Lars Nielsen's Discoveries

ArgumentException “Value does not fall within the expected range” error using fields on SPListItem

Lars Nielsen — Tue, 24 Jan 2012 19:56:21 +0000

Here’s a pointer that might save a few minutes (or hours!) of frustration if you hit the same problem as me. I was trying to extract the column (field) values from an SPListItem. Easy enough, you just use code like this:


public void ProcessItem(SPListItem item)
{
    string name = item.Name;
    string title = item.Title;
    int id = item.ID;
    // Do something else here
}

OK so far so good. I had this code within my project and ran it. As soon as it got the line where it assigns the name variable it threw an exception:

System.ArgumentException: Value does not fall within the expected range

I put a breakpoint in the function and ran it with the VS debugger attached to the W3WP process to hit the breakpoint. I began to check the item object in Visual Studio, and tried out a few expressions:

item.Name and item.Title both threw the exception
item["Name"] and item["Title"] both threw the exception as well
item.ID and item["ID"] returned the correct values
item.Fields.ContainsField(“Name”), item.Fields.ContainsField(“Title”) and item.Fields.ContainsField(“ID”) returned true

How weird. The ID column was behaving like I expected, but Title and Name were not. All three fields were in the Fields property on the item, but I couldn’t extract their column values without getting the ArgumentException. It took a while and a fair bit of chasing around on forums to figure this out. What was happening is that my SPListItem objects were coming out of the results of a query in SharePoint, using an SPQuery object. It turns out that this is the behaviour you see if you forget to include the columns that you want in the ViewFields property of SPQuery. This property should have a list of FieldRef XML elements so I had to add elements like this:

Check the MSDN documentation for more information about using SPQuery and ViewFields.

SQL Transaction Logs filling up

Lars Nielsen — Sat, 17 Sep 2011 19:31:45 +0000

This issue started with someone telling me that they couldn’t update a document in their team site – they got an error saying that they might be a problem with the library or the server might be busy.

I then tried uploading a file to the document library and got an error:

Unable to complete this operation. Please contact your administrator. at Microsoft.SharePoint.Library.SPRequestInternalClass.PutFile(String bstrUrl, String bstrWebRelativeUrl, Object varFile, PutFileOpt PutFileOpt, String bstrCreatedBy, String bstrModifiedBy, Int32 iCreatedByID, Int32 iModifiedByID, Object varTimeCreated, Object varTimeLastModified, Object varProperties, String bstrCheckinComment, UInt32& pdwVirusCheckStatus, String& pVirusCheckMessage) at Microsoft.SharePoint.Library.SPRequest.PutFile(String bstrUrl, String bstrWebRelativeUrl

I then tried updating a list item on the site and got an error:

Exception from HRESULT: 0x80040E14

This was happening not just on this site but on all the parent sites as well, right up to the root of the site collection. Basically it was impossible to do updates on the site collection. I found this page http://fnoja.blogspot.com/2007/12/exception-from-hresult-0x80040e14.html and found plenty of errors in the SQL Server Database server. There was a low space warning on the drive related to transaction logs, and errors in the event log related to the SQL Transaction logs being full. Indeed the LDF files for the content databases were huge because they weren’t being cleared out properly during the backup. Shrinking down the log files solved the problem. The best way to do this is to is to use DBCC in Server Management Studio – remember to back up the log file first See http://support.microsoft.com/kb/907511. Note that you should not use AutoShrink with SharePoint databases. The moral is: always keep an wary eye on your SharePoint database server…

Remove a lock on a file without rebooting

Lars Nielsen — Fri, 02 Sep 2011 18:51:51 +0000

I often used the superb range of tools from SysInternals created by Mark Russinovich. I first heard about Mark when he exposed the Sony Rootkit copy-protection scandal. It’s perhaps ironic that Sony, a company that once attempted to infect everyone’s PC with a rootkit to prevent them copying CD’s, should now be the subject of almost weekly security breaches of its own.

Anyway, one of the SysInternals tools I use most is Process Explorer. This tool is a great way to see what’s going on with the RAM in your system and which processes are most resource-heavy. One handly use for Process Explorer is when you have a file that’s locked for some reason and you can’t delete or move it. This can happen, for example, with text files that are not closed properly by an application, or a DLL that’s locked by a service process. Using Process Explorer, you can search for a file by filename, and it will find all the processes that have a handle onto that file. If the process is a service, then restarting the service will probably release the lock. If it’s another process, you might need to kill it to release the lock. After that you should be able delete, move, or rename the file in question.

Showing recent updates with the Lightning Conductor web part

Lars Nielsen — Tue, 26 Jul 2011 22:07:53 +0000

I love using the Lightning Conductor web part. It’s so flexible and easy to use that it can often replace a data view web part with something easier to configure and that end users can understand. Does anyone really believe that the Data View Web Part and SharePoint Designer are end user tools? They’re for developers only in my organisation. But the Lightning Conductor Web Part is easy enough to understand so end users who understand SharePoint lists, libraries and columns, as well as filtering and sorting, can figure it out.

I recently wanted to configure a LCWP to show the 10 most recently updated documents in a site, across multiple document libraries. It’s easy enough to configure it to pull data from multiple document libraries. At first I tried setting the Item Limit in the web part properties pane to be 10, and then sorting the results by Modified date order. But this failed to pull back some of the more recent documents.

Then I realised this was because the web part was running a query on the site first, but limited to 10 items, and then sorting those 10 results. I wanted to query all the documents, then sort them and then pick the top 10. To do this, you leave the Item Limit blank but instead set the Page Size property to be 10 in the web part properties, and uncheck the “Show Page buttons” box. It’s probably a good idea also to filter the query as well, for example only select documents updated in the last week. You can do this by setting a condition “greater than or equal” and use the Days Offset setting and set it to Today-7 days. Otherwise the web part would need to deal with a potentially large result set (and the corresponding SQL traffic in querying across multiple lists) and you may get slow performance as the number of documents in the site grows.

Send E-mail to a SharePoint group in SharePoint Designer Workflow

Lars Nielsen — Thu, 07 Jul 2011 21:00:41 +0000

When creating a workflow in SharePoint Designer, of course you can use the send e-mail activity to send e-mail notifications to people during the workflow. The dialog box for the send e-mail activity allows you to choose the recipient of the e-mail. You can either choose individual addresses from the address book, use one of the columns on the workflow item, or choose one of the SharePoint groups associated with the site.

I wanted to use this latter option (mailing a SharePoint group) to notify people when an item in a list had been flagged for deletion. So I set up a special group on the site called “Deleted Item Notification”, for people who would get that e-mail. In my workflow, I set the e-mail activity to send mail to the group Deleted Item Notification. I put myself into the Deleted Item Notification group to test the workflow, flagged an item for deletion and fired off the workflow. Then I waited for my notification e-mail to arrive. I waited and waited and.. nothing happened. I checked the workflow history and it had errored out trying to send the e-mail.

I changed the workflow to e-mail my account directly and it worked fine. So the problem was sending mail to a group. It took a while to find a solution to this – the clue was in this post.

I had to change the group settings for the Deleted Item Notification group to allow everyone to see the membership of the group. I also had to grant that group read access onto the site that was hosting the workflow – I think it would also have worked if I just allowed it to read the list to which my workflow was attached. After those changes, the workflow sent e-mail correctly.

Workflow fails to compile: Resource identifier has already been used in this assembly

Lars Nielsen — Fri, 24 Jun 2011 21:23:43 +0000

Creating a workflow using WSPBuilder I suddenly hit a problem that my solution threw an error when I tried to build it. The error was in the CSC file (the C Sharp compiler) of the form:

Resource identifier has already been used in this assembly

The problem became more apparent looking at the line that Visual Studio used to invoke the CSC (C# compiler). The command line (shown in the output window in Visual Studio) contained two /resource switches which had duplicate entries that referenced the class in my assembly. The error showed up as CS1508 and this MSDN article has the details. The problem happens when you use declarative conditions in the if/else condition in a workflow.

I finally found this great post from Chris Mann explains how to fix this. A simple change the to .csproj file is all you need.

Equal height columns using div and CSS

Lars Nielsen — Thu, 26 May 2011 19:59:30 +0000

Finally now SharePoint is using DIV elements in the v4.master page, rather than using tables for layout as SharePoint 2007 did (and which is not xhtml compatible and has a load of other problems as well).

Using DIV’s rather than tables for layout is a great idea, but some things become more difficult to do with CSS and DIV’s. One of the more common issue is now to get a screen layout with two columns of equal height. With tables in HTML it was easy – just create two cells in a row and they will be sized to be the same height by the browser. DIV’s however are given a height dynamically to contain their content so it becomes quite difficult to create even a simple layout with 2 columns (DIV’s) of the same height.

There are plenty of possible solutions to this, and this post lists some of them. In my work recently I was working on some CSS that seemed to be following this technique. I guess it works, but it’s not exactly straightforward, and I was having trouble getting it stripped down to just two columns and working in a fixed width design rather than flexible width. So in the end I found this simpler alternative and used that as the basis for my CSS. It works fine in all the main browsers (although there is a problem with IE 6 and 7 – see below) and doesn’t involve the complexity of defining widths and relative positions in several places. It uses a similar trick, in making the background of the whole wrapper to be whatever you want for the background of the left hand column, and the other columns sit to the right and cover the background with another background. In this, I managed to create a layout with 2 columns of equal height with quite simple and basic CSS.

Here’s the code for a page to show the technique:






div test




    
    
    	
    	Left
Left
Left
Left
Left
Left
Left
Left

    	Left Left Left Left Left Left Left Left Left Left Left Left Left Left Left Left Left Left Left Left Left Left Left
 		
    
 
    
    
        
        
		Right Right Right Right Right Right Right Right Right Right Right Right Right Right Right Right
		Right Right Right Right Right Right Right Right Right Right Right Right Right Right Right Right
		Right Right Right Right Right Right Right Right Right Right Right Right Right Right Right Right
		Right Right Right Right Right Right Right Right Right Right Right Right Right Right Right Right
		Right Right Right Right Right Right Right Right Right Right Right Right Right Right Right Right
		Right Right Right Right Right Right Right Right Right Right Right Right Right Right Right Right
		Right Right Right Right Right Right Right Right Right Right Right Right Right Right Right Right
		Right Right Right Right Right Right Right Right Right Right Right Right Right Right Right Right
		Right Right Right Right Right Right Right Right Right Right Right Right Right Right Right Right
		Right Right Right Right Right Right Right Right Right Right Right Right Right Right Right Right
		Right Right Right Right Right Right Right Right Right Right Right Right Right Right Right Right
		Right Right Right Right Right Right Right Right Right Right Right Right Right Right Right Right
		Right Right Right Right Right Right Right Right Right Right Right Right Right Right Right Right
		Right Right Right Right Right Right Right Right Right Right Right Right Right Right Right Right
		Right Right Right Right Right Right Right Right Right Right Right Right Right Right Right Right
		Right Right Right Right Right Right Right Right Right Right Right Right Right Right Right Right
		Right Right Right Right Right Right Right Right Right Right Right Right Right Right Right Right
		Right Right Right Right Right Right Right Right Right Right Right Right Right Right Right Right
		Right Right Right Right Right Right Right Right Right Right Right Right Right Right Right Right
		Right Right Right Right Right Right Right Right Right Right Right Right Right Right Right Right
		Right Right Right Right Right Right Right Right Right Right Right Right Right Right Right Right
		Right Right Right Right Right Right Right Right Right Right Right Right Right Right Right Right
		Right Right Right Right Right Right Right Right Right Right Right Right Right Right Right Right

Update

This technique seems to cause problems with IE6 and IE7 in that the colour of the left hand DIV bleeds over into the right hand div at the bottom if the left hand div is higher than the right hand one, and you resize the browser to be less high than the DIV’s themselves. Other browsers seem to be OK. In the case of SharePoint 2010, it doesn’t officially support IE 6 anyway, and nearly everyone who uses IE 7 has upgraded to IE 8 at least, so in practice this may not be too much of a limitation.

My Links loses the orange gradient background on hover – A CSS case-sensitivity issue

Lars Nielsen — Fri, 06 May 2011 19:39:55 +0000

In trying to create CSS to style a custom master page in SharePoint, I discovered a strange quirk (you’ll see the significance of that word later) in the way SharePoint out-of-the-box styles and controls work.

The “My Links” link at the top of the page next to the “Welcome” box is rendered by the following line in a master page:

If you look at the HTML that comes out on the page, it’s marked up with HTML like this:

I’m sure your first thought on seeing this is: yuck! It’s the familiar wonky SharePoint HTML, but I want to point out this class assignment in the inner tag:

class=”ms-SPLink ms-hovercellinactive”

and this attribute:

hoverActive=”ms-SPLink ms-hovercellactive”

What happens here is that as the user runs the mouse over “My Links” the onmouseover attribute calls the Javascript function MMU_PopMenuIfShowing. That reads the hoverActive attribute and then dynamically changes the CSS class for the SPAN tag from “ms-SPLink ms-hovercellinactive” to “ms-SPLink ms-hovercellactive”, and that’s what creates the orange gradient background on “My Links” when you hover over it. Now if we look in the out-of-the-box core.css file, which is always on every SharePoint page, it has a CSS definitions for classes ms-hovercellinactive and ms-hovercellactive – here’s the extract from core.css:

.ms-HoverCellInActive,.ms-SpLinkButtonInActive
{
border:none;
margin:1px;
color:#4c4c4c;
vertical-align:top;
background-color:transparent;
}
.ms-HoverCellActive,.ms-SpLinkButtonActive
{
border:#6f9dd9 1px solid;
vertical-align:top;
background-color:#ffbb47;
background-image:url("/_layouts/images/menubuttonhover.gif");
}

So core.css defines the class as “ms-HoverCellActive” but in the HTML SPAN tag the class attribute is “ms-hovercellactive”. Does it matter that they have different case? Well, yes and no. Normally the SharePoint default.master page has no DOCTYPE definition at the top. This puts IE (8, not sure about 9) into Quirks Mode, and in Quirks Mode it ignores the case of CSS classes so it will render correctly. However, best practice in web design is to include a DOCTYPE and to take IE out of Quirks Mode. In my master page I’m using:

As soon as you put a DOCTYPE at the start of the page, and take IE out of Quirks mode, it starts to treat CSS class names as case-sensitive and expects the case to match. When you hover over “My Links” you lose the orange gradient background. To fix this, you need to define CSS for ms-hovercellactive and ms-hovercellinactive explicitly, in your own custom CSS file. You can either replicate what’s in core.css, or define your own CSS for these classes.

Task form not showing in Manage Form Templates – The specified form cannot be found

Lars Nielsen — Fri, 29 Apr 2011 20:14:07 +0000

I hit a problem recently developing a workflow in Visual Studio for SharePoint. I had created a custom task form in InfoPath and then I decided to change the task form a bit. When I re-deployed the workflow I found I couldn’t get the new version of the form to pick up; each time you edited a task it used the old version.

I changed everything about the form – the name and the GUID and then redeployed it. This time when I tried to edit the task I got an error:

The specified form cannot be found

On reading around I found I need to look in the Manage Form Templates page in Central Administration to see if my task form has been marked as Workflow enabled. The old form was there but not the new form with its new name.

The solution was quite simple but took me ages to spot. These two vital lines had been deleted from the element in feature.xml for the workflow feature:


ReceiverAssembly="Microsoft.Office.Workflow.Feature, Version=12.0.0.0, Culture=neutral, PublicKeyToken=71e9bce111e9429c"
 ReceiverClass="Microsoft.Office.Workflow.Feature.WorkflowFeatureReceiver"

These two lines of code are the feature receiver that makes SharePoint run its own internal code to install the form. Without them, the new task form was not being installed properly when the workflow feature was activated.

Crawl log errors and resolutions in Search

Lars Nielsen — Fri, 15 Apr 2011 21:58:11 +0000

I was recently going through the crawl log and fixing errors – here are some of the error messages in the crawl log, and resolutions:

Error in the Microsoft Windows SharePoint Services Protocol Handler

This was an attempt to crawl a non-existent ASPX page. Click on the link and you get a 404 error. I’m not sure why it’s trying to crawl a deleted page. Perhaps a broken link somewhere?

The crawler could not communicate with the server. Check that the server is available and that the firewall access is configured correctly

This was at attempt to crawl a broken ASPX page – in this case a view of a calendar that gave a .NET error when you click on the link in the crawl log. The resolution is to delete the ASPX view page from the list using SharePoint designer.

The parameter is incorrect

The crawler was trying to index a Word 2007 Template File (.dotx) instead of a regular Word .docx file.

Error in the Site Data Web Service

The target site was present (you can open it in SharePoint Designer) but its welcome page was pointing to a non-existent page so you could not browse to it. Resolution is to go to the Site Settings page directly by appending “/_layouts/settings.aspx” onto the end of the site root URL, e.g. “http://mydomain/site/subsite/_layouts/settings.aspx” and click on “Welcome Page” and reset the welcome page to point to a real page so that you can browse to the site.

Finally, I had a problem that the crawler was crawling content OK. I could find documents and pages by their URL in the crawl log, and the log said that they had been crawled successfully. But when you searched for the items by keyword in the search box, no results were returned. The problem seemed to be that the index not updating properly. I used the Reset All Crawled Content page in Search Administration in Central Administration, followed by a full crawl, and everything was OK.