On 07.01.2013 20:03, Michael Van Canneyt wrote:
- Some statements end implicitly a transaction. Ex:
http://dev.mysql.com/doc/refman/5.0/en/implicit-commit.html or all
oracle DDL statements. sqldb is not aware that the transaction is
terminated and its state doesn't match the db state.
- You can't make a method for every database command that has to run
outside a transaction.
Why not ?
Of course you can when you have all the time of the world. A method
that bypasses sql transaction control covers all cases once it is
released. Any custom connector method is trunk first and then wait for
next release to get it "production ready".
Eh ? That is the same for the bypass ?
As far as I understand it: the bypass could "immediately" be used for
any affacted SQL command once trunk is released (or the fixes version
with that bypass applied). If however in the other approach some
commands are forgotten the user has to wait again until the next version
is released.
Regards,
Sven
--
_______________________________________________
Lazarus mailing list
[email protected]
http://lists.lazarus.freepascal.org/mailman/listinfo/lazarus
