Snippets and info

Compilier directive attributes.

2006-02-08T14:55:00.000Z

Add the attribute [MethodImplAttribute(MethodImplOptions.NoInlining)] to a c# method to stop the compiler inlining it.

How to create an event for a ASP.Net Page

2005-08-18T16:44:00.000Z

To create an Event for a Asp.Net Page or Control you need the following elements:

1. The event object

private static readonly object EventFoo = new object();

2. The EventHandler (that you will add delegates to)
[Description(""), Category("Action")]
public event EventHandler Foo
{
add
{
base.Events.AddHandler (FooPage.EventFoo, value);
}
remove
{
base.Events.RemoveHandler (FooPage.EventFoo, value);
}
}

3. OnFoo is the method that will be invoked whenever the event is to be fired. protected virtual void OnFoo (EventArgs e)
{
if (this.Events != null)
{
EventHandler handler = (EventHandler) Events[FooPage.EventFoo];
if (handler != null)
{
handler(this, e);
}
}
}

Now you can add handlers to Foo as you would with any other Event (for example Page_Load)

Slashdot | Could Microsoft Buy Red Hat?

2005-05-16T21:20:00.000Z

Spotted on the Slashdot blog Slashdot | Could Microsoft Buy Red Hat?: "Re:GPL? (Score:5, Funny)
by Anonymous Coward on Monday May 16, @02:37PM (#12545816)

If Microsoft bought Red Hat, wouldn't this mean that Microsoft would be under the GPL when it comes to releasing software?

I'm glad you asked that, because I collect stupid questions and I hadn't seen that one before.

Stock, Fonts and more...

2005-04-22T09:34:00.000Z

Just some links...

Showing background colors in Internet Explorer via CSS

2004-12-08T11:29:00.000Z

A common question is how to force background colors to show when printing web sites. Normally the answer is; you can't but by using an IE filter such as the following the desired effect can be reached:

filter:progid:DXImageTransform.Microsoft.Gradient(gradientType=0,startColorStr=#ff6600,endColorStr=#ff6600);

This background will be visible in Print Preview

This background will Not be visible in Print Preview

Some things to remember:

It's IE only (but then that is 9 out of 10 browsers overall and for most non techie commercial sites even higher).
You may need to define a height to the container tag (via the tag or css).
This will not work on all tags, for example div, p, h1 etc..., table all work but tr's will not.

Migrate SQL Server Script

2004-11-22T10:57:00.000Z

Here's a nice script that a colleague found takes the grunt work out of transfering data from one database to another. After running this (and changing the SourceDatabase and TargetDatabase variables) you'll get a set of insert scripts that you can copy n' paste (with a little clean up if there were any schema changes between the two databases) and run on your server (at your own risk ;) ).

Certainly takes the grunt work out of migrating a database.

DECLARE @data_source varchar(8000)
DECLARE @data_target varchar(8000)

SELECT @data_source = 'SourceDataBase'
SELECT @data_target = 'TargetDataBase'

DECLARE @table_name varchar(8000)
DECLARE table_cursor CURSOR FOR
SELECT sysobjects.name FROM sysobjects WHERE sysobjects.xtype = 'U' AND sysobjects.name <> 'dtproperties'

OPEN table_cursor
FETCH NEXT FROM table_cursor INTO @table_name
WHILE @@FETCH_STATUS = 0
BEGIN
DECLARE @column_list varchar(8000)
SELECT @column_list = ''
DECLARE @column_identity bit
SELECT @column_identity = 0

DECLARE @column_name varchar(8000)
DECLARE @column_status tinyint

DECLARE column_cursor CURSOR FOR
SELECT syscolumns.name, syscolumns.status FROM sysobjects INNER JOIN syscolumns ON sysobjects.id = syscolumns.id WHERE sysobjects.xtype = 'U' AND sysobjects.name = @table_name

OPEN column_cursor
FETCH NEXT FROM column_cursor INTO @column_name, @column_status
WHILE @@FETCH_STATUS = 0
BEGIN
SELECT @column_list = (@column_list + '[' + @column_name + '], ')
IF ((@column_status & 128) = 128) SELECT @column_identity = 1
FETCH NEXT FROM column_cursor INTO @column_name, @column_status
END

CLOSE column_cursor
DEALLOCATE column_cursor
IF (@column_identity = 1) PRINT 'SET IDENTITY_INSERT ['+ @table_name + '] ON' PRINT 'INSERT INTO [' + @data_target + ']..[' + @table_name + '] (' + LEFT(@column_list, LEN(@column_list) - 1) + ') SELECT ' + LEFT(@column_list, LEN(@column_list) - 1) + ' FROM [' + @data_source + ']..[' + @table_name + ']'
IF (@column_identity = 1) PRINT 'SET IDENTITY_INSERT ['+ @table_name + '] OFF'
FETCH NEXT FROM table_cursor INTO @table_name
END
CLOSE table_cursor
DEALLOCATE table_cursor

Korby Parnell's WebLog - Visual SourceSafe

2004-10-28T00:38:00.000Z

Interesting site on Visual Sourcesafe - hopefully will be useful to the 'Bugaloo' project. Korby Parnell's WebLog - Visual SourceSafe

.NET XML Web Services

2004-10-27T14:48:00.000Z

An amazing collection of Web services (although perhaps they should have used the spellcheck webservice on their site content ;) ). .NET XML Web Services Repertory

NASA World Wind

2004-10-24T12:26:00.000Z

Every so often you see a really good idea done well - here's one of them:
NASA World Wind

Visual Studio / C# Express Tweaks Part 2: Project Assembly Settings

2004-07-25T00:00:00.000Z

Following on from earlier where I talked about using FXCop with Visual C# Express (or indeed any other version of Visual Studio), I'm now going to quickly run through the top issues that FXCop will report with newly made assemblies (and indeed most mature assemblies). Specifically I'm talking about these:

CriticalError : Assemblies should have the CLSCompliantAttribute.
CriticalError : Assemblies should be marked with the ComVisible attribute.
CriticalError : Assemblies should have Permission Requests.

Where the rain gets in

So what are these issues and why do you get them free with your new projects? Well tackling them in order "CLSCompliant" does what it says on the tin - it enforces the Common Language Specification rules that allow an assembly to produce valid Common Language Runtime code (or IL - Intermediate Language). You might now be thinking: "Isn't C# built on top of the CLR , why does it need to say it's compliant?", well the answer is that C# isn't hundred percent compliant and in fact allows you to do some CLR no-nos. By specifying CLS Compliance Visual Studio (or at least your Compiler) will enforce the slightly stricter rules of the CLS.

Next up; ComVisible, basically this deals with Net objects in the Win32 world and to be honest if you don't know anything about this you should simply declare the attribute as false, as is indeed the Microsoft recommendation (it recommends marking the Assembly invisible and individual 'known good' Methods Visible).

Lastly Permission Requests; this is the implementation of an underused feature of the .Net Framework; the ability to declare the permissions that a program requires in order to run properly. Let me explain; in normal execution a .Net program will be constantly monitored by the framework, if your program attempts to execute a command it does not have sufficient rights for (due to where the program was run from or the user's security settings etc...) the framework will step in a stop execution. So do Permission Requests get you round this? No, but they allow you to declare up front what you want to be able to do and give the user an opportunity to trust your application and grant those rights (which is infinitely better than your program failing mid process). Now this is a very simple way of looking at security and the subject deserves a proper understanding (which I'm not going to do here), especially when you get to the cases of denying rights to third party code and of course "Click Once" (the .Net internet deployment platform).

So why aren't these preset for your new application? Short answer because the alternative is for Microsoft to second guess what you want from your program (think- "It looks like you're writing a letter?"), long answer - well that's what the MSDN is for.

Fixing a Hole.

So how do I sort these issues out? Well firstly ensure that they are issues. Occasionally you may wish to produce non CLS compliant code etc... Assuming this is not the case we can easily knock these off of FXCop's list.

The first thing to realise is that all of these settings are contained within your AssemblyInfo file that will be either in your project root or within the project Properties subfolder depending on your version (C# Express 2005 uses the Properties subfolder). AssemblyInfo contains metadata about the DLL or EXE that will be produced by your code you will see that it is prefilled with values regarding version numbering and such like. Now to the 'fixes':

Marking an Assembly as CLSCompliant:
[assembly: System.CLSCompliant(true)]

Marking an Assembly invisible to Com services:
[assembly:System.Runtime.InteropServices.ComVisible(false)]

Finally Assembly Permissions; to properly set security permissions you must work out every time you use a framework feature that relies upon a such a permission. So how do you work that out? Well fortunately Microsoft supply a tool called PermCalc with studio (there is an application called PermView for earlier versions), that will do the hard work for you. Less good is the fact it is a command line tool, but if you've read my previous article you'll know the goodness that is the External Tools section of the Visual Studio/C# Express Tool's menu:

Fire up Tools>ExternalTools... and fill in the following:

Title
Get Debug Permissions

Command*:
C:\Program Files\Microsoft Visual Studio 8\Common7\IDE\PermCalc.exe

Argument**:
$(TargetDir)$(TargetName)$(TargetExt) -nologo -library

Check the "Use Output Window" option

*Change paths to reflect your own folder structure this is the default location
**If you're are setting permissions for a library rather than an exe add -Library to the command (you may want two commands, one for each mode)

Also note you'll need to do an initial 'build' for this to work (so that the assemblies are compilied).

Security permissions are complicated by the nature of what they do however once we know what permissions are needed they are easy enough to apply. Here are some examples of permissions set in the assemblyInfo.cs:

First add this using directive (with the others at the top of the file)
using System.Security.Permissions

Now we can add some permissions, lets use File IO permissions as an example.

Add request to access C:\
[assembly:FileIOPermissionAttribute(SecurityAction.RequestMinimum, Read="c:\\")]

Some quick notes: RequestMinimum states that for this program to run we need this permission (i.e. we can't run without it), other settings are RequestOptional (which states we can optionally run without the permission - note if you allow code that relies on this permission to be run when the request was denied a SecurityException will be thrown) also there is the RequestRefuse which will stop a assembly from having a permission even if it would normally be granted.

Also note the: Read="c:\\" section this is called a flag. Flags are present on many security permissions and allow for fine grain control of permissions (IOPermissions also feature the Write flag but visit the msdn for details on other flags). If you're wondering why the value is "c:\\" not "c:\" its because '\' is an escape character and thus needs itself to be escaped (confusing eh? again MSDN is your friend).

Now in addition to specific security permissions there are also PermissionSets which allow you to specify permissions on a large scale basis. Three main categories are available; Nothing, Execution, FullTrust. I'm not going to go into these in detail but here is how to request that you application is fully trusted at runtime.

A request for full permissions
[assembly:PermissionSetAttribute(SecurityAction.RequestMinimum, Name="FullTrust")]

Due to the user's security settings this code may or may not run; typically code marked "FullTrust" will run from local machines, but not from Intranets and never from the Internet (unless that user is a very trusting individual).

Remember these settings are on a assembly by assembly basis i.e. for each project in your solution (if you made an empty project and thus don't have a AssemblyInfo.cs simply copy another one to the project, edit the appropriate settings and include it in the project).

In closing, this advice is equally valid for those using VB.net just remember to alter the syntax to reflect language differences. I hope this helps those with difficulties or those who are just curious about the inner workings of .Net assemblies.

Visual Studio / C# Express Tweaks Part 1: Microsoft FXCop

2004-07-24T23:30:00.000Z

Well I've been playing with Visual C# Express for a few days now - and I'm sold, generics, partial classes and the new IDE features are great (especially refactoring). One thing that is unfortunate with the Express release is that FXCop isn't integrated like it is for it's bigger brother (VS2005). However the trick of adding FXCop to the External Tools list (in the tools window) still works.

For those who don't know you can simply add it via Tools>ExternalTools... then filling the details as below

Title:
FXCop Project
Command:
C:\Program Files\Microsoft FxCop 1.30\FxCopCmd.exe
Arguments:
/f:$(TargetPath) /c /s
Finally check the "Use Output Window" option

More interestingly, You can also target FXCop projects from the command line. This approach offers some distinct advantages over targetting specific assemblies; namely, the abilty to target multiple assemblies, track new issues from build to build and being able to refresh your main FXCop project file without leaving the comfort of your IDE.

To do this simply add a new External Tool entry with arguements like: Arguments:
/p:$(SolutionDir)FXCop\$(SolutionFileName)_Analysis.fxcop /u /c /s

These arguments will look for the main FXCop with in a FXCop subfolder within the solution's folder so for the solution "Application" it would look for "Application.sln_Analysis.fxcop" within "\FXCop". I've made a small exe that will build a FXCop project for the current solution but you can simply build it via the main FXCop program then changing the filepath arguments accordingly.

Channel 9 Previews MS User Interface Research

2004-07-24T20:38:00.000Z

Microsoft's VIBE team have been featured on Channel9 (MS's new community channel for it's employee's); there's some very interesting features previewed here which I hope will be appearing on my desktop soon!

VIBE Video

Agile Development answers to criticisms.

2004-07-24T11:42:00.000Z

There is a good article on Agile methods, it touches on a lot of things I haven’t seen before might be of interest.

Answers to a Sceptics view of Agile Programming While it won't answer every objection to Agile methods you come across, it is interesting and provides a good begining for justifying agile methods to management.

Unit Tests For the Databases

2004-07-23T10:08:00.000Z

First WebForms get them now databases! Interesting stuff: Database Unit testing using Enterprise Services Edit: On the same site Unit Tests for abstract classes

Test First Development - the 'Killer Feature'

2004-07-23T10:06:00.000Z

I've been using Unit Testing for some time now and after some protesting I've eventually come round to being something of an advocate of Unit Testing. The thing that finally sold me was the usefulness and productiveness of unit tests when used with a Test-First methodology.

Test-First development itself was born out of the whole Xtreme Programming/Agile programming wave of a few years ago, and basically says that you don't code then test, but write your tests then implement the features needed to make them pass. This to me is a great example of what Agile Programming does best; namely to be practical in approach yet retain empathy with the developer real world situation. So, how does it do that? Well, quite simply developers hate doing tests, sure it's something that most developers will agree is a worthwhile endeavour but even developers who advocate proper testing procedures and QA baulk at actually doing those same tests themselves (can you say NIMBY?); mainly because its tedious, it isn’t any sort of ‘challenge’, I've got a deadline & this won't help me hit it and anyway “shouldn’t we have a tester who does this?”

Now, in test driven development we write the tests up front, before a line of proper code has been written, you might think this is worse but actually writing test frameworks first gives you several real advantages; you to fully understand what you are going to do to implement a feature or a piece of business logic – questions such as: what properties does a class need to expose and what granularity is needed for the methods available to the end user are sometimes only obvious when you actually try and use it from the outside. But the killer feature of testing up front and the one that sells it the test lazy developer is that up front testing adds value to your work; the ability to use those tests to help YOUR work rather than prove that your work is good enough to your 'friendly' QA representative. Test-first leverages the effort invested in tests when you need it most, i.e. when your code is experiencing churn (modifying/adding/refactoring code), and there are few times where code experiences more churn then when it is first being written. The ability to write a method, run the tests and know immediately that your code fulfils it’s requirements (and can handle exceptional circumstances) is a great boost to personal morale & productivity and implicitly leads to good quality code and importantly provides a great way of debugging those issues that do crop up.

It’s nice to see that development tools are finally reflecting the new methodologies that have appeared over the last few years and giving real improvements after a realatively dry spell and (especially in Visual Studio’s case) reliance on improvements in the underlying platform rather than new thinking in the IDE. Thinking of Visual Studio specifically, the 2005 edition’s Team features include a NUnit like testing framework (although as you may imagine it’s more tightly integrated into the IDE than NUnit). This enhanced framework facilitates Test first programming by allowing tests to refer to non-existent member (merely failing the test with a “member not found”), and for that member’s implementation stub to be generated by Studio’s refactor options. It’s just a shame that those features aren’t available in the lower end Studio options (and the new Express editions).

Array and ArrayList Properties

2004-07-20T09:26:00.000Z

Often you want to expose a typed array, but also want the flexibility of an ArrayList underneath, with c# we can easily do this via these few lines: private ArrayList myIntArray = new ArrayList(); public int MyIntArray { get{ return (int[])myIntArray.ToArray(typeof(int));} } But beware, for efficency you shouldn't do this too often (each call will create a new Array - unless you do some caching logic) and also you shouldn't add a 'set' method for the same reason (the hit with a set is generally worse), instead I'd recommend exposing the functionality needed (perhaps Add() and Clear()). As soon as you start needing more of the functionality of the ArrayList you should consider exposing the actual ArrayList itself. Of course this should be a moot point once Generics are part of the released (non beta) Framework, then you can expose what are effectively typed ArrayLists.

Sealed By Default or not?

2004-06-25T13:42:00.000Z

There was a question on Eric Gunnerson's C# Compendium as to whether c# classes should be sealed by default - here's my take on it.

They appear similar in concept but to me its about making your intentions as a class designer clear: what your saying when you declare a virtual? and compare that to the statement your making when you declare a class sealed.

By declaring a virtual you are stating a granular level of behaviour in a very specific way. This is different to what you state when using the same concept at class level.

If you could declare a class 'extensible' you could be implying that the class as a whole is implicitly overridable - which clearly isn't the case in C# (this of course is different to Java). Furthermore you are stating your extensibility level twice (if method A is virtual then the class X must also be extensible).

By having a class explicitly declared as sealed you imply that the class is a final implementation and cannot be extended (which is true). You aren't doubling your extensibility declaration and your intentions are therefore made clear to the user of your class.

The right way to do an javascript Anchor tag

2004-06-04T12:45:00.000Z

I was wondering what the current consensus was so had a look around, the result - confused but after sifting around I got this:

<a href="#" onclick="javascript:function()">

This means you get your onclick for browsers that support javascript and the return false stops the href been called for those that don't. Aditionally you can navigate to a page that offers alternate functionality by setting the href (i.e. a serverside page: href="function.aspx").

Additionally you should at the end of the onclick statement return a boolean which determines whether a default action should be performed or not (i.e should the href be followed). In most cases you won't want the default action to be performed so you would return false (some people recommend returning nothing - which is treated as false but I don't know if this behaviour is guarenteed to be true in all implementations now or in the future).

So the final result is <a href="#" onclick="javascript:function();return false;">

Tools for C#

2004-05-13T08:43:00.000Z

Possibly fed up with the old 'dogfood' arguement here's a nice listing of c# tools from the people who should know... -C# Tools-

SQL transaction example (especially for facades)

2004-05-12T15:15:00.000Z

This is the bare bones of a transaction wrapper for a series of database access methods you'll need a connection string from somewhere, some real methods (instead of the dummy MyMethods) - but this shows the usage of SqlTransaction and the necessary cleanup code (to do a rollback and release resources).

//these are declared outside of the try so they are available in the ctach/finally blocks
SqlConnection sqlConnection = null;
SqlTransaction sqlTransaction = null;
try
{
/*put all your connection opening code (connection string etc in one place - I call mine GetDataConnection)*/
sqlConnection = GetDataConnection() ;
sqlTransaction = sqlConnection.BeginTransaction();

MyMethod(fooInput, sqlConnection, sqlTransaction)
MyMethod2(fooInput, sqlConnection, sqlTransaction)

sqlTransaction.Commit();
}
catch(Exception e)
{
if (sqlTransaction!=null) {sqlTransaction.Rollback()};
throw new Exception("Failure in method x", e) ;
}
finally
{
if(sqlConnection!=null && sqlConnection.State==ConnectionState.Open)
{sqlConnection.Close();}
}

those 'MyMethod' methods will usually use SqlCommand in some manner such as:
GetProductSales(int productId, SqlConnection, SqlCommand)
{
const string sql_GetProductName = ("SELECT ProductSales " +
"FROM Products " +
"WHERE ProductId = {0} " ;

string sqlString = string.Format(sql_GetProductName, productId );
SqlCommand sqlCommand = new SqlCommand(sqlString, sqlConnection, sqlTransaction)

return Convert.ToInt32(sqlCommand.ExecuteScalar());

}
Once you are more comfortable with what's going on you can (and should) use Parameters, DataReaders and if your not object mapping DataSets and DataAdapters.

Managing BLOB Data Fields

2004-05-10T12:20:00.000Z

Get a bunch of really smart guys together build a complex advanced database system - then what do they do? name a core datatype 'Blob' *sigh* Managing BLOB Data Fields in ADO (warning: this example includes VB code).

Build a Quick Visual Studio.NET Search Macro

2004-05-10T11:32:00.000Z

Want something that can save you at least ten minutes a day? (well for me at least) -Build a Quick Visual Studio.NET Search Macro- This puppy saves the usual copy paste to google that most of us go through - although I'd recommend losing the error Message Box (why thank you for pointing out I'm wrong...), and customising the search string to suit your tastes (i.e. choose your region's google page etc...). I'm now gonna have a cup of tea to celebrate my extra ten minutes - productivity lives!

RegEx Tag Matching

2004-05-07T09:34:00.000Z

RegEx (Regular Expressions) aren't the most fun in the world but they are powerful and surprisingly quick - here's one I made matches multiple white space characters and html/xml etc tags from a string (i use it in replace mode to strip html to unformatted text for things like on the fly word counts).

(\s\s.*?\s)|&nbsp|<(.|\n)*?>|&(.|\n)*?

Lutz Roeder's Programming.NET C# VB CLR WinFX

2004-05-07T08:15:00.000Z

Lutz Roeder's Programming.NET site has something for any .Net designer.

Maintaining an arraylist through viewstate (plus auto databinding)

2004-05-06T15:48:00.000Z

Here's how to add an arraylist, in c#, to a UserControl so you expose a Items property to the parent (which they can iterate through etc...). It uses viewstate to maintain the list and this example also shows how you can sync changes to the arraylist to a DataGrid called dataGrid (whenever a change is made to the arraylist the DataGrid is refreshed and rebound).

///

/// ArrayList of Items.
///

public ArrayList Items
{

get {return (ArrayList) ViewState["Items"];}

set
{
ViewState["Items"] = value;
dataGrid.DataSource = Items;
dataGrid.DataBind();}
}

Note: don't put heavyweight objects into the arraylist or you'll end up with huge client html pages, also this could be adapted to use normal arrays or even better generic data stores.

Other Character Codes

2004-05-06T13:45:00.000Z

I mentioned whitespace character codes earlier but see Web Monkey's Character Reference for a fuller list of other special character codes (XHTML, XSLT, XML and HTML) check this list.

Got no space?

2004-05-06T11:34:00.007Z

Ever needed whitespace in an xslt? While the (non-breaking space) name tag doesn't exist in xslt's there are alternatives: You can use a literal space like this (bit clunky but possibly more correct than the following method). <xsl:text> </xsl:text> Alternatively use the decimal versions of spaces just like your old (you could make a new tag mapping to these values but I wouldn't recommend it): Non-breaking space character is ; Tab character is ; Newline character is ;

JavaScript Library

2004-05-06T08:18:00.000Z

JavaScript Kit some good scripts on here.

DataGrids Anonymous

2004-05-05T14:41:00.000Z

Anybody who has started using knows the pain they can cause but they are a powerful tool if you know what to do with them - this tutorial should certainly help to that end -An Examination of the DataGrid
You could also find this from the MSDN handy with pesky ASP.Net DataGrids Top Questions about the DataGrid Web Server Control

Web Service Method's with Session

2004-04-30T13:20:00.000Z

Trying to use session with an ASP.Net Web Service - won't work? Well you need to add (EnableSession=true) to you Web Method Attribute i.e. [ WebMethod(EnableSession=true)] .

Snippets and info

2004-04-30T13:09:00.000Z

This is really a dumping ground for useful bits I'd otherwise forget but if anything helps you - so much the better. Here's a few links to start with.
A good clear start in C# for anyone (well almost)
Adam's Blog
Joel On Software: Some good advice to be found here.