Hello,

an interesting case with Firebird 2.5 here. Its the same across all
architectures (SS, CS, SC).

Two sessions via isql to the same database.

- Session 1 creates or alters a SP

set echo on;

SET TERM ^^ ;
CREATE OR ALTER PROCEDURE P_1 returns (
I Integer)
AS
begin
i = 9;
suspend;
end ^^
SET TERM ; ^^

commit;

- Session 2 executes

SQL> select * from p_1;

I
============
9

Which is correct. In the same session:

SQL> commit;

Starting a new transaction with snapshot isolation.

SQL> set transaction snapshot;


- Session 1 now alters the stored procedure to

set echo on;

SET TERM ^^ ;
CREATE OR ALTER PROCEDURE P_1 returns (
I Integer)
AS
begin
i = 10;
suspend;
end ^^
SET TERM ; ^^

commit;


- Transaction of session 2 is still running (in snapshot isolation). I
then execute another :

SQL> select * from p_1;

I
============
10

Which IMHO is wrong.


With Firebird 2.1, the last select * from p_1 gives me 9 as result,
which is IMHO correct, because I started a snapshot isolation
transaction *before* the CREATE OR ALTER PROCEDURE has been executed by
session 1.

So, anything wrong with snapshot isolation in Firebird 2.5? Is this
reproducable?

Thanks!



--
Best Regards,
Thomas Steinmaurer
LogManager Series - Logging/Auditing Suites supporting
InterBase, Firebird, Advantage Database, MS SQL Server and
NexusDB V2
Upscene Productions
http://www.upscene.com
My blog:
http://blog.upscene.com/thomas/