| *** mordred1 is now known as Guest923 | 12:07 | |
| opendevreview | Merged openstack/diskimage-builder master: allow passing arbitrary build options in containerfile element https://review.opendev.org/c/openstack/diskimage-builder/+/902366 | 16:17 |
|---|---|---|
| *** atmark_ is now known as atmark | 17:16 | |
| * JayF would <3 an ordered list of the line of hands in the air | 17:41 | |
| JayF | for jitsi | 17:42 |
| clarkb | JayF: they accept feature requests on their github issue tracker I think | 17:42 |
| JayF | jitsi is the correct target project for that, yeah? | 17:43 |
| clarkb | https://github.com/jitsi/jitsi-meet/issues/new/choose | 17:43 |
| fungi | "get in line, ye scurvy scallywags!" | 17:43 |
| clarkb | JayF: yup specifically "jitsi-meet" the top level project I linked I think | 17:44 |
| JayF | I like hand raising flows just because I'm *awful* at figuring out when people are done speaking on video | 17:44 |
| clarkb | Like openstack they have a lot of different components too but I think for high level feature requests like that the best thing is a top level request then they can farm it out/break it down as necessary | 17:44 |
| fungi | this shokz bluetooth headset is amazing. i walked into the kitchen to get something and then realized after a minute that i forgot to bring my netbook with me | 17:45 |
| JayF | clarkb: https://github.com/jitsi/jitsi-meet/issues/14625 thanks, I filed it here | 18:30 |
| clarkb | looks great. Hopefully they agree | 18:34 |
| fungi | for those interested in such things, there's some ongoing preliminary discussion about the details of the jit implemented in python 3.13: https://discuss.python.org/t/50756 | 18:36 |
| clarkb | almost looks like gcc won't be a valid compiler anymore? Or if it is it isn't useful since you're compiling with llvm anyway so may as well use clang? | 18:38 |
| fungi | if you want the jit then compile that way, right | 18:38 |
| fungi | it'll mainly be a decision distros need to make for themselves | 18:39 |
| fungi | the performance gains aren't phenomenal, but they're enough that it might sway decisions | 18:40 |
| JayF | I'll note it's not the default in gentoo yet, which implies to me there are packages we ship that can't pass tests with jit enabled -- but that's just conjecturing based on past gentoo-y behaviors. | 18:41 |
| fungi | JayF: gentoo is packaging python 3.13 alpha versions already? | 18:43 |
| JayF | > (3.13) **3.13.0_alpha4^t **3.13.0_alpha5^t | 18:43 |
| fungi | i'm building 3.13.0a6 locally from source, but not with clang yet | 18:44 |
| JayF | the ** indicates it's hard-masked, meaning you have to opt into having it installed | 18:44 |
| JayF | and it's a separate slot | 18:44 |
| JayF | part of why gentoo is great for python dev is it's trivial to install "n" versions of a tool against "n" pythons | 18:44 |
| JayF | e.g. I have tox installed for all pythons, 3.8-3.12, and can run from the one I prfer | 18:44 |
| JayF | Gentoo devs actually filed a handful of the 3.12 support bugs around the community :D | 18:45 |
| fungi | makes sense. debian doesn't usually start packaging new python versions in experimental until they reach the beta phase, hence my surprise | 18:45 |
| JayF | given we build all the things in gentoo, they have to more or less go through every package checking compatibility | 18:45 |
| JayF | so they try to get started as early as possible (and portage is written in python so early adopting new python there for portage devs is useful) | 18:46 |
| fungi | very cool | 18:46 |
Generated by irclog2html.py 2.17.3 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!