| 00:22:06
| * phated | quit (Remote host closed the connection) |
| 00:22:43
| * phated | joined |
| 00:23:07
| * phated | quit (Read error: Connection reset by peer) |
| 00:47:51
| * creationix | joined |
| 02:11:47
| * phated | joined |
| 02:16:27
| * phated | quit (Ping timeout: 268 seconds) |
| 03:59:56
| * phated | joined |
| 04:04:08
| * phated | quit (Ping timeout: 246 seconds) |
| 04:23:44
| * phated | joined |
| 04:28:08
| * phated | quit (Ping timeout: 240 seconds) |
| 05:16:57
| * pfrazee | quit (Quit: Connection closed for inactivity) |
| 08:29:28
| <emilbayes> | https://blog.acolyer.org/2017/06/09/node-fz-fuzzing-the-server-side-event-driven-architecture/ |
| 08:40:23
| <emilbayes> | TL;DR: They modified libuv to fuzz the event loop and reorder, delay or otherwise randomize how work is handled, but within the guarantees that node generally has (except that timers can be indefinitely delayed, which broke a lot of tests) |
| 08:41:00
| <emilbayes> | according to the blog post they will publish node.fz as soon as they have updated to latest libuv (?) / node |
| 10:21:35
| * creationix | quit (Remote host closed the connection) |
| 10:25:10
| * mylesborins | quit (Quit: farewell for now) |
| 10:25:40
| * mylesborins | joined |
| 12:56:36
| <robertkowalski> | mafintosh feross turns out we spend most of our time in lookups and announcements. this flamegraph is when we already cached lookups: https://cldup.com/afR-Y74pIL.png |
| 12:56:57
| <robertkowalski> | in both cases a good amount of time is spend in BEncode |
| 12:59:09
| <robertkowalski> | s/spend/spent |
| 13:12:40
| <mafintosh> | bencode is prob pretty optimal already |
| 13:37:11
| * creationix | joined |
| 14:41:55
| * pfrazee | joined |
| 21:15:52
| * phated | joined |
| 23:10:51
| * creationix | quit (Remote host closed the connection) |