=== erkules_ is now known as erkules [09:59] slangasek: or stgraber: please review https://code.launchpad.net/~xnox/upstart/bug-1199778/+merge/174372 to unblock SRU. stgraber - wasn't there something else that needs to go in the SRU? (we now pulled the last one) [09:59] in the mean time I'm preparing saucy upload. [10:00] stgraber: lp:~ubuntu-core-dev/ubuntu/raring/upstart/raring === Md_ is now known as Md [13:17] xnox: the respawn fix [13:18] stgraber: ack. So we need to cherry-pick the respawn fix, above chroot-reexec-fix. Can you review above merge-proposal, if you are in the mood for upstart subtle bugs =) [13:25] xnox: I'll take a look once I'm fully awake ;) currently still at the processing IRC backlog and e-mails phase ;) [14:52] ok, I think I'm as awake as I'll be today, taking a look at that branch now [15:32] xnox: should test_session_upgrade do some checks of the unserialized data or is checking the return code sufficient? [15:33] stgraber: i think it should be checking unserialized data for a complete test. [15:33] stgraber: checking the return code is sufficient test for the bug in question: "upstart upgraded failed with active chroot session" [15:34] xnox: ok, can you add some tests to at least ensure those lists and hashes contain stuff post-unserialization? I don't think we need a ton of asserts but just checking that they aren't empty anymore would be nice [15:34] right, I see what you mean. [15:39] stgraber: and that caught a bug =) [15:43] xnox: yay! [17:04] stgraber: so, pushed more asserts. As it turns out chroot-sessions only have CONF_JOB_DIR, unlike normal system init which has CONF_FILE & CONF_JOB_DIR. [17:04] stgraber: so it checks that the correct session got deserialised with proper conf sources and that none of the job classes belong to a session. [17:05] stgraber: since we don't deserialise those at the moment. [17:12] stgraber: right, I'm off to my lunch break.... [17:12] * xnox scratches head [17:12] will be back to poke upstart or android a bit more.