[02:48] *** lizmat_ joined
[02:49] *** vrurg joined
[02:51] *** vrurg_ left
[02:51] *** lizmat left
[03:12] *** apogee_ntv left
[03:12] *** apogee_ntv joined
[07:29] *** lizmat_ left
[07:29] *** lizmat joined
[10:56] *** camelia left
[10:57] *** nine left
[11:00] *** nine joined
[11:18] *** camelia joined
[14:08] *** rakkable left
[14:09] *** rakkable joined
[15:07] <timo> all this time we never turned on -flto huh?

[15:09] <[Coke]> 1809 ^^ if someone can force push that, it should trigger testing.

[15:28] <timo> the code as it is in the PR is wrong, but I don't think we have any tests that would trigger a crash

[15:32] <[Coke]> ah

[15:33] <timo> then MD fixed it to not be potentially crashy and it was as slow as what we have in the main branch right now ;)

[19:39] <timo> it seems like we lost the optimization that used to turn a boolification on an iterator into a specialized function that doesn't check the type first, and boolifying the iter object seems to be a surprising amount of time spent in the example from 1809