[Firebird-net-provider] Question: Why must I have a transaction object supplied for command object?

[Jason V. <jas...@ya...>]

Here's the situation:

I want to perform a simple SELECT * FROM SALES, read
the data and still be able to use the open connection
to continue reading.

I'm developing simultainous versions of the app using
both FbConnection/FbCommand/FbReader and
SqlConnection/SqlCommand/SqlReader from application
level objects that abstract to them.

When I perform the "ExecutReader" on the command with
SqlCommand, I can do that without supplying a
SqlTransaction object, and continue to use the
connection on subsequent readers and such.

With FbCommand, however, it throws an exception if the
transaction isn't supplied (which the source clearly
shows it requires, without option).

This means I can't use the connection until that
transaction is committed or rolled back. I can't
commit the transaction, because there's nothing to
commit - I'm only reading the data. I can't rollback
either, both throw exceptions.

Related to this question, I'm downloading the mono C#
compiler tools, but I wondered how (and perhaps why
not) I might build the source for the .NET provider
within VS.NET2003?  I want to pull out the exception
throw when transaction is found to be null - I don't
quite understand any reason the transaction is always
required.

Perhaps there's something else about this I simply
don't understand.

Many thanks......

Jason


__________________________________
Do you Yahoo!?
SBC Yahoo! DSL - Now only $29.95 per month!
http://sbc.yahoo.com