Reduce amount of initial connects during non-SQLite test-RDBMS lock-grabs
authorPeter Rabbitson <ribasushi@cpan.org>
Wed, 7 Oct 2015 11:39:46 +0000 (13:39 +0200)
committerPeter Rabbitson <ribasushi@cpan.org>
Thu, 15 Oct 2015 15:33:55 +0000 (17:33 +0200)
commit0f6d86e4c025273bdbe5d6527e44d7b42b7fede8
treef646b85de97969ede87ae96efe055c175e773149
parent821edc0964a64b9d20b7d02c4a738b87e806f32d
Reduce amount of initial connects during non-SQLite test-RDBMS lock-grabs

I am not sure why I wrote things this way back in 8d6b1478d, perhaps I didn't
properly think through that sqlt_type will be correctly determined no matter
what. As a result any MySQL test would connect at least once, and under the
right circumstances consume enough per-connection-thread memory to get a
low-RAM system (like TravisCI) to start OOM-killing things indiscriminately.
t/lib/DBICTest/BaseSchema.pm