I use the content network sparingly and in campaigns which are for the content network only. However, I've never felt google chooses the sites to display my ads very well so review the sites regularly. The things I'm mostly interested in are a) which sites my ads appeared on I don't think they should (or I don't want them appearing on) and b) which sites are costing a lot but delivering no conversions. The following perl script processes a TSV version of the placement report (with specific fields) into a more easily managed list I can use to add domain exceptions to my campaign.
It appears that DBD::Oracle restricts the RowCacheSize to 128, why? I know how much memory I can afford to use for retrieving rows better than DBD::Oracle and as far as I can see this could be set a lot higher.
Just spent a few hours trying to find out why the RowCacheSize attribute in DBD::Oracle does not seem to have any affect at all. The following code demonstrates the problem:
I frequent perl monks a bit and late yesterday an issue came in with DBD::ODBC and Microsoft Access. The perl monks node and rt 46597 detail it all quite well. It appears DBD::ODBC was not following the DBI specification (as it was originally written) to the letter with respect to sticky bound parameter types but since then Tim Bunce has modified the DBI pod slightly to allow DBDs to enable a bound parameter to be rebound with a different type. The change to DBD::ODBC was fairly small but by the time I'd added a test case, bumped the version to 1.21_1 and tested a dozen or so ODBC drivers this consumed a fair amount of time.
Todays unicode problems are with JSON::XS. We are using JSON extensively in the project I am working on and we have a lot of unicode strings. JSON::XS really is the best and fastest perl JSON parser (I've tried them all) and Marc Lehmann who wrote JSON::XS is very knowledgeable and usually quick to respond to any query.