[rt.cpan.org #18387] Apparent transaction state corruption leads to silent failures
" via RT" <[email protected]>
| Newsgroups | gmane.comp.db.postgresql.dbdpg |
|---|---|
| Message-ID | <[email protected]> |
Queue: DBD-Pg
Ticket <URL: http://rt.cpan.org/Ticket/Display.html?id=18387 >
On Tue Mar 28 10:02:29 2006, JONJ wrote:
> Huh. Yeah, the DBI docs make more sense there. But maybe I was
> misunderstanding the DBD::Pg docs on the subject. They mention "BEGIN"
> et al. (SQL statements), and not the begin_work method, which are two
> different things.
They're just describing how DBD::Pg implements AutoCommit=1 mode.
> If AutoCommit is on, and one does a begin_work, some other statements,
> and a commit or rollback, is everything after that point auto-committed
> again until the next begin_work call?
Yep.