I upgraded to LLBLGen 3.1 last week and I noticed that the validation is stricter than 2.6. I had to make a few adjustments, but nothing major. When I tried to Refresh the Relational Model Data from the Database, it said that I had some pending changes and gave me a Database Schema Update Script to run on our database to bring it into sync with the designer.
That makes sense to me because I did make changes, however not all the changes in that script seem to be related to things that I changed. There are several tables that the update script is trying to add the identity property to the Primary Key field. Our application is huge and those tables have been around for years and I have no idea where those tables are used in our legacy code or why they don't have the identity property set.
My question is essentially this: Do I have to add the identity property to those fields? If so, why? If not, why does the LLBLGen designer think I do?
Here is a simplified version of the script for just one table:
ALTER TABLE [dbo].[ActivityLink] DROP CONSTRAINT [PK_72d0ad741b5a0f6d307b2c0e9ce]
ALTER TABLE [dbo].[ActivityLink]
ALTER COLUMN [ActivityLinkID] [int] IDENTITY (1,1) NOT NULL
ALTER TABLE [dbo].[ActivityLink] WITH NOCHECK
ADD CONSTRAINT [PK_72d0ad741b5a0f6d307b2c0e9ce] PRIMARY KEY CLUSTERED
(
[ActivityLinkID]
) ON [PRIMARY]
LLBLGen Pro v3.1 Build: September 30th, 2011
We use SQL Server 2008