[sf-perl] Great meeting last night!
not.com at gmail.com
Thu Jan 28 10:28:42 PST 2010
I had a couple thoughts from Joe's talk, none of which are about the
One was about how seeding rand with a known constant can still give
different results with different perls sometimes (seemed to be
constant across platforms, and usually across versions, but there was
one case of 5.11.3 giving different results). There was a section in
Mark Jason Dominus' "Higher-Order Perl" on the problems with getting
"rand" repeatable, and on rolling your own. Look in Chapter 4
"Iterators" at http://hop.perl.plover.com/book/ and search for "Random
As for writing tests that touch a database- here's a sane way giving
people the option to run. Have the tests get a database connection
string & user, password from environment variables, and document them
in the README.
More information about the SanFrancisco-pm