| *** edcragg has quit IRC | 03:28 | |
| *** fay_ has joined #baserock | 04:27 | |
| *** faybrocklebank has quit IRC | 04:29 | |
| *** gtristan has quit IRC | 04:54 | |
| *** gtristan has joined #baserock | 05:23 | |
| *** rdale has joined #baserock | 10:20 | |
| *** rdale has quit IRC | 10:28 | |
| *** edcragg has joined #baserock | 11:12 | |
| gtristan | paulsherwood, https://people.gnome.org/~tvb/sample.ogv | 11:18 |
|---|---|---|
| gtristan | I'm a believer ;-) | 11:18 |
| * gtristan thinks he could get better performance with instances: 3 / max-jobs: 3 | 11:20 | |
| paulsherwood | gtristan: what's your machine spec? | 11:28 |
| gtristan | quad core very fast i7 (up to 3.9GHz) and plenty of SSD | 11:29 |
| gtristan | but I dont use half of it with this build setup | 11:29 |
| gtristan | because configure scripts and shell scripts and such on the emulators are slow | 11:30 |
| paulsherwood | well i reckon max-jobs: 4, and instances: 2 or 3 | 11:30 |
| gtristan | I'm doing max-jobs 4 right now but I only really leverage 3 | 11:30 |
| paulsherwood | did you manage to take a look at my latest attempt at 'claim' ? | 11:30 |
| gtristan | have to make the network interface virtio | 11:30 |
| gtristan | I did | 11:30 |
| gtristan | I didnt get to test it yet | 11:31 |
| gtristan | I like the patch, it make claim() cleaner | 11:31 |
| gtristan | and it certainly wont break my fix, as the point of my fix was to move yeild outside of the exception handling of flock() | 11:31 |
| paulsherwood | ack | 11:31 |
| gtristan | I only dont understand why your fix, fixes cntl-c :) | 11:31 |
| paulsherwood | me neither! :) | 11:31 |
| gtristan | I would like to understand that | 11:32 |
| gtristan | heh | 11:32 |
| paulsherwood | richard_maw turned me onto contexmanagers and with statements ages ago... now i can't resist that pattern :) | 11:32 |
| gtristan | Oh, as the not-python guy that I am, that took me a solid hour to figure out | 11:33 |
| gtristan | I presume it will come more naturally with practice ;-) | 11:33 |
| paulsherwood | :) | 11:33 |
| gtristan | I was reading on generator functions to understand what yeild does, but clearly it's used differently with that contextmanager decorator | 11:34 |
| paulsherwood | my understanding is it just hands control to the code_block | 11:34 |
| * paulsherwood waves his hands as if magic happens | 11:35 | |
| gtristan | sigh... I have to kill this build, I forgot to include 'make' in build essential.... but breaks my heart | 11:35 |
| gtristan | it's gotten so far ! | 11:35 |
| paulsherwood | yup... you're soooo close | 11:35 |
| * gtristan is building sqlite3 & libffi | 11:35 | |
| gtristan | perl and cmake took a good hour | 11:35 |
| gtristan | but the rest are going by pretty quickly | 11:36 |
| paulsherwood | gtristan: maybe spin up an aws server, build there? | 11:36 |
| paulsherwood | (this would also force you to reproduce your env :-)) | 11:36 |
| gtristan | I guess glibc, gcc, perl and cmake are the tough ones, the former two take hours | 11:36 |
| gtristan | well, good point... lets try that later in the week | 11:37 |
| gtristan | should be able to speed things up with virtio network | 11:38 |
| gtristan | and Rob Landley says... he saw a 20% performance gain in configure by using a statically linked busybox | 11:38 |
| gtristan | because of how memory paging works in qemu | 11:39 |
| gtristan | that may not be plausible as some things require better tools than what busybox provides | 11:39 |
| *** gtristan has quit IRC | 12:09 | |
| *** gtristan has joined #baserock | 13:02 | |
| *** locallycompact has joined #baserock | 13:52 | |
| *** cosm has quit IRC | 15:02 | |
| * paulsherwood waves from somewhere over the atlantic | 15:12 | |
| *** cosm has joined #baserock | 15:14 | |
| *** rdale has joined #baserock | 15:18 | |
| *** cosm has quit IRC | 15:44 | |
| *** cosm has joined #baserock | 15:59 | |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!