Rob Kinyon rob.kin...@gmail.com writes:
MVCC is, afaik, the only way to do properly serializable transactions.
Oracle does this, but no-one else does.
PostgreSQL uses MVCC as well, and although it supports serializable
transactions, the default isolation level is read committed.
--
ilmari
A
IMHO you are doing the right thing, but i can advice you a futher
improvement and very important - perfomance boost for your module.
You can do shorcuts shorter:
in setup_finalize of Catalyst:
foreach my $source ($schema-sources) {
next unless index($source, '::') == -1;
Hi,
On Fri, Feb 5, 2010 at 2:05 PM, Oleg Pronin syber@gmail.com wrote:
IMHO you are doing the right thing, but i can advice you a futher
improvement and very important - perfomance boost for your module.
You can do shorcuts shorter:
in setup_finalize of Catalyst:
foreach my
Thanks to the list for tips. I can't say the process is running that
much faster (its just a lot of data to sling around) but postgres can
now handle multiple processes at once so I'm seeing a definite speed
up with multiple jobs running.
However, I'm still seeing something interesting, as
James Marca wrote:
From reading the docs, starting with
http://search.cpan.org/~frew/DBIx-Class-0.08115/lib/DBIx/Class/Storage/DBI.pm#DBIx::Class_specific_connection_attributes
and then linking to http://search.cpan.org/perldoc?DBI#prepare_cached,
I see this for 'prepare_cached':
Like prepare