[12:15] aciba: i don't think raw ctypes can parse C Header files to extract constants from [12:53] aciba: looks like NetBSD's kern.boottime uses 1, 83, not 1, 21 [12:56] meena: good catch, I thought ctypes could load any symbol exposed in the header file, but C constants are preprocessed, so no luck there. [12:57] meena: I guess we have to somehow specify the second value depending on the specific BSD version to make it really compatible with all [13:04] yupp, about to do that [13:08] aciba: thanks for making me test this on NetBSD ;) [13:09] meena: thank you! [13:26] aciba: i think the boottime patch is ready, I'll try to fix the DMI unittests now, unless i fall asleep before [13:29] meena: sweet, thanks [14:16] meena: o/~ To be able to have my PR just work, I would like to have the updated ds-identify included, and I would also add that first in the list of /etc/rc.local ( already tested ). I can go ahead and add the call to it in /etc/rc.local... but was just wondering if you think that will be implemented soon? [14:16] Code_Bleu: i just pushed unittest fixes and extra unittests! [14:17] Once updated in repo, I'd have to just re-run the build-on-openbsd to get the new file...I guess [14:17] 👍 [14:19] Code_Bleu: yeah, the boottime stuff just got merged by aciba [14:20] and it works on NetBSD now too :O [14:21] 💥 [14:21] But i had to make the code a little uglier. [14:21] I just pushed up the ds-identify inclusion into the build-on-openbsd file in my PR [14:21] Next time i have to do this, I'll just send a patch to OpenBSD to add sysctlbyname() [14:32] time for some upgrades. === meena1 is now known as meena [14:36] meena: I had looked up the mypy failure you have on your PR and some quick google searches show this is/was a bug with mypy and some put a comment in the code to ignore false positive. There is still the shell lint issue too - https://github.com/canonical/cloud-init/actions/runs/7115635902/job/19372249769?pr=4654: [14:47] Code_Bleu: fixed tests. Thanks for pointing those out [14:48] * meena works better in peer-to-peer kinda setups [14:48] "pair programming" i think it's called. [14:50] 👍 [14:51] need to fix up the commit messages to say (#4654) [14:57] meena: Now with the updates to the sysctl_field, you think we can add the board_name, and chassis_asset_tag to this and get rid of the ERRORs with ds-identify? That would just leave the blkid one - https://github.com/canonical/cloud-init/pull/4654/files#diff-8e27dfc14b3ffceee1e29d3cf58dc25309c0ac070879febd47c578de04ea2dddR219 [14:57] -ubottu:#cloud-init- Pull 4654 in canonical/cloud-init "dmi: support OpenBSD native reading" [Open] [14:58] ERROR: Unknown field board_name. Cannot call sysctl. [14:58] ERROR: Unknown field chassis_asset_tag. Cannot call sysctl. [14:58] ERROR: failed running [127]: blkid -c /dev/null -o export [15:06] Code_Bleu: good enough. [15:06] but the blkid stuff needs to be fixed [15:07] I'm gonna add OpenBSD DMI tests and then declare this done [15:12] done. At least I think so. Let's see what CI says [15:25] okay, caught all the mypy issues now, i think [16:48] meena: I updated logic in build-on-openbsd again. I noticed on a fresh install of OpenBSD using ISO and no extra stuff from the build tool I was using with QEMU, it did not have /etc/rc.local file. So, I changed the logic to append if exists and cloud-init is not in the file, or create if doesn't exist with making sure the 'export PATH' is included - [16:48] https://github.com/canonical/cloud-init/pull/4660/files#diff-ee4bc86ff002c271617337b508f60599f501b91a4859239ce9923304cae10f60 [16:48] -ubottu:#cloud-init- Pull 4660 in canonical/cloud-init "fix(openbsd): services & build tool" [Open] [17:10] Code_Bleu: I'm working on disklabel support for openbsd/netbsd [17:11] and not FreeBSD, because we don't have disklabel. We only have bsdlabel, which works differently [17:20] meena: cool [17:20] meena: hey, the ctype thing is broke for me again :( [17:20] :( [17:20] good thing it was merged before you noticed [17:22] sorry, I only noticed after running through this on a new install in CloudStack. I didn't realize there was changes after it worked before for me [17:22] Traceback (most recent call last):¬ [17:22] File "/usr/local/lib/python3.10/site-packages/cloud_init-23.4+16.g41e02f1c-py3.10.egg/cloudinit/util.py", line 2118, in uptime¬ [17:22] uptime_str = str(time.time() - boottime())¬ [17:22] File "/usr/local/lib/python3.10/site-packages/cloud_init-23.4+16.g41e02f1c-py3.10.egg/cloudinit/util.py", line 2093, in boottime¬ [17:22] libc.sysctlbyname(¬ [17:22] File "/usr/local/lib/python3.10/ctypes/__init__.py", line 387, in __getattr__¬ [17:22] func = self.__getitem__(name)¬ [17:22] File "/usr/local/lib/python3.10/ctypes/__init__.py", line 392, in __getitem__¬ [17:22] func = self._FuncPtr((name_or_ordinal, self))¬ [17:22] AttributeError: Unable to resolve symbol¬ [17:22] Y [17:22] sorry for spam... [17:23] Code_Bleu: looks like cache issue [17:23] delete the .pyc [17:23] ah..ok :fingerscrossed: [17:24] Code_Bleu: here: 🤞 [17:27] yeh, apparently I don't know the code for that ...or my weechat plugin for emojis isn't setup for it 😄 [17:31] Code_Bleu: possibly: :crossed_fingers: [17:32] meena: ok, sike... I was kiding...no issue with the ctype stuff 😉 [17:32] \o/ [17:32] removing the pyc files didn't work the first time...so i just did a pip uninstall cloud-init and that did it [17:33] meena: AWS Open Source director mentions FreeBSD AMIs in passing: https://www.theregister.com/2023/12/04/david_nalley_interview/ [17:33] "FreeBSD is also an option. “I met with the FreeBSD release manager here, talking about AMI’s [Amazon Machine Images] for FreeBSD." [17:33] Colin Parcival works on the FreeBSD AMIs, and on FIRECRACKER [17:34] and he recently took over FreeBSD release management [17:34] yeah I guess it was him being referred to [17:45] minimal: I guessing you're not an "above and beyond" type of person? 😉 😆 [17:45] yah know...minimal and all [17:49] oi! I resemble that remark! ;-) [19:10] meena: I have some comments on my PR for ya when you get a min. [19:19] Code_Bleu: i think this has been the fastest bedtime in weeks, so I have time *now* [19:25] Code_Bleu: my comment was just a suggestion [19:25] I didn't mark it as Changes Requested [19:28] meena: ok, since I'm more of a n00b than you...I still would like to know if there is a better way, but my mind is on several things today at once and I'm multitasking like a fool. So, since it's just a suggestion...I'll leave it as is for now and change later if needed [19:30] Code_Bleu: just add a comment and/or mark my comment as resolved [19:31] meena: will do. thanks again for all of your input [19:31] Code_Bleu: I'm great at hand-holding. [19:33] meena: well, I thought I was a bit better than "hand holding" , but ok 😄 [19:34] Code_Bleu: we all have great aspirations and then we spend four hours trying to find where the missing semicolon goes [19:35] * meena certainly has done that [19:39] lol [22:23] I'm curious why the network activators / renderers code were made a configuration option rather than hard-coded in the distro [23:07] Code_Bleu: I just sent you more questions by a clueless Linux user :-P [23:26] holmanb: and I answered two of them, kind of [23:27] more for Code_Bleu's benefit then yours, really [23:34] +1 thanks meena [23:49] holmanb: I'll breifly looked over the comments. It's gonna take some time to get back with you. It's been one of those days. I thought for sure this would have been a done deal this time :/ [23:50] Code_Bleu: it's looking pretty close [23:50] Code_Bleu: there are just a couple of things that need to change, and a couple of things that I need to make sure that I understand before landing [23:50] holmanb: in a nutshell, the sysvinit tmpl files are not being used because we are using /etc/rc.local, but I left those in there anyway and they do work if you manually run them with rcctl. [23:51] holmanb: I do need to go for now...but I will try to update the comments on GH later tonight...or sometime tomorrow.