[Haskell-cafe] Re: Strictness, order of IO operations: NewCGI & HDBC

[oleg at pobox.com]

Tim Smith wrote:
> Has anyone found out how to lift bracket into another monad?

Yes, please see the thread `Re: Control.Exceptions and MonadIO'
staring at
  http://www.haskell.org/pipermail/haskell-cafe/2006-April/015444.html

There is also a Haskell' ticket: 
	http://hackage.haskell.org/trac/haskell-prime/ticket/110

We have used this CatchableMonadIO in takusen and the other project,
with success. One wish is to have HUnit work in this CatchableMonadIO
rather than just IO (currently Alistair Bayley has to maintain an
adjusted version of HUnit for that purpose).

Regarding the issue of proper disconnections from the database server:
it was the thrust for the development of Takusen is to statically
prevent errors like forgetting to disconnect from the database or to
use the database connection after disconnecting (as well as related
issues of using a cursor after closing it, or forgetting to close the
cursor or to dispose the resultset). As to performance: Takusen can
read and process 2 million rows from the database, in the same space
it takes to fetch and process 10 rows. In fact, doing this is one of
the performance tests.