About the author

Miron Abramson
Me
Software Engineer,
CTO at PixeliT
and .NET addicted for long time.
Open source projects:
MbCompression - Compression library

Recent comments

Authors

Disclaimer

The opinions expressed herein are my own personal opinions and do not represent my employer's view in anyway.

© Copyright 2014

Creative Commons License

Blog Flux Directory
Technology Blogs - Blog Top Sites

HttpCompression on IIS7 (Using MbCompression library)

If You have full control on the server, you probably don't need my compression library MbCompression, But if you are using share hosting, sure you do.

I just uploaded the configuration for the compression library to work on IIS7.  Check the latest source code from here

Currently rated 4.6 by 7 people

  • Currently 4.571429/5 Stars.
  • 1
  • 2
  • 3
  • 4
  • 5

Posted by Miron on Thursday, July 23, 2009 8:46 AM
Permalink | Comments (5) | Post RSSRSS comment feed

Cropping image using jQuery, Jcrop and ASP.NET

Letting the client cropping his images on the site was never an easy task. With jQuery and the plugin Jcrop by Deep Liquid it is easier than ever to do it in ASP.NET site. Here is how to do it:

First we need to include jQuery library,Jcrop script file & Jquery css file in the page. Then, we need to tell Jcrop wich element is our image to crop, and what method to be fire when we are changing the cropping area:

//
//  Initialize Jcrop
//
$(function() {
$('#theImage').Jcrop({
        onChange: showCoords,
        onSelect: showCoords
    });
});

//
//  Will fire every move of the cropping area
//
function showCoords(c) {
    $('#x1').val(c.x);
    $('#y1').val(c.y);
    $('#x2').val(c.x2);
    $('#y2').val(c.y2);
    $('#w').val(c.w);
    $('#h').val(c.h);
};

x1,y1,x2,y2 are inputs from where the code behind will read the values of the cropping area.

The code behind that will actualy crop the image:

 protected void btnCrop_Click(object sender, EventArgs e)
{
    int X1 = Convert.ToInt32(Request.Form["x1"]);
    int Y1 = Convert.ToInt32(Request["y1"]);
    int X2 = Convert.ToInt32(Request.Form["x2"]);
    int Y2 = Convert.ToInt32(Request.Form["y2"]);
    int X = System.Math.Min(X1, X2);
    int Y = System.Math.Min(Y1, Y2);
    int w = Convert.ToInt32(Request.Form["w"]);
    int h = Convert.ToInt32(Request.Form["h"]);

    // That can be any image (jpg,jpeg,png,gif) from anywhere in the server
    string originalFile = Server.MapPath("~/images/miautito.jpg");


    using (Image img = Image.FromFile(originalFile))
    {
        using (System.Drawing.Bitmap _bitmap = new System.Drawing.Bitmap(w, h))
        {
            _bitmap.SetResolution(img.HorizontalResolution, img.VerticalResolution);
            using (Graphics _graphic = Graphics.FromImage(_bitmap))
            {
                _graphic.InterpolationMode = System.Drawing.Drawing2D.InterpolationMode.HighQualityBicubic;
                _graphic.SmoothingMode = System.Drawing.Drawing2D.SmoothingMode.HighQuality;
                _graphic.PixelOffsetMode = System.Drawing.Drawing2D.PixelOffsetMode.HighQuality;
                _graphic.CompositingQuality = System.Drawing.Drawing2D.CompositingQuality.HighQuality;
                _graphic.DrawImage(img, 0, 0, w, h);
                _graphic.DrawImage(img, new Rectangle(0, 0, w, h), X, Y, w, h, GraphicsUnit.Pixel);

                string extension = Path.GetExtension(originalFile);
                string croppedFileName = Guid.NewGuid().ToString();
                string path = Server.MapPath("~/cropped/");


                // If the image is a gif file, change it into png
                if (extension.EndsWith("gif", StringComparison.OrdinalIgnoreCase))
                {
                    extension = ".png";
                }

                string newFullPathName = string.Concat(path, croppedFileName, extension);

                using (EncoderParameters encoderParameters = new EncoderParameters(1))
                {
                    encoderParameters.Param[0] = new EncoderParameter(Encoder.Quality, 100L);
                    _bitmap.Save(newFullPathName, GetImageCodec(extension), encoderParameters);
                }
            }
        }
    }
}

Method to find the right codec to save the image in the maximum quality:

/// <summary>
/// Find the right codec
/// </summary>
/// <param name="extension"></param>
/// <returns></returns>
public static ImageCodecInfo GetImageCodec(string extension)
{
    extension = extension.ToUpperInvariant();
    ImageCodecInfo[] codecs = ImageCodecInfo.GetImageEncoders();
    foreach (ImageCodecInfo codec in codecs)
    {
        if (codec.FilenameExtension.Contains(extension))
        {
            return codec;
        }
    }
    return codecs[1];
}

You can see some demos of the Jquery Here, and actual demo using the above code in here: ASP.NET Demo.

Complete working source is available in the link below:

Crop.zip (165.24 kb)

Currently rated 4.6 by 17 people

  • Currently 4.647059/5 Stars.
  • 1
  • 2
  • 3
  • 4
  • 5

Categories: ASP.NET | C# | Client side | Server side
Posted by Miron on Monday, April 20, 2009 8:17 AM
Permalink | Comments (5) | Post RSSRSS comment feed

Why I think I will stick with WebForms

The MVC concept is clear and logic and the guys at Microsoft did a very good job with the new MVC Framework , But I think I will stick with the 'old fashioned' WebForms and this is why:

1. I don't need anybody to 'force' me to seperate my code into layers. I do it any way in every site I build. I have a DataAccess layer, a Business logic layer and a 'View' layer, and I know exactly when and where to use each layer.

2. I build web sites. I don't need my code to work with any other 'View' other than HTML, and if I do, as I mentioned before, my Business logic layer is in a seperate dll so I can reuse it.

3. One of the first things I do when I start a new project is to disable the ViewState in the web.config, which leaves me with a hidden field with about 50 chars length (aroung 50bytes) - I can live with that. That will not slow down my site. If the page is not doing any POST, I can remove the <form runat="serve"> from it, and then there will be no ViewState at all.

4. I heavily use JavaScript and Ajax (usually with jQuery & my own js 'framework') to give better user experiance.

5. The only 'Databinding controls' I use are Repeater and ListView (and DropDownList) that are rendering exactly the HTML I want.

6. The only server control (other than  'Databinding controls') I use are HyperLink,Literal and PlaceHolder that not render any extra HTML other then what I want.

7. I heavily use inline expression

8. I use short 2-3 chars names for my controls.

9. I do use NUint (truly, only when I must)

10. I use httpcompression to improve performance.

11. I use Url rewriting if needed.

12. My code is well organized and very easy to maintain.

13. I know all the little secrets of the ASP.NET framework and know how to take it to it's limits

 

I bet my sites have no worse performance than MVC sites or any other technology  and I'm sure it will take me less time to build a well organized site with all the points above.

I built and designed tens of sites. From small, medium to very large, and never had a performance problem that couldn't be solved by using the above points.

Use MVC if you want to be cool. I will stay with my oldy friend - WebForm

Why should I replace a winning horse ?

Currently rated 3.5 by 21 people

  • Currently 3.523809/5 Stars.
  • 1
  • 2
  • 3
  • 4
  • 5

Tags:
Categories: ASP.NET | C# | Performance
Posted by Miron on Friday, February 13, 2009 6:26 AM
Permalink | Comments (21) | Post RSSRSS comment feed

Generate thumbnails on the fly - Yet another version

There are many posts how to generate thumbnails using ASP.NET. Some good some less. Be sure to get distance from the versions that use Image.GetThumbnailImage method.

I'm using my own version that going with me from project to project and it is an HttpHandler that registered in the web.config. You can specified maximum height, maximum width, both, specified quality (1-100) and if the image will be cached or not. It is working well with .jpg, .bmp, .png and .gif. The code is too long to write it here, but it is really easy to undestand and really easy to implement and use it.

To use it, add the class to your dll or just drop it in the App_Code folder,register the handler in the web.config as the following:

<add verb="*" path="image.axd" type="Miron.Web.ImageHandler" validate="false"/>

In your site, to generate a thumbnail with size 150X150 in 90% quality (The image will be cached automatically):

<img src="image.axd?src=~/images/photo.jpg&amp;w=150&amp;h=150&amp;q=90">

If you want the image to not be stored in the cache:

<img src="image.axd?src=~/images/photo.jpg&amp;w=150&amp;h=150&amp;q=90&amp;nocache">

 

You can see it in action in the gallery I did to my new nephew Noam: http://mironabramson.com/noam

Hope it will be useful.

ImageHandler.zip (2.85 kb)

Currently rated 4.2 by 6 people

  • Currently 4.166667/5 Stars.
  • 1
  • 2
  • 3
  • 4
  • 5

Categories: ASP.NET | C#
Posted by Miron on Sunday, February 08, 2009 7:51 AM
Permalink | Comments (6) | Post RSSRSS comment feed

Fast version of the Activator.CreateInstance method using IL

We all know (I guess) the method Activator.CreateInstance(Type) and the generic version:  Activator.CreateInstance<T>()  that are used to create an instance of a specified Type.

In now days all are talking about 'Entity-relationship model' , 'Object-relational mapping' and that stuff, that basically it means getting a IDataReader, loop over it and fill with the data a List with specfied type of objects and return a List instead of DataSet or DataTable (or similar mechanism). While looping over the IDataReader, every iterate, an instance of the specified type needs to be create. It can be done using one of the methods above, but it can be improved by usind IL and some help using some cache. Here is how can it be done:

public static class FastObjectFactory
{
    private static readonly Hashtable creatorCache = Hashtable.Synchronized(new Hashtable());
    private readonly static Type coType = typeof(CreateObject);
    public delegate object CreateObject();

    /// <summary>
    /// Create an object that will used as a 'factory' to the specified type T 
   /// <returns></returns>
    public static CreateObject CreateObjectFactory<T>() where T : class
    {
        Type t = typeof(T);
        FastObjectFactory.CreateObject c = creatorCache[t] as FastObjectFactory.CreateObject;
        if (c == null)
        {
            lock (creatorCache.SyncRoot)
            {
                c = creatorCache[t] as FastObjectFactory.CreateObject;
                if (c != null)
                {
                    return c;
                }
                DynamicMethod dynMethod = new DynamicMethod("DM$OBJ_FACTORY_" + t.Name, typeof(object), null, t);
                ILGenerator ilGen = dynMethod.GetILGenerator();

                ilGen.Emit(OpCodes.Newobj, t.GetConstructor(Type.EmptyTypes));
                ilGen.Emit(OpCodes.Ret);
                c = (CreateObject)dynMethod.CreateDelegate(coType);
                creatorCache.Add(t, c);
            }
        }
        return c;
    }
}

Note the static HashTable is been used as a cache. The first  time we create a delegate 'CreateObjec' for the given type it is 'slow', so all the point here, is to cache this delegate. The next time we need to create an object from the given type, the 'CreateObjec' delegate will be used from the cache. Without caching it, the whole story worth nothing.

Here are some benchmarks:

1 Object:



100 Objects:

1000 Objects:

Comparing to the generic constraint : new()

Below there is the small project I used for the benchmarks:

FastObjectFactory.zip (50.35 kb)

Currently rated 5.0 by 6 people

  • Currently 5/5 Stars.
  • 1
  • 2
  • 3
  • 4
  • 5

Categories: C# | Performance | IL
Posted by Miron on Saturday, August 09, 2008 12:52 PM
Permalink | Comments (19) | Post RSSRSS comment feed

High performance css minifier

It is known that minifying the JavaScript and CSS files can really reduce the files size and improve the general site performance. Lot of sites (BlogEngine.NET among them)  do it on run-time and not it the build time. At this point, I want to recommend the JavaScript minifier JSMIN by Douglas Crockford. It does the job very good, and by far, much faster than all the RegularExpression \ Replace minifiers. Because in my Compression project MbCompression I do the minifying on run-time, I decided to use jsmin minifier.

The Css minifier

All the Css minifiers I found are using Regular Expression\ Replace to remove the unneeded characters. This is working fine, but have a realy bad performance, special on run-time. The speed is slow (special using the Regular Expression), and another important thing is, every Replace creates a very big string in the memory! (Strings are immutable, remember?), so if you have several Replaces, it creates several big strings in the server memory!

To improve performance, I took  JSMIN idea, that are not using any strings in the memory, and perform the minifying much faster, and created a CSS minifier that produce a small CSS file, much faster, and with much less memory overhead.

This minifier was tested on several CSS files. Feel free to download it and use it. If there is any file that not been minify correctly, send it to me, and I will try to improve the minifier.

CssMinifier.cs (9.59 kb)

Currently rated 4.9 by 12 people

  • Currently 4.916667/5 Stars.
  • 1
  • 2
  • 3
  • 4
  • 5

Categories: C# | Client side | Performance
Posted by Miron on Sunday, June 29, 2008 4:48 AM
Permalink | Comments (19) | Post RSSRSS comment feed

Create your own new Type and use it on run-time (C#)

It is not something you will use on daily bases. It doesn't have good performance. But one day, you will have to use it so it's good to know that it is possible. As you can understand from the title, I'm talking about creating a new type with fields and properties (it can also have methods), create instance/s from it and use it. It will not be 'type safe', of course, and you will be able to use it - read and set its values only with reflection. but, all the DataBindingControls (GridView, FormView etc...) are binding the data using reflection, so they will be happy to bind and use your new created objects from your own type you created on run-time.

Lets cut the crap and jump into the code:

Let's say you got an xml from a webservice. and you want to create an object from it. Something like that:

You receive an xml like this:

<root>
    <column name="Name">Miron</column>
    <column name="LastName">Abramson</column>
    <column name="Blog">www.blog.mironabramson.com</column>
</root> 

You want to make a 'match' type that looks like this:

public class MyType
{
    public string Name{ get; set; }
    public string LastName{ get; set; }
    public string Blog{ get; set; }
}

and than create an object from that type and fill it with your data. The prolem is that you don't know what will be the values of the xml attributes and fields. So you need to create the object on run-time:

private object CreateOurNewObject()
{
    string _xml = "<root>" +
        "<column name=\"Name\">Miron</column>" +
        "<column name=\"LastName\">Abramson</column>" +
        "<column name=\"Blog\">www.blog.mironabramson.com</column>" +
        "</root>";

    XmlDocument xmlDoc = new XmlDocument();
    xmlDoc.LoadXml(_xml);

    // create a dynamic assembly and module
    AssemblyName assemblyName = new AssemblyName();
    assemblyName.Name = "tmpAssembly";
    AssemblyBuilder assemblyBuilder = Thread.GetDomain().DefineDynamicAssembly(assemblyName, AssemblyBuilderAccess.Run);
    ModuleBuilder module = assemblyBuilder.DefineDynamicModule("tmpModule");

    // create a new type builder
    TypeBuilder typeBuilder = module.DefineType("BindableRowCellCollection", TypeAttributes.Public | TypeAttributes.Class);

    // Loop over the attributes that will be used as the properties names in out new type
    foreach (XmlNode node in xmlDoc.SelectSingleNode("root").ChildNodes)
    {
        string propertyName = node.Attributes["name"].Value;

        // Generate a private field
        FieldBuilder field = typeBuilder.DefineField("_" + propertyName, typeof(string), FieldAttributes.Private);
        // Generate a public property
        PropertyBuilder property =
            typeBuilder.DefineProperty(propertyName,
                             PropertyAttributes.None,
                             typeof(string),
                             new Type[] { typeof(string) });

        // The property set and property get methods require a special set of attributes:

        MethodAttributes GetSetAttr =
            MethodAttributes.Public |
            MethodAttributes.HideBySig;

        // Define the "get" accessor method for current private field.
        MethodBuilder currGetPropMthdBldr =
            typeBuilder.DefineMethod("get_value",
                                       GetSetAttr,
                                       typeof(string),
                                       Type.EmptyTypes);

        // Intermediate Language stuff...
        ILGenerator currGetIL = currGetPropMthdBldr.GetILGenerator();
        currGetIL.Emit(OpCodes.Ldarg_0);
        currGetIL.Emit(OpCodes.Ldfld, field);
        currGetIL.Emit(OpCodes.Ret);

        // Define the "set" accessor method for current private field.
        MethodBuilder currSetPropMthdBldr =
            typeBuilder.DefineMethod("set_value",
                                       GetSetAttr,
                                       null,
                                       new Type[] { typeof(string) });

        // Again some Intermediate Language stuff...
        ILGenerator currSetIL = currSetPropMthdBldr.GetILGenerator();
        currSetIL.Emit(OpCodes.Ldarg_0);
        currSetIL.Emit(OpCodes.Ldarg_1);
        currSetIL.Emit(OpCodes.Stfld, field);
        currSetIL.Emit(OpCodes.Ret);

        // Last, we must map the two methods created above to our PropertyBuilder to
        // their corresponding behaviors, "get" and "set" respectively.
        property.SetGetMethod(currGetPropMthdBldr);
        property.SetSetMethod(currSetPropMthdBldr);
    }

    // Generate our type
    Type generetedType = typeBuilder.CreateType();

    // Now we have our type. Let's create an instance from it:
    object generetedObject = Activator.CreateInstance(generetedType);

    // Loop over all the generated properties, and assign the values from our XML:
    PropertyInfo[] properties = generetedType.GetProperties();

    int propertiesCounter = 0;

    // Loop over the values that we will assign to the properties
    foreach (XmlNode node in xmlDoc.SelectSingleNode("root").ChildNodes)
    {
        string value = node.InnerText;
        properties[propertiesCounter].SetValue(generetedObject, value, null);
        propertiesCounter++;
    }
  
    //Yoopy ! Return our new genereted object.
    return generetedObject;
}

Mazal Tov!!!

We create our type and instance from it on run-time !!!

In the file bellow, there is full working exmple of the code

 

MyTypeRunTime.zip (4.15 kb)

Currently rated 4.4 by 29 people

  • Currently 4.379311/5 Stars.
  • 1
  • 2
  • 3
  • 4
  • 5

Posted by Miron on Monday, June 09, 2008 10:43 AM
Permalink | Comments (24) | Post RSSRSS comment feed

Sorting a collection using Linq and 'SortExpression' string

Already happened to you that you had a collection of object from type 'X' with some properties, and you had to sort it one time by property 'ID', and another time by property 'Name' ? You wished that you can sort it by just using a 'Sort Expression' ? If still not, I'm sure this moment will arrive sooner or later. Let me save you some time and an headache.

This is how it can be done: 

 public static IEnumerable<T> Sort<T>(this IEnumerable<T> source, string sortExpression)
{
    string[] sortParts = sortExpression.Split(' ');
    var param = Expression.Parameter(typeof(T), string.Empty);
    try
    {
        var property = Expression.Property(param, sortParts[0]);
        var sortLambda = Expression.Lambda<Func<T, object>>(Expression.Convert(property, typeof(object)), param);

        if (sortParts.Length > 1 && sortParts[1].Equals("desc", StringComparison.OrdinalIgnoreCase))
        {
            return source.AsQueryable<T>().OrderByDescending<T, object>(sortLambda);
        }
        return source.AsQueryable<T>().OrderBy<T, object>(sortLambda);
    }
    catch (ArgumentException)
    {
        return source;
    }
}

Just drop it in a static class, and you will be able to sort any collection that implement the interface IEnumerable.

Lets say you have a class 'User':

public class User
{
    public int ID { get; set; }
    public string Name { get; set; }
}

and a List<User> collection: users. You can sort it however you want:

IEnumerable<User> sortedUsersIEnumerable = users.Sort<User>("ID desc"); 

Or

List<User> sortedUsersList = users.Sort<User>("Name").ToList();

I really think this extension should be 'built-in' part of the 'Linq'. 

Extensions.cs (1.08 kb)

Currently rated 4.5 by 10 people

  • Currently 4.500001/5 Stars.
  • 1
  • 2
  • 3
  • 4
  • 5

Categories: ASP.NET | C# | Server side
Posted by Miron on Wednesday, May 07, 2008 1:59 PM
Permalink | Comments (7) | Post RSSRSS comment feed

Uniform Distribution algorithm in C#

Background
Where I work, one of the serviced we give to our clients is Medical articles.With the time, we have tones of articles. Millions. All PDF files, and all are in the same folder.

The Problem
It became impossible to open the folder with Windows explorer, trying to search, copy or move files. I had to find a way to reorder it, and divide it into 1000 folders that every folder will have around the same number of files, and when a request for a specific file will come, I will be able to know in witch folder it is.

The Solution 
I contacted my brother Ari (his site is outdated) for help. He have PhD from the "Electrical Engineering department" at the Technion institute.  He is the smartest guy I have known. So, that is his Algorithm, I just implemented it in C#. It's not too complicate, but it does the job perfect.

The algorithm gets a string and maps it into the set 0-999 (or any given range) with uniform distribution.

/// <summary>
/// Get the position in the range of the specified string
/// </summary>
/// <param name="input"></param>
/// <returns></returns>
public int GetPosition(string input)
{
    if (string.IsNullOrEmpty(input))
    {
        return 0;
    }

    input = hasher.ComputeHash(input).Replace("-",string.Empty);
    double Sum = 0.0;
    int Aj;
    double Tj;
    double Pj;

    for (int j = 0; j < input.Length; j++)
    {
        Aj = (int)input[j];
        Tj = (Math.PI * (1 + j%5) / 2);
        Pj = Math.Pow(Aj,Tj);
        Sum += Math.Round((Math.Ceiling(Pj) - Pj) * Range) * Tj;
    }
    return ((int)(Sum % Range)) + LowerValue;
}

To use it, just create an object from type Mapper, and call the 'GetPosition' method for any needed string:

Mapper aMap = new Mapper(RANGE);
int position = aMap .GetPosition(input)); 

The Results 

I run the algorithm few times with random strings 10 chars long and map them into some ranges of 'folders'. Here are the results:

1,000,000 random strings into range of 1000: the folder with the maximum # of files contained 1156 files, and the minimum  contained 886 files.
1,000,000 random strings into range of 100: the folder with the maximum # of files contained 10,699 files, and the minimum  contained 8800 files.
1,000,000 random strings into range of 10: the folder with the maximum # of files contained 106,873 files, and the minimum  contained 98,747 files.

Very nice results. Maybe not perfect, but certainly good enough for this kind of purpose!

Special thanks to my brother Ari for his help.

Source code:

AriMap.zip (2.34 kb)

Currently rated 5.0 by 1 people

  • Currently 5/5 Stars.
  • 1
  • 2
  • 3
  • 4
  • 5

Categories: C#
Posted by Miron on Wednesday, April 02, 2008 8:58 AM
Permalink | Comments (0) | Post RSSRSS comment feed

Extensions can make us free from testing against null

The concept of the Extensions in the last frameworks is great. They are extremely easy to write and use, give us infinit options to add needed methods we just wished to have in the framework,  and the thing that most catched my eyes is the fact that the extensions are "attached" to an instance of an object, but can be excute with a 'null' object. It is kind of weird.

Excuting the following code will thow an exception (of course):

string s;
s = null;
s = s.Replace("bla", "alb");

But, if we add the following extension:

public static string EReplace(this string s, string oldValue, string newValue,
    bool ignoreCase)
{
    if (s == null)
        return s;

    if (ignoreCase)
    {
         return Regex.Replace(s, oldValue, newValue, RegexOptions.IgnoreCase);
    }
    return s.Replace(oldValue, newValue);

 and try to excute the code:

string s;
s = null;
s = s.EReplace("bla", "alb",true);

the code will run without any exception.

Conclusion:

Using extensions and validate input, will make us free from testing against null any object before excuting a method, and this way, avoid errors that can be throw exceptions on runtime environment.

Currently rated 4.0 by 1 people

  • Currently 4/5 Stars.
  • 1
  • 2
  • 3
  • 4
  • 5

Categories: C#
Posted by Miron on Saturday, March 22, 2008 5:32 AM
Permalink | Comments (1) | Post RSSRSS comment feed