| *** tristan <tristan!tristan@223.62.212.129> has joined #buildstream | 04:53 | |
| *** ChanServ sets mode: +o tristan | 04:53 | |
| juergbi | nanonyme: by default mtime is not captured. buildbox-casd does support optional capture of mtime, which buildstream uses for workspace support but not otherwise | 06:03 |
|---|---|---|
| juergbi | (iirc) | 06:03 |
| nanonyme | juergbi: okay, I was just trying to figure out if local source import cache keys were affected by mtime | 06:56 |
| juergbi | ah, no, only content. capturing mtime by default for local files would typically be a problem (e.g. git not storing mtimes either) | 06:57 |
| nanonyme | juergbi: good. Then it's some other reason why for me cache keys aren't persisting as expected | 07:17 |
| *** tristan <tristan!tristan@223.62.212.129> has quit IRC | 09:17 | |
| *** tristan <tristan!tristan@223.62.188.129> has joined #buildstream | 13:49 | |
| *** ChanServ sets mode: +o tristan | 13:49 | |
| *** tristan <tristan!tristan@223.62.188.129> has quit IRC | 13:57 | |
| *** tristan <tristan!tristan@223.62.188.41> has joined #buildstream | 16:03 | |
| *** ChanServ sets mode: +o tristan | 16:03 | |
| *** tristan <tristan!tristan@223.62.188.41> has quit IRC | 16:39 | |
| nanonyme | juergbi: I don't suppose you have had another chance to look at the copy_file_range changes? | 17:33 |
| juergbi | nanonyme: I've commented. It looks fine to me if we can fix the couple of things I've mentioned | 18:25 |
| abderrahim[m] | <nanonyme> "juergbi: good. Then it's some..." <- I've had this happen to when python created __pycache__ directories in a local source. It could be something similar | 18:29 |
| abderrahim[m] | * python created `__pycache__, * __pycache__` directories | 18:29 |
| juergbi | nanonyme: isn't the hasattr("os"...) change wrong? | 18:44 |
| nanonyme | It is. I wonder where that came from. Maybe stale local changes on different machine. | 18:45 |
| nanonyme | @abderrahim:gnome.org: the problem was that I had a bug with freedesktop-sdk-binary-seed such that one of the things in buildstrap had sandbox arch not set so it defaulted to x86_64 | 18:46 |
| nanonyme | @abderrahim:gnome.org: this was problematic because I'm native building things and transferring over CAS | 18:47 |
| nanonyme | Erm, s/buildstrap/bootstrap/ | 18:48 |
| * nanonyme is too tired | 18:48 | |
| juergbi | nanonyme: ah, another small thing, the helper function should be private, i.e. rename it to _copy_file_range | 18:51 |
| nanonyme | Done | 18:52 |
| nanonyme | juergbi: could also be written more efficiently with Python 3.8+ syntax. But I guess we support older Python atm. | 18:55 |
| nanonyme | juergbi: haha. Linter failed because I used global :D | 19:00 |
| nanonyme | @abderrahim:gnome.org: thanks for the ground-work. This build https://gitlab.com/freedesktop-sdk/freedesktop-sdk-binary-seed/-/jobs/3591745470 would not probably exist in this form without it | 19:18 |
| nanonyme | (that is, the "let's fix CAS keys" feature) | 19:46 |
| nanonyme | juergbi: looks like there's some random failure in ostree tracking test | 20:14 |
Generated by irclog2html.py 2.17.3 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!