves represented at each level in this hierarchy? Well that’s what the _Keys tables are for. This is probably best illustrated by way of a simple example – how can I query my own data repository to find the databases on my own PC for which monitoring data has been collected? Like this: SELECT clstr._Name AS cluster_name, srvr._Name AS instance_name, db._Name AS database_name FROM data.Cluster_SqlServer_Database_Keys db JOIN data.Cluster_SqlServer_Keys srvr ON db.ParentId = srvr.Id -- Note here how the parent of a Database is a Server JOIN data.Cluster_Keys clstr ON srvr.ParentId = clstr.Id -- Note here how the parent of a Server is a Cluster WHERE clstr._Name = 'dev-chrisl2' -- This is the hostname of my own PC ORDER BY clstr._Name, srvr._Name, db._Name;  cluster_nameinstance_namedatabase_name 1dev-chrisl2SqlMonitorData 2dev-chrisl2master 3dev-chrisl2model 4dev-chrisl2msdb 5dev-chrisl2mssqlsystemresource 6dev-chrisl2tempdb 7dev-chrisl2sql2005SqlMonitorData 8dev-chrisl2sql2005TestDatabase 9dev-chrisl2sql2005master 10dev-chrisl2sql2005model 11dev-chrisl2sql2005msdb 12dev-chrisl2sql2005mssqlsystemresource 13dev-chrisl2sql2005tempdb 14dev-chrisl2sql2008SqlMonitorData 15dev-chrisl2sql2008master 16dev-chrisl2sql2008model 17dev-chrisl2sql2008msdb 18dev-chrisl2sql2008mssqlsystemresource 19dev-chrisl2sql2008tempdb These results show that I have three SQL Server instances on my machine (a default instance, one named sql2005 and one named sql2008), and each instance has the usual set of system databases, along with a database named SqlMonitorData. Basically, this is where I test SQL Monitor on different versions of SQL Server, when I’m developing. There are a few important things we can learn from this query: Each _Keys table has a column named Id. This is the primary key. Each _Keys table has a column named ParentId. A foreign key relationship is defined between each _Keys table and its parent _Keys table in the hierarchy. There are two exceptions to this, Cluster_Keys and Group_Keys, because clusters and groups live at the root level of the object hierarchy. Each _Keys table has a column named _Name. This is used to uniquely identify objects in the table within the scope of the same shared parent object. Actually, that last item isn’t always true. In some cases, the _Name column is actually called something else. For example, the data.Cluster_Machine_Services_Keys table has a column named _ServiceName instead of _Name (sorry for the inconsistency). In other cases, a name isn’t sufficient to uniquely identify an object. For example, right now my PC has multiple processes running, all sharing the same name, Chrome (one for each tab open in my web-browser). In such cases, multiple columns are used to uniquely identify an object within the scope of the same shared parent object. Well, that’s it for now. I’ve given you enough information for you to explore the _Keys tables to see how objects are stored in your own data repositories. In a future post, I’ll try to explain how monitoring data is stored for each object, using the _StableSamples and _UnstableSamples tables. If you have any questions about this post, or suggestions for future posts, just submit them in the comments section below.

  Read the article

• How can I switch an existing set of Subversion repositories to use ActiveDirectory?

  - by jpierson

  I have a set of private Subversion repositories on a Windows Server 2003 box which developers access via SVNServe over the svn:// protocol. Currently we have been using the authz and passwd files for each repository to control access however with the growing number of repositories and developers I'm considering switching to using their credentials from ActiveDirectory. We run in an all Microsoft shop and use IIS instead of Apache on all of our web servers so I would prefer to continue to use SVNServe if possible. Besides it being possible, I'm also concerned about how to migrate our repositories so that the history for the existing users map to the correct ActiveDirectory accounts. Keep in mind also that I'm not the network administrator and I'm not terrible familiar with ActiveDirectory so I'll probably have to go through some other people to get the changes made in ActiveDirectory if necessary. What are my options? UPDATE 1: It appears from the SVN documentation that by using SASL I should be able to get SVNServe to authenticate using ActiveDirectory. To clarify, the answer that I'm looking for is how to go about configuring SVNServe (if possible) to use ActiveDirectory for authentication and then how to modify an existing repository to remap existing svn users to their ActiveDirectory domain login accounts. UPDATE 2: It appears that the SASL support in SVNServe works off of a plugin model and the documentation only shows as an example. Looking at the Cyrus SASL Library it looks like a number of authentication "mechanisms" are supported but I'm not sure which one is to be used for ActiveDirectory support nor can I find any documentation about such matters. UPDATE 3: Ok, well it looks like in order to communication with ActiveDirectory I'm looking to use saslauthd instead of sasldb for the *auxprop_plugin* property. Unfortunately it appears that according to some posts (possibly outdated and inaccurate) saslauthd does not build on Windows and such endeavors are considered a work in progress. UPDATE 4: The lastest post I've found on this topic makes it sound as though the proper binaries () are available through the MIT Kerberos Library but it sounds like the author of this post on Nabble.com is still having issues getting things working. UPDATE 5: It looks like from the TortoiseSVN discussions and also this post on svn.haxx.se that even if saslgssapi.dll or whatever necessary binaries are available and configured on the Windows server that the clients will also need the same customization in order to work with these repositories. If this is true, we will only be able to get ActiveDirectory support from a windows client only if changes are made in these clients such as TortoiseSVN and CollabNet build of the client binaries to support such authentication schemes. Although thats what these posts suggest, this is contradictory from what I originally assumed from other reading in that being SASL compatible should require no changes on the client but instead only that the server be setup to handle the authentication mechanism. After reading a bit more carefully in the document about Cyrus SASL in Subversion section 5 states "1.5+ clients with Cyrus SASL support will be able to authenticate against 1.5+ servers with SASL enabled, provided at least one of the mechanisms supported by the server is also supported by the client." So clearly GSSAPI support (which I understand is required for Active Directory) must be available within the client and the server. I have to say, I'm learning way too much about the internals of how Subversion handles authentication than I ever wanted to and I juts simply want to get an answer about whether I can have Active Directory authentication support when using SVNServe on a Windows server and accessing this from Windows clients. According to the official documentation it seems that this is possible however you can see that the configuration is not trivial if even possible at all.

  Read the article

• Dynamically creating a Generic Type at Runtime

  - by Rick Strahl

  I learned something new today. Not uncommon, but it's a core .NET runtime feature I simply did not know although I know I've run into this issue a few times and worked around it in other ways. Today there was no working around it and a few folks on Twitter pointed me in the right direction. The question I ran into is: How do I create a type instance of a generic type when I have dynamically acquired the type at runtime? Yup it's not something that you do everyday, but when you're writing code that parses objects dynamically at runtime it comes up from time to time. In my case it's in the bowels of a custom JSON parser. After some thought triggered by a comment today I realized it would be fairly easy to implement two-way Dictionary parsing for most concrete dictionary types. I could use a custom Dictionary serialization format that serializes as an array of key/value objects. Basically I can use a custom type (that matches the JSON signature) to hold my parsed dictionary data and then add it to the actual dictionary when parsing is complete. Generic Types at Runtime One issue that came up in the process was how to figure out what type the Dictionary<K,V> generic parameters take. Reflection actually makes it fairly easy to figure out generic types at runtime with code like this: if (arrayType.GetInterface("IDictionary") != null) { if (arrayType.IsGenericType) { var keyType = arrayType.GetGenericArguments()[0]; var valueType = arrayType.GetGenericArguments()[1]; … } } The GetArrayType method gets passed a type instance that is the array or array-like object that is rendered in JSON as an array (which includes IList, IDictionary, IDataReader and a few others). In my case the type passed would be something like Dictionary<string, CustomerEntity>. So I know what the parent container class type is. Based on the the container type using it's then possible to use GetGenericTypeArguments() to retrieve all the generic types in sequential order of definition (ie. string, CustomerEntity). That's the easy part. Creating a Generic Type and Providing Generic Parameters at RunTime The next problem is how do I get a concrete type instance for the generic type? I know what the type name and I have a type instance is but it's generic, so how do I get a type reference to keyvaluepair<K,V> that is specific to the keyType and valueType above? Here are a couple of things that come to mind but that don't work (and yes I tried that unsuccessfully first): Type elementType = typeof(keyvalue<keyType, valueType>); Type elementType = typeof(keyvalue<typeof(keyType), typeof(valueType)>); The problem is that this explicit syntax expects a type literal not some dynamic runtime value, so both of the above won't even compile. I turns out the way to create a generic type at runtime is using a fancy bit of syntax that until today I was completely unaware of: Type elementType = typeof(keyvalue<,>).MakeGenericType(keyType, valueType); The key is the type(keyvalue<,>) bit which looks weird at best. It works however and produces a non-generic type reference. You can see the difference between the full generic type and the non-typed (?) generic type in the debugger: The nonGenericType doesn't show any type specialization, while the elementType type shows the string, CustomerEntity (truncated above) in the type name. Once the full type reference exists (elementType) it's then easy to create an instance. In my case the parser parses through the JSON and when it completes parsing the value/object it creates a new keyvalue<T,V> instance. Now that I know the element type that's pretty trivial with: // Objects start out null until we find the opening tag resultObject = Activator.CreateInstance(elementType); Here the result object is picked up by the JSON array parser which creates an instance of the child object (keyvalue<K,V>) and then parses and assigns values from the JSON document using the types  key/value property signature. Internally the parser then takes each individually parsed item and adds it to a list of  List<keyvalue<K,V>> items. Parsing through a Generic type when you only have Runtime Type Information When parsing of the JSON array is done, the List needs to be turned into a defacto Dictionary<K,V>. This should be easy since I know that I'm dealing with an IDictionary, and I know the generic types for the key and value. The problem is again though that this needs to happen at runtime which would mean using several Convert.ChangeType() calls in the code to dynamically cast at runtime. Yuk. In the end I decided the easier and probably only slightly slower way to do this is a to use the dynamic type to collect the items and assign them to avoid all the dynamic casting madness: else if (IsIDictionary) { IDictionary dict = Activator.CreateInstance(arrayType) as IDictionary; foreach (dynamic item in items) { dict.Add(item.key, item.value); } return dict; } This code creates an instance of the generic dictionary type first, then loops through all of my custom keyvalue<K,V> items and assigns them to the actual dictionary. By using Dynamic here I can side step all the explicit type conversions that would be required in the three highlighted areas (not to mention that this nested method doesn't have access to the dictionary item generic types here). Static <- -> Dynamic Dynamic casting in a static language like C# is a bitch to say the least. This is one of the few times when I've cursed static typing and the arcane syntax that's required to coax types into the right format. It works but it's pretty nasty code. If it weren't for dynamic that last bit of code would have been a pretty ugly as well with a bunch of Convert.ChangeType() calls to litter the code. Fortunately this type of type convulsion is rather rare and reserved for system level code. It's not every day that you create a string to object parser after all :-)© Rick Strahl, West Wind Technologies, 2005-2011Posted in .NET  CSharp   Tweet (function() { var po = document.createElement('script'); po.type = 'text/javascript'; po.async = true; po.src = 'https://apis.google.com/js/plusone.js'; var s = document.getElementsByTagName('script')[0]; s.parentNode.insertBefore(po, s); })();

  Read the article

• A simple Dynamic Proxy

  - by Abhijeet Patel

  Frameworks such as EF4 and MOQ do what most developers consider "dark magic". For instance in EF4, when you use a POCO for an entity you can opt-in to get behaviors such as "lazy-loading" and "change tracking" at runtime merely by ensuring that your type has the following characteristics: The class must be public and not sealed. The class must have a public or protected parameter-less constructor. The class must have public or protected properties Adhere to this and your type is magically endowed with these behaviors without any additional programming on your part. Behind the scenes the framework subclasses your type at runtime and creates a "dynamic proxy" which has these additional behaviors and when you navigate properties of your POCO, the framework replaces the POCO type with derived type instances. The MOQ framework does simlar magic. Let's say you have a simple interface:   public interface IFoo      {          int GetNum();      }   We can verify that the GetNum() was invoked on a mock like so:   var mock = new Mock<IFoo>(MockBehavior.Default);   mock.Setup(f => f.GetNum());   var num = mock.Object.GetNum();   mock.Verify(f => f.GetNum());   Beind the scenes the MOQ framework is generating a dynamic proxy by implementing IFoo at runtime. the call to moq.Object returns the dynamic proxy on which we then call "GetNum" and then verify that this method was invoked. No dark magic at all, just clever programming is what's going on here, just not visible and hence appears magical! Let's create a simple dynamic proxy generator which accepts an interface type and dynamically creates a proxy implementing the interface type specified at runtime.     public static class DynamicProxyGenerator   {       public static T GetInstanceFor<T>()       {           Type typeOfT = typeof(T);           var methodInfos = typeOfT.GetMethods();           AssemblyName assName = new AssemblyName("testAssembly");           var assBuilder = AppDomain.CurrentDomain.DefineDynamicAssembly(assName, AssemblyBuilderAccess.RunAndSave);           var moduleBuilder = assBuilder.DefineDynamicModule("testModule", "test.dll");           var typeBuilder = moduleBuilder.DefineType(typeOfT.Name + "Proxy", TypeAttributes.Public);              typeBuilder.AddInterfaceImplementation(typeOfT);           var ctorBuilder = typeBuilder.DefineConstructor(                     MethodAttributes.Public,                     CallingConventions.Standard,                     new Type[] { });           var ilGenerator = ctorBuilder.GetILGenerator();           ilGenerator.EmitWriteLine("Creating Proxy instance");           ilGenerator.Emit(OpCodes.Ret);           foreach (var methodInfo in methodInfos)           {               var methodBuilder = typeBuilder.DefineMethod(                   methodInfo.Name,                   MethodAttributes.Public | MethodAttributes.Virtual,                   methodInfo.ReturnType,                   methodInfo.GetParameters().Select(p => p.GetType()).ToArray()                   );               var methodILGen = methodBuilder.GetILGenerator();               methodILGen.EmitWriteLine("I'm a proxy");               if (methodInfo.ReturnType == typeof(void))               {                   methodILGen.Emit(OpCodes.Ret);               }               else               {                   if (methodInfo.ReturnType.IsValueType || methodInfo.ReturnType.IsEnum)                   {                       MethodInfo getMethod = typeof(Activator).GetMethod(/span>"CreateInstance",new Type[]{typeof((Type)});                                               LocalBuilder lb = methodILGen.DeclareLocal(methodInfo.ReturnType);                       methodILGen.Emit(OpCodes.Ldtoken, lb.LocalType);                       methodILGen.Emit(OpCodes.Call, typeofype).GetMethod("GetTypeFromHandle"));  ));                       methodILGen.Emit(OpCodes.Callvirt, getMethod);                       methodILGen.Emit(OpCodes.Unbox_Any, lb.LocalType);                                                              }                 else                   {                       methodILGen.Emit(OpCodes.Ldnull);                   }                   methodILGen.Emit(OpCodes.Ret);               }               typeBuilder.DefineMethodOverride(methodBuilder, methodInfo);           }                     Type constructedType = typeBuilder.CreateType();           var instance = Activator.CreateInstance(constructedType);           return (T)instance;       }   }   Dynamic proxies are created by calling into the following main types: AssemblyBuilder, TypeBuilder, Modulebuilder and ILGenerator. These types enable dynamically creating an assembly and emitting .NET modules and types in that assembly, all using IL instructions. Let's break down the code above a bit and examine it piece by piece                Type typeOfT = typeof(T);              var methodInfos = typeOfT.GetMethods();              AssemblyName assName = new AssemblyName("testAssembly");              var assBuilder = AppDomain.CurrentDomain.DefineDynamicAssembly(assName, AssemblyBuilderAccess.RunAndSave);              var moduleBuilder = assBuilder.DefineDynamicModule("testModule", "test.dll");              var typeBuilder = moduleBuilder.DefineType(typeOfT.Name + "Proxy", TypeAttributes.Public);   We are instructing the runtime to create an assembly caled "test.dll"and in this assembly we then emit a new module called "testModule". We then emit a new type definition of name "typeName"Proxy into this new module. This is the definition for the "dynamic proxy" for type T                 typeBuilder.AddInterfaceImplementation(typeOfT);               var ctorBuilder = typeBuilder.DefineConstructor(                         MethodAttributes.Public,                         CallingConventions.Standard,                         new Type[] { });               var ilGenerator = ctorBuilder.GetILGenerator();               ilGenerator.EmitWriteLine("Creating Proxy instance");               ilGenerator.Emit(OpCodes.Ret);   The newly created type implements type T and defines a default parameterless constructor in which we emit a call to Console.WriteLine. This call is not necessary but we do this so that we can see first hand that when the proxy is constructed, when our default constructor is invoked.   var methodBuilder = typeBuilder.DefineMethod(                      methodInfo.Name,                      MethodAttributes.Public | MethodAttributes.Virtual,                      methodInfo.ReturnType,                      methodInfo.GetParameters().Select(p => p.GetType()).ToArray()                      );   We then iterate over each method declared on type T and add a method definition of the same name into our "dynamic proxy" definition     if (methodInfo.ReturnType == typeof(void))   {       methodILGen.Emit(OpCodes.Ret);   }   If the return type specified in the method declaration of T is void we simply return.     if (methodInfo.ReturnType.IsValueType || methodInfo.ReturnType.IsEnum)   {                               MethodInfo getMethod = typeof(Activator).GetMethod("CreateInstance",                                                         new Type[]{typeof(Type)});                               LocalBuilder lb = methodILGen.DeclareLocal(methodInfo.ReturnType);                                                     methodILGen.Emit(OpCodes.Ldtoken, lb.LocalType);       methodILGen.Emit(OpCodes.Call, typeof(Type).GetMethod("GetTypeFromHandle"));       methodILGen.Emit(OpCodes.Callvirt, getMethod);       methodILGen.Emit(OpCodes.Unbox_Any, lb.LocalType);   }   If the return type in the method declaration of T is either a value type or an enum, then we need to create an instance of the value type and return that instance the caller. In order to accomplish that we need to do the following: 1) Get a handle to the Activator.CreateInstance method 2) Declare a local variable which represents the Type of the return type(i.e the type object of the return type) specified on the method declaration of T(obtained from the MethodInfo) and push this Type object onto the evaluation stack. In reality a RuntimeTypeHandle is what is pushed onto the stack. 3) Invoke the "GetTypeFromHandle" method(a static method in the Type class) passing in the RuntimeTypeHandle pushed onto the stack previously as an argument, the result of this invocation is a Type object (representing the method's return type) which is pushed onto the top of the evaluation stack. 4) Invoke Activator.CreateInstance passing in the Type object from step 3, the result of this invocation is an instance of the value type boxed as a reference type and pushed onto the top of the evaluation stack. 5) Unbox the result and place it into the local variable of the return type defined in step 2   methodILGen.Emit(OpCodes.Ldnull);   If the return type is a reference type then we just load a null onto the evaluation stack   methodILGen.Emit(OpCodes.Ret);   Emit a a return statement to return whatever is on top of the evaluation stack(null or an instance of a value type) back to the caller     Type constructedType = typeBuilder.CreateType();   var instance = Activator.CreateInstance(constructedType);   return (T)instance;   Now that we have a definition of the "dynamic proxy" implementing all the methods declared on T, we can now create an instance of the proxy type and return that out typed as T. The caller can now invoke the generator and request a dynamic proxy for any type T. In our example when the client invokes GetNum() we get back "0". Lets add a new method on the interface called DayOfWeek GetDay()   public interface IFoo      {          int GetNum();          DayOfWeek GetDay();      }   When GetDay() is invoked, the "dynamic proxy" returns "Sunday" since that is the default value for the DayOfWeek enum This is a very trivial example of dynammic proxies, frameworks like MOQ have a way more sophisticated implementation of this paradigm where in you can instruct the framework to create proxies which return specified values for a method implementation.

  Read the article

• Setting up RADIUS + LDAP for WPA2 on Ubuntu

  - by Morten Siebuhr

  I'm setting up a wireless network for ~150 users. In short, I'm looking for a guide to set RADIUS server to authenticate WPA2 against a LDAP. On Ubuntu. I got a working LDAP, but as it is not in production use, it can very easily be adapted to whatever changes this project may require. I've been looking at FreeRADIUS, but any RADIUS server will do. We got a separate physical network just for WiFi, so not too many worries about security on that front. Our AP's are HP's low end enterprise stuff - they seem to support whatever you can think of. All Ubuntu Server, baby! And the bad news: I now somebody less knowledgeable than me will eventually take over administration, so the setup has to be as "trivial" as possible. So far, our setup is based only on software from the Ubuntu repositories, with exception of our LDAP administration web application and a few small special scripts. So no "fetch package X, untar, ./configure"-things if avoidable. UPDATE 2009-08-18: While I found several useful resources, there is one serious obstacle: Ignoring EAP-Type/tls because we do not have OpenSSL support. Ignoring EAP-Type/ttls because we do not have OpenSSL support. Ignoring EAP-Type/peap because we do not have OpenSSL support. Basically the Ubuntu version of FreeRADIUS does not support SSL (bug 183840), which makes all the secure EAP-types useless. Bummer. But some useful documentation for anybody interested: http://vuksan.com/linux/dot1x/802-1x-LDAP.html http://tldp.org/HOWTO/html_single/8021X-HOWTO/#confradius UPDATE 2009-08-19: I ended up compiling my own FreeRADIUS package yesterday evening - there's a really good recipe at http://www.linuxinsight.com/building-debian-freeradius-package-with-eap-tls-ttls-peap-support.html (See the comments to the post for updated instructions). I got a certificate from http://CACert.org (you should probably get a "real" cert if possible) Then I followed the instructions at http://vuksan.com/linux/dot1x/802-1x-LDAP.html. This links to http://tldp.org/HOWTO/html_single/8021X-HOWTO/, which is a very worthwhile read if you want to know how WiFi security works. UPDATE 2009-08-27: After following the above guide, I've managed to get FreeRADIUS to talk to LDAP: I've created a test user in LDAP, with the password mr2Yx36M - this gives an LDAP entry roughly of: uid: testuser sambaLMPassword: CF3D6F8A92967E0FE72C57EF50F76A05 sambaNTPassword: DA44187ECA97B7C14A22F29F52BEBD90 userPassword: {SSHA}Z0SwaKO5tuGxgxtceRDjiDGFy6bRL6ja When using radtest, I can connect fine: > radtest testuser "mr2Yx36N" sbhr.dk 0 radius-private-password Sending Access-Request of id 215 to 130.225.235.6 port 1812 User-Name = "msiebuhr" User-Password = "mr2Yx36N" NAS-IP-Address = 127.0.1.1 NAS-Port = 0 rad_recv: Access-Accept packet from host 130.225.235.6 port 1812, id=215, length=20 > But when I try through the AP, it doesn't fly - while it does confirm that it figures out the NT and LM passwords: ... rlm_ldap: sambaNTPassword -> NT-Password == 0x4441343431383745434139374237433134413232463239463532424542443930 rlm_ldap: sambaLMPassword -> LM-Password == 0x4346334436463841393239363745304645373243353745463530463736413035 [ldap] looking for reply items in directory... WARNING: No "known good" password was found in LDAP. Are you sure that the user is configured correctly? [ldap] user testuser authorized to use remote access rlm_ldap: ldap_release_conn: Release Id: 0 ++[ldap] returns ok ++[expiration] returns noop ++[logintime] returns noop [pap] Normalizing NT-Password from hex encoding [pap] Normalizing LM-Password from hex encoding ... It is clear that the NT and LM passwords differ from the above, yet the message [ldap] user testuser authorized to use remote access - and the user is later rejected...

  Read the article

• Azure, don't give me multiple VMs, give me one elastic VM

  - by FransBouma

  Yesterday, Microsoft revealed new major features for Windows Azure (see ScottGu's post). It all looks shiny and great, but after reading most of the material describing the new features, I still find the overall idea behind all of it flawed: why should I care on how much VMs my web app runs? Isn't that a problem to solve for the Windows Azure engineers / software? And what if I need the file system, why can't I simply get a virtual filesystem ? To illustrate my point, let's use a real example: a product website with a customer system/database and next to it a support site with accompanying database. Both are written in .NET, using ASP.NET and use a SQL Server database each. The product website offers files to download by customers, very simple. You have a couple of options to host these websites: Buy a server, place it in a rack at an ISP and run the sites on that server Use 'shared hosting' with an ISP, which means your sites' appdomains are running on the same machine, as well as the files stored, and the databases are hosted in the same server as the other shared databases. Hire a VM, install your OS of choice at an ISP, and host the sites on that VM, basically the same as the first option, except you don't have a physical server At some cloud-vendor, either host the sites 'shared' or in a VM. See above. With all of those options, scalability is a problem, even the cloud-based ones, though not due to the same reasons: The physical server solution has the obvious problem that if you need more power, you need to buy a bigger server or more servers which requires you to add replication and other overhead Shared hosting solutions are almost always capped on memory usage / traffic and database size: if your sites get too big, you have to move out of the shared hosting environment and start over with one of the other solutions The VM solution, be it a VM at an ISP or 'in the cloud' at e.g. Windows Azure or Amazon, in theory allows scaling out by simply instantiating more VMs, however that too introduces the same overhead problems as with the physical servers: suddenly more than 1 instance runs your sites. If a cloud vendor offers its services in the form of VMs, you won't gain much over having a VM at some ISP: the main problems you have to work around are still there: when you spin up more than one VM, your application must be completely stateless at any moment, including the DB sub system, because what's in memory in instance 1 might not be in memory in instance 2. This might sounds trivial but it's not. A lot of the websites out there started rather small: they were perfectly runnable on a single machine with normal memory and CPU power. After all, you don't need a big machine to run a website with even thousands of users a day. Moving these sites to a multi-VM environment will cause a problem: all the in-memory state they use, all the multi-page transitions they use while keeping state across the transition, they can't do that anymore like they did that on a single machine: state is something of the past, you have to store every byte of state in either a DB or in a viewstate or in a cookie somewhere so with the next request, all state information is available through the request, as nothing is kept in-memory. Our example uses a bunch of files in a file system. Using multiple VMs will require that these files move to a cloud storage system which is mounted in each VM so we don't have to store the files on each VM. This might require different file paths, but this change should be minor. What's perhaps less minor is the maintenance procedure in place on the new type of cloud storage used: instead of ftp-ing into a VM, you might have to update the files using different ways / tools. All in all this makes moving an existing website which was written for an environment that's based around a VM (namely .NET with its CLR) overly cumbersome and problematic: it forces you to refactor your website system to be able to be used 'in the cloud', which is caused by the limited way how e.g. Windows Azure offers its cloud services: in blocks of VMs. Offer a scalable, flexible VM which extends with my needs Instead, cloud vendors should offer simply one VM to me. On that VM I run the websites, store my DB and my files. As it's a virtual machine, how this machine is actually ran on physical hardware (e.g. partitioned), I don't care, as that's the problem for the cloud vendor to solve. If I need more resources, e.g. I have more traffic to my server, way more visitors per day, the VM stretches, like I bought a bigger box. This frees me from the problem which comes with multiple VMs: I don't have any refactoring to do at all: I can simply build my website as if it runs on my local hardware server, upload it to the VM offered by the cloud vendor, install it on the VM and I'm done. "But that might require changes to windows!" Yes, but Microsoft is Windows. Windows Azure is their service, they can make whatever change to what they offer to make it look like it's windows. Yet, they're stuck, like Amazon, in thinking in VMs, which forces developers to 'think ahead' and gamble whether they would need to migrate to a cloud with multiple VMs in the future or not. Which comes down to: gamble whether they should invest time in code / architecture which they might never need. (YAGNI anyone?) So the VM we're talking about, is that a low-level VM which runs a guest OS, or is that VM a different kind of VM? The flexible VM: .NET's CLR ? My example websites are ASP.NET based, which means they run inside a .NET appdomain, on the .NET CLR, which is a VM. The only physical OS resource the sites need is the file system, however this too is accessed through .NET. In short: all the websites see is what .NET allows the websites to see, the world as the websites know it is what .NET shows them and lets them access. How the .NET appdomain is run physically, that's the concern of .NET, not mine. This begs the question why Windows Azure doesn't offer virtual appdomains? Or better: .NET environments which look like one machine but could be physically multiple machines. In such an environment, no change has to be made to the websites to migrate them from a local machine or own server to the cloud to get proper scaling: the .NET VM will simply scale with the need: more memory needed, more CPU power needed, it stretches. What it offers to the application running inside the appdomain is simply increasing, but not fragmented: all resources are available to the application: this means that the problem of how to scale is back to where it should be: with the cloud vendor. "Yeah, great, but what about the databases?" The .NET application communicates with the database server through a .NET ADO.NET provider. Where the database is located is not a problem of the appdomain: the ADO.NET provider has to solve that. I.o.w.: we can host the databases in an environment which offers itself as a single resource and is accessible through one connection string without replication overhead on the outside, and use that environment inside the .NET VM as if it was a single DB. But what about memory replication and other problems? This environment isn't simple, at least not for the cloud vendor. But it is simple for the customer who wants to run his sites in that cloud: no work needed. No refactoring needed of existing code. Upload it, run it. Perhaps I'm dreaming and what I described above isn't possible. Yet, I think if cloud vendors don't move into that direction, what they're offering isn't interesting: it doesn't solve a problem at all, it simply offers a way to instantiate more VMs with the guest OS of choice at the cost of me needing to refactor my website code so it can run in the straight jacket form factor dictated by the cloud vendor. Let's not kid ourselves here: most of us developers will never build a website which needs a truck load of VMs to run it: almost all websites created by developers can run on just a few VMs at most. Yet, the most expensive change is right at the start: moving from one to two VMs. As soon as you have refactored your website code to run across multiple VMs, adding another one is just as easy as clicking a mouse button. But that first step, that's the problem here and as it's right there at the beginning of scaling the website, it's particularly strange that cloud vendors refuse to solve that problem and leave it to the developers to solve that. Which makes migrating 'to the cloud' particularly expensive.

  Read the article

• Fun with Aggregates

  - by Paul White

  There are interesting things to be learned from even the simplest queries.  For example, imagine you are given the task of writing a query to list AdventureWorks product names where the product has at least one entry in the transaction history table, but fewer than ten. One possible query to meet that specification is: SELECT p.Name FROM Production.Product AS p JOIN Production.TransactionHistory AS th ON p.ProductID = th.ProductID GROUP BY p.ProductID, p.Name HAVING COUNT_BIG(*) < 10; That query correctly returns 23 rows (execution plan and data sample shown below): The execution plan looks a bit different from the written form of the query: the base tables are accessed in reverse order, and the aggregation is performed before the join.  The general idea is to read all rows from the history table, compute the count of rows grouped by ProductID, merge join the results to the Product table on ProductID, and finally filter to only return rows where the count is less than ten. This ‘fully-optimized’ plan has an estimated cost of around 0.33 units.  The reason for the quote marks there is that this plan is not quite as optimal as it could be – surely it would make sense to push the Filter down past the join too?  To answer that, let’s look at some other ways to formulate this query.  This being SQL, there are any number of ways to write logically-equivalent query specifications, so we’ll just look at a couple of interesting ones.  The first query is an attempt to reverse-engineer T-SQL from the optimized query plan shown above.  It joins the result of pre-aggregating the history table to the Product table before filtering: SELECT p.Name FROM ( SELECT th.ProductID, cnt = COUNT_BIG(*) FROM Production.TransactionHistory AS th GROUP BY th.ProductID ) AS q1 JOIN Production.Product AS p ON p.ProductID = q1.ProductID WHERE q1.cnt < 10; Perhaps a little surprisingly, we get a slightly different execution plan: The results are the same (23 rows) but this time the Filter is pushed below the join!  The optimizer chooses nested loops for the join, because the cardinality estimate for rows passing the Filter is a bit low (estimate 1 versus 23 actual), though you can force a merge join with a hint and the Filter still appears below the join.  In yet another variation, the < 10 predicate can be ‘manually pushed’ by specifying it in a HAVING clause in the “q1” sub-query instead of in the WHERE clause as written above. The reason this predicate can be pushed past the join in this query form, but not in the original formulation is simply an optimizer limitation – it does make efforts (primarily during the simplification phase) to encourage logically-equivalent query specifications to produce the same execution plan, but the implementation is not completely comprehensive. Moving on to a second example, the following query specification results from phrasing the requirement as “list the products where there exists fewer than ten correlated rows in the history table”: SELECT p.Name FROM Production.Product AS p WHERE EXISTS ( SELECT * FROM Production.TransactionHistory AS th WHERE th.ProductID = p.ProductID HAVING COUNT_BIG(*) < 10 ); Unfortunately, this query produces an incorrect result (86 rows): The problem is that it lists products with no history rows, though the reasons are interesting.  The COUNT_BIG(*) in the EXISTS clause is a scalar aggregate (meaning there is no GROUP BY clause) and scalar aggregates always produce a value, even when the input is an empty set.  In the case of the COUNT aggregate, the result of aggregating the empty set is zero (the other standard aggregates produce a NULL).  To make the point really clear, let’s look at product 709, which happens to be one for which no history rows exist: -- Scalar aggregate SELECT COUNT_BIG(*) FROM Production.TransactionHistory AS th WHERE th.ProductID = 709;   -- Vector aggregate SELECT COUNT_BIG(*) FROM Production.TransactionHistory AS th WHERE th.ProductID = 709 GROUP BY th.ProductID; The estimated execution plans for these two statements are almost identical: You might expect the Stream Aggregate to have a Group By for the second statement, but this is not the case.  The query includes an equality comparison to a constant value (709), so all qualified rows are guaranteed to have the same value for ProductID and the Group By is optimized away. In fact there are some minor differences between the two plans (the first is auto-parameterized and qualifies for trivial plan, whereas the second is not auto-parameterized and requires cost-based optimization), but there is nothing to indicate that one is a scalar aggregate and the other is a vector aggregate.  This is something I would like to see exposed in show plan so I suggested it on Connect.  Anyway, the results of running the two queries show the difference at runtime: The scalar aggregate (no GROUP BY) returns a result of zero, whereas the vector aggregate (with a GROUP BY clause) returns nothing at all.  Returning to our EXISTS query, we could ‘fix’ it by changing the HAVING clause to reject rows where the scalar aggregate returns zero: SELECT p.Name FROM Production.Product AS p WHERE EXISTS ( SELECT * FROM Production.TransactionHistory AS th WHERE th.ProductID = p.ProductID HAVING COUNT_BIG(*) BETWEEN 1 AND 9 ); The query now returns the correct 23 rows: Unfortunately, the execution plan is less efficient now – it has an estimated cost of 0.78 compared to 0.33 for the earlier plans.  Let’s try adding a redundant GROUP BY instead of changing the HAVING clause: SELECT p.Name FROM Production.Product AS p WHERE EXISTS ( SELECT * FROM Production.TransactionHistory AS th WHERE th.ProductID = p.ProductID GROUP BY th.ProductID HAVING COUNT_BIG(*) < 10 ); Not only do we now get correct results (23 rows), this is the execution plan: I like to compare that plan to quantum physics: if you don’t find it shocking, you haven’t understood it properly :)  The simple addition of a redundant GROUP BY has resulted in the EXISTS form of the query being transformed into exactly the same optimal plan we found earlier.  What’s more, in SQL Server 2008 and later, we can replace the odd-looking GROUP BY with an explicit GROUP BY on the empty set: SELECT p.Name FROM Production.Product AS p WHERE EXISTS ( SELECT * FROM Production.TransactionHistory AS th WHERE th.ProductID = p.ProductID GROUP BY () HAVING COUNT_BIG(*) < 10 ); I offer that as an alternative because some people find it more intuitive (and it perhaps has more geek value too).  Whichever way you prefer, it’s rather satisfying to note that the result of the sub-query does not exist for a particular correlated value where a vector aggregate is used (the scalar COUNT aggregate always returns a value, even if zero, so it always ‘EXISTS’ regardless which ProductID is logically being evaluated). The following query forms also produce the optimal plan and correct results, so long as a vector aggregate is used (you can probably find more equivalent query forms): WHERE Clause SELECT p.Name FROM Production.Product AS p WHERE ( SELECT COUNT_BIG(*) FROM Production.TransactionHistory AS th WHERE th.ProductID = p.ProductID GROUP BY () ) < 10; APPLY SELECT p.Name FROM Production.Product AS p CROSS APPLY ( SELECT NULL FROM Production.TransactionHistory AS th WHERE th.ProductID = p.ProductID GROUP BY () HAVING COUNT_BIG(*) < 10 ) AS ca (dummy); FROM Clause SELECT q1.Name FROM ( SELECT p.Name, cnt = ( SELECT COUNT_BIG(*) FROM Production.TransactionHistory AS th WHERE th.ProductID = p.ProductID GROUP BY () ) FROM Production.Product AS p ) AS q1 WHERE q1.cnt < 10; This last example uses SUM(1) instead of COUNT and does not require a vector aggregate…you should be able to work out why :) SELECT q.Name FROM ( SELECT p.Name, cnt = ( SELECT SUM(1) FROM Production.TransactionHistory AS th WHERE th.ProductID = p.ProductID ) FROM Production.Product AS p ) AS q WHERE q.cnt < 10; The semantics of SQL aggregates are rather odd in places.  It definitely pays to get to know the rules, and to be careful to check whether your queries are using scalar or vector aggregates.  As we have seen, query plans do not show in which ‘mode’ an aggregate is running and getting it wrong can cause poor performance, wrong results, or both. © 2012 Paul White Twitter: @SQL_Kiwi email: [email protected]

  Read the article

• Intelligent "Subtraction" of one text logfile from another

  - by Vi

  Example: Application generates large text log file A with many different messages. It generates similarly large log file B when does not function correctly. I want to see what messages in file B are essentially new, i.e. to filter-out everything from A. Trivial prototype is: Sort | uniq both files Join files sort | uniq -c grep -v "^2" This produces symmetric difference and inconvenient. How to do it better? (including non-symmetric difference and preserving of messages order in B) Program should first analyse A and learn which messages are common, then analyse B showing with messages needs attention. Ideally it should automatically disregard things like timestamps, line numbers or other volatile things. Example. A: 0:00:00.234 Received buffer 0x324234 0:00:00.237 Processeed buffer 0x324234 0:00:00.238 Send buffer 0x324255 0:00:03.334 Received buffer 0x324255 0:00:03.337 Processeed buffer 0x324255 0:00:03.339 Send buffer 0x324255 0:00:05.171 Received buffer 0x32421A 0:00:05.173 Processeed buffer 0x32421A 0:00:05.178 Send buffer 0x32421A B: 0:00:00.134 Received buffer 0x324111 0:00:00.137 Processeed buffer 0x324111 0:00:00.138 Send buffer 0x324111 0:00:03.334 Received buffer 0x324222 0:00:03.337 Processeed buffer 0x324222 0:00:03.338 Error processing buffer 0x324222 0:00:03.339 Send buffer 0x3242222 0:00:05.271 Received buffer 0x3242FA 0:00:05.273 Processeed buffer 0x3242FA 0:00:05.278 Send buffer 0x3242FA 0:00:07.280 Send buffer 0x3242FA failed Result: 0:00:03.338 Error processing buffer 0x324222 0:00:07.280 Send buffer 0x3242FA failed One of ways of solving it can be something like that: Split each line to logical units: 0:00:00.134 Received buffer 0x324111,0:00:00.134,Received,buffer,0x324111,324111,Received buffer, \d:\d\d:\d\d\.\d\d\d, \d+:\d+:\d+.\d+, 0x[0-9A-F]{6}, ... It should find individual words, simple patterns in numbers, common layouts (e.g. "some date than text than number than text than end_of_line"), also handle combinations of above. As it is not easy task, user assistance (adding regexes with explicit "disregard that","make the main factor","don't split to parts","consider as date/number","take care of order/quantity of such messages" rules) should be supported (but not required) for it. Find recurring units and "categorize" lines, filter out too volatile things like timestamps, addresses or line numbers. Analyse the second file, find things that has new logical units (one-time or recurring), or anything that will "amaze" the system which has got used to the first file. Example of doing some bit of this manually: $ cat A | head -n 1 0:00:00.234 Received buffer 0x324234 $ cat A | egrep -v "Received buffer" | head -n 1 0:00:00.237 Processeed buffer 0x324234 $ cat A | egrep -v "Received buffer|Processeed buffer" | head -n 1 0:00:00.238 Send buffer 0x324255 $ cat A | egrep -v "Received buffer|Processeed buffer|Send buffer" | head -n 1 $ cat B | egrep -v "Received buffer|Processeed buffer|Send buffer" 0:00:03.338 Error processing buffer 0x324222 0:00:07.280 Send buffer 0x3242FA failed This is a boring thing (there are a lot of message types); also I can accidentally include some too broad pattern. Also it can't handle complicated things like interrelation between messages. I know that it is AI-related. May be there are already developed tools?

  Read the article

• C#/.NET Little Wonders: The Timeout static class

  - by James Michael Hare

  Once again, in this series of posts I look at the parts of the .NET Framework that may seem trivial, but can help improve your code by making it easier to write and maintain. The index of all my past little wonders posts can be found here. When I started the “Little Wonders” series, I really wanted to pay homage to parts of the .NET Framework that are often small but can help in big ways.  The item I have to discuss today really is a very small item in the .NET BCL, but once again I feel it can help make the intention of code much clearer and thus is worthy of note. The Problem - Magic numbers aren’t very readable or maintainable In my first Little Wonders Post (Five Little Wonders That Make Code Better) I mention the TimeSpan factory methods which, I feel, really help the readability of constructed TimeSpan instances. Just to quickly recap that discussion, ask yourself what the TimeSpan specified in each case below is 1: // Five minutes? Five Seconds? 2: var fiveWhat1 = new TimeSpan(0, 0, 5); 3: var fiveWhat2 = new TimeSpan(0, 0, 5, 0); 4: var fiveWhat3 = new TimeSpan(0, 0, 5, 0, 0); You’d think they’d all be the same unit of time, right?  After all, most overloads tend to tack additional arguments on the end.  But this is not the case with TimeSpan, where the constructor forms are:     TimeSpan(int hours, int minutes, int seconds);     TimeSpan(int days, int hours, int minutes, int seconds);     TimeSpan(int days, int hours, int minutes, int seconds, int milliseconds); Notice how in the 4 and 5 parameter version we suddenly have the parameter days slipping in front of hours?  This can make reading constructors like those above much harder.  Fortunately, there are TimeSpan factory methods to help make your intention crystal clear: 1: // Ah! Much clearer! 2: var fiveSeconds = TimeSpan.FromSeconds(5); These are great because they remove all ambiguity from the reader!  So in short, magic numbers in constructors and methods can be ambiguous, and anything we can do to clean up the intention of the developer will make the code much easier to read and maintain. Timeout – Readable identifiers for infinite timeout values In a similar way to TimeSpan, let’s consider specifying timeouts for some of .NET’s (or our own) many methods that allow you to specify timeout periods. For example, in the TPL Task class, there is a family of Wait() methods that can take TimeSpan or int for timeouts.  Typically, if you want to specify an infinite timeout, you’d just call the version that doesn’t take a timeout parameter at all: 1: myTask.Wait(); // infinite wait But there are versions that take the int or TimeSpan for timeout as well: 1: // Wait for 100 ms 2: myTask.Wait(100); 3:  4: // Wait for 5 seconds 5: myTask.Wait(TimeSpan.FromSeconds(5); Now, if we want to specify an infinite timeout to wait on the Task, we could pass –1 (or a TimeSpan set to –1 ms), which what the .NET BCL methods with timeouts use to represent an infinite timeout: 1: // Also infinite timeouts, but harder to read/maintain 2: myTask.Wait(-1); 3: myTask.Wait(TimeSpan.FromMilliseconds(-1)); However, these are not as readable or maintainable.  If you were writing this code, you might make the mistake of thinking 0 or int.MaxValue was an infinite timeout, and you’d be incorrect.  Also, reading the code above it isn’t as clear that –1 is infinite unless you happen to know that is the specified behavior. To make the code like this easier to read and maintain, there is a static class called Timeout in the System.Threading namespace which contains definition for infinite timeouts specified as both int and TimeSpan forms: Timeout.Infinite An integer constant with a value of –1 Timeout.InfiniteTimeSpan A static readonly TimeSpan which represents –1 ms (only available in .NET 4.5+) This makes our calls to Task.Wait() (or any other calls with timeouts) much more clear: 1: // intention to wait indefinitely is quite clear now 2: myTask.Wait(Timeout.Infinite); 3: myTask.Wait(Timeout.InfiniteTimeSpan); But wait, you may say, why would we care at all?  Why not use the version of Wait() that takes no arguments?  Good question!  When you’re directly calling the method with an infinite timeout that’s what you’d most likely do, but what if you are just passing along a timeout specified by a caller from higher up?  Or perhaps storing a timeout value from a configuration file, and want to default it to infinite? For example, perhaps you are designing a communications module and want to be able to shutdown gracefully, but if you can’t gracefully finish in a specified amount of time you want to force the connection closed.  You could create a Shutdown() method in your class, and take a TimeSpan or an int for the amount of time to wait for a clean shutdown – perhaps waiting for client to acknowledge – before terminating the connection.  So, assume we had a pub/sub system with a class to broadcast messages: 1: // Some class to broadcast messages to connected clients 2: public class Broadcaster 3: { 4: // ... 5:  6: // Shutdown connection to clients, wait for ack back from clients 7: // until all acks received or timeout, whichever happens first 8: public void Shutdown(int timeout) 9: { 10: // Kick off a task here to send shutdown request to clients and wait 11: // for the task to finish below for the specified time... 12:  13: if (!shutdownTask.Wait(timeout)) 14: { 15: // If Wait() returns false, we timed out and task 16: // did not join in time. 17: } 18: } 19: } We could even add an overload to allow us to use TimeSpan instead of int, to give our callers the flexibility to specify timeouts either way: 1: // overload to allow them to specify Timeout in TimeSpan, would 2: // just call the int version passing in the TotalMilliseconds... 3: public void Shutdown(TimeSpan timeout) 4: { 5: Shutdown(timeout.TotalMilliseconds); 6: } Notice in case of this class, we don’t assume the caller wants infinite timeouts, we choose to rely on them to tell us how long to wait.  So now, if they choose an infinite timeout, they could use the –1, which is more cryptic, or use Timeout class to make the intention clear: 1: // shutdown the broadcaster, waiting until all clients ack back 2: // without timing out. 3: myBroadcaster.Shutdown(Timeout.Infinite); We could even add a default argument using the int parameter version so that specifying no arguments to Shutdown() assumes an infinite timeout: 1: // Modified original Shutdown() method to add a default of 2: // Timeout.Infinite, works because Timeout.Infinite is a compile 3: // time constant. 4: public void Shutdown(int timeout = Timeout.Infinite) 5: { 6: // same code as before 7: } Note that you can’t default the ShutDown(TimeSpan) overload with Timeout.InfiniteTimeSpan since it is not a compile-time constant.  The only acceptable default for a TimeSpan parameter would be default(TimeSpan) which is zero milliseconds, which specified no wait, not infinite wait. Summary While Timeout.Infinite and Timeout.InfiniteTimeSpan are not earth-shattering classes in terms of functionality, they do give you very handy and readable constant values that you can use in your programs to help increase readability and maintainability when specifying infinite timeouts for various timeouts in the BCL and your own applications. Technorati Tags: C#,CSharp,.NET,Little Wonders,Timeout,Task

  Read the article

• how to make bridge networking with KVM work in Fedora19

  - by netllama

  I'm attempting to get several virtual machines setup on a Fedora-19 host system, with the traditional bridge network devices (br0, br1, etc). I've done this many times before with older versions of Fedora (16, 14, etc), and it just works. However, for reasons that I cannot figure out, the bridge doesn't seem to be working in Fedora19. While I can successfully connect to the outside world (local network + internet) from inside a VM, nothing can communicate with the VM from outside (local network). I'm referring to something as trivial as pinging. From inside the VM, I can ping anything successfully (0% packet loss). However, from outside the VM (on the host, or any other system on the same network), I see 100% packet loss when pinging the IP address of the VM. My first question is simply, does anyone else have this working successfully in F19? And if so, what steps did you need to follow? I'm not using NetworkManager at all, its all the network service. There are no firewalls involved anywhere (iptables & firewall services are currently disabled). Here's the current host configuration: # brctl show bridge name bridge id STP enabled interfaces br0 8000.38eaa792efe5 no em2 vnet1 br1 8000.38eaa792efe6 no em3 br2 8000.38eaa792efe7 no em4 vnet0 virbr0 8000.525400db3ebf yes virbr0-nic # more /etc/sysconfig/network-scripts/ifcfg-em2 TYPE=Ethernet BRIDGE="br0" NAME=em2 DEVICE="em2" UUID=aeaa839e-c89c-4d6e-9daa-79b6a1b919bd ONBOOT=yes HWADDR=38:EA:A7:92:EF:E5 NM_CONTROLLED="no" # more /etc/sysconfig/network-scripts/ifcfg-br0 TYPE=Bridge NM_CONTROLLED="no" BOOTPROTO=dhcp NAME=br0 DEVICE="br0" ONBOOT=yes # ifconfig em2 ;ifconfig br0 em2: flags=4163<UP,BROADCAST,RUNNING,MULTICAST> mtu 1500 inet6 fe80::3aea:a7ff:fe92:efe5 prefixlen 64 scopeid 0x20<link> ether 38:ea:a7:92:ef:e5 txqueuelen 1000 (Ethernet) RX packets 100093 bytes 52354831 (49.9 MiB) RX errors 0 dropped 0 overruns 0 frame 0 TX packets 25321 bytes 15791341 (15.0 MiB) TX errors 0 dropped 0 overruns 0 carrier 0 collisions 0 device memory 0xf7d00000-f7e00000 br0: flags=4163<UP,BROADCAST,RUNNING,MULTICAST> mtu 1500 inet 10.31.99.226 netmask 255.255.252.0 broadcast 10.31.99.255 inet6 fe80::3aea:a7ff:fe92:efe5 prefixlen 64 scopeid 0x20<link> ether 38:ea:a7:92:ef:e5 txqueuelen 0 (Ethernet) RX packets 19619 bytes 1963328 (1.8 MiB) RX errors 0 dropped 0 overruns 0 frame 0 TX packets 11 bytes 1074 (1.0 KiB) TX errors 0 dropped 0 overruns 0 carrier 0 collisions 0 Relevant section from /etc/libvirt/qemu/foo.xml (one of the VMs with this problem): <interface type='bridge'> <mac address='52:54:00:26:22:9d'/> <source bridge='br0'/> <model type='virtio'/> <address type='pci' domain='0x0000' bus='0x00' slot='0x03' function='0x0'/> </interface> # ps -ef | grep qemu qemu 1491 1 82 13:25 ? 00:42:09 /usr/bin/qemu-system-x86_64 -machine accel=kvm -name cuda-linux64-build5 -S -machine pc-0.13,accel=kvm,usb=off -cpu SandyBridge,+pdpe1gb,+osxsave,+dca,+pcid,+pdcm,+xtpr,+tm2,+est,+smx,+vmx,+ds_cpl,+monitor,+dtes64,+pbe,+tm,+ht,+ss,+acpi,+ds,+vme -m 16384 -smp 6,sockets=6,cores=1,threads=1 -uuid 6e930234-bdfd-044d-2787-22d4bbbe30b1 -no-user-config -nodefaults -chardev socket,id=charmonitor,path=/var/lib/libvirt/qemu/cuda-linux64-build5.monitor,server,nowait -mon chardev=charmonitor,id=monitor,mode=control -rtc base=localtime -no-shutdown -device piix3-usb-uhci,id=usb,bus=pci.0,addr=0x1.0x2 -drive file=/var/lib/libvirt/images/cuda-linux64-build5.img,if=none,id=drive-virtio-disk0,format=raw,cache=writeback -device virtio-blk-pci,scsi=off,bus=pci.0,addr=0x4,drive=drive-virtio-disk0,id=virtio-disk0,bootindex=1 -netdev tap,fd=25,id=hostnet0,vhost=on,vhostfd=26 -device virtio-net-pci,netdev=hostnet0,id=net0,mac=52:54:00:26:22:9d,bus=pci.0,addr=0x3 -chardev pty,id=charserial0 -device isa-serial,chardev=charserial0,id=serial0 -vnc 127.0.0.1:1 -vga cirrus -device virtio-balloon-pci,id=balloon0,bus=pci.0,addr=0x5 I can provide additional information, if requested. thanks!

  Read the article

• C#/.NET Little Wonders: Using &lsquo;default&rsquo; to Get Default Values

  - by James Michael Hare

  Once again, in this series of posts I look at the parts of the .NET Framework that may seem trivial, but can help improve your code by making it easier to write and maintain. The index of all my past little wonders posts can be found here. Today’s little wonder is another of those small items that can help a lot in certain situations, especially when writing generics.  In particular, it is useful in determining what the default value of a given type would be. The Problem: what’s the default value for a generic type? There comes a time when you’re writing generic code where you may want to set an item of a given generic type.  Seems simple enough, right?  We’ll let’s see! Let’s say we want to query a Dictionary<TKey, TValue> for a given key and get back the value, but if the key doesn’t exist, we’d like a default value instead of throwing an exception. So, for example, we might have a the following dictionary defined: 1: var lookup = new Dictionary<int, string> 2: { 3: { 1, "Apple" }, 4: { 2, "Orange" }, 5: { 3, "Banana" }, 6: { 4, "Pear" }, 7: { 9, "Peach" } 8: }; And using those definitions, perhaps we want to do something like this: 1: // assume a default 2: string value = "Unknown"; 3:  4: // if the item exists in dictionary, get its value 5: if (lookup.ContainsKey(5)) 6: { 7: value = lookup[5]; 8: } But that’s inefficient, because then we’re double-hashing (once for ContainsKey() and once for the indexer).  Well, to avoid the double-hashing, we could use TryGetValue() instead: 1: string value; 2:  3: // if key exists, value will be put in value, if not default it 4: if (!lookup.TryGetValue(5, out value)) 5: { 6: value = "Unknown"; 7: } But the “flow” of using of TryGetValue() can get clunky at times when you just want to assign either the value or a default to a variable.  Essentially it’s 3-ish lines (depending on formatting) for 1 assignment.  So perhaps instead we’d like to write an extension method to support a cleaner interface that will return a default if the item isn’t found: 1: public static class DictionaryExtensions 2: { 3: public static TValue GetValueOrDefault<TKey, TValue>(this Dictionary<TKey, TValue> dict, 4: TKey key, TValue defaultIfNotFound) 5: { 6: TValue value; 7:  8: // value will be the result or the default for TValue 9: if (!dict.TryGetValue(key, out value)) 10: { 11: value = defaultIfNotFound; 12: } 13:  14: return value; 15: } 16: } 17:  So this creates an extension method on Dictionary<TKey, TValue> that will attempt to get a value using the given key, and will return the defaultIfNotFound as a stand-in if the key does not exist. This code compiles, fine, but what if we would like to go one step further and allow them to specify a default if not found, or accept the default for the type?  Obviously, we could overload the method to take the default or not, but that would be duplicated code and a bit heavy for just specifying a default.  It seems reasonable that we could set the not found value to be either the default for the type, or the specified value. So what if we defaulted the type to null? 1: public static TValue GetValueOrDefault<TKey, TValue>(this Dictionary<TKey, TValue> dict, 2: TKey key, TValue defaultIfNotFound = null) // ... No, this won’t work, because only reference types (and Nullable<T> wrapped types due to syntactical sugar) can be assigned to null.  So what about a calling parameterless constructor? 1: public static TValue GetValueOrDefault<TKey, TValue>(this Dictionary<TKey, TValue> dict, 2: TKey key, TValue defaultIfNotFound = new TValue()) // ... No, this won’t work either for several reasons.  First, we’d expect a reference type to return null, not an “empty” instance.  Secondly, not all reference types have a parameter-less constructor (string for example does not).  And finally, a constructor cannot be determined at compile-time, while default values can. The Solution: default(T) – returns the default value for type T Many of us know the default keyword for its uses in switch statements as the default case.  But it has another use as well: it can return us the default value for a given type.  And since it generates the same defaults that default field initialization uses, it can be determined at compile-time as well. For example: 1: var x = default(int); // x is 0 2:  3: var y = default(bool); // y is false 4:  5: var z = default(string); // z is null 6:  7: var t = default(TimeSpan); // t is a TimeSpan with Ticks == 0 8:  9: var n = default(int?); // n is a Nullable<int> with HasValue == false Notice that for numeric types the default is 0, and for reference types the default is null.  In addition, for struct types, the value is a default-constructed struct – which simply means a struct where every field has their default value (hence 0 Ticks for TimeSpan, etc.). So using this, we could modify our code to this: 1: public static class DictionaryExtensions 2: { 3: public static TValue GetValueOrDefault<TKey, TValue>(this Dictionary<TKey, TValue> dict, 4: TKey key, TValue defaultIfNotFound = default(TValue)) 5: { 6: TValue value; 7:  8: // value will be the result or the default for TValue 9: if (!dict.TryGetValue(key, out value)) 10: { 11: value = defaultIfNotFound; 12: } 13:  14: return value; 15: } 16: } Now, if defaultIfNotFound is unspecified, it will use default(TValue) which will be the default value for whatever value type the dictionary holds.  So let’s consider how we could use this: 1: lookup.GetValueOrDefault(1); // returns “Apple” 2:  3: lookup.GetValueOrDefault(5); // returns null 4:  5: lookup.GetValueOrDefault(5, “Unknown”); // returns “Unknown” 6:  Again, do not confuse a parameter-less constructor with the default value for a type.  Remember that the default value for any type is the compile-time default for any instance of that type (0 for numeric, false for bool, null for reference types, and struct will all default fields for struct).  Consider the difference: 1: // both zero 2: int i1 = default(int); 3: int i2 = new int(); 4:  5: // both “zeroed” structs 6: var dt1 = default(DateTime); 7: var dt2 = new DateTime(); 8:  9: // sb1 is null, sb2 is an “empty” string builder 10: var sb1 = default(StringBuilder()); 11: var sb2 = new StringBuilder(); So in the above code, notice that the value types all resolve the same whether using default or parameter-less construction.  This is because a value type is never null (even Nullable<T> wrapped types are never “null” in a reference sense), they will just by default contain fields with all default values. However, for reference types, the default is null and not a constructed instance.  Also it should be noted that not all classes have parameter-less constructors (string, for instance, doesn’t have one – and doesn’t need one). Summary Whenever you need to get the default value for a type, especially a generic type, consider using the default keyword.  This handy word will give you the default value for the given type at compile-time, which can then be used for initialization, optional parameters, etc. Technorati Tags: C#,CSharp,.NET,Little Wonders,default

  Read the article

• C#: LINQ vs foreach - Round 1.

  - by James Michael Hare

  So I was reading Peter Kellner's blog entry on Resharper 5.0 and its LINQ refactoring and thought that was very cool.  But that raised a point I had always been curious about in my head -- which is a better choice: manual foreach loops or LINQ?    The answer is not really clear-cut.  There are two sides to any code cost arguments: performance and maintainability.  The first of these is obvious and quantifiable.  Given any two pieces of code that perform the same function, you can run them side-by-side and see which piece of code performs better.   Unfortunately, this is not always a good measure.  Well written assembly language outperforms well written C++ code, but you lose a lot in maintainability which creates a big techncial debt load that is hard to offset as the application ages.  In contrast, higher level constructs make the code more brief and easier to understand, hence reducing technical cost.   Now, obviously in this case we're not talking two separate languages, we're comparing doing something manually in the language versus using a higher-order set of IEnumerable extensions that are in the System.Linq library.   Well, before we discuss any further, let's look at some sample code and the numbers.  First, let's take a look at the for loop and the LINQ expression.  This is just a simple find comparison:       // find implemented via LINQ     public static bool FindViaLinq(IEnumerable<int> list, int target)     {         return list.Any(item => item == target);     }         // find implemented via standard iteration     public static bool FindViaIteration(IEnumerable<int> list, int target)     {         foreach (var i in list)         {             if (i == target)             {                 return true;             }         }           return false;     }   Okay, looking at this from a maintainability point of view, the Linq expression is definitely more concise (8 lines down to 1) and is very readable in intention.  You don't have to actually analyze the behavior of the loop to determine what it's doing.   So let's take a look at performance metrics from 100,000 iterations of these methods on a List<int> of varying sizes filled with random data.  For this test, we fill a target array with 100,000 random integers and then run the exact same pseudo-random targets through both searches.                       List<T> On 100,000 Iterations     Method      Size     Total (ms)  Per Iteration (ms)  % Slower     Any         10       26          0.00046             30.00%     Iteration   10       20          0.00023             -     Any         100      116         0.00201             18.37%     Iteration   100      98          0.00118             -     Any         1000     1058        0.01853             16.78%     Iteration   1000     906         0.01155             -     Any         10,000   10,383      0.18189             17.41%     Iteration   10,000   8843        0.11362             -     Any         100,000  104,004     1.8297              18.27%     Iteration   100,000  87,941      1.13163             -   The LINQ expression is running about 17% slower for average size collections and worse for smaller collections.  Presumably, this is due to the overhead of the state machine used to track the iterators for the yield returns in the LINQ expressions, which seems about right in a tight loop such as this.   So what about other LINQ expressions?  After all, Any() is one of the more trivial ones.  I decided to try the TakeWhile() algorithm using a Count() to get the position stopped like the sample Pete was using in his blog that Resharper refactored for him into LINQ:       // Linq form     public static int GetTargetPosition1(IEnumerable<int> list, int target)     {         return list.TakeWhile(item => item != target).Count();     }       // traditionally iterative form     public static int GetTargetPosition2(IEnumerable<int> list, int target)     {         int count = 0;           foreach (var i in list)         {             if(i == target)             {                 break;             }               ++count;         }           return count;     }   Once again, the LINQ expression is much shorter, easier to read, and should be easier to maintain over time, reducing the cost of technical debt.  So I ran these through the same test data:                       List<T> On 100,000 Iterations     Method      Size     Total (ms)  Per Iteration (ms)  % Slower     TakeWhile   10       41          0.00041             128%     Iteration   10       18          0.00018             -     TakeWhile   100      171         0.00171             88%     Iteration   100      91          0.00091             -     TakeWhile   1000     1604        0.01604             94%     Iteration   1000     825         0.00825             -     TakeWhile   10,000   15765       0.15765             92%     Iteration   10,000   8204        0.08204             -     TakeWhile   100,000  156950      1.5695              92%     Iteration   100,000  81635       0.81635             -     Wow!  I expected some overhead due to the state machines iterators produce, but 90% slower?  That seems a little heavy to me.  So then I thought, well, what if TakeWhile() is not the right tool for the job?  The problem is TakeWhile returns each item for processing using yield return, whereas our for-loop really doesn't care about the item beyond using it as a stop condition to evaluate. So what if that back and forth with the iterator state machine is the problem?  Well, we can quickly create an (albeit ugly) lambda that uses the Any() along with a count in a closure (if a LINQ guru knows a better way PLEASE let me know!), after all , this is more consistent with what we're trying to do, we're trying to find the first occurence of an item and halt once we find it, we just happen to be counting on the way.  This mostly matches Any().       // a new method that uses linq but evaluates the count in a closure.     public static int TakeWhileViaLinq2(IEnumerable<int> list, int target)     {         int count = 0;         list.Any(item =>             {                 if(item == target)                 {                     return true;                 }                   ++count;                 return false;             });         return count;     }     Now how does this one compare?                         List<T> On 100,000 Iterations     Method         Size     Total (ms)  Per Iteration (ms)  % Slower     TakeWhile      10       41          0.00041             128%     Any w/Closure  10       23          0.00023             28%     Iteration      10       18          0.00018             -     TakeWhile      100      171         0.00171             88%     Any w/Closure  100      116         0.00116             27%     Iteration      100      91          0.00091             -     TakeWhile      1000     1604        0.01604             94%     Any w/Closure  1000     1101        0.01101             33%     Iteration      1000     825         0.00825             -     TakeWhile      10,000   15765       0.15765             92%     Any w/Closure  10,000   10802       0.10802             32%     Iteration      10,000   8204        0.08204             -     TakeWhile      100,000  156950      1.5695              92%     Any w/Closure  100,000  108378      1.08378             33%     Iteration      100,000  81635       0.81635             -     Much better!  It seems that the overhead of TakeAny() returning each item and updating the state in the state machine is drastically reduced by using Any() since Any() iterates forward until it finds the value we're looking for -- for the task we're attempting to do.   So the lesson there is, make sure when you use a LINQ expression you're choosing the best expression for the job, because if you're doing more work than you really need, you'll have a slower algorithm.  But this is true of any choice of algorithm or collection in general.     Even with the Any() with the count in the closure it is still about 30% slower, but let's consider that angle carefully.  For a list of 100,000 items, it was the difference between 1.01 ms and 0.82 ms roughly in a List<T>.  That's really not that bad at all in the grand scheme of things.  Even running at 90% slower with TakeWhile(), for the vast majority of my projects, an extra millisecond to save potential errors in the long term and improve maintainability is a small price to pay.  And if your typical list is 1000 items or less we're talking only microseconds worth of difference.   It's like they say: 90% of your performance bottlenecks are in 2% of your code, so over-optimizing almost never pays off.  So personally, I'll take the LINQ expression wherever I can because they will be easier to read and maintain (thus reducing technical debt) and I can rely on Microsoft's development to have coded and unit tested those algorithm fully for me instead of relying on a developer to code the loop logic correctly.   If something's 90% slower, yes, it's worth keeping in mind, but it's really not until you start get magnitudes-of-order slower (10x, 100x, 1000x) that alarm bells should really go off.  And if I ever do need that last millisecond of performance?  Well then I'll optimize JUST THAT problem spot.  To me it's worth it for the readability, speed-to-market, and maintainability.

  Read the article

• techniques for an AI for a highly cramped turn-based tactics game

  - by Adam M.

  I'm trying to write an AI for a tactics game in the vein of Final Fantasy Tactics or Vandal Hearts. I can't change the game rules in any way, only upgrade the AI. I have experience programming AI for classic board games (basically minimax and its variants), but I think the branching factor is too great for the approach to be reasonable here. I'll describe the game and some current AI flaws that I'd like to fix. I'd like to hear ideas for applicable techniques. I'm a decent enough programmer, so I only need the ideas, not an implementation (though that's always appreciated). I'd rather not expend effort chasing (too many) dead ends, so although speculation and brainstorming are good and probably helpful, I'd prefer to hear from somebody with actual experience solving this kind of problem. For those who know it, the game is the land battle mini-game in Sid Meier's Pirates! (2004) and you can skim/skip the next two paragraphs. For those who don't, here's briefly how it works. The battle is turn-based and takes place on a 16x16 grid. There are three terrain types: clear (no hindrance), forest (hinders movement, ranged attacks, and sight), and rock (impassible, but does not hinder attacks or sight). The map is randomly generated with roughly equal amounts of each type of terrain. Because there are many rock and forest tiles, movement is typically very cramped. This is tactically important. The terrain is not flat; higher terrain gives minor bonuses. The terrain is known to both sides. The player is always the attacker and the AI is always the defender, so it's perfectly valid for the AI to set up a defensive position and just wait. The player wins by killing all defenders or by getting a unit to the city gates (a tile on the other side of the map). There are very few units on each side, usually 4-8. Because of this, it's crucial not to take damage without gaining some advantage from it. Units can take multiple actions per turn. All units on one side move before any units on the other side. Order of execution is important, and interleaving of actions between units is often useful. Units have melee and ranged attacks. Melee attacks vary widely in strength; ranged attacks have the same strength but vary in range. The main challenges I face are these: Lots of useful move combinations start with a "useless" move that gains no immediate advantage, or even loses advantage, in order to set up a powerful flank attack in the future. And, since the player units are stronger and have longer range, the AI pretty much always has to take some losses before they can start to gain kills. The AI must be able to look ahead to distinguish between sacrificial actions that provide a future benefit and those that don't. Because the terrain is so cramped, most of the tactics come down to achieving good positioning with multiple units that work together to defend an area. For instance, two defenders can often dominate a narrow pass by positioning themselves so an enemy unit attempting to pass must expose itself to a flank attack. But one defender in the same pass would be useless, and three units can defend a slightly larger pass. Etc. The AI should be able to figure out where the player must go to reach the city gates and how to best position its few units to cover the approaches, shifting, splitting, or combining them appropriately as the player moves. Because flank attacks are extremely deadly (and engineering flank attacks is key to the player strategy), the AI should be competent at moving its units so that they cover each other's flanks unless the sacrifice of a unit would give a substantial benefit. They should also be able to force flank attacks on players, for instance by threatening a unit from two different directions such that responding to one threat exposes the flank to the other. The AI should attack if possible, but sometimes there are no good ways to approach the player's position. In that case, the AI should be able to recognize this and set up a defensive position of its own. But the AI shouldn't be vulnerable to a trivial exploit where the player repeatedly opens and closes a hole in his defense and shoots at the AI as it approaches and retreats. That is, the AI should ideally be able to recognize that the player is capable of establishing a solid defense of an area, even if the defense is not currently in place. (I suppose if a good unit allocation algorithm existed, as needed for the second bullet point, the AI could run it on the player units to see where they could defend.) Because it's important to choose a good order of action and interleave actions between units, it's not as simple as just finding the best move for each unit in turn. All of these can be accomplished with a minimax search in theory, but the search space is too large, so specialized techniques are needed. I thought about techniques such as influence mapping, but I don't see how to use the technique to great effect. I thought about assigning goals to the units. This can help them work together in some limited way, and the problem of "how do I accomplish this goal?" is easier to solve than "how do I win this battle?", but assigning good goals is a hard problem in itself, because it requires knowing whether the goal is achievable and whether it's a good use of resources. So, does anyone have specific ideas for techniques that can help cleverize this AI? Update: I found a related question on Stackoverflow: http://stackoverflow.com/questions/3133273/ai-for-a-final-fantasy-tactics-like-game The selected answer gives a decent approach to choosing between alternative actions, but it doesn't seem to have much ability to look into the future and discern beneficial sacrifices from wasteful ones. It also focuses on a single unit at a time and it's not clear how it could be extended to support cooperation between units in defending or attacking.

  Read the article

• ASP.NET MVC JavaScript Routing

  - by zowens

  Have you ever done this sort of thing in your ASP.NET MVC view? The weird thing about this isn’t the alert function, it’s the code block containing the Url formation using the ASP.NET MVC UrlHelper. The terrible thing about this experience is the obvious lack of IntelliSense and this ugly inline JavaScript code. Inline JavaScript isn’t portable to other pages beyond the current page of execution. It is generally considered bad practice to use inline JavaScript in your public-facing pages. How ludicrous would it be to copy and paste the entire jQuery code base into your pages…? Not something you’d ever consider doing. The problem is that your URLs have to be generated by ASP.NET at runtime and really can’t be copied to your JavaScript code without some trickery. How about this? Does the hard-coded URL bother you? It really bothers me. The typical solution to this whole routing in JavaScript issue is to just hard-code your URLs into your JavaScript files and call it done. But what if your URLs change? You have to now go an track down the places in JavaScript and manually replace them. What if you get the pattern wrong? Do you have tests around it? This isn’t something you should have to worry about.   The Solution To Our Problems The solution is to port routing over to JavaScript. Does that sound daunting to you? It’s actually not very hard, but I decided to create my own generator that will do all the work for you. What I have created is a very basic port of the route formation feature of ASP.NET routing. It will generate the formatted URLs based on your routing patterns. Here’s how you’d do this: Does that feel familiar? It looks a lot like something you’d do inside of your ASP.NET MVC views… but this is inside of a JavaScript file… just a plain ol’ .js file.  Your first question might be why do you have to have that “.toUrl()” thing. The reason is that I wanted to make POST and GET requests dead simple. Here’s how you’d do a POST request (and the same would work with a GET request):   The first parameter is extra data passed to the post request and the second parameter is a function that handles the success of the POST request. If you’re familiar with jQuery’s Ajax goodness, you’ll know how to use it. (if not, check out http://api.jquery.com/jQuery.Post/ and the parameters are essentially the same). But we still haven’t gotten rid of the magic strings. We still have controller names and action names represented as strings. This is going to blow your mind… If you’ve seen T4MVC, this will look familiar. We’re essentially doing the same sort of thing with my JavaScript router, but we’re porting the concept to JavaScript. The good news is that parameters to the controllers are directly reflected in the action function, just like T4MVC. And the even better news… IntlliSense is easily transferred to the JavaScript version if you’re using Visual Studio as your JavaScript editor. The additional data parameter gives you the ability to pass extra routing data to the URL formatter.   About the Magic You may be wondering how this all work. It’s actually quite simple. I’ve built a simple jQuery pluggin (called routeManager) that hangs off the main jQuery namespace and routes all the URLs. Every time your solution builds, a routing file will be generated with this pluggin, all your route and controller definitions along with your documentation. Then by the power of Visual Studio, you get some really slick IntelliSense that is hard to live without. But there are a few steps you have to take before this whole thing is going to work. First and foremost, you need a reference to the JsRouting.Core.dll to your projects containing controllers or routes. Second, you have to specify your routes in a bit of a non-standard way. See, we can’t just pull routes out of your App_Start in your Global.asax. We force you to build a route source like this: The way we determine the routes is by pulling in all RouteSources and generating routes based upon the mapped routes. There are various reasons why we can’t use RouteCollection (different post for another day)… but in this case, you get the same route mapping experience. Converting the RouteSource to a RouteCollection is trivial (there’s an extension method for that). Next thing you have to do is generate a documentation XML file. This is done by going to the project settings, going to the build tab and clicking the checkbox. (this isn’t required, but nice to have). The final thing you need to do is hook up the generation mechanism. Pop open your project file and look for the AfterBuild step. Now change the build step task to look like this: The “PathToOutputExe” is the path to the JsRouting.Output.exe file. This will change based on where you put the EXE. The “PathToOutputJs” is a path to the output JavaScript file. The “DicrectoryOfAssemblies” is a path to the directory containing controller and routing DLLs. The JsRouting.Output.exe executable pulls in all these assemblies and scans them for controllers and route sources.   Now that wasn’t too bad, was it :)   The State of the Project This is definitely not complete… I have a lot of plans for this little project of mine. For starters, I need to look at the generation mechanism. Either I will be creating a utility that will do the project file manipulation or I will go a different direction. I’d like some feedback on this if you feel partial either way. Another thing I don’t support currently is areas. While this wouldn’t be too hard to support, I just don’t use areas and I wanted something up quickly (this is, after all, for a current project of mine). I’ll be adding support shortly. There are a few things that I haven’t covered in this post that I will most certainly be covering in another post, such as routing constraints and how these will be translated to JavaScript. I decided to open source this whole thing, since it’s a nice little utility I think others should really be using. Currently we’re using ASP.NET MVC 2, but it should work with MVC 3 as well. I’ll upgrade it as soon as MVC 3 is released. Along those same lines, I’m investigating how this could be put on the NuGet feed. Show me the Bits! OK, OK! The code is posted on my GitHub account. Go nuts. Tell me what you think. Tell me what you want. Tell me that you hate it. All feedback is welcome! https://github.com/zowens/ASP.NET-MVC-JavaScript-Routing

  Read the article

• Using Node.js as an accelerator for WCF REST services

  - by Elton Stoneman

  Node.js is a server-side JavaScript platform "for easily building fast, scalable network applications". It's built on Google's V8 JavaScript engine and uses an (almost) entirely async event-driven processing model, running in a single thread. If you're new to Node and your reaction is "why would I want to run JavaScript on the server side?", this is the headline answer: in 150 lines of JavaScript you can build a Node.js app which works as an accelerator for WCF REST services*. It can double your messages-per-second throughput, halve your CPU workload and use one-fifth of the memory footprint, compared to the WCF services direct.   Well, it can if: 1) your WCF services are first-class HTTP citizens, honouring client cache ETag headers in request and response; 2) your services do a reasonable amount of work to build a response; 3) your data is read more often than it's written. In one of my projects I have a set of REST services in WCF which deal with data that only gets updated weekly, but which can be read hundreds of times an hour. The services issue ETags and will return a 304 if the client sends a request with the current ETag, which means in the most common scenario the client uses its local cached copy. But when the weekly update happens, then all the client caches are invalidated and they all need the same new data. Then the service will get hundreds of requests with old ETags, and they go through the full service stack to build the same response for each, taking up threads and processing time. Part of that processing means going off to a database on a separate cloud, which introduces more latency and downtime potential.   We can use ASP.NET output caching with WCF to solve the repeated processing problem, but the server will still be thread-bound on incoming requests, and to get the current ETags reliably needs a database call per request. The accelerator solves that by running as a proxy - all client calls come into the proxy, and the proxy routes calls to the underlying REST service. We could use Node as a straight passthrough proxy and expect some benefit, as the server would be less thread-bound, but we would still have one WCF and one database call per proxy call. But add some smart caching logic to the proxy, and share ETags between Node and WCF (so the proxy doesn't even need to call the servcie to get the current ETag), and the underlying service will only be invoked when data has changed, and then only once - all subsequent client requests will be served from the proxy cache.   I've built this as a sample up on GitHub: NodeWcfAccelerator on sixeyed.codegallery. Here's how the architecture looks:     The code is very simple. The Node proxy runs on port 8010 and all client requests target the proxy. If the client request has an ETag header then the proxy looks up the ETag in the tag cache to see if it is current - the sample uses memcached to share ETags between .NET and Node. If the ETag from the client matches the current server tag, the proxy sends a 304 response with an empty body to the client, telling it to use its own cached version of the data. If the ETag from the client is stale, the proxy looks for a local cached version of the response, checking for a file named after the current ETag. If that file exists, its contents are returned to the client as the body in a 200 response, which includes the current ETag in the header. If the proxy does not have a local cached file for the service response, it calls the service, and writes the WCF response to the local cache file, and to the body of a 200 response for the client. So the WCF service is only troubled if both client and proxy have stale (or no) caches.   The only (vaguely) clever bit in the sample is using the ETag cache, so the proxy can serve cached requests without any communication with the underlying service, which it does completely generically, so the proxy has no notion of what it is serving or what the services it proxies are doing. The relative path from the URL is used as the lookup key, so there's no shared key-generation logic between .NET and Node, and when WCF stores a tag it also stores the "read" URL against the ETag so it can be used for a reverse lookup, e.g:   Key Value /WcfSampleService/PersonService.svc/rest/fetch/3 "28cd4796-76b8-451b-adfd-75cb50a50fa6" "28cd4796-76b8-451b-adfd-75cb50a50fa6" /WcfSampleService/PersonService.svc/rest/fetch/3    In Node we read the cache using the incoming URL path as the key and we know that "28cd4796-76b8-451b-adfd-75cb50a50fa6" is the current ETag; we look for a local cached response in /caches/28cd4796-76b8-451b-adfd-75cb50a50fa6.body (and the corresponding .header file which contains the original service response headers, so the proxy response is exactly the same as the underlying service). When the data is updated, we need to invalidate the ETag cache – which is why we need the reverse lookup in the cache. In the WCF update service, we don't need to know the URL of the related read service - we fetch the entity from the database, do a reverse lookup on the tag cache using the old ETag to get the read URL, update the new ETag against the URL, store the new reverse lookup and delete the old one.   Running Apache Bench against the two endpoints gives the headline performance comparison. Making 1000 requests with concurrency of 100, and not sending any ETag headers in the requests, with the Node proxy I get 102 requests handled per second, average response time of 975 milliseconds with 90% of responses served within 850 milliseconds; going direct to WCF with the same parameters, I get 53 requests handled per second, mean response time of 1853 milliseconds, with 90% of response served within 3260 milliseconds. Informally monitoring server usage during the tests, Node maxed at 20% CPU and 20Mb memory; IIS maxed at 60% CPU and 100Mb memory.   Note that the sample WCF service does a database read and sleeps for 250 milliseconds to simulate a moderate processing load, so this is *not* a baseline Node-vs-WCF comparison, but for similar scenarios where the  service call is expensive but applicable to numerous clients for a long timespan, the performance boost from the accelerator is considerable.     * - actually, the accelerator will work nicely for any HTTP request, where the URL (path + querystring) uniquely identifies a resource. In the sample, there is an assumption that the ETag is a GUID wrapped in double-quotes (e.g. "28cd4796-76b8-451b-adfd-75cb50a50fa6") – which is the default for WCF services. I use that assumption to name the cache files uniquely, but it is a trivial change to adapt to other ETag formats.

  Read the article

• The Shift: how Orchard painlessly shifted to document storage, and how it’ll affect you

  - by Bertrand Le Roy

  We’ve known it all along. The storage for Orchard content items would be much more efficient using a document database than a relational one. Orchard content items are composed of parts that serialize naturally into infoset kinds of documents. Storing them as relational data like we’ve done so far was unnatural and requires the data for a single item to span multiple tables, related through 1-1 relationships. This means lots of joins in queries, and a great potential for Select N+1 problems. Document databases, unfortunately, are still a tough sell in many places that prefer the more familiar relational model. Being able to x-copy Orchard to hosters has also been a basic constraint in the design of Orchard. Combine those with the necessity at the time to run in medium trust, and with license compatibility issues, and you’ll find yourself with very few reasonable choices. So we went, a little reluctantly, for relational SQL stores, with the dream of one day transitioning to document storage. We have played for a while with the idea of building our own document storage on top of SQL databases, and Sébastien implemented something more than decent along those lines, but we had a better way all along that we didn’t notice until recently… In Orchard, there are fields, which are named properties that you can add dynamically to a content part. Because they are so dynamic, we have been storing them as XML into a column on the main content item table. This infoset storage and its associated API are fairly generic, but were only used for fields. The breakthrough was when Sébastien realized how this existing storage could give us the advantages of document storage with minimal changes, while continuing to use relational databases as the substrate. public bool CommercialPrices { get { return this.Retrieve(p => p.CommercialPrices); } set { this.Store(p => p.CommercialPrices, value); } } This code is very compact and efficient because the API can infer from the expression what the type and name of the property are. It is then able to do the proper conversions for you. For this code to work in a content part, there is no need for a record at all. This is particularly nice for site settings: one query on one table and you get everything you need. This shows how the existing infoset solves the data storage problem, but you still need to query. Well, for those properties that need to be filtered and sorted on, you can still use the current record-based relational system. This of course continues to work. We do however provide APIs that make it trivial to store into both record properties and the infoset storage in one operation: public double Price { get { return Retrieve(r => r.Price); } set { Store(r => r.Price, value); } } This code looks strikingly similar to the non-record case above. The difference is that it will manage both the infoset and the record-based storages. The call to the Store method will send the data in both places, keeping them in sync. The call to the Retrieve method does something even cooler: if the property you’re looking for exists in the infoset, it will return it, but if it doesn’t, it will automatically look into the record for it. And if that wasn’t cool enough, it will take that value from the record and store it into the infoset for the next time it’s required. This means that your data will start automagically migrating to infoset storage just by virtue of using the code above instead of the usual: public double Price { get { return Record.Price; } set { Record.Price = value; } } As your users browse the site, it will get faster and faster as Select N+1 issues will optimize themselves away. If you preferred, you could still have explicit migration code, but it really shouldn’t be necessary most of the time. If you do already have code using QueryHints to mitigate Select N+1 issues, you might want to reconsider those, as with the new system, you’ll want to avoid joins that you don’t need for filtering or sorting, further optimizing your queries. There are some rare cases where the storage of the property must be handled differently. Check out this string[] property on SearchSettingsPart for example: public string[] SearchedFields { get { return (Retrieve<string>("SearchedFields") ?? "") .Split(new[] {',', ' '}, StringSplitOptions.RemoveEmptyEntries); } set { Store("SearchedFields", String.Join(", ", value)); } } The array of strings is transformed by the property accessors into and from a comma-separated list stored in a string. The Retrieve and Store overloads used in this case are lower-level versions that explicitly specify the type and name of the attribute to retrieve or store. You may be wondering what this means for code or operations that look directly at the database tables instead of going through the new infoset APIs. Even if there is a record, the infoset version of the property will win if it exists, so it is necessary to keep the infoset up-to-date. It’s not very complicated, but definitely something to keep in mind. Here is what a product record looks like in Nwazet.Commerce for example: And here is the same data in the infoset: The infoset is stored in Orchard_Framework_ContentItemRecord or Orchard_Framework_ContentItemVersionRecord, depending on whether the content type is versionable or not. A good way to find what you’re looking for is to inspect the record table first, as it’s usually easier to read, and then get the item record of the same id. Here is the detailed XML document for this product: <Data> <ProductPart Inventory="40" Price="18" Sku="pi-camera-box" OutOfStockMessage="" AllowBackOrder="false" Weight="0.2" Size="" ShippingCost="null" IsDigital="false" /> <ProductAttributesPart Attributes="" /> <AutoroutePart DisplayAlias="camera-box" /> <TitlePart Title="Nwazet Pi Camera Box" /> <BodyPart Text="[...]" /> <CommonPart CreatedUtc="2013-09-10T00:39:00Z" PublishedUtc="2013-09-14T01:07:47Z" /> </Data> The data is neatly organized under each part. It is easy to see how that document is all you need to know about that content item, all in one table. If you want to modify that data directly in the database, you should be careful to do it in both the record table and the infoset in the content item record. In this configuration, the record is now nothing more than an index, and will only be used for sorting and filtering. Of course, it’s perfectly fine to mix record-backed properties and record-less properties on the same part. It really depends what you think must be sorted and filtered on. In turn, this potentially simplifies migrations considerably. So here it is, the great shift of Orchard to document storage, something that Orchard has been designed for all along, and that we were able to implement with a satisfying and surprising economy of resources. Expect this code to make its way into the 1.8 version of Orchard when that’s available.

  Read the article

• CodePlex Daily Summary for Sunday, March 28, 2010

  CodePlex Daily Summary for Sunday, March 28, 2010New ProjectsFeed Tracker: Feed Tracker allows you to track your favorite feeds (RSS 2.0 and Atom 1.0) and open them up directly in your browser.FIM 2010 Resource Management Client: The Forefront Identity Management 2010 Resource Management Client is a library to communicate with the FIM 2010 web service. The development langu...Infection Protection: A game about controlling disease outbreak in a city. Developed for OGPC 2010, using Qt.OrthoLab: Homepage of Orthocone open-source laboratory.Paragliding ThermalMarker: Paragliding / Hanggliding Windows Application that receives waypoint files and returns only the thermals that get triggered more often in a place.RSSFalls: RssFalls makes it easier for developers to download RSS or Podcast enclosures.String Library for C++ Language: StrLib++ is a string library for C++ language. for now it support only ANSI strings, later Unicode support will added for UT8, UTF16 and UTF32 for...Sweeper: Sweeper is a Visual Studio 2008 add-in for C# that takes care of many of the trivial code-formatting issues that developers run into - particularly...System.Common: A .Net library that provides methods, properties and more that the .Net Framework doesn't provide.Tiveriad: The framework is designed to help you more easily build modular Windows applicationT-Shirts Online: Online shop build in Silverlight 4 using DIBS as payment module.New ReleasesArkSwitch: ArkSwitch v1.1.3: This release has some important changes. Thanks to MichyPrima for helping with some of the code. 1. Improved theming to more easily support multip...Catharsis: Catharsis 2.5 on catarsa.com: The Catharsis framework has finally its own portal http://catarsa.com The latest release version is 2.5 - string names of properties are not any ...EffiProz - A Pure C# Database: EffiProz CF 1.0: EffiProz for .Net compact framework.Encrypted Notes: Encrypted Notes 1.6: This is the latest version of Encrypted Notes (1.6). It has an installer - it will create a directory 'CPascoe' in My Documents. Once you have ext...Extend SmallBasic: Teaching Extensions v.009: Added Pentagon Crazy Recipe QuizGapi.NET - .NET (C#) wrapper for Google API: Gapi.NET 0.5.0.0: - Fixed some minor bugs. - Add minor features. - Performance improvement. See code check-ins for detailed informationHouseFly experimental controls: HouseFly experimental control: Alpha version of HouseFly experimental controlsiTuner - The iTunes Companion: iTuner 1.2.3738 Beta 2: V1.2 allows you to synchronize one or more iTunes playlists to a USB MP3 player. Beta 2 resolves all known issues. This continues the evolution ye...jQuery Library for SharePoint Web Services: SPServices 0.5.4: IMPORTANT NOTE: This release is in an alpha state. You should only download it if you know what you are getting and are interested in testing it f...JSINQ - LINQ to Objects for JavaScript: JSINQ 1.0: This is the first stable release of JSINQ. It is fully compatible with the 0.9 beta release. It contains the following new features: Now supports ...MSBuild Mercurial Tasks: 1.0.0 Beta: First release of the application. This version integrates all the basic functionalities of Mercurial as defined in the Use Case 1.Open Portal Foundation: Open Portal Foundation V1.4.2: What's news? noscript template was updated naming convention for layout autogenerated controls now use the "master" prefix. The documentation ce...Open Portal Foundation: Open Portal Foundation V1.4.4: What's news? ASP .NET Master page support for custom aspx integrated pages New usercontrols for ASCX, ASPX and Master page integration : Link: f...Paint.NET PSD Plugin: 1.5.0: RLE compression is now working fully on save. File sizes are now competitive with Photoshop's. Saving takes about twice as long with RLE compressi...Paragliding ThermalMarker: ThermalMarker_Alfa0.1: Release Alfa 1Simple Service Locator: Simple Service Locator v0.7: The Simple Service Locator is an easy-to-use Inversion of Control library that is a complete implementation of the Common Service Locator interface...String Library for C++ Language: Release 0.9: version 0.9 beta release DO NOT USE IN SERIOUS PROJECTS this release use default application heap, and because visual studio is using special debug...Sweeper: Sweeper Alpha 1: SweeperA Visual Studio Add-in for C# Code Formatting - Visual Studio 2008 Includes: A UI for options, Enable or disable any specific task you want ...T-Shirts Online: 1.0: First release of the online shop.Twilio Server Library for .NET (TSL.NET): v0.1.0 Beta: This is the first release of TSL.NET. This v0.1.0 release is a Beta. Subsequent builds will be posted as v0.1.x and release-candidate Betas will be...Vr30 OS: Facebook 1.0: Connect you to Facebook without your web browser.Vr30 OS: SkyBlog 1.0: SkyBlog without web browser.Vr30 OS: YouTube 1.0: Youtube without web browserWeb Image Resize Handler: Web Image Resize, Zoom, Rotate and Greyscale v.1.0: Efficient Web Image Resize, Zoom, Rotate and Greyscale cacheing handler for ASP.Net.WinXound: WinXound 3.3.0 Beta 1 for Mac OsX: This is the first Beta release for Apple Mac OsX (Universal Binary). DEBUG HELP NEEDED ! Please signal bugs, suggestions or feedback to: stefano_b...Most Popular ProjectsMetaSharpRawrWBFS ManagerASP.NET Ajax LibraryMicrosoft SQL Server Product Samples: DatabaseSilverlight ToolkitAJAX Control ToolkitLiveUpload to FacebookWindows Presentation Foundation (WPF)ASP.NETMost Active ProjectsRawrjQuery Library for SharePoint Web ServicesManaged Extensibility FrameworkBlogEngine.NETMicrosoft Biology Foundationpatterns & practices: Composite WPF and SilverlightLINQ to TwitterFarseer Physics EngineTable2ClassNB_Store - Free DotNetNuke Ecommerce Catalog Module

  Read the article

• SortedDictionary and SortedList

  - by Simon Cooper

  Apart from Dictionary<TKey, TValue>, there's two other dictionaries in the BCL - SortedDictionary<TKey, TValue> and SortedList<TKey, TValue>. On the face of it, these two classes do the same thing - provide an IDictionary<TKey, TValue> interface where the iterator returns the items sorted by the key. So what's the difference between them, and when should you use one rather than the other? (as in my previous post, I'll assume you have some basic algorithm & datastructure knowledge) SortedDictionary We'll first cover SortedDictionary. This is implemented as a special sort of binary tree called a red-black tree. Essentially, it's a binary tree that uses various constraints on how the nodes of the tree can be arranged to ensure the tree is always roughly balanced (for more gory algorithmical details, see the wikipedia link above). What I'm concerned about in this post is how the .NET SortedDictionary is actually implemented. In .NET 4, behind the scenes, the actual implementation of the tree is delegated to a SortedSet<KeyValuePair<TKey, TValue>>. One example tree might look like this: Each node in the above tree is stored as a separate SortedSet<T>.Node object (remember, in a SortedDictionary, T is instantiated to KeyValuePair<TKey, TValue>): class Node { public bool IsRed; public T Item; public SortedSet<T>.Node Left; public SortedSet<T>.Node Right; } The SortedSet only stores a reference to the root node; all the data in the tree is accessed by traversing the Left and Right node references until you reach the node you're looking for. Each individual node can be physically stored anywhere in memory; what's important is the relationship between the nodes. This is also why there is no constructor to SortedDictionary or SortedSet that takes an integer representing the capacity; there are no internal arrays that need to be created and resized. This may seen trivial, but it's an important distinction between SortedDictionary and SortedList that I'll cover later on. And that's pretty much it; it's a standard red-black tree. Plenty of webpages and datastructure books cover the algorithms behind the tree itself far better than I could. What's interesting is the comparions between SortedDictionary and SortedList, which I'll cover at the end. As a side point, SortedDictionary has existed in the BCL ever since .NET 2. That means that, all through .NET 2, 3, and 3.5, there has been a bona-fide sorted set class in the BCL (called TreeSet). However, it was internal, so it couldn't be used outside System.dll. Only in .NET 4 was this class exposed as SortedSet. SortedList Whereas SortedDictionary didn't use any backing arrays, SortedList does. It is implemented just as the name suggests; two arrays, one containing the keys, and one the values (I've just used random letters for the values): The items in the keys array are always guarenteed to be stored in sorted order, and the value corresponding to each key is stored in the same index as the key in the values array. In this example, the value for key item 5 is 'z', and for key item 8 is 'm'. Whenever an item is inserted or removed from the SortedList, a binary search is run on the keys array to find the correct index, then all the items in the arrays are shifted to accomodate the new or removed item. For example, if the key 3 was removed, a binary search would be run to find the array index the item was at, then everything above that index would be moved down by one: and then if the key/value pair {7, 'f'} was added, a binary search would be run on the keys to find the index to insert the new item, and everything above that index would be moved up to accomodate the new item: If another item was then added, both arrays would be resized (to a length of 10) before the new item was added to the arrays. As you can see, any insertions or removals in the middle of the list require a proportion of the array contents to be moved; an O(n) operation. However, if the insertion or removal is at the end of the array (ie the largest key), then it's only O(log n); the cost of the binary search to determine it does actually need to be added to the end (excluding the occasional O(n) cost of resizing the arrays to fit more items). As a side effect of using backing arrays, SortedList offers IList Keys and Values views that simply use the backing keys or values arrays, as well as various methods utilising the array index of stored items, which SortedDictionary does not (and cannot) offer. The Comparison So, when should you use one and not the other? Well, here's the important differences: Memory usage SortedDictionary and SortedList have got very different memory profiles. SortedDictionary... has a memory overhead of one object instance, a bool, and two references per item. On 64-bit systems, this adds up to ~40 bytes, not including the stored item and the reference to it from the Node object. stores the items in separate objects that can be spread all over the heap. This helps to keep memory fragmentation low, as the individual node objects can be allocated wherever there's a spare 60 bytes. In contrast, SortedList... has no additional overhead per item (only the reference to it in the array entries), however the backing arrays can be significantly larger than you need; every time the arrays are resized they double in size. That means that if you add 513 items to a SortedList, the backing arrays will each have a length of 1024. To conteract this, the TrimExcess method resizes the arrays back down to the actual size needed, or you can simply assign list.Capacity = list.Count. stores its items in a continuous block in memory. If the list stores thousands of items, this can cause significant problems with Large Object Heap memory fragmentation as the array resizes, which SortedDictionary doesn't have. Performance Operations on a SortedDictionary always have O(log n) performance, regardless of where in the collection you're adding or removing items. In contrast, SortedList has O(n) performance when you're altering the middle of the collection. If you're adding or removing from the end (ie the largest item), then performance is O(log n), same as SortedDictionary (in practice, it will likely be slightly faster, due to the array items all being in the same area in memory, also called locality of reference). So, when should you use one and not the other? As always with these sort of things, there are no hard-and-fast rules. But generally, if you: need to access items using their index within the collection are populating the dictionary all at once from sorted data aren't adding or removing keys once it's populated then use a SortedList. But if you: don't know how many items are going to be in the dictionary are populating the dictionary from random, unsorted data are adding & removing items randomly then use a SortedDictionary. The default (again, there's no definite rules on these sort of things!) should be to use SortedDictionary, unless there's a good reason to use SortedList, due to the bad performance of SortedList when altering the middle of the collection.

  Read the article

• Windows can see Ubuntu Server printer, but can't print to it

  - by Mike

  I have an old desktop that I'm trying to set up as a home backup/print server. Backup was trivial, but am having issues getting the printing to work. The printer is connected to the server running Ubuntu Server 9.10 (no gui). If I access the printer via http://hostname:631/printers/, I am able to print a test page, so I know the printer is working; however, I am having no luck from Windows. Windows can see the printer when browsed via \hostname\, but I am unable to connect. Windows says "Windows cannot connect to the printer" without indicating why. Any suggestions? From /etc/samba/smb.conf: [global] workgroup = WORKGROUP dns proxy = no security = user username map = /etc/samba/smbusers encrypt passwords = true passdb backend = tdbsam obey pam restrictions = yes unix password sync = yes passwd program = /usr/bin/passwd %u passwd chat = *Enter\snew\s*\spassword:* %n\n *Retype\snew\s*\spassword:* %n\n *password\supdated\ssuccessfully* . pam password change = yes map to guest = bad user load printers = yes printing = cups printcap name = cups [printers] comment = All Printers browseable = no path = /var/spool/samba writable = no printable = yes guest ok = yes read only = yes create mask = 0700 [print$] comment = Printer Drivers path = /var/lib/samba/printers browseable = yes read only = yes guest ok = yes From /etc/cups/cupsd.conf: LogLevel warn SystemGroup lpadmin Port 631 Listen /var/run/cups/cups.sock Browsing On BrowseOrder allow,deny BrowseAllow all BrowseRemoteProtocols CUPS BrowseAddress @LOCAL BrowseLocalProtocols CUPS dnssd DefaultAuthType Basic <Location /> Order allow,deny Allow all </Location> <Location /admin> Order allow,deny Allow all </Location> <Location /admin/conf> AuthType Default Require user @SYSTEM Order allow,deny Allow all </Location> <Policy default> <Limit Send-Document Send-URI Hold-Job Release-Job Restart-Job Purge-Jobs Set-Job-Attributes Create-Job-Subscription Renew-Subscription Cancel-Subscription Get-Notifications Reprocess-Job Cancel-Current-Job Suspend-Current-Job Resume-Job CUPS-Move-Job CUPS-Get-Document> Require user @OWNER @SYSTEM Order deny,allow </Limit> <Limit CUPS-Add-Modify-Printer CUPS-Delete-Printer CUPS-Add-Modify-Class CUPS-Delete-Class CUPS-Set-Default CUPS-Get-Devices> AuthType Default Require user @SYSTEM Order deny,allow </Limit> <Limit Pause-Printer Resume-Printer Enable-Printer Disable-Printer Pause-Printer-After-Current-Job Hold-New-Jobs Release-Held-New-Jobs Deactivate-Printer Activate-Printer Restart-Printer Shutdown-Printer Startup-Printer Promote-Job Schedule-Job-After CUPS-Accept-Jobs CUPS-Reject-Jobs> AuthType Default Require user @SYSTEM Order deny,allow </Limit> <Limit CUPS-Authenticate-Job> Require user @OWNER @SYSTEM Order deny,allow </Limit> <Limit All> Order deny,allow </Limit> </Policy> <Policy authenticated> <Limit Create-Job Print-Job Print-URI> AuthType Default Order deny,allow </Limit> <Limit Send-Document Send-URI Hold-Job Release-Job Restart-Job Purge-Jobs Set-Job-Attributes Create-Job-Subscription Renew-Subscription Cancel-Subscription Get-Notifications Reprocess-Job Cancel-Current-Job Suspend-Current-Job Resume-Job CUPS-Move-Job CUPS-Get-Document> AuthType Default Require user @OWNER @SYSTEM Order deny,allow </Limit> <Limit CUPS-Add-Modify-Printer CUPS-Delete-Printer CUPS-Add-Modify-Class CUPS-Delete-Class CUPS-Set-Default> AuthType Default Require user @SYSTEM Order deny,allow </Limit> <Limit Pause-Printer Resume-Printer Enable-Printer Disable-Printer Pause-Printer-After-Current-Job Hold-New-Jobs Release-Held-New-Jobs Deactivate-Printer Activate-Printer Restart-Printer Shutdown-Printer Startup-Printer Promote-Job Schedule-Job-After CUPS-Accept-Jobs CUPS-Reject-Jobs> AuthType Default Require user @SYSTEM Order deny,allow </Limit> <Limit Cancel-Job CUPS-Authenticate-Job> AuthType Default Require user @OWNER @SYSTEM Order deny,allow </Limit> <Limit All> Order deny,allow </Limit> </Policy>

  Read the article

• Elegance, thy Name is jQuery

  - by SGWellens

  So, I'm browsing though some questions over on the Stack Overflow website and I found a good jQuery question just a few minutes old. Here is a link to it. It was a tough question; I knew that by answering it, I could learn new stuff and reinforce what I already knew: Reading is good, doing is better. Maybe I could help someone in the process too. I cut and pasted the HTML from the question into my Visual Studio IDE and went back to Stack Overflow to reread the question. Dang, someone had already answered it! And it was a great answer. I never even had a chance to start analyzing the issue. Now I know what a one-legged man feels like in an ass-kicking contest. Nevertheless, since the question and answer were so interesting, I decided to dissect them and learn as much as possible. The HTML consisted of some divs separated by h3 headings.  Note the elements are laid out sequentially with no programmatic grouping: <h3 class="heading">Heading 1</h3> <div>Content</div> <div>More content</div> <div>Even more content</div><h3 class="heading">Heading 2</h3> <div>some content</div> <div>some more content</div><h3 class="heading">Heading 3</h3> <div>other content</div></form></body>  The requirement was to wrap a div around each h3 heading and the subsequent divs grouping them into sections. Why? I don't know, I suppose if you screen-scrapped some HTML from another site, you might want to reformat it before displaying it on your own. Anyways… Here is the marvelously, succinct posted answer: $('.heading').each(function(){ $(this).nextUntil('.heading').andSelf().wrapAll('<div class="section">');}); I was familiar with all the parts except for nextUntil and andSelf. But, I'll analyze the whole answer for completeness. I'll do this by rewriting the posted answer in a different style and adding a boat-load of comments: function Test(){ // $Sections is a jQuery object and it will contain three elements var $Sections = $('.heading'); // use each to iterate over each of the three elements $Sections.each(function () { // $this is a jquery object containing the current element // being iterated var $this = $(this); // nextUntil gets the following sibling elements until it reaches // an element with the CSS class 'heading' // andSelf adds in the source element (this) to the collection $this = $this.nextUntil('.heading').andSelf(); // wrap the elements with a div $this.wrapAll('<div class="section" >'); });}  The code here doesn't look nearly as concise and elegant as the original answer. However, unless you and your staff are jQuery masters, during development it really helps to work through algorithms step by step. You can step through this code in the debugger and examine the jQuery objects to make sure one step is working before proceeding on to the next. It's much easier to debug and troubleshoot when each logical coding step is a separate line of code. Note: You may think the original code runs much faster than this version. However, the time difference is trivial: Not enough to worry about: Less than 1 millisecond (tested in IE and FF). Note: You may want to jam everything into one line because it results in less traffic being sent to the client. That is true. However, most Internet servers now compress HTML and JavaScript by stripping out comments and white space (go to Bing or Google and view the source). This feature should be enabled on your server: Let the server compress your code, you don't need to do it. Free Career Advice: Creating maintainable code is Job One—Maximum Priority—The Prime Directive. If you find yourself suddenly transferred to customer support, it may be that the code you are writing is not as readable as it could be and not as readable as it should be. Moving on… I created a CSS class to enhance the results: .section{ background-color: yellow; border: 2px solid black; margin: 5px;} Here is the rendered output before:   …and after the jQuery code runs.   Pretty Cool! But, while playing with this code, the logic of nextUntil began to bother me: What happens in the last section? What stops elements from being collected since there are no more elements with the .heading class? The answer is nothing.  In this case it stopped collecting elements because it was at the end of the page.  But what if there were additional HTML elements? I added an anchor tag and another div to the HTML: <h3 class="heading">Heading 1</h3> <div>Content</div> <div>More content</div> <div>Even more content</div><h3 class="heading">Heading 2</h3> <div>some content</div> <div>some more content</div><h3 class="heading">Heading 3</h3> <div>other content</div><a>this is a link</a><div>unrelated div</div> </form></body> The code as-is will include both the anchor and the unrelated div. This isn't what we want.   My first attempt to correct this used the filter parameter of the nextUntil function: nextUntil('.heading', 'div')  This will only collect div elements. But it merely skipped the anchor tag and it still collected the unrelated div:   The problem is we need a way to tell the nextUntil function when to stop. CSS selectors to the rescue! nextUntil('.heading, a')  This tells nextUntil to stop collecting elements when it gets to an element with a .heading class OR when it gets to an anchor tag. In this case it solved the problem. FYI: The comma operator in a CSS selector allows multiple criteria.   Bingo! One final note, we could have broken the code down even more: We could have replaced the andSelf function here: $this = $this.nextUntil('.heading, a').andSelf(); With this: // get all the following siblings and then add the current item$this = $this.nextUntil('.heading, a');$this.add(this);  But in this case, the andSelf function reads real nice. In my opinion. Here's a link to a jsFiddle if you want to play with it. I hope someone finds this useful Steve Wellens CodeProject

  Read the article

• Developing a Cost Model for Cloud Applications

  - by BuckWoody

  Note - please pay attention to the date of this post. As much as I attempt to make the information below accurate, the nature of distributed computing means that components, units and pricing will change over time. The definitive costs for Microsoft Windows Azure and SQL Azure are located here, and are more accurate than anything you will see in this post: http://www.microsoft.com/windowsazure/offers/  When writing software that is run on a Platform-as-a-Service (PaaS) offering like Windows Azure / SQL Azure, one of the questions you must answer is how much the system will cost. I will not discuss the comparisons between on-premise costs (which are nigh impossible to calculate accurately) versus cloud costs, but instead focus on creating a general model for estimating costs for a given application. You should be aware that there are (at this writing) two billing mechanisms for Windows and SQL Azure: “Pay-as-you-go” or consumption, and “Subscription” or commitment. Conceptually, you can consider the former a pay-as-you-go cell phone plan, where you pay by the unit used (at a slightly higher rate) and the latter as a standard cell phone plan where you commit to a contract and thus pay lower rates. In this post I’ll stick with the pay-as-you-go mechanism for simplicity, which should be the maximum cost you would pay. From there you may be able to get a lower cost if you use the other mechanism. In any case, the model you create should hold. Developing a good cost model is essential. As a developer or architect, you’ll most certainly be asked how much something will cost, and you need to have a reliable way to estimate that. Businesses and Organizations have been used to paying for servers, software licenses, and other infrastructure as an up-front cost, and power, people to the systems and so on as an ongoing (and sometimes not factored) cost. When presented with a new paradigm like distributed computing, they may not understand the true cost/value proposition, and that’s where the architect and developer can guide the conversation to make a choice based on features of the application versus the true costs. The two big buckets of use-types for these applications are customer-based and steady-state. In the customer-based use type, each successful use of the program results in a sale or income for your organization. Perhaps you’ve written an application that provides the spot-price of foo, and your customer pays for the use of that application. In that case, once you’ve estimated your cost for a successful traversal of the application, you can build that into the price you charge the user. It’s a standard restaurant model, where the price of the meal is determined by the cost of making it, plus any profit you can make. In the second use-type, the application will be used by a more-or-less constant number of processes or users and no direct revenue is attached to the system. A typical example is a customer-tracking system used by the employees within your company. In this case, the cost model is often created “in reverse” - meaning that you pilot the application, monitor the use (and costs) and that cost is held steady. This is where the comparison with an on-premise system becomes necessary, even though it is more difficult to estimate those on-premise true costs. For instance, do you know exactly how much cost the air conditioning is because you have a team of system administrators? This may sound trivial, but that, along with the insurance for the building, the wiring, and every other part of the system is in fact a cost to the business. There are three primary methods that I’ve been successful with in estimating the cost. None are perfect, all are demand-driven. The general process is to lay out a matrix of: components units cost per unit and then multiply that times the usage of the system, based on which components you use in the program. That sounds a bit simplistic, but using those metrics in a calculation becomes more detailed. In all of the methods that follow, you need to know your application. The components for a PaaS include computing instances, storage, transactions, bandwidth and in the case of SQL Azure, database size. In most cases, architects start with the first model and progress through the other methods to gain accuracy. Simple Estimation The simplest way to calculate costs is to architect the application (even UML or on-paper, no coding involved) and then estimate which of the components you’ll use, and how much of each will be used. Microsoft provides two tools to do this - one is a simple slider-application located here: http://www.microsoft.com/windowsazure/pricing-calculator/  The other is a tool you download to create an “Return on Investment” (ROI) spreadsheet, which has the advantage of leading you through various questions to estimate what you plan to use, located here: https://roianalyst.alinean.com/msft/AutoLogin.do?d=176318219048082115  You can also just create a spreadsheet yourself with a structure like this: Program Element Azure Component Unit of Measure Cost Per Unit Estimated Use of Component Total Cost Per Component Cumulative Cost               Of course, the consideration with this model is that it is difficult to predict a system that is not running or hasn’t even been developed. Which brings us to the next model type. Measure and Project A more accurate model is to actually write the code for the application, using the Software Development Kit (SDK) which can run entirely disconnected from Azure. The code should be instrumented to estimate the use of the application components, logging to a local file on the development system. A series of unit and integration tests should be run, which will create load on the test system. You can use standard development concepts to track this usage, and even use Windows Performance Monitor counters. The best place to start with this method is to use the Windows Azure Diagnostics subsystem in your code, which you can read more about here: http://blogs.msdn.com/b/sumitm/archive/2009/11/18/introducing-windows-azure-diagnostics.aspx This set of API’s greatly simplifies tracking the application, and in fact you can use this information for more than just a cost model. After you have the tracking logs, you can plug the numbers into ay of the tools above, which should give a representative cost or in some cases a unit cost. The consideration with this model is that the SDK fabric is not a one-to-one comparison with performance on the actual Windows Azure fabric. Those differences are usually smaller, but they do need to be considered. Also, you may not be able to accurately predict the load on the system, which might lead to an architectural change, which changes the model. This leads us to the next, most accurate method for a cost model. Sample and Estimate Using standard statistical and other predictive math, once the application is deployed you will get a bill each month from Microsoft for your Azure usage. The bill is quite detailed, and you can export the data from it to do analysis, and using methods like regression and so on project out into the future what the costs will be. I normally advise that the architect also extrapolate a unit cost from those metrics as well. This is the information that should be reported back to the executives that pay the bills: the past cost, future projected costs, and unit cost “per click” or “per transaction”, as your case warrants. The challenge here is in the model itself - statistical methods are not foolproof, and the larger the sample (in this case I recommend the entire population, not a smaller sample) is key. References and Tools Articles: http://blogs.msdn.com/b/patrick_butler_monterde/archive/2010/02/10/windows-azure-billing-overview.aspx http://technet.microsoft.com/en-us/magazine/gg213848.aspx http://blog.codingoutloud.com/2011/06/05/azure-faq-how-much-will-it-cost-me-to-run-my-application-on-windows-azure/ http://blogs.msdn.com/b/johnalioto/archive/2010/08/25/10054193.aspx http://geekswithblogs.net/iupdateable/archive/2010/02/08/qampa-how-can-i-calculate-the-tco-and-roi-when.aspx   Other Tools: http://cloud-assessment.com/ http://communities.quest.com/community/cloud_tools

  Read the article

• Elegance, thy Name is jQuery

  - by SGWellens

  So, I'm browsing though some questions over on the Stack Overflow website and I found a good jQuery question just a few minutes old. Here is a link to it. It was a tough question; I knew that by answering it, I could learn new stuff and reinforce what I already knew: Reading is good, doing is better. Maybe I could help someone in the process too. I cut and pasted the HTML from the question into my Visual Studio IDE and went back to Stack Overflow to reread the question. Dang, someone had already answered it! And it was a great answer. I never even had a chance to start analyzing the issue. Now I know what a one-legged man feels like in an ass-kicking contest. Nevertheless, since the question and answer were so interesting, I decided to dissect them and learn as much as possible. The HTML consisted of some divs separated by h3 headings.  Note the elements are laid out sequentially with no programmatic grouping: <h3 class="heading">Heading 1</h3> <div>Content</div> <div>More content</div> <div>Even more content</div><h3 class="heading">Heading 2</h3> <div>some content</div> <div>some more content</div><h3 class="heading">Heading 3</h3> <div>other content</div></form></body>  The requirement was to wrap a div around each h3 heading and the subsequent divs grouping them into sections. Why? I don't know, I suppose if you screen-scrapped some HTML from another site, you might want to reformat it before displaying it on your own. Anyways… Here is the marvelously, succinct posted answer: $('.heading').each(function(){ $(this).nextUntil('.heading').andSelf().wrapAll('<div class="section">');}); I was familiar with all the parts except for nextUntil and andSelf. But, I'll analyze the whole answer for completeness. I'll do this by rewriting the posted answer in a different style and adding a boat-load of comments: function Test(){ // $Sections is a jQuery object and it will contain three elements var $Sections = $('.heading'); // use each to iterate over each of the three elements $Sections.each(function () { // $this is a jquery object containing the current element // being iterated var $this = $(this); // nextUntil gets the following sibling elements until it reaches // an element with the CSS class 'heading' // andSelf adds in the source element (this) to the collection $this = $this.nextUntil('.heading').andSelf(); // wrap the elements with a div $this.wrapAll('<div class="section" >'); });}  The code here doesn't look nearly as concise and elegant as the original answer. However, unless you and your staff are jQuery masters, during development it really helps to work through algorithms step by step. You can step through this code in the debugger and examine the jQuery objects to make sure one step is working before proceeding on to the next. It's much easier to debug and troubleshoot when each logical coding step is a separate line. Note: You may think the original code runs much faster than this version. However, the time difference is trivial: Not enough to worry about: Less than 1 millisecond (tested in IE and FF). Note: You may want to jam everything into one line because it results in less traffic being sent to the client. That is true. However, most Internet servers now compress HTML and JavaScript by stripping out comments and white space (go to Bing or Google and view the source). This feature should be enabled on your server: Let the server compress your code, you don't need to do it. Free Career Advice: Creating maintainable code is Job One—Maximum Priority—The Prime Directive. If you find yourself suddenly transferred to customer support, it may be that the code you are writing is not as readable as it could be and not as readable as it should be. Moving on… I created a CSS class to see the results: .section{ background-color: yellow; border: 2px solid black; margin: 5px;} Here is the rendered output before:   …and after the jQuery code runs.   Pretty Cool! But, while playing with this code, the logic of nextUntil began to bother me: What happens in the last section? What stops elements from being collected since there are no more elements with the .heading class? The answer is nothing.  In this case it stopped because it was at the end of the page.  But what if there were additional HTML elements? I added an anchor tag and another div to the HTML: <h3 class="heading">Heading 1</h3> <div>Content</div> <div>More content</div> <div>Even more content</div><h3 class="heading">Heading 2</h3> <div>some content</div> <div>some more content</div><h3 class="heading">Heading 3</h3> <div>other content</div><a>this is a link</a><div>unrelated div</div> </form></body> The code as-is will include both the anchor and the unrelated div. This isn't what we want.   My first attempt to correct this used the filter parameter of the nextUntil function: nextUntil('.heading', 'div')  This will only collect div elements. But it merely skipped the anchor tag and it still collected the unrelated div:   The problem is we need a way to tell the nextUntil function when to stop. CSS selectors to the rescue: nextUntil('.heading, a')  This tells nextUntil to stop collecting sibling elements when it gets to an element with a .heading class OR when it gets to an anchor tag. In this case it solved the problem. FYI: The comma operator in a CSS selector allows multiple criteria.   Bingo! One final note, we could have broken the code down even more: We could have replaced the andSelf function here: $this = $this.nextUntil('.heading, a').andSelf(); With this: // get all the following siblings and then add the current item$this = $this.nextUntil('.heading, a');$this.add(this);  But in this case, the andSelf function reads real nice. In my opinion. Here's a link to a jsFiddle if you want to play with it. I hope someone finds this useful Steve Wellens CodeProject

  Read the article

• Superclass Sensitive Actions

  - by Geertjan

  I've created a small piece of functionality that enables you to create actions for Java classes in the IDE. When the user right-clicks on a Java class, they will see one or more actions depending on the superclass of the selected class. To explain this visually, here I have "BlaTopComponent.java". I right-click on its node in the Projects window and I see "This is a TopComponent": Indeed, when you look at the source code of "BlaTopComponent.java", you'll see that it implements the TopComponent class. Next, in the screenshot below, you see that I have right-click a different class. In this case, there's an action available because the selected class implements the ActionListener class. Then, take a look at this one. Here both TopComponent and ActionListener are superclasses of the current class, hence both the actions are available to be invoked: Finally, here's a class that subclasses neither TopComponent nor ActionListener, hence neither of the actions that I created for doing something that relates to TopComponents or ActionListeners is available, since those actions are irrelevant in this context: How does this work? Well, it's a combination of my blog entries "Generic Node Popup Registration Solution" and "Showing an Action on a TopComponent Node". The cool part is that the definition of the two actions that you see above is remarkably trivial: import java.awt.event.ActionEvent; import java.awt.event.ActionListener; import javax.swing.JOptionPane; import org.openide.loaders.DataObject; import org.openide.util.Utilities; public class TopComponentSensitiveAction implements ActionListener { private final DataObject context; public TopComponentSensitiveAction() { context = Utilities.actionsGlobalContext().lookup(DataObject.class); } @Override public void actionPerformed(ActionEvent ev) { //Do something with the context: JOptionPane.showMessageDialog(null, "TopComponent: " + context.getNodeDelegate().getDisplayName()); } } The above is the action that will be available if you right-click a Java class that extends TopComponent. This, in turn, is the action that will be available if you right-click a Java class that implements ActionListener: import java.awt.event.ActionEvent; import java.awt.event.ActionListener; import javax.swing.JOptionPane; import org.openide.loaders.DataObject; import org.openide.util.Utilities; public class ActionListenerSensitiveAction implements ActionListener { private final DataObject context; public ActionListenerSensitiveAction() { context = Utilities.actionsGlobalContext().lookup(DataObject.class); } @Override public void actionPerformed(ActionEvent ev) { //Do something with the context: JOptionPane.showMessageDialog(null, "ActionListener: " + context.getNodeDelegate().getDisplayName()); } } Indeed, the classes, at this stage are the same. But, depending on what I want to do with TopComponents or ActionListeners, I now have a starting point, which includes access to the DataObject, from where I can get down into the source code, as shown here. This is how the two ActionListeners that you see defined above are registered in the layer, which could ultimately be done via annotations on the ActionListeners, of course: <folder name="Actions"> <folder name="Tools"> <file name="org-netbeans-sbas-impl-TopComponentSensitiveAction.instance"> <attr stringvalue="This is a TopComponent" name="displayName"/> <attr name="instanceCreate" methodvalue="org.netbeans.sbas.SuperclassSensitiveAction.create"/> <attr name="type" stringvalue="org.openide.windows.TopComponent"/> <attr name="delegate" newvalue="org.netbeans.sbas.impl.TopComponentSensitiveAction"/> </file> <file name="org-netbeans-sbas-impl-ActionListenerSensitiveAction.instance"> <attr stringvalue="This is an ActionListener" name="displayName"/> <attr name="instanceCreate" methodvalue="org.netbeans.sbas.SuperclassSensitiveAction.create"/> <attr name="type" stringvalue="java.awt.event.ActionListener"/> <attr name="delegate" newvalue="org.netbeans.sbas.impl.ActionListenerSensitiveAction"/> </file> </folder> </folder> <folder name="Loaders"> <folder name="text"> <folder name="x-java"> <folder name="Actions"> <file name="org-netbeans-sbas-impl-TopComponentSensitiveAction.shadow"> <attr name="originalFile" stringvalue="Actions/Tools/org-netbeans-sbas-impl-TopComponentSensitiveAction.instance"/> <attr intvalue="150" name="position"/> </file> <file name="org-netbeans-sbas-impl-ActionListenerSensitiveAction.shadow"> <attr name="originalFile" stringvalue="Actions/Tools/org-netbeans-sbas-impl-ActionListenerSensitiveAction.instance"/> <attr intvalue="160" name="position"/> </file> </folder> </folder> </folder> </folder> The most important parts of the layer registration are the lines that are highlighted above. Those lines connect the layer to the generic action that delegates back to the action listeners defined above, as follows: public final class SuperclassSensitiveAction extends AbstractAction implements ContextAwareAction { private final Map map; //This method is called from the layer, via "instanceCreate", //magically receiving a map, which contains all the attributes //that are defined in the layer for the file: static SuperclassSensitiveAction create(Map map) { return new SuperclassSensitiveAction(Utilities.actionsGlobalContext(), map); } public SuperclassSensitiveAction(Lookup context, Map m) { super(m.get("displayName").toString()); this.map = m; String superclass = m.get("type").toString(); //Enable the menu item only if //we're dealing with a class of type superclass: JavaSource javaSource = JavaSource.forFileObject( context.lookup(DataObject.class).getPrimaryFile()); try { javaSource.runUserActionTask(new ScanTask(this, superclass), true); } catch (IOException ex) { Exceptions.printStackTrace(ex); } //Hide the menu item if it isn't enabled: putValue(DynamicMenuContent.HIDE_WHEN_DISABLED, true); } @Override public void actionPerformed(ActionEvent ev) { ActionListener delegatedAction = (ActionListener)map.get("delegate"); delegatedAction.actionPerformed(ev); } @Override public Action createContextAwareInstance(Lookup actionContext) { return new SuperclassSensitiveAction(actionContext, map); } private class ScanTask implements Task<CompilationController> { private SuperclassSensitiveAction action = null; private String superclass; private ScanTask(SuperclassSensitiveAction action, String superclass) { this.action = action; this.superclass = superclass; } @Override public void run(final CompilationController info) throws Exception { info.toPhase(Phase.ELEMENTS_RESOLVED); new EnableIfGivenSuperclassMatches(info, action, superclass).scan( info.getCompilationUnit(), null); } } private static class EnableIfGivenSuperclassMatches extends TreePathScanner<Void, Void> { private CompilationInfo info; private final AbstractAction action; private final String superclassName; public EnableIfGivenSuperclassMatches(CompilationInfo info, AbstractAction action, String superclassName) { this.info = info; this.action = action; this.superclassName = superclassName; } @Override public Void visitClass(ClassTree t, Void v) { Element el = info.getTrees().getElement(getCurrentPath()); if (el != null) { TypeElement te = (TypeElement) el; List<? extends TypeMirror> interfaces = te.getInterfaces(); if (te.getSuperclass().toString().equals(superclassName)) { action.setEnabled(true); } else { action.setEnabled(false); } for (TypeMirror typeMirror : interfaces) { if (typeMirror.toString().equals(superclassName)){ action.setEnabled(true); } } } return null; } } } This is a pretty cool solution and, as you can see, very generic. Create a new ActionListener, register it in the layer so that it maps to the generic class above, and make sure to set the type attribute, which defines the superclass to which the action should be sensitive.

  Read the article

• NUMA-aware placement of communication variables

  - by Dave

  For classic NUMA-aware programming I'm typically most concerned about simple cold, capacity and compulsory misses and whether we can satisfy the miss by locally connected memory or whether we have to pull the line from its home node over the coherent interconnect -- we'd like to minimize channel contention and conserve interconnect bandwidth. That is, for this style of programming we're quite aware of where memory is homed relative to the threads that will be accessing it. Ideally, a page is collocated on the node with the thread that's expected to most frequently access the page, as simple misses on the page can be satisfied without resorting to transferring the line over the interconnect. The default "first touch" NUMA page placement policy tends to work reasonable well in this regard. When a virtual page is first accessed, the operating system will attempt to provision and map that virtual page to a physical page allocated from the node where the accessing thread is running. It's worth noting that the node-level memory interleaving granularity is usually a multiple of the page size, so we can say that a given page P resides on some node N. That is, the memory underlying a page resides on just one node. But when thinking about accesses to heavily-written communication variables we normally consider what caches the lines underlying such variables might be resident in, and in what states. We want to minimize coherence misses and cache probe activity and interconnect traffic in general. I don't usually give much thought to the location of the home NUMA node underlying such highly shared variables. On a SPARC T5440, for instance, which consists of 4 T2+ processors connected by a central coherence hub, the home node and placement of heavily accessed communication variables has very little impact on performance. The variables are frequently accessed so likely in M-state in some cache, and the location of the home node is of little consequence because a requester can use cache-to-cache transfers to get the line. Or at least that's what I thought. Recently, though, I was exploring a simple shared memory point-to-point communication model where a client writes a request into a request mailbox and then busy-waits on a response variable. It's a simple example of delegation based on message passing. The server polls the request mailbox, and having fetched a new request value, performs some operation and then writes a reply value into the response variable. As noted above, on a T5440 performance is insensitive to the placement of the communication variables -- the request and response mailbox words. But on a Sun/Oracle X4800 I noticed that was not the case and that NUMA placement of the communication variables was actually quite important. For background an X4800 system consists of 8 Intel X7560 Xeons . Each package (socket) has 8 cores with 2 contexts per core, so the system is 8x8x2. Each package is also a NUMA node and has locally attached memory. Every package has 3 point-to-point QPI links for cache coherence, and the system is configured with a twisted ladder "mobius" topology. The cache coherence fabric is glueless -- there's not central arbiter or coherence hub. The maximum distance between any two nodes is just 2 hops over the QPI links. For any given node, 3 other nodes are 1 hop distant and the remaining 4 nodes are 2 hops distant. Using a single request (client) thread and a single response (server) thread, a benchmark harness explored all permutations of NUMA placement for the two threads and the two communication variables, measuring the average round-trip-time and throughput rate between the client and server. In this benchmark the server simply acts as a simple transponder, writing the request value plus 1 back into the reply field, so there's no particular computation phase and we're only measuring communication overheads. In addition to varying the placement of communication variables over pairs of nodes, we also explored variations where both variables were placed on one page (and thus on one node) -- either on the same cache line or different cache lines -- while varying the node where the variables reside along with the placement of the threads. The key observation was that if the client and server threads were on different nodes, then the best placement of variables was to have the request variable (written by the client and read by the server) reside on the same node as the client thread, and to place the response variable (written by the server and read by the client) on the same node as the server. That is, if you have a variable that's to be written by one thread and read by another, it should be homed with the writer thread. For our simple client-server model that means using split request and response communication variables with unidirectional message flow on a given page. This can yield up to twice the throughput of less favorable placement strategies. Our X4800 uses the QPI 1.0 protocol with source-based snooping. Briefly, when node A needs to probe a cache line it fires off snoop requests to all the nodes in the system. Those recipients then forward their response not to the original requester, but to the home node H of the cache line. H waits for and collects the responses, adjudicates and resolves conflicts and ensures memory-model ordering, and then sends a definitive reply back to the original requester A. If some node B needed to transfer the line to A, it will do so by cache-to-cache transfer and let H know about the disposition of the cache line. A needs to wait for the authoritative response from H. So if a thread on node A wants to write a value to be read by a thread on node B, the latency is dependent on the distances between A, B, and H. We observe the best performance when the written-to variable is co-homed with the writer A. That is, we want H and A to be the same node, as the writer doesn't need the home to respond over the QPI link, as the writer and the home reside on the very same node. With architecturally informed placement of communication variables we eliminate at least one QPI hop from the critical path. Newer Intel processors use the QPI 1.1 coherence protocol with home-based snooping. As noted above, under source-snooping a requester broadcasts snoop requests to all nodes. Those nodes send their response to the home node of the location, which provides memory ordering, reconciles conflicts, etc., and then posts a definitive reply to the requester. In home-based snooping the snoop probe goes directly to the home node and are not broadcast. The home node can consult snoop filters -- if present -- and send out requests to retrieve the line if necessary. The 3rd party owner of the line, if any, can respond either to the home or the original requester (or even to both) according to the protocol policies. There are myriad variations that have been implemented, and unfortunately vendor terminology doesn't always agree between vendors or with the academic taxonomy papers. The key is that home-snooping enables the use of a snoop filter to reduce interconnect traffic. And while home-snooping might have a longer critical path (latency) than source-based snooping, it also may require fewer messages and less overall bandwidth. It'll be interesting to reprise these experiments on a platform with home-based snooping. While collecting data I also noticed that there are placement concerns even in the seemingly trivial case when both threads and both variables reside on a single node. Internally, the cores on each X7560 package are connected by an internal ring. (Actually there are multiple contra-rotating rings). And the last-level on-chip cache (LLC) is partitioned in banks or slices, which with each slice being associated with a core on the ring topology. A hardware hash function associates each physical address with a specific home bank. Thus we face distance and topology concerns even for intra-package communications, although the latencies are not nearly the magnitude we see inter-package. I've not seen such communication distance artifacts on the T2+, where the cache banks are connected to the cores via a high-speed crossbar instead of a ring -- communication latencies seem more regular.

  Read the article

• Fun with Aggregates

  - by Paul White

  There are interesting things to be learned from even the simplest queries.  For example, imagine you are given the task of writing a query to list AdventureWorks product names where the product has at least one entry in the transaction history table, but fewer than ten. One possible query to meet that specification is: SELECT p.Name FROM Production.Product AS p JOIN Production.TransactionHistory AS th ON p.ProductID = th.ProductID GROUP BY p.ProductID, p.Name HAVING COUNT_BIG(*) < 10; That query correctly returns 23 rows (execution plan and data sample shown below): The execution plan looks a bit different from the written form of the query: the base tables are accessed in reverse order, and the aggregation is performed before the join.  The general idea is to read all rows from the history table, compute the count of rows grouped by ProductID, merge join the results to the Product table on ProductID, and finally filter to only return rows where the count is less than ten. This ‘fully-optimized’ plan has an estimated cost of around 0.33 units.  The reason for the quote marks there is that this plan is not quite as optimal as it could be – surely it would make sense to push the Filter down past the join too?  To answer that, let’s look at some other ways to formulate this query.  This being SQL, there are any number of ways to write logically-equivalent query specifications, so we’ll just look at a couple of interesting ones.  The first query is an attempt to reverse-engineer T-SQL from the optimized query plan shown above.  It joins the result of pre-aggregating the history table to the Product table before filtering: SELECT p.Name FROM ( SELECT th.ProductID, cnt = COUNT_BIG(*) FROM Production.TransactionHistory AS th GROUP BY th.ProductID ) AS q1 JOIN Production.Product AS p ON p.ProductID = q1.ProductID WHERE q1.cnt < 10; Perhaps a little surprisingly, we get a slightly different execution plan: The results are the same (23 rows) but this time the Filter is pushed below the join!  The optimizer chooses nested loops for the join, because the cardinality estimate for rows passing the Filter is a bit low (estimate 1 versus 23 actual), though you can force a merge join with a hint and the Filter still appears below the join.  In yet another variation, the < 10 predicate can be ‘manually pushed’ by specifying it in a HAVING clause in the “q1” sub-query instead of in the WHERE clause as written above. The reason this predicate can be pushed past the join in this query form, but not in the original formulation is simply an optimizer limitation – it does make efforts (primarily during the simplification phase) to encourage logically-equivalent query specifications to produce the same execution plan, but the implementation is not completely comprehensive. Moving on to a second example, the following query specification results from phrasing the requirement as “list the products where there exists fewer than ten correlated rows in the history table”: SELECT p.Name FROM Production.Product AS p WHERE EXISTS ( SELECT * FROM Production.TransactionHistory AS th WHERE th.ProductID = p.ProductID HAVING COUNT_BIG(*) < 10 ); Unfortunately, this query produces an incorrect result (86 rows): The problem is that it lists products with no history rows, though the reasons are interesting.  The COUNT_BIG(*) in the EXISTS clause is a scalar aggregate (meaning there is no GROUP BY clause) and scalar aggregates always produce a value, even when the input is an empty set.  In the case of the COUNT aggregate, the result of aggregating the empty set is zero (the other standard aggregates produce a NULL).  To make the point really clear, let’s look at product 709, which happens to be one for which no history rows exist: -- Scalar aggregate SELECT COUNT_BIG(*) FROM Production.TransactionHistory AS th WHERE th.ProductID = 709;   -- Vector aggregate SELECT COUNT_BIG(*) FROM Production.TransactionHistory AS th WHERE th.ProductID = 709 GROUP BY th.ProductID; The estimated execution plans for these two statements are almost identical: You might expect the Stream Aggregate to have a Group By for the second statement, but this is not the case.  The query includes an equality comparison to a constant value (709), so all qualified rows are guaranteed to have the same value for ProductID and the Group By is optimized away. In fact there are some minor differences between the two plans (the first is auto-parameterized and qualifies for trivial plan, whereas the second is not auto-parameterized and requires cost-based optimization), but there is nothing to indicate that one is a scalar aggregate and the other is a vector aggregate.  This is something I would like to see exposed in show plan so I suggested it on Connect.  Anyway, the results of running the two queries show the difference at runtime: The scalar aggregate (no GROUP BY) returns a result of zero, whereas the vector aggregate (with a GROUP BY clause) returns nothing at all.  Returning to our EXISTS query, we could ‘fix’ it by changing the HAVING clause to reject rows where the scalar aggregate returns zero: SELECT p.Name FROM Production.Product AS p WHERE EXISTS ( SELECT * FROM Production.TransactionHistory AS th WHERE th.ProductID = p.ProductID HAVING COUNT_BIG(*) BETWEEN 1 AND 9 ); The query now returns the correct 23 rows: Unfortunately, the execution plan is less efficient now – it has an estimated cost of 0.78 compared to 0.33 for the earlier plans.  Let’s try adding a redundant GROUP BY instead of changing the HAVING clause: SELECT p.Name FROM Production.Product AS p WHERE EXISTS ( SELECT * FROM Production.TransactionHistory AS th WHERE th.ProductID = p.ProductID GROUP BY th.ProductID HAVING COUNT_BIG(*) < 10 ); Not only do we now get correct results (23 rows), this is the execution plan: I like to compare that plan to quantum physics: if you don’t find it shocking, you haven’t understood it properly :)  The simple addition of a redundant GROUP BY has resulted in the EXISTS form of the query being transformed into exactly the same optimal plan we found earlier.  What’s more, in SQL Server 2008 and later, we can replace the odd-looking GROUP BY with an explicit GROUP BY on the empty set: SELECT p.Name FROM Production.Product AS p WHERE EXISTS ( SELECT * FROM Production.TransactionHistory AS th WHERE th.ProductID = p.ProductID GROUP BY () HAVING COUNT_BIG(*) < 10 ); I offer that as an alternative because some people find it more intuitive (and it perhaps has more geek value too).  Whichever way you prefer, it’s rather satisfying to note that the result of the sub-query does not exist for a particular correlated value where a vector aggregate is used (the scalar COUNT aggregate always returns a value, even if zero, so it always ‘EXISTS’ regardless which ProductID is logically being evaluated). The following query forms also produce the optimal plan and correct results, so long as a vector aggregate is used (you can probably find more equivalent query forms): WHERE Clause SELECT p.Name FROM Production.Product AS p WHERE ( SELECT COUNT_BIG(*) FROM Production.TransactionHistory AS th WHERE th.ProductID = p.ProductID GROUP BY () ) < 10; APPLY SELECT p.Name FROM Production.Product AS p CROSS APPLY ( SELECT NULL FROM Production.TransactionHistory AS th WHERE th.ProductID = p.ProductID GROUP BY () HAVING COUNT_BIG(*) < 10 ) AS ca (dummy); FROM Clause SELECT q1.Name FROM ( SELECT p.Name, cnt = ( SELECT COUNT_BIG(*) FROM Production.TransactionHistory AS th WHERE th.ProductID = p.ProductID GROUP BY () ) FROM Production.Product AS p ) AS q1 WHERE q1.cnt < 10; This last example uses SUM(1) instead of COUNT and does not require a vector aggregate…you should be able to work out why :) SELECT q.Name FROM ( SELECT p.Name, cnt = ( SELECT SUM(1) FROM Production.TransactionHistory AS th WHERE th.ProductID = p.ProductID ) FROM Production.Product AS p ) AS q WHERE q.cnt < 10; The semantics of SQL aggregates are rather odd in places.  It definitely pays to get to know the rules, and to be careful to check whether your queries are using scalar or vector aggregates.  As we have seen, query plans do not show in which ‘mode’ an aggregate is running and getting it wrong can cause poor performance, wrong results, or both. © 2012 Paul White Twitter: @SQL_Kiwi email: [email protected]

  Read the article

< Previous Page | 42 43 44 45 46 47 48 49 50 51  | Next Page >