>Here is the jist of my problem (i haven't tried to code a sample, hoping someone can shed some light)
>
>MsSql server table foo( pk_foo idenity, cFid1 char(10))
>
>vfp view v_foo
>use v_foo nodata
>append blank
>replace cFid1 with 'a'
>tableupdate()
>SQLExec( lnHandle, 'SELECT @@IDENTITY AS nId', 'id' )
>select v_foo
>replace pk_foo with id.nId && This dirtys the view's buffer
>SETFLDSTATE('pk_foo',1,'v_foo') && This cleans the buffer flag
>? GetFldState( -1 ) && shows '111'
>? GetNextModified(0) && shows 1
>goto 1 && Framework does this for me
>tableupdate() && error: Invalid call issued while executing a SQLMORERESULTS() sequence.
>
>
>I think the problem is GetNextModified is not checking the same thing GetFldState(-1) does.
>
>Does anyone know what I am talking about?

IIRC you're exactly correct. SETFLDSTATE() does not work entirely as expected; it seems to "fool" only GETFLDSTATE(), not other functions that rely on the "true" modified status. This behaviour was noted in HackFox3 and I don't think it's been modified/fixed in later versions of VFP.