I migrated some content to an SSI and when I viewed the page it was different. The file was SHTML.

I had a div that was supposed to hug the top of the page. After, it was moved to SSI, it was pushed down 1 line.

Tools Used:

Fiddler
Notepad ++

The cause:

Shtml SSI streams the bytes of the included file byte for byte. The included file was saved as UTF-8 With BOM.  The BOM bytes were being sent to the client and rendered as a space. A div following a space will move to the next line because a div is a block element.

I saw the extra bytes when I viewed the raw hex response in fiddler.
When I used Notepad++ to save the file as UTF-8 without BOM the problem disappeared.

ASP server side includes don’t suffer from this, only SHTML.

For some reason I couldn’t find any sites on the web about this. The 24 hour hosting chat support were clueless. Save yourself some time and save include files as UTF-8 without BOM.

First, get Log4net.
The easiest way to do this is to open the Tools, Library Package Manager, Package Manager Console then type
Install-Package Log4Net

Next add the log4net config to your App.config.

WATCH OUT!   <configSections> must be the first child element of <configuration>

WATCH OUT!  Target the 4.0 framework. My project was defaulted to 4.0 Client Profile and would not build

<configSections>
        <section name="log4net" type="log4net.Config.Log4NetConfigurationSectionHandler,log4net, Version=1.2.10.0, Culture=neutral, PublicKeyToken=1b44e1d426115821" />
     </configSections>
    <!-- Log4net Logging Setup -->
    <log4net debug="false">
        <appender name="LogFileAppender" type="log4net.Appender.FileAppender,log4net" >
            <param name="File" value="Errors.log" />
            <param name="AppendToFile" value="true" />
            <layout type="log4net.Layout.PatternLayout,log4net">
                <param name="ConversionPattern" value="%n%n%n%d [%t] %-5p %c  - %m%n%n" />
            </layout>
        </appender>
        <root>
            <priority value="ALL" />
            <appender-ref ref="LogFileAppender" />
        </root>
        <category name="DesktopLogger.Form1">
            <priority value="ALL" />
        </category>
    </log4net>

Notice this line

<param name="File" value="Errors.log" />

I left the path off. That will put the log in the same folder where the executable was run from. We could set it to

value=”c:\\temp\\Errors.log”  or what ever you need.

Now that setup is complete we have three steps to using the logger. 1.) load the configuration 2.) create a log category 3.) log something  

-----------------------

1.) Load the configuration by calling

 log4net.Config.XmlConfigurator.Configure();

on the line just before

InitializeComponent();

2.) Add the log category as a property of the form itself, not in a function

    public partial class Form1 : Form
    {
        private readonly ILog log = LogManager.GetLogger("YourAppName.Form1.cs");

3.) Log errors by calling the Error method of your log object

            int x = 10;
            int y = 10;
            try
            {
                int z= x/(y-10);
            }
            catch (Exception ex) {
                log.Error("onload", ex);
            }

Lets look at the log file entry

2011-08-10 11:07:14,584 [9] ERROR YourAppName.Form1.cs  - onload

System.DivideByZeroException: Attempted to divide by zero.
   at ProPayMigration.Form1..ctor() in C:\dev\local\ProPayMigration\ProPayMigration\Form1.cs:line 60

YourAppName.Form1.cs is the category name we set when the log is initialized

onload is the string I passed in when I called the method to create the entry

It turns out that this is caused by not using the correct content-type when the request was made. The request.form isn’t parsed unless the content-type is application/x-www-form-urlencoded.

The way to read the raw data is to read the Request’s InputStream.

Here are the steps.
Create a stream and set it to the Reqest.InputStream
Read it into a byte array
Encode the Byte array to a string.

            string sPostData = String.Empty;
            if (Request.RequestType=="POST")
            {
                System.IO.Stream myStream = Request.InputStream;

                // Read the file into the byte array.
                byte[] input = new byte[Request.InputStream.Length];
                myStream.Read(input, 0, (int)Request.InputStream.Length - 1);

                //encode to a string
                System.Text.UTF8Encoding enc = new System.Text.UTF8Encoding();
                sPostData = enc.GetString(input);
            }
            if (sPostData!=string.Empty) {

               //parse and process away

            }

I want the module also handle redirects for form submits that use both get and post.

Get is easy, you just pull off the querry string and tack it back on when you’ve found the new page. Post isn’t so easy.

In html when you use <form method=”post” things change completely behind the scenes. Because of this a 301 message won’t do.  When a client receives this message it does a get on the new page and the form data is lost.

Server.Transfer will not work for me because I need the module to redirect to a full url. This causes a runtime error  when server.transfer is used even if the full url is the same server.

The best solution I’ve found is to use 307 temporarily moved if the original method was “post”.

According to the HTTP 1.1 RFC a browser must confirm re-posting to the new location.  If the user confirms the new location is used. I tested this on 5 browsers, Firefox, Opera, Safari, IE, and Chrome.

Only Firefox and Opera followed the RFC. IE, Safari and Chrome simply redirect the post as if nothing happened.

Here is the source from the module…

using System;
using System.Linq;
using System.Runtime.Serialization;
using System.Web;
using Sample.Data.Web;

namespace Sample.Web.HttpHandlers
{

public class RedirectModule : IHttpModule
{
#region under the hood settings

public void Init(HttpApplication app)
{
app.BeginRequest += new EventHandler(OnBeginRequest);
}

public void Dispose() { }

#endregion

public void OnBeginRequest(Object sender, EventArgs e)
{
HttpApplication context = sender as HttpApplication;
try
{
string requestMethod = context.Request.RequestType;
string requestUrl = context.Request.Url.ToString().ToLower();
string protocol = context.Request.Url.Scheme;
string host = context.Request.Url.Authority;
string file = context.Request.Url.LocalPath;
string query = context.Request.Url.Query;

string strippedRequest = protocol + "://" + host + file;

using (SampleWebEntities db = new SampleWebEntities())
{
var redirect = (
from u in db.UrlRedirects
where   u.RequestServer.ToLower().Contains(host.ToLower()) &&
u.RequestUrl.ToLower()==file
select u

).FirstOrDefault();

if (redirect != null)
{
context.Response.Clear();

if (requestMethod == "POST")
{
context.Response.Status = "307 Temporary Redirect";
}
else
{
context.Response.Status = "301 Moved Permanently";
}

context.Response.AddHeader("Location", redirect.ResponseUrl + query);
context.Response.AddHeader("Cache-Control", "no-cache");
context.Response.End();

}
}
}
catch (Exception ex)
{
//TODO: Handle Exception
string str = ex.ToString();
}

}//OnBeginRequest

}//class

}//Namespace

After about 10 years my garage door quit working. After a couple of turns of the screw driver I see that the innerds are full of gear dust. My drive gear is stripped!

To fix or replace? That is the question. I quickly discover that a new gear is only $12 on ebay and that includes grease and free shipping. Done.

When it arrives I come to a step in the instructions that simply says “remove roll pin from under the drive gear”.

Hmm. That seems to be a problem. As it turns out you need a Roll Punch to remove a Roll Pin.  A tool which I do not own. Home Depot, no. Lowes, no.  Sears, scoreboard! $20 bucks.

My garage door opener is now in full repair.  I could probably return the set but I’ll probably need it in another 10 years.

Drive Gear Stripped

I want to display the first 25 characters of a text field in a grid.

You can do this a few ways such as, truncate when the datasource selects it, alter it on rowdatabound and probably a few others. But I wanted to try to find a way using the a format string.

<asp:Label ID="amanotesLabel"
runat="server"
Text='<%#String.Format("{0}",Eval("notes")).Substring(0,25)%>'

it throws an error when the string is less than 25 characters which can be corrected like so….
There are 25 spaces after the {0}

<asp:Label ID="amanotesLabel"

runat="server"
Text='<%#String.Format("{0}                         ",Eval("notes")).Substring(0,25)%>'

So there you have it, no events needed.

I had a generic list, I wanted to  make a copy of it. So the aproach I used was to make the new list = the first one.

Enter the wierdness; any mods to the second list showed up in the first one. So I guess it used a pointer or passed it by reference instead of  making a copy like I expected.

Live and learn I guess. The work around was to initialize it blank and call the add range function passing it the first list.

Here is an example of what I’m talking about. You can past this into Linqpad as C# Statements and play around more.

Returns:

]]> http://www.npathweb.com/2009/10/19/copy-a-generic-list-list-by-ref-by-value-wierdness/feed/ 0 http://www.npathweb.com/2009/10/09/search-your-entire-server-for-a-string-across-all-databases-mssql-2000/ http://www.npathweb.com/2009/10/09/search-your-entire-server-for-a-string-across-all-databases-mssql-2000/#comments Fri, 09 Oct 2009 21:39:42 +0000 B.Harding http://www.npathweb.com/?p=94 You can use this query to search your entire server for a string. It will search views, stored procs, triggers, everything.

All the sql text in your database is stored in system tables under each database. There are a few hardships to overcome in order to search.

1.) Text is split into 4000 character chunks. So every 4k chunk has a colID to note its position in the stack of rows .  The names are stored in sysobjects and the text is stored in syscomments. The issue is when your search hits in the second chunk. You need to pull in all chunks when any one finds a hit.

2.) Each database has it’s own system tables. For one search to check every database’s syscomments you need to loop. My example solution didn’t use a cursor.

I pulled a list of database names out of master.dbo.sysdatabases and put it in a temp table. That is what I used for my loop. I  just jet the top row, work it, delete it, and get the next one from the top.

In the loop I execute sql constructed dynamically based on the database name of in the current loop iteration. It searches for all object Ids that have text matching the search term. Then gets all chucks for that object.  To help me narrow down the search I also search on object type. V for view, p for proc  etc…Leave type ” to get all types.

If you paste this in your query window and set the search strings you should get results.  I clipped the text to 3990 because I kept getting errors that the row was too large. I should probably look into it but this will do for now.

I’m dealing with MSSql Server 2000 here and this works.  I plan to make this a stored proc and call it from asp but right now it’s 5:00pm on Friday and this can wait ’till Monday.

Create Table #dblist (NAME VARCHAR(80))

createtable #results (name nvarchar(80),text varchar(4000))

insert into #dblist

exec(‘select name from master..sysdatabases order by name’)

declare @ObjectType as nvarchar(80)
declare @search as nvarchar(80)

set @search = ‘update’ –change this
set @ObjectType=‘p’  –searches only procedures
–set to ” for all types

declare @dbname as nvarchar(80)

set @dbname= (select top 1 name from #dblist)

while @dbname<>”

begin

insert into #results
exec(
‘select id ‘ +
‘into #found ‘ +
‘from ['+@dbname+']..syscomments ‘+
‘where [text] like ”%’+ @search +‘%” ‘ +

‘select ”’+ @dbname+‘:” + so.name as name, left(sc.text, 3990) ‘ +
‘from [' + @dbname + ']..syscomments sc ‘ +
‘inner join [' + @dbname + ']..sysobjects so on sc.id= so.id ‘ +
‘where sc.id in (select id from #found) ‘+
‘and xtype like ”%’ + @ObjectType+ ‘%” ‘+
‘order by name, colid ‘+
‘drop table #found’

)

delete from #dblist where name= @dbname

set @dbname= (select top 1 name from #dblist)

end

select * from #results

drop table #dblist

drop table #results

]]> http://www.npathweb.com/2009/10/09/search-your-entire-server-for-a-string-across-all-databases-mssql-2000/feed/ 1 http://www.npathweb.com/2009/09/29/83/ http://www.npathweb.com/2009/09/29/83/#comments Tue, 29 Sep 2009 13:34:07 +0000 B.Harding http://www.npathweb.com/?p=83 Lets say you have an  ip address in a string such as  ”127.1.2.3″

This code will store the information in an int64 (C#) or bigint (SQL) in with three places per octet.

127001002003

The highest this number could go is 255,255,255,255.00 which is too large for a 32 bit integer. Hence the Int64. In SQL an 64 bit Integer is big int so that is what you’ll need to store in an MS-SQL Database.

I worked this up as a work-around on a project and present it here for your enjoyment.

using System.Text.RegularExpressions;

public static Int64 convertIP(string sAddress) {

Match ipaddress = Regex.Match(sAddress,“^([\\d]*).([\\d]*).([\\d]*).([\\d]*)”);

Int64 iAddress = 0;

Int64 ip1 =Convert.ToInt64(ipaddress.Groups[1].ToString()) * 1000000000;
Int64 ip2 =Convert.ToInt64(ipaddress.Groups[2].ToString()) * 1000000;
Int64 ip3 =Convert.ToInt64(ipaddress.Groups[3].ToString()) * 1000;
Int64 ip4 =Convert.ToInt64(ipaddress.Groups[4].ToString());

iAddress = ip1 + ip2 + ip3 + ip4;

return iAddress;

}

Extra credit: This number is now 127,001,002,003. You can use string formating to display it as 127.001.002.003 by using the format string N0 (Number, zero decimal places) and changing the number group separator from the default comma to a period. All with the raw int64, no function required.

Int64 iAddress = convertIP(Request.UserHostAddress);

Response.Write(iAddress).ToString(“N0″, new NumberFormatInfo { NumberGroupSeparator =“.” }));

Some things you might not know about Scott Guthrie

1. When Scott Guthrie throws exceptions, it’s across the room.
2. All arrays Scott Guthrie declares are of infinite size, because Scott Guthrie knows no bounds.
3. Scott Guthrie doesn’t have disk latency because the hard drive knows to hurry up.
4. Scott Guthrie writes code that optimizes itself.
5. Scott Guthrie can’t test for equality because he has no equal.
6. Scott Guthrie doesn’t need garbage collection because he doesn’t call .Dispose(), he calls .DropKick().
7. Scott Guthrie’s first program was kill -9.
8. Scott Guthrie burst the dot com bubble.
9. All browsers support the hex definitions #scott and #guthrie for the colors black and blue.
10. MySpace actually isn’t your space, it’s Scott’s (he just lets you use it).
11. Scott Guthrie can write infinite recursion functions … and have them return.
12. Scott Guthrie can solve the Towers of Hanoi in one move.
13. The only pattern Scott Guthrie knows is God Object.
14. Scott Guthrie finished World of Warcraft.
15. Project managers never ask Scott Guthrie for estimations … ever.
16. Scott Guthrie doesn’t use web standards as the web will conform to him.
17. “It works on my machine” always holds true for Scott Guthrie.
18. Whiteboards are white because Scott Guthrie scared them that way.
19. Scott Guthrie doesn’t do Burn Down charts, he does Smack Down charts.
20. Scott Guthrie can delete the Recycling Bin.
21. Scott Guthrie’s red polo shirt can type 140 wpm.
22. Scott Guthrie can unit test entire applications with a single assert.
23. Scott Guthrie doesn’t bug hunt as that signifies a probability of failure, he goes bug killing.
24. Scott Guthrie’s keyboard doesn’t have a Ctrl key because nothing controls Scott Guthrie.
25. When Scott Guthrie is web surfing, websites get the message “Warning: Internet Explorer has deemed this user to be malicious or dangerous. Proceed?”.