Hi, its been some times i'm reading, first time i'm writing :)
first let me present myself : Christophe, 33, working on a e-crm application web based that deals with emails. This is app is written mostly in Perl with some C++. We're using mp1, Apache 1.3, Embperl and Postgres. The Perl libs are full object oriented. Yes i know i'm a lucky guy :) I need some explications on how to deals with transactions over differents objects and method calls with mod_perl and ApacheDBI. My problem is simple : i need to do some stuff that deals with many method call on differents objects, each of these objects asking for their own database object retrieve from ApacheDBI. My question is : is it safe to use transactions of multiple objects, doing a commit or rollback at the end as i'm sure i will stay in the same interpreter space and that no other user can ask for the same database handle to do some stuff while i'm working? or can ApacheDBI give the same database handle to another user and mess with my transaction? btw, ApacheDBI is always returning me a handle with the default parameters i used when issuing the connection ( {AutoCommit => 1} for example), even if i changed the parameters with local->{dbh}->{AutoCommit} = 0. Its seems i can't write in the cached handle hash but what i only get is a local copy of this cached handle. Am i right with this point? Is there another way to process? patch ApacheDBI to write in the %Connected hash that deals with the cached database handles? is this safe? Thanks a lot with you answers, qui éclaireront ma lanterne as we say in french. Christophe -- Reporting bugs: http://perl.apache.org/bugs/ Mail list info: http://perl.apache.org/maillist/modperl.html