Create ASP.NET virtual directories in IIS Express using windows credentials to another server (website project)

Moving from developing ASP.NET sites against the full IIS to IIS Express led me to migrating the many virtual directories that were defined for the site. There was an added dimension to the problem, in that the the virtual directories map to central development resources server to keep the many GB of images and tens of thousands of pdf files off the developer SSD drives, for space reasons (and sanity).

To my relief it turns out to be totally possible for website projects (as oppose to website application projects), but I didn’t on performing a quick internet search, find any reference to credentials.

The virtual directories can be set up as part of the site definition in the .vs folder within the applicationhost.config of the website solution. By editing that file, under the following node:

<site
  <application path= …
      <virtualDirectory path= …

Luckily rather than in notepad, I opened up this file in Visual Studio, and found intellisense gave me the answer I had been looking for, and it was a good one!

virtualdirectory

You can see there are userName and password attributes available.

So add a virtualDirectory node for each virtual directory, following this pattern, one after another, after the application node.

<virtualDirectory path="/imageitems" physicalPath="\\resServer\Resources\ProdRes"  
userName="devaccess@mydomain.net" password="ourpassword123" />

We use a read only, low rights user for this, so I’m not too bothered about the password getting into the config file.

Convert ASP.NET “Website Project” from IIS to IIS Express

I had this problem to solve and thought I’d document my solution here, as I didn’t find anything appropriate documented when I tried searching the search engines.

Website Project vs Website Application Project

There are two broad ways ASP.NET websites can be setup as applications.

  • Web site projects
  • Web site application projects

The quickest way to tell the difference is that an application project uses a visual studio project file to define the files that are in the project, whereas for a web site project, any file under the folder is classed as part of the project.

For more information look at: Web Application Projects versus Web Site Projects in Visual Studio (msdn)

Web site projects are opened by going to File->Open Website, rather than File->Open Project, obviously there are many ways to open projects.

I labour this point as on the forums it is almost the norm for questions about website projects get people answering as if it were a website application project, understandable as that is the form of project most developers will experience.

Finally I decided to tackle the problem of being required to have local administrator rights to develop with IIS hosted sites. Running with local admin rights in this day and age is becoming more and more risky and something that is getting less and less necessary so long as you have an alternative elevated user to use when the need arises to perform operations that require it (like installing software). 

These days developing sites against IIS Express is the norm, back in the day we outgrew Cassini, a small developer webserver that used to host sites, as we had url rewriting, host headers, custom url rewrite modules, virtual folders mapped with custom credentials etc. It turns out that now in 2017 all this is possible using IIS Express. How do you move the site though?

These are the steps I performed that worked for me:

Remove the IIS website

In IIS manager I removed the website, so IIS is no longer managing that folder.

Edit the solution file

I edited the solution file to remove the website project definition node from it.

Add the website back in

I then opened the solution and used the add existing website option to add the website folder from source control back into the solution.

Define virtual directories

I then edited the applicationhost.config file that resides in the solution .vs folder, defining the virtual directories that were previously defined in IIS. I did this under the <virtualDirectory path= section for the site.

Added https binding

I added a binding to the <bindings> section for https,
<binding protocol="https" bindingInformation="*:44340:localhost" />

Custom rewriting 

The custom URL rewriter and rewrite rules seemed to come over by themselves so no action was needed.

Custom Host name

I followed the instructions on USING A CUSTOM HOSTNAME WITH IIS EXPRESS AND VISUAL STUDIO 2015 (VS2015)

to get the custom hostname so the site responds to custom host headers correctly.

<binding protocol=”http” bindingInformation=”*:80:mywebapp.127.0.0.1.xip.io” />

Success

I was then amazed to see the site come up first time on rebuilding it and launching it.

Installing Dynamics GP Intrastat Module

If you are looking for the Intrastat module, it is part of the Main GP install, I think fromGP2013 when more of the product extras were bundled in - but certainly from GP 2015. When running the GP installer the various features that can be installed are offered in the installer, find Enhanced Intrastat, click on the icon to change if it is installed or not by the installer. This can also be performed retrospectively to add the feature to an existing install.

intrastat installer Dynamics GP

You must switch the module on in company options and use the intrastat setup window to get it going. Look for the user guide on line or on the installation media for full details.

SQL to extract contiguous ranges for maintenance tasks in Dynamics GP

From time to time I find some SQL that really makes me smile and sit back and stare at it with awe. Today is one of those days, let me show you…

Many of the maintenance windows in Dynamics GP ask for ranges of numbers to be entered. They consist of a start number and then an end number, start and end points for which you would like to execute a operation over.

It is a common thing in GP is to use a GP macro mail merge to automate the user interface where a repetitive operations are required.  If a macro mail merge is being used for one of these windows, then it requires the start and end numbers over which we would like to perform the operation.

Let move to something more tangible and the real world example I’m working on. However don’t get tied down by my example, there are many other range windows in GP that this principle would work on too.

Remove Sales History Windows Dynamics GP

I want to surgically remove 901,872 historical sales documents from GP, using a macro. I am using a macro so that we are safe, the GP biz logic gets applied by the UI. I could use SQL but always best to try macro first so to benefit from the safety the UI gives us. This could take some time, supplying the document numbers and one by one deleting them.

Master numbers are a number id that ties together a string of sales documents that are related, it is held against all sales documents. For efficiency we notice that we can remove documents by master number, so all related documents, quote->order->invoice->return will be deleted at the same time. This would mean looping through the remove operation of window much less, as one master number removes many documents.

This brings us to 446,741 unique master numbers that need removing from 901,872 sales documents (yes I’ve checked and only three master numbers in my range link to related documents outside my range, so I’ve taken those out of this list). I could macro the window to set the "from" and "to" to be the same master number, thus removing one set of related documents at a time, then moving to the next number and then the next. It would work, but testing showed I’d be here all week doing it. Besides I can see somehow that I can be more clever than that!

GP is more efficient on these types of operation if you can supply ranges of numbers to work on, so really I want to look at the list of 446,741 numbers and find all the contiguous ranges within the sequence. Then identifying the start and end numbers to put in the UI start and end fields. Although it is worth remembering sometimes there will just be on number in a particular range, when there are gaps on both sides of the master number.

Hitting Google I found the following post, where they use recursive CTEs to find the ranges. Restricting my sequence to only 10,000 records, it took about three mins to run. Letting it run on the full sequence, I ran out of patience (and probably server resources) to let it get to end. I did try to tune it with some physical tables indexed to try and improve performance, but ultimately recursive CTE’s are terrible for large data sets like this.

How to find contiguous ranges with SQL

I went looking again and found another solution in this post.

How to find the boundaries of groups of contiguous sequential numbers?

This is genius insight in its script design. I do use ROW_NUMBER() a lot myself for all sorts of weird problem solving, but had never seen it applied like this before.

I adapted the snippet of the SQL to my problem and came up with this:

DECLARE @MNTable TABLE (MSTRNUMB INT PRIMARY KEY (MSTRNUMB));

INSERT INTO @MNTable
SELECT DISTINCT MSTRNUMB AS MSTRNUMB
FROM [dbo].[SOP30200] --WHERE Blah blah...
ORDER BY MSTRNUMB;

WITH CTE
AS (
SELECT ROW_NUMBER() OVER (
ORDER BY MSTRNUMB
) - MSTRNUMB AS Groups
,MSTRNUMB
FROM @MNTable
)
SELECT MIN(MSTRNUMB) AS [From]
,MAX(MSTRNUMB) AS [To]
FROM CTE
GROUP BY Groups
ORDER BY MIN(MSTRNUMB)
 
So from a query that never finished processing we go to this:
 

query times

Really, it is now down to a few seconds, to get all the results!

sqlresults

We have now only 3,760 row ranges, from 446,741 master numbers, so I now only need to get the macro to iterate the remove window 3,760 times, that is a much more acceptable!

So the mail merge macro now takes each row of this result set and uses the from and to values from the query to populate the boxes for the master number and then sets the report options and finally then processes. Job done.

Back to the SQL for a moment…

The SQL is fascinating for this though, looking at the CTE data we see how it works.

sql result cte

The row number increments but the subtracting of the ROW_NUMBER from the master number sequence causes a banding or grouping number to be created. So long as the sequence number is only one greater than the previous number, then the banding number is maintained the same in the group. This is because the relative difference between the ROW_NUMBER and the master number has not changed. If the master number jumps up because of a gap it will have gone up by more than the row number, so the grouping will number will be changed to a new grouping number.

This grouping number is then subsequently used to gather the max and min values for each section. I love the efficiency and simplicity of this solution. It also shows how many ways there can be to solve a problem.

 

Further improvements

I can see further improvements are possible. If we in-fill the sequence of master numbers before processing them with dummy master numbers, where documents don’t exist in the database, then this will further reduce the number of times we have to loop. In that case it would allow cases where the max of one range to the min of the next contain no documents to be combined into one range.

I’ve ran out of time and leave it to the reader to investigate that optimisation.