Search Results

Search found 11680 results on 468 pages for 'convenience methods'.

Page 15/468 | < Previous Page | 11 12 13 14 15 16 17 18 19 20 21 22  | Next Page >

  • Method flags as arguments or as member variables?

    - by Martin
    I think the title "Method flags as arguments or as member variables?" may be suboptimal, but as I'm missing any better terminology atm., here goes: I'm currently trying to get my head around the problem of whether flags for a given class (private) method should be passed as function arguments or via member variable and/or whether there is some pattern or name that covers this aspect and/or whether this hints at some other design problems. By example (language could be C++, Java, C#, doesn't really matter IMHO): class Thingamajig { private ResultType DoInternalStuff(FlagType calcSelect) { ResultType res; for (... some loop condition ...) { ... if (calcSelect == typeA) { ... } else if (calcSelect == typeX) { ... } else if ... } ... return res; } private void InteralStuffInvoker(FlagType calcSelect) { ... DoInternalStuff(calcSelect); ... } public void DoThisStuff() { ... some code ... InternalStuffInvoker(typeA); ... some more code ... } public ResultType DoThatStuff() { ... some code ... ResultType x = DoInternalStuff(typeX); ... some more code ... further process x ... return x; } } What we see above is that the method InternalStuffInvoker takes an argument that is not used inside this function at all but is only forwarded to the other private method DoInternalStuff. (Where DoInternalStuffwill be used privately at other places in this class, e.g. in the DoThatStuff (public) method.) An alternative solution would be to add a member variable that carries this information: class Thingamajig { private ResultType DoInternalStuff() { ResultType res; for (... some loop condition ...) { ... if (m_calcSelect == typeA) { ... } ... } ... return res; } private void InteralStuffInvoker() { ... DoInternalStuff(); ... } public void DoThisStuff() { ... some code ... m_calcSelect = typeA; InternalStuffInvoker(); ... some more code ... } public ResultType DoThatStuff() { ... some code ... m_calcSelect = typeX; ResultType x = DoInternalStuff(); ... some more code ... further process x ... return x; } } Especially for deep call chains where the selector-flag for the inner method is selected outside, using a member variable can make the intermediate functions cleaner, as they don't need to carry a pass-through parameter. On the other hand, this member variable isn't really representing any object state (as it's neither set nor available outside), but is really a hidden additional argument for the "inner" private method. What are the pros and cons of each approach?

    Read the article

  • Simple question about what methodology to pick for my information system [closed]

    - by Neenee Kale
    Possible Duplicate: I need help on methodologies for information system project I will be implementing a student information system for parents for my final year project. I have to choose the best suitable methodology which i could use through out my project. could you please recommend me any methodologies i could use please. Also i would like to ask is Agile system development a methodology?

    Read the article

  • Null Or Empty Coalescing

    - by Latest Microsoft Blogs
    In my last blog post, I wrote about the proper way to check for empty enumerations and proposed an IsNullOrEmpty method for collections which sparked a lot of discussion. This post covers a similar issue, but from a different angle. A very long time ago Read More......(read more)

    Read the article

  • Should all, none, or some overriden methods call Super?

    - by JoJo
    When designing a class, how do you decide when all overridden methods should call super or when none of the overridden methods should call super? Also, is it considered bad practice if your code logic requires a mixture of supered and non-supered methods like the Javascript example below? ChildClass = new Class.create(ParentClass, { /** * @Override */ initialize: function($super) { $super(); this.foo = 99; }, /** * @Override */ methodOne: function($super) { $super(); this.foo++; }, /** * @Override */ methodTwo: function($super) { this.foo--; } }); After delving into the iPhone and Android SDKs, I noticed that super must be called on every overridden method, or else the program will crash because something wouldn't get initialized. When deriving from a template/delegate, none of the methods are supered (obviously). So what exactly are these "je ne sais quoi" qualities that determine whether a all, none, or some overriden methods should call super?

    Read the article

  • Requriing static class setter to be called before Constructor, bad design?

    - by roverred
    I have a class, say Foo, and every instance of Foo will need and contain the same List object, myList. Since every class instance will share the same List Object, I thought it would be good to make myList static and use a static function to set myList before the constructor is called. I was wondering if this was bad, because this requires the setter to be called before the constructor. If the person doesn't, the program will crash. Alternative way would be passing myList every time. Thanks.

    Read the article

  • Group method parameter or individual parameter?

    - by Nassign
    I would like to ask on method parameters design consideration. I am usually deciding between using individual variables as parameters versus grouping them to a class or dictionary as one parameter. Is there such a rule when you should use individual parameter against using a class or a dictionary to group the parameter? Individual parameter - Straight forward, strongly typed Dictionary parameter - Very extensible, like HTTP request but cannot be strongly typed. Class parameter - Extensible by adding member to the class parameter, strongly typed. I am looking for a design reference on when to use which? Note: I am not sure if this question is valid in programmers but I definitely think it would be closed in stackoverflow, If it is still not valid, please point me to the proper page.

    Read the article

  • Should all, none, or some overridden methods call Super?

    - by JoJo
    When designing a class, how do you decide when all overridden methods should call super or when none of the overridden methods should call super? Also, is it considered bad practice if your code logic requires a mixture of supered and non-supered methods like the Javascript example below? ChildClass = new Class.create(ParentClass, { /** * @Override */ initialize: function($super) { $super(); this.foo = 99; }, /** * @Override */ methodOne: function($super) { $super(); this.foo++; }, /** * @Override */ methodTwo: function($super) { this.foo--; } }); After delving into the iPhone and Android SDKs, I noticed that super must be called on every overridden method, or else the program will crash because something wouldn't get initialized. When deriving from a template/delegate, none of the methods are supered (obviously). So what exactly are these "je ne sais quoi" qualities that determine whether a all, none, or some overriden methods should call super?

    Read the article

  • "static" as a semantic clue about statelessness?

    - by leoger
    this might be a little philosophical but I hope someone can help me find a good way to think about this. I've recently undertaken a refactoring of a medium sized project in Java to go back and add unit tests. When I realized what a pain it was to mock singletons and statics, I finally "got" what I've been reading about them all this time. (I'm one of those people that needs to learn from experience. Oh well.) So, now that I'm using Spring to create the objects and wire them around, I'm getting rid of static keywords left and right. (If I could potentially want to mock it, it's not really static in the same sense that Math.abs() is, right?) The thing is, I had gotten into the habit of using static to denote that a method didn't rely on any object state. For example: //Before import com.thirdparty.ThirdPartyLibrary.Thingy; public class ThirdPartyLibraryWrapper { public static Thingy newThingy(InputType input) { new Thingy.Builder().withInput(input).alwaysFrobnicate().build(); } } //called as... ThirdPartyLibraryWrapper.newThingy(input); //After public class ThirdPartyFactory { public Thingy newThingy(InputType input) { new Thingy.Builder().withInput(input).alwaysFrobnicate().build(); } } //called as... thirdPartyFactoryInstance.newThingy(input); So, here's where it gets touchy-feely. I liked the old way because the capital letter told me that, just like Math.sin(x), ThirdPartyLibraryWrapper.newThingy(x) did the same thing the same way every time. There's no object state to change how the object does what I'm asking it to do. Here are some possible answers I'm considering. Nobody else feels this way so there's something wrong with me. Maybe I just haven't really internalized the OO way of doing things! Maybe I'm writing in Java but thinking in FORTRAN or somesuch. (Which would be impressive since I've never written FORTRAN.) Maybe I'm using staticness as a sort of proxy for immutability for the purposes of reasoning about code. That being said, what clues should I have in my code for someone coming along to maintain it to know what's stateful and what's not? Perhaps this should just come for free if I choose good object metaphors? e.g. thingyWrapper doesn't sound like it has state indepdent of the wrapped Thingy which may itself be mutable. Similarly, a thingyFactory sounds like it should be immutable but could have different strategies that are chosen among at creation. I hope I've been clear and thanks in advance for your advice!

    Read the article

  • How to sell logistical procedures that require less time to perform but more finesse?

    - by foampile
    I am working with a group where part of the responsibilities is managing a certain set of configuration files which, of course, have the same skeleton/structure across different environments but different values (like server, user, this setting, that setting etc.). Pretty classic scenario... The problem is that everyone just goes and modifies final, environment-specific files and basically repeats the work for every environment. Personally, I am offended to have to peform repeatable, mundane tasks in this day and age when we have technologies to automate it all. So I devised a very simple procedure of abstracting the files into templates, stubbing env-specific values with parameters and then wrote a simple Perl script that, given a template and an environment matrix with env-specific values for each param, produces the final file. So this is nothing special, cutting-edge or revolutionary -- I am pretty sure that 20 years ago efficient places did their CM like that. However, that requires that changes are made at the template level and then distributed across different environments using the script and not making changes in the final environment-specific files. This is where I am encountering resentment as they feel "comfortable" doing it their old, manual, repeated labor way. Personally, I don't have a problem with them working hard rather than smart but the problem is when I have to build on top of someone else's changes, I have to merge their changes into my template from a specific file, which takes time and is grueling. So my question is how to go about selling my method, which makes it so much faster in an environment that is resentful to change and where most things have to be done at the level of the least competent team member?

    Read the article

  • Best Method of function parameter validation

    - by Aglystas
    I've been dabbling with the idea of creating my own CMS for the experience and because it would be fun to run my website off my own code base. One of the decisions I keep coming back to is how best to validate incoming parameters for functions. This is mostly in reference to simple data types since object validation would be quite a bit more complex. At first I debated creating a naming convention that would contain information about what the parameters should be, (int, string, bool, etc) then I also figured I could create options to validate against. But then in every function I still need to run some sort of parameter validation that parses the parameter name to determine what the value can be then validate against it, granted this would be handled by passing the list of parameters to function but that still needs to happen and one of my goals is to remove the parameter validation from the function itself so that you can only have the actual function code that accomplishes the intended task without the additional code for validation. Is there any good way of handling this, or is it so low level that typically parameter validation is just done at the start of the function call anyway, so I should stick with doing that.

    Read the article

  • Ruby - when to use instance variables vs parameters between methods?

    - by Michael Durrant
    I'm writing several methods that call other methods. To pass the information I have a couple of choices: Pass the information as parameters Set instance variables so that other methods can access them When should I choose one option over the other? It seems that the first option is good as it is very specific about what is being passed. the downside seems to be that a lot of values are being passed around. The second method doesn't require passing all the values around but seems to lead to a lot of magic where methods set instance variables 'somewhere' Should I always be very explicit about gets passed to other methods in the class? Are there exceptions so this?

    Read the article

  • How to name a method that both performs a task and returns a boolean as a status?

    - by Limbo Exile
    If there is a method bool DoStuff() { try { // doing stuff... return true; } catch (Exception ex) { return false; } } should it rather be called IsStuffDone()? Both names could be misinterpreted by the user: If the name is DoStuff() why does it return a boolean? If the name is IsStuffDone() it is not clear whether the method performs a task or only checks its result. Is there a convention for this case? Or an alternative approach, as this one is considered flawed? For example in languages that have output parameters, like C#, a boolean status variable could be passed to the method as one and the method's return type would be void.

    Read the article

  • Do we ethically have the right to use the MAC Address for verification purposes?

    - by Matt Ridge
    I am writing a program, or starting at the very beginning of it, and I am thinking of purchase verification systems as a final step. I will be catering to Macs, PCs, and possibly Linux if all is said and done. I will also be programming this for smartphones as well using C++ and Objective-C. (I am writing a blueprint before going head first into it) That being said, I am not asking for help on doing it yet, but what I’m looking for is a realistic measurement for what could be expected as a viable and ethical option for purchase verification systems. Apple through the Apple Store, and some other stores out there have their own "You bought it" check. I am looking to use a three prong verification system. Email/password 16 to 32 character serial number using alpha/numeric and symbols with Upper and lowercase variants. MAC Address. The first two are in my mind ok, but I have to ask on an ethical standpoint, is a MAC Address to lock the software to said hardware unethical, or is it smart? I understand if an Ethernet card changes if not part of the logic board, or if the logic board changes so does the MAC address, so if that changes it will have to be re-verified, but I have to ask with how everything is today... Is it ethical to actually use the MAC address as a validation key or no? Should I be forward with this kind of verification system or should I keep it hidden as a secret? Yes I know hackers and others will find ways of knowing what I am doing, but in reality this is why I am asking. I know no verification is foolproof, but making it so that its harder to break is something I've always been interested in, and learning how to program is bringing up these questions, because I don't want to assume one thing and find out it's not really accepted in the programming world as a "you shouldn't do that" maneuver... Thanks in advance... I know this is my first programming question, but I am just learning how to program, and I am just making sure I'm not breaking some ethical programmer credo I shouldn't...

    Read the article

  • Retrieving recent tweets using LINQ

    - by brian_ritchie
    There are a few different APIs for accessing Twitter from .NET.  In this example, I'll use linq2twitter.  Other APIs can be found on Twitter's development site. First off, we'll use the LINQ provider to pull in the recent tweets. .csharpcode, .csharpcode pre { font-size: small; color: black; font-family: Consolas, "Courier New", Courier, Monospace; background-color: #ffffff; /*white-space: pre;*/ } .csharpcode pre { margin: 0em; } .csharpcode .rem { color: #008000; } .csharpcode .kwrd { color: #0000ff; } .csharpcode .str { color: #006080; } .csharpcode .op { color: #0000c0; } .csharpcode .preproc { color: #cc6633; } .csharpcode .asp { background-color: #ffff00; } .csharpcode .html { color: #800000; } .csharpcode .attr { color: #ff0000; } .csharpcode .alt { background-color: #f4f4f4; width: 100%; margin: 0em; } .csharpcode .lnum { color: #606060; } 1: public static Status[] GetLatestTweets(string screenName, int numTweets) 2: { 3: try 4: { 5: var twitterCtx = new LinqToTwitter.TwitterContext(); 6: var list = from tweet in twitterCtx.Status 7: where tweet.Type == StatusType.User && 8: tweet.ScreenName == screenName 9: orderby tweet.CreatedAt descending 10: select tweet; 11: // using Take() on array because it was failing against the provider 12: var recentTweets = list.ToArray().Take(numTweets).ToArray(); 13: return recentTweets; 14: } 15: catch 16: { 17: return new Status[0]; 18: } 19: } Once they have been retrieved, they would be placed inside an MVC model. Next, the tweets need to be formatted for display. I've defined an extension method to aid with date formatting: .csharpcode, .csharpcode pre { font-size: small; color: black; font-family: Consolas, "Courier New", Courier, Monospace; background-color: #ffffff; /*white-space: pre;*/ } .csharpcode pre { margin: 0em; } .csharpcode .rem { color: #008000; } .csharpcode .kwrd { color: #0000ff; } .csharpcode .str { color: #006080; } .csharpcode .op { color: #0000c0; } .csharpcode .preproc { color: #cc6633; } .csharpcode .asp { background-color: #ffff00; } .csharpcode .html { color: #800000; } .csharpcode .attr { color: #ff0000; } .csharpcode .alt { background-color: #f4f4f4; width: 100%; margin: 0em; } .csharpcode .lnum { color: #606060; } 1: public static class DateTimeExtension 2: { 3: public static string ToAgo(this DateTime date2) 4: { 5: DateTime date1 = DateTime.Now; 6: if (DateTime.Compare(date1, date2) >= 0) 7: { 8: TimeSpan ts = date1.Subtract(date2); 9: if (ts.TotalDays >= 1) 10: return string.Format("{0} days", (int)ts.TotalDays); 11: else if (ts.Hours > 2) 12: return string.Format("{0} hours", ts.Hours); 13: else if (ts.Hours > 0) 14: return string.Format("{0} hours, {1} minutes", 15: ts.Hours, ts.Minutes); 16: else if (ts.Minutes > 5) 17: return string.Format("{0} minutes", ts.Minutes); 18: else if (ts.Minutes > 0) 19: return string.Format("{0} mintutes, {1} seconds", 20: ts.Minutes, ts.Seconds); 21: else 22: return string.Format("{0} seconds", ts.Seconds); 23: } 24: else 25: return "Not valid"; 26: } 27: } Finally, here is the piece of the view used to render the tweets. .csharpcode, .csharpcode pre { font-size: small; color: black; font-family: Consolas, "Courier New", Courier, Monospace; background-color: #ffffff; /*white-space: pre;*/ } .csharpcode pre { margin: 0em; } .csharpcode .rem { color: #008000; } .csharpcode .kwrd { color: #0000ff; } .csharpcode .str { color: #006080; } .csharpcode .op { color: #0000c0; } .csharpcode .preproc { color: #cc6633; } .csharpcode .asp { background-color: #ffff00; } .csharpcode .html { color: #800000; } .csharpcode .attr { color: #ff0000; } .csharpcode .alt { background-color: #f4f4f4; width: 100%; margin: 0em; } .csharpcode .lnum { color: #606060; } 1: <ul class="tweets"> 2: <% 3: foreach (var tweet in Model.Tweets) 4: { 5: %> 6: <li class="tweets"> 7: <span class="tweetTime"><%=tweet.CreatedAt.ToAgo() %> ago</span>: 8: <%=tweet.Text%> 9: </li> 10: <%} %> 11: </ul>  

    Read the article

  • Pass structure as an argument in c# method

    - by MegaMind
    I want to know if it is possible to pass a Structure as a parameter in c# method and if possible, is it a good practice to do so? I have a c# method which is taking six arguments, i really hate that. I could create a carrier class for that and pass it as an argument, but i want to know if structure could do the job. I want to mention here that few arguments to that method are of ref type and few are of value type.

    Read the article

  • Requring static class setter to be called before constructor, bad design?

    - by roverred
    I have a class, say Foo, and every instance of Foo will need and contain the same List object, myList. Since every class instance will share the same List Object, I thought it would be good to make myList static and use a static function to set myList before the constructor is called. I was wondering if this was bad, because this requires the setter to be called before the constructor? If the person doesn't, the program will crash. Alternative way would be passing myList every time.

    Read the article

  • Should I have a separate method for Update(), Insert(), etc., or have a generic Query() that would be able to handle all of these?

    - by Prayos
    I'm currently trying to write a class library for a connection to a database. Looking over it, there are several different types of queries: Select From, Update, Insert, etc. My question is, what is the best practice for writing these queries in a C# application? Should I have a separate method for each of them(i.e. Update(), Insert()), or have a generic Query() that would be able to handle all of these? Thanks for any and all help!

    Read the article

  • Metro: Namespaces and Modules

    - by Stephen.Walther
    The goal of this blog entry is to describe how you can use the Windows JavaScript (WinJS) library to create namespaces. In particular, you learn how to use the WinJS.Namespace.define() and WinJS.Namespace.defineWithParent() methods. You also learn how to hide private methods by using the module pattern. Why Do We Need Namespaces? Before we do anything else, we should start by answering the question: Why do we need namespaces? What function do they serve? Do they just add needless complexity to our Metro applications? After all, plenty of JavaScript libraries do just fine without introducing support for namespaces. For example, jQuery has no support for namespaces and jQuery is the most popular JavaScript library in the universe. If jQuery can do without namespaces, why do we need to worry about namespaces at all? Namespaces perform two functions in a programming language. First, namespaces prevent naming collisions. In other words, namespaces enable you to create more than one object with the same name without conflict. For example, imagine that two companies – company A and company B – both want to make a JavaScript shopping cart control and both companies want to name the control ShoppingCart. By creating a CompanyA namespace and CompanyB namespace, both companies can create a ShoppingCart control: a CompanyA.ShoppingCart and a CompanyB.ShoppingCart control. The second function of a namespace is organization. Namespaces are used to group related functionality even when the functionality is defined in different physical files. For example, I know that all of the methods in the WinJS library related to working with classes can be found in the WinJS.Class namespace. Namespaces make it easier to understand the functionality available in a library. If you are building a simple JavaScript application then you won’t have much reason to care about namespaces. If you need to use multiple libraries written by different people then namespaces become very important. Using WinJS.Namespace.define() In the WinJS library, the most basic method of creating a namespace is to use the WinJS.Namespace.define() method. This method enables you to declare a namespace (of arbitrary depth). The WinJS.Namespace.define() method has the following parameters: · name – A string representing the name of the new namespace. You can add nested namespace by using dot notation · members – An optional collection of objects to add to the new namespace For example, the following code sample declares two new namespaces named CompanyA and CompanyB.Controls. Both namespaces contain a ShoppingCart object which has a checkout() method: // Create CompanyA namespace with ShoppingCart WinJS.Namespace.define("CompanyA"); CompanyA.ShoppingCart = { checkout: function (){ return "Checking out from A"; } }; // Create CompanyB.Controls namespace with ShoppingCart WinJS.Namespace.define( "CompanyB.Controls", { ShoppingCart: { checkout: function(){ return "Checking out from B"; } } } ); // Call CompanyA ShoppingCart checkout method console.log(CompanyA.ShoppingCart.checkout()); // Writes "Checking out from A" // Call CompanyB.Controls checkout method console.log(CompanyB.Controls.ShoppingCart.checkout()); // Writes "Checking out from B" In the code above, the CompanyA namespace is created by calling WinJS.Namespace.define(“CompanyA”). Next, the ShoppingCart is added to this namespace. The namespace is defined and an object is added to the namespace in separate lines of code. A different approach is taken in the case of the CompanyB.Controls namespace. The namespace is created and the ShoppingCart object is added to the namespace with the following single line of code: WinJS.Namespace.define( "CompanyB.Controls", { ShoppingCart: { checkout: function(){ return "Checking out from B"; } } } ); Notice that CompanyB.Controls is a nested namespace. The top level namespace CompanyB contains the namespace Controls. You can declare a nested namespace using dot notation and the WinJS library handles the details of creating one namespace within the other. After the namespaces have been defined, you can use either of the two shopping cart controls. You call CompanyA.ShoppingCart.checkout() or you can call CompanyB.Controls.ShoppingCart.checkout(). Using WinJS.Namespace.defineWithParent() The WinJS.Namespace.defineWithParent() method is similar to the WinJS.Namespace.define() method. Both methods enable you to define a new namespace. The difference is that the defineWithParent() method enables you to add a new namespace to an existing namespace. The WinJS.Namespace.defineWithParent() method has the following parameters: · parentNamespace – An object which represents a parent namespace · name – A string representing the new namespace to add to the parent namespace · members – An optional collection of objects to add to the new namespace The following code sample demonstrates how you can create a root namespace named CompanyA and add a Controls child namespace to the CompanyA parent namespace: WinJS.Namespace.define("CompanyA"); WinJS.Namespace.defineWithParent(CompanyA, "Controls", { ShoppingCart: { checkout: function () { return "Checking out"; } } } ); console.log(CompanyA.Controls.ShoppingCart.checkout()); // Writes "Checking out" One significant advantage of using the defineWithParent() method over the define() method is the defineWithParent() method is strongly-typed. In other words, you use an object to represent the base namespace instead of a string. If you misspell the name of the object (CompnyA) then you get a runtime error. Using the Module Pattern When you are building a JavaScript library, you want to be able to create both public and private methods. Some methods, the public methods, are intended to be used by consumers of your JavaScript library. The public methods act as your library’s public API. Other methods, the private methods, are not intended for public consumption. Instead, these methods are internal methods required to get the library to function. You don’t want people calling these internal methods because you might need to change them in the future. JavaScript does not support access modifiers. You can’t mark an object or method as public or private. Anyone gets to call any method and anyone gets to interact with any object. The only mechanism for encapsulating (hiding) methods and objects in JavaScript is to take advantage of functions. In JavaScript, a function determines variable scope. A JavaScript variable either has global scope – it is available everywhere – or it has function scope – it is available only within a function. If you want to hide an object or method then you need to place it within a function. For example, the following code contains a function named doSomething() which contains a nested function named doSomethingElse(): function doSomething() { console.log("doSomething"); function doSomethingElse() { console.log("doSomethingElse"); } } doSomething(); // Writes "doSomething" doSomethingElse(); // Throws ReferenceError You can call doSomethingElse() only within the doSomething() function. The doSomethingElse() function is encapsulated in the doSomething() function. The WinJS library takes advantage of function encapsulation to hide all of its internal methods. All of the WinJS methods are defined within self-executing anonymous functions. Everything is hidden by default. Public methods are exposed by explicitly adding the public methods to namespaces defined in the global scope. Imagine, for example, that I want a small library of utility methods. I want to create a method for calculating sales tax and a method for calculating the expected ship date of a product. The following library encapsulates the implementation of my library in a self-executing anonymous function: (function (global) { // Public method which calculates tax function calculateTax(price) { return calculateFederalTax(price) + calculateStateTax(price); } // Private method for calculating state tax function calculateStateTax(price) { return price * 0.08; } // Private method for calculating federal tax function calculateFederalTax(price) { return price * 0.02; } // Public method which returns the expected ship date function calculateShipDate(currentDate) { currentDate.setDate(currentDate.getDate() + 4); return currentDate; } // Export public methods WinJS.Namespace.define("CompanyA.Utilities", { calculateTax: calculateTax, calculateShipDate: calculateShipDate } ); })(this); // Show expected ship date var shipDate = CompanyA.Utilities.calculateShipDate(new Date()); console.log(shipDate); // Show price + tax var price = 12.33; var tax = CompanyA.Utilities.calculateTax(price); console.log(price + tax); In the code above, the self-executing anonymous function contains four functions: calculateTax(), calculateStateTax(), calculateFederalTax(), and calculateShipDate(). The following statement is used to expose only the calcuateTax() and the calculateShipDate() functions: // Export public methods WinJS.Namespace.define("CompanyA.Utilities", { calculateTax: calculateTax, calculateShipDate: calculateShipDate } ); Because the calculateTax() and calcuateShipDate() functions are added to the CompanyA.Utilities namespace, you can call these two methods outside of the self-executing function. These are the public methods of your library which form the public API. The calculateStateTax() and calculateFederalTax() methods, on the other hand, are forever hidden within the black hole of the self-executing function. These methods are encapsulated and can never be called outside of scope of the self-executing function. These are the internal methods of your library. Summary The goal of this blog entry was to describe why and how you use namespaces with the WinJS library. You learned how to define namespaces using both the WinJS.Namespace.define() and WinJS.Namespace.defineWithParent() methods. We also discussed how to hide private members and expose public members using the module pattern.

    Read the article

  • Controller Action Methods with different signatures

    - by Narsil
    I am trying to get my URLs in files/id format. I am guessing I should have two Index methods in my controller, one with a parameter and one with not. But I get this error message in browser below. Anyway here is my controller methods: public ActionResult Index() { return Content("Index "); } // // GET: /Files/5 public ActionResult Index(int id) { File file = fileRepository.GetFile(id); if (file == null) return Content("Not Found"); else return Content(file.FileID.ToString()); } Error: Server Error in '/' Application. The current request for action 'Index' on controller type 'FilesController' is ambiguous between the following action methods: System.Web.Mvc.ActionResult Index() on type FileHosting.Controllers.FilesController System.Web.Mvc.ActionResult Index(Int32) on type FileHosting.Controllers.FilesController Description: An unhandled exception occurred during the execution of the current web request. Please review the stack trace for more information about the error and where it originated in the code. Exception Details: System.Reflection.AmbiguousMatchException: The current request for action 'Index' on controller type 'FilesController' is ambiguous between the following action methods: System.Web.Mvc.ActionResult Index() on type FileHosting.Controllers.FilesController System.Web.Mvc.ActionResult Index(Int32) on type FileHosting.Controllers.FilesController Source Error: An unhandled exception was generated during the execution of the current web request. Information regarding the origin and location of the exception can be identified using the exception stack trace below. Stack Trace: [AmbiguousMatchException: The current request for action 'Index' on controller type 'FilesController' is ambiguous between the following action methods: System.Web.Mvc.ActionResult Index() on type FileHosting.Controllers.FilesController System.Web.Mvc.ActionResult Index(Int32) on type FileHosting.Controllers.FilesController] System.Web.Mvc.ActionMethodSelector.FindActionMethod(ControllerContext controllerContext, String actionName) +396292 System.Web.Mvc.ReflectedControllerDescriptor.FindAction(ControllerContext controllerContext, String actionName) +62 System.Web.Mvc.ControllerActionInvoker.FindAction(ControllerContext controllerContext, ControllerDescriptor controllerDescriptor, String actionName) +13 System.Web.Mvc.ControllerActionInvoker.InvokeAction(ControllerContext controllerContext, String actionName) +99 System.Web.Mvc.Controller.ExecuteCore() +105 System.Web.Mvc.ControllerBase.Execute(RequestContext requestContext) +39 System.Web.Mvc.ControllerBase.System.Web.Mvc.IController.Execute(RequestContext requestContext) +7 System.Web.Mvc.<c_DisplayClass8.b_4() +34 System.Web.Mvc.Async.<c_DisplayClass1.b_0() +21 System.Web.Mvc.Async.<c__DisplayClass81.<BeginSynchronous>b__7(IAsyncResult _) +12 System.Web.Mvc.Async.WrappedAsyncResult1.End() +59 System.Web.Mvc.MvcHandler.EndProcessRequest(IAsyncResult asyncResult) +44 System.Web.Mvc.MvcHandler.System.Web.IHttpAsyncHandler.EndProcessRequest(IAsyncResult result) +7 System.Web.CallHandlerExecutionStep.System.Web.HttpApplication.IExecutionStep.Execute() +8677678 System.Web.HttpApplication.ExecuteStep(IExecutionStep step, Boolean& completedSynchronously) +155 Updated Code: public ActionResult Index(int? id) { if (id.HasValue) { File file = fileRepository.GetFile(id.Value); if (file == null) return Content("Not Found"); else return Content(file.FileID.ToString()); } else return Content("Index"); } It's still not the thing I want. URLs have to be in files?id=3 format. I want files/3 routes from global.asax routes.MapRoute( "Default", // Route name "{controller}/{action}/{id}", // URL with parameters new { controller = "Home", action = "Index", id = UrlParameter.Optional } // Parameter defaults ); //files/3 //it's the one I wrote routes.MapRoute("Files", "{controller}/{id}", new { controller = "Files", action = "Index", id = UrlParameter.Optional} ); I tried adding a new route after reading Jeff's post but I can't get it working. It still works with files?id=2 though.

    Read the article

  • Today’s Performance Tip: Views are for Convenience, Not Performance!

    - by Jonathan Kehayias
    I tweeted this last week on twitter and got a lot of retweets so I thought that I’d blog the story behind the tweet. Most vendor databases have views in them, and when people want to retrieve data from a database, it seems like the most common first stop they make are the vendor supplied Views.  This post is in no way a bash against the usage or creation of Views in a SQL Server Database, I have created them before to simplify code and compartmentalize commonly required queries so that there...(read more)

    Read the article

  • Objective-C: How to access methods in other classes

    - by Adam
    I have what I know is a simple question, but after many searches in books and on the Internet, I can't seem to come up with a solution. I have a standard iPhone project that contains, among other things, a ViewController. My app works just fine at this point. I now want to create a generic class (extending NSObject) that will have some basic utility methods. Let's call this class Util.m (along with the associated .h file). I create the Util class (and .h file) in my project, and now I want to access the methods contained in that class class from my ViewController. Here's an example of a simple version of Util.h #import <Foundation/Foundation.h> @interface Util : NSObject { } - (void)myMethod; @end Then the Util.m file would look something like this: #import "Util.h" @implementation Util - (void)myMethod { NSLog(@"myMethod Called"); } @end Now that my Util class is created, I want to call the "myMethod" method from my ViewController. In my ViewController's .h file, I do the following: #import "Util.h" @interface MyViewController : UIViewController { Util *utils; } @property (assign) Util *utils; @end Finally, in the ViewController.m, I do the following: #import "Util.h" @implementation MyViewController @synthesize utils; - (void)viewDidLoad { [super viewDidLoad]; utils.myMethod; //this doesn't work [utils myMethod]; //this doesn't work either NSLog(@"utils = %@", utils); //in the console, this prints "utils = (null)" } What am I doing wrong? I'd like to not only be able to directly reference other classes/methods in a simple util class like this, but I'd also like to directly reference other ViewControllers and their properties and methods as well. I'm stumped! Please Help.

    Read the article

  • Use of 'super' keyword when accessing non-overridden superclass methods

    - by jonny
    I'm trying to get the hang of inheritance in Java and have learnt that when overriding methods (and hiding fields) in sub classes, they can still be accessed from the super class by using the 'super' keyword. What I want to know is, should the 'super' keyword be used for non-overridden methods? Is there any difference (for non-overridden methods / non-hidden fields)? I've put together an example below. public class Vehicle { public int tyreCost; public Vehicle(int tyreCost) { this.tyreCost = tyreCost; } public int getTyreCost() { return tyreCost; } } and public class Car extends Vehicle { public int wheelCount; public Vehicle(int tyreCost, int wheelCount) { super(tyreCost); this.wheelCount = wheelCount; } public int getTotalTyreReplacementCost() { return getTyreCost() * wheelCount; } } Specifically, given that getTyreCost() hasn't been overridden, should getTotalTyreReplacementCost() use getTyreCost(), or super.getTyreCost() ? I'm wondering whether super should be used in all instances where fields or methods of the superclass are accessed (to show in the code that you are accessing the superclass), or only in the overridden/hidden ones (so they stand out).

    Read the article

  • What are the interets of synthetic methods?

    - by romaintaz
    Problem One friend suggested an interesting problem. Given the following code: public class OuterClass { private String message = "Hello World"; private class InnerClass { private String getMessage() { return message; } } } From an external class, how may I print the message variable content? Of course, changing the accessibility of methods or fields is not allowed. (the source here, but it is a french blog) Solution The code to solve this problem is the following: try { Method m = OuterClass.class.getDeclaredMethod("access$000", OuterClass.class); OuterClass outerClass = new OuterClass(); System.out.println(m.invoke(outerClass, outerClass)); } catch (Exception e) { e.printStackTrace(); } Note that the access$000 method name is not really standard (even if this format is the one that is strongly recommanded), and some JVM will name this method access$0. Thus, a better solution is to check for synthetic methods: Method method = null; int i = 0; while ((method == null) && (i < OuterClass.class.getDeclaredMethods().length)) { if (OuterClass.class.getDeclaredMethods()[i].isSynthetic()) { method = OuterClass.class.getDeclaredMethods()[i]; } i++; } if (method != null) { try { System.out.println(method.invoke(null, new OuterClass())); } catch (Exception e) { e.printStackTrace(); } } So the interesting point in this problem is to highlight the use of synthetic methods. With these methods, I can access a private field as it was done in the solution. Of course, I need to use reflection, and I think that the use of this kind of thing can be quite dangerous... Question What is the interest - for me, as a developer - of a synthetic method? What can be a good situation where using the synthetic can be useful?

    Read the article

  • auto-document exceptions on methods in C#/.NET

    - by Sarah Vessels
    I would like some tool, preferably one that plugs into VS 2008/2010, that will go through my methods and add XML comments about the possible exceptions they can throw. I don't want the <summary> or other XML tags to be generated for me because I'll fill those out myself, but it would be nice if even on private/protected methods I could see which exceptions could be thrown. Otherwise I find myself going through the methods and hovering on all the method calls within them to see the list of exceptions, then updating that method's <exception list to include those. Maybe a VS macro could do this? From this: private static string getConfigFilePath() { return Path.Combine(Environment.CurrentDirectory, CONFIG_FILE); } To this: /// <exception cref="System.ArgumentException"/> /// <exception cref="System.ArgumentNullException"/> /// <exception cref="System.IO.IOException"/> /// <exception cref="System.IO.DirectoryNotFoundException"/> /// <exception cref="System.Security.SecurityException"/> private static string getConfigFilePath() { return Path.Combine(Environment.CurrentDirectory, CONFIG_FILE); } Update: it seems like the tool would have to go through the methods recursively, e.g., method1 calls method2 which calls method3 which is documented as throwing NullReferenceException, so both method2 and method1 are documented by the tool as also throwing NullReferenceException. The tool would also need to eliminate duplicates, like if two calls within a method are documented as throwing DirectoryNotFoundException, the method would only list <exception cref="System.IO.DirectoryNotFoundException"/> once.

    Read the article

  • Using separate model methods to manage transactions

    - by DCrystal
    If i’ve got 2(or more) model methods which do (for example, in billing system) enrolling/withdrawing, and one controller’s method that calls 2(or more) of these model methods. Is it a good way(maybe, any suggestions how to do it better) to write/use 2model methods like these: public function start_transaction(){ $this->db->trans_start(); } public function end_transaction(){ $this->db->trans_complete(); } And call in controller’s method: public function smth(){ //something $this->model->start_transaction(); $this->model->enroll(); //something else $this->model->withdraw(); $this->model->end_transaction(); } Will transaction be reversed, if model's withdraw() method fails? Thanks.

    Read the article

< Previous Page | 11 12 13 14 15 16 17 18 19 20 21 22  | Next Page >