ADO Script Stops Working After Switch To SQL Server From Access

I was trying to migrate an ASP application from Access to SQL Server 2005 express, and found surprisingly that some code stops working which would seem to be independent of the data source. For example:

qtext = "SELECT MainText, TextType FROM MessageText WHERE [etc. etc. -- the query is OK]"
rsMain.open qtext, dbcon, , , adCmdText
if not (rsMain.BOF and rsMain.EOF) then

rsMain.MoveFirst
do while not rsMain.EOF

select case rsMain.fields(1).value

case 1

session("headline") = rsMain.fields(0).value
case 2

introtext = rsMain.fields(0).value
case 6

helptext = rsMain.fields(0).value
end select
rsMain.MoveNext
loop
end if
rsMain.Close


"rsMain" is an ADO recorset which defaults to a forward-only cursor as opened above; dbcon is an ADO database connection. Nothing fancy; and it has been working for years as expected with Access. But the recordset seems to behave differently when SQL Server is the data source (even though the whole point of ADO is to provide a level of abstraction, and I expected it not to change.) In the code above, checking the value of "rsMain.Fields(1).value" causes "rsMain.Fields(0).value" to disappear or become inaccessible. As a simple debug exercise, I tried response.writing "rsMain.Fields(0).value", a block of text, prior to the select case block, and it works fine once (as long as rsMain.Fields(1).value has not been accessed), but then a second write comes up empty.


Is it reasonable to expect that ADO recordsets and cursors function differently if SQL server is the data source vs. Access? Is this documented anywhere?