This post explains nicely why i just deleted every web in my whole site collection.

Basically OpenWeb() with no parameter passed will return the first valid web in the URL string passed to open the SPSite object! So it kept walking up the hierarchy, returning every level to my recursive delete function.

Ah well, at least it wasn’t production….

• Team sites now have a pages library
• Wiki linking is built into pages. You can now wiki link to pages, lists, items and docs
• Silverlight is built-in to SharePoint’s ‘Create’ dialog. Apparently SL can be made accessible. Not sure if this is the case with this Silverlight dialog. Need to check accessibility
• Pages check for changes every minute and auto save if you have made any.
• Built-in Rich Text Editor has multiple levels of undo
• Versioning is on by default.
• New master pages exist. v4 is the latest. Changes here are carried through to site settings pages. Also have a minimal (no ribbon), simple (for error pages) and default (legacy)
• Accessibility is covered by everything being AA and XHTML. Not sure if this applies to central admin though.
• Text editors all render to XHTML
• You can add buttons to the SP Ribbon. I tried this and it’s a nightmare….
• You can associate more forms with lists, rather than New, Edit & Display. This includes InfoPath forms. Done through SPD
• XLSTViewWebPart is used for all list data now. More efficient and custom XSL can be supplied.
• Ajax settings can be configured on lists, i.e. refresh interval, manual refresh button, enable/disable Ajax etc.
• Lists can support inline editing (enabled in SPD) and a rating column (think star’s *****) , although InfoPath doesn’t support the latter in it’s designer.
• 2 versions of InfoPath are being banded about, a designer and a ‘Filler’
• A new post processing synchronous event now exists for lists that can be used to give feedback to a user after an item has been added.
• Cancelled events can redirect to a custom page that displays within a SP Modal Dialog
• Event submitters can be determined. In WSS sometimes the event owner would be assigned to a workflow/system account. MSF holds details on who raised what event.
• Built-in SP Explorer doesn’t expose bound list events. This is poor. No visibility if a list has 34 events already bound to it.
• Lists now support referential integrity (master detail) and cascading deletes.
• List joins can be made through queries (join up to 6 lists). Limit of 800kb for return results
• Projected fields now exist within lists, i.e. on a child list display a field from the parent, i.e. parent:company name
• MSF (Microsoft SharePoint Foundation) now supports 50m list items
• Throttling available from central admin for list queries. Can be time based (i.e. only allow this quota between 9-5pm). Settings available to allow/deny overrides via the object model. Administrator has a lot of control on query results.
• Unique column constraints available on column types. Company name can be made unique. Requires indexing and depends on SQL Collation
• Linq To SharePoint removes CAML Pain, but you need to generate SP Class Entities via a tool in the 14 hive called SP Metal. Basically point SP Metal at your SP Site and it generates a load of strongly typed C# classes.
• Lists are still stored as STP files, despite site templates being converted to WSP’s.
• Columns now have validation. Syntax similar to Excel, i.e. [CompanyEmployees] >= 50

Client Object Model

• Consistent interface across .NET, Silverlight, Javascript. Same (to a degree) method names. Obviously some differences in type support, i.e. JS’s NAN
• ‘Client Context’ used to batch operations
• All communications go through a client.svc WCF Service
• Requests appear to be sent as XML, responses as JSON
• New Silverlight webpart. XAP files deployed to webparts folder. XAP file is simple a zip for cross platform operation
• LINQ To SP not provided within Silverlight. ADO.NET Data services used instead
• Uploading an office theme for use in SP2010 doesnt work in this beta.
• SharePoint ‘Web application service’ service under central admin is used to sync your WFE servers. It’s dangerous though because it starts deleting your IIS files when you stop it. It will recreate them when restarting but any web.config changes etc you’ve made are gone

Workflows

• Workflows can now run against a site & list items. Event handlers needed to initialise
• Workflows can listen to notifications from external systems before resuming excution. Uses EDE (external data exchange). Long winded and complicated process
• Workflows can now be added & deployed via the new Feature packager
• Workflows can be created in Visio, exported to SPD and then again brought into VS2010. You can move between Visio & SPD easy enough but it seems that when it enters VS2010 it can’t be exported again to the others
• No version control in SPD other than SP2010’s list version control.
• New workflow events (started, completed)
• New visio web plugin displays visually the workflow path. It’s ok. Doesnt center workflow properly on load. Bit like page viewer web part
• SSP is gone, replaced by Service Providers. These can be load balanced & moved across servers as required.
• New Word automation service can be used to convert docs to PDF, Word 2003 etc. Word Open XML SDK intended for doc manipulation. Will still need Aspose.Net for Word 2003 docs
• Developers can create their own Service Providers. Looks reasonably complicated. Specific usage scenarios, i.e. intense calculations etc

BCS

• BCS is the new BDC
• Entities are now ECT’s, External Content Types
• BCS is now supports read/write to external systems
• SPD used quite well to create connections and CRUD screens
• SP2010 now has external lists that can expose data from external systems. Appears that workflows now support for these list types
• SSO details now stored in new Secure Store Service
• Groove (SP Workspace) seemed decent enough. Change I made in Groove went back to SP2010 and then on out to external system (SQL DB)

ECM

• Folders now supported in Page Libraries
• new content deployment option ’snapshot’. Time based snapshot sent between servers. Requires Enterprise SQL ‘05 or ‘08.
• New web analytics & reports. Looked ok but definitely not google contenders. Didn’t appear to be a means to export
• End users can now tag metadata to content themselves in addition to standard method
  Record management now possible ‘outside’ record center
• New ‘Unique Document Id’ service. Every document given a unique ID. Stays with it even if moved. Apparently has limited configuration options. This ID when used in conjunction with a specific URL (by sticking the ID on the query string) will take people to the document regardless if it is moved within the same document library or across libraries within a site collection
• Hold actions no longer tied to record center
• New document set. Content type composed on sub content types. ‘Document Set’ feature needed to be activated for it to work.
• Doc sets have versions and workflows can be run against them and constituents. Possible use is WF to check that all docs have been completed before moving to next stage.

Search

• 3 Levels of Search for SP Foundation Server
  • • Basic – Search Server Express 2010
    • Intermediate – SharePoint search 2010 (Index outside SP & Filestores)
    • Advanced – FAST can accept ‘push’ updates from code
• JDBC Support now included
• SP2010 has search connectors for Exchange, Notes, Documentum, FileNet, Opentext & custom .NET connector
• Support for search attachments
• Search web parts now unsealed. Can be overridden.
• New ability to override a query before its executed and results before display to users
• Search can have multiple levels of ranking models. Can be manipulated via PowerShell

BI

• BI Includes status indicators (as before, i.e. red amber green), chart web part, Performance Point Services, Excel Services, BI Search, Visio Services (Visual representation of KPI’s in Visio)
• Reporting Services doesn’t ship with 2010. Separate download
• BI Search adds Excel & Reporting services to search. Search for ‘delivery’ & spreadsheets with query are returned & filtered on query
• BI Centre template – Starting point with BI Portals
• New server product ‘Gemini’ provides in memory olap manipulation. Seamless for developers, no need to touch it.
• Little change is PerformancePoint Services
• Office Excel Web app looks decent.

Sandbox Solutions

• Site collection owners have power to deploy WSP’s to the sites sandbox, which is called ’solution gallery’
  Sandbox basically uses a cut down SP2010 Object model.
• Code executes in a separate process ‘Microsoft SharePoint Foundation User Code Service’
• A validation framework exists that can scan & validate wsp’s that are uploaded to the sandbox. Validators can revoke permission for the wsp to execute
• Sandboxed solutions are isolated to their own site collection.
  Two versions of the Microsoft.SharePoint.dll exist. The full version and the subset.
• When you create a new VS2010 project, you’re prompted if you want full trust of sandbox. If you select sandbox it ’still’ uses the full dll. No compile time checking that you should be using a restricted sharepoint dll.
• Sandbox apps use quotas. These are configurable and cover a varied number of scenarios such as process memory, db calls etc. For example, 20 db calls = 1 resource point. It you hit the max quote your web part/code is halted by SharePoint
• A timer job resets quotas at the end of the day.
• Doesn’t appear to be a means to determine what caused your quota to max out
• Quotas set via central admin or powershell
• I feel that setting quotas could be a dark art. Users * web parts * queries/cpu executions etc
• No means to promote from sandbox to live. Would require a removal & redeployment
• At runtime, sandbox wsps are retrieved from the solution gallery. Any required dll’s are extracted to a temp folder, loaded into memory and immediately deleted.
• It doesn’t look like you can debug sandboxed apps. There’s A) no place for PDB files and B) the user code service is a SP Service under central admin. No way to attach to it as far as I can see.
• Sandboxed apps can be load balanced

If you want to customise the default forms that come with SharePoint lists (DispForm.aspx, EditForm.aspx, NewForm.aspx) then you should deploy the list as a feature so you can modify the actual file in the 12 hive instead of the unghosted version in the database. This saves you reconfiguring the list in every environment you deploy to which is fantastic for a Dev –> UAT –> Production transition.

Preamble:

SharePoint features are great for many reasons I won’t go into here  but one that i will.

Most people are already using the SharePoint solution deployment approach but if you’re one of the unconverted, there’s plenty of information out there in google land on how Solutions + WSPBuilder + Powershell = Win.

Setting the Scene:

I recently had to develop a custom SharePoint list. The project structure consisted of:

• List template Feature
• Content Type Feature
• Content Type association feature (to join the dots). [Note 1 – if you’re already lost then see Andrew Connell’s blogs on feature deployment]

What i wanted to do then was to change the list ‘EditForm.aspx’ and ‘NewForm.aspx’ to be completely custom. So I fired up SharePoint designer and added a DataView webpart.

[Note 2 – there’s many ways to make a hash of this – the way that i find works is this one: http://social.msdn.microsoft.com/Forums/en-US/sharepointcustomization/thread/06873fe5-19a9-4b32-91a9-31164afdd4bb – i.e. don't delete the ListView webpart, just hide it]

So now i can implement my custom forms in any way i like and add in all kinds of extra functionality. That’s a blog post for another time though.

The issue:

As the form is now unghosted (i.e. it has been customised from the list definition [filesystem] so it exists in the database), all form changes need to be done in SharePoint Designer and will have to be repeated every time you redeploy the list.

There’s other reasons why you shouldn’t unghost pages – performance (not precompiled), potential upgrade problems, etc..

Is using SPD in each environment so bad?

It’s not terrible, it just takes a lot longer, is easier to make a mess of and I don’t like the SPD development environment either. i just find it slow and annoying. After all, it is doing round trips to the database every time you call or save a page. And it often attempts to nanny you by ‘cleaning up’ your HTML behind the scenes. The most obvious example of this is the removal of the ‘~’ character in relative URLs as it attempts to resolve these as absolute.

So I’d rather work in a proper code editor environment and have my changes directly applied.

The Solution:

As you’ve probably worked out by now, it’s by making sure your pages never get unghosted.

The key to this is to deploy your list feature as normal, but make changes to the EditForm.aspx directly in the list definition feature in the 12 hive.

How to do this is to use SPD on your dev environment, get the page exactly how you want it and then go deploy the feature.

As SharePoint feature GUIDs are non-deterministic (sigh) the instance of the list you create will have a random GUID which you won’t know at this point. It simple become a cut and paste exercise to get the new list GUID and insert it in your EditForm to replace the existing one.

The easiest way to get the list GUID is:

Site Actions –> Site Settings –> Audience Targetting Settings

The URL will now show you the list GUID in a non-html encoded format:

And voila! You can now deploy this list to any environment as many times as you like and the only manual work is to replace the GUID once the list instance has been created! 2 words – AweSome.

• Add the column name as a mapped property.
• Navigate to a search results page and Edit the page to access the web parts
• In the Search Core Results Web Part select Edit and Modify the Web Part
• In the Search Core Results settings window navigate to the Results Query Options
• Select the Selected columns and enter the column name to be displayed, in this case <Column Name=”DocumentSubject”/> and Click OK to save the change
• In Data View Properties click XSL Editor
• Navigate to the section beginning  <div class=”srch-Description”>
• To remove the synopsis from the search result, delete the section from as highlighted in the screenshot above
• Under the tag <div class=”Srch-Description”> enter the following line for the column name to appear in the search result
  <xsl:value-of select”=documentsubject”/> </div>
• Click Save to save the changes, click Apply and Ok the see the changes in the search results page.

The Search Result will now bring back the Document Subject for the document.

Event Type: Error
Event Source: Office Server Search
Event Category: Search service
Event ID: 10038
Date:                       8/8/2009
Time: 9:53:02 AM
User: N/A
Computer: WFE1
Description:
Query machine WFE1′ has been taken out of rotation due to this error: The
content index is corrupt.   0xc0041800.  It will be retried in 15 seconds.
Component: c9dba2c2-0077-46f7-b352-aeeddb6ac2fd

In the case where there are multiple front end servers in the farm, end users will remain unaffected since their search queries will be served by the other front end server, but this issue can cause crawls to hang and will log errors to the application logs every time a crawl is attempted.

This happened on one of our production environments recently, and the first thought was that we should reset the indexes and recrawl the content (which would have worked, but would have taken about 3 days to re-crawl everything and we wanted it working by Monday morning).  So in trying to avoid a recrawl, we did the following:

• Open the Operations tab of the Central Administration site
• Select Services on Servers and choose the offending WFE from the list
• Stop the Office Server Search Service on the WFE
• Start the service again, but give a new location for the index file
• The next time the crawl runs, the index will be propagated to the new location and bring the server back into rotation.

Hope this helps avoid many hours of re-crawling content.