| minimal | doublechecked to be sure (23.3 fine, git main shows the mentioned output). | 00:22 |
|---|---|---|
| minimal | can't see anything obvious in a diff between 23.3 and main code | 00:22 |
| minimal | meena: looks like my issue may be #4487 | 11:26 |
| meena | minimal: it's still bizarre. | 11:28 |
| meena | why is it trying to render all these other files? | 11:31 |
| meena | minimal: is there a difference between python versions? | 11:31 |
| minimal | meena: yeah I mentioned in a note I added to the issue that as I'm specifying openrc init then I don't know why it treies to template the others | 11:43 |
| minimal | I tested building 23.3 on the *same* build environment (i.e. versions of python etc) and it doesn't behave the same | 11:44 |
| meena | The systemd stuff has been around a lot longer than that tho. So are my changes throwing it off then? | 11:47 |
| minimal | the packaging still completes fine, I just find those YAML messages strange | 11:48 |
| minimal | I'm wondering if their the recent logging changes have made this visible but it always silently happened | 11:49 |
| meena | Ah, i didn't consider that | 11:53 |
| dch | meena: why do you have `{{ prefix }}` in the ds-identify rc.d template? | 13:57 |
| dch | does that get populated by jinja at some point pre packaging? | 13:58 |
| meena | dch: it gets rendered when you package it. | 13:58 |
| dch | good good | 13:58 |
| meena | dch: i test like this: python3 ./setup.py install --distro=freebsd --prefix=/usr/local --init-system=sysvinit_freebsd --root=/tmp/tmp.p03232blah | 13:58 |
| meena | shellcheck refuses to check a jinja templated shellscripts! | 13:59 |
| meena | dch: anyway, i updated the PR if you wanna test it | 14:00 |
| dch | i see theres a = now thanks! | 14:00 |
| meena | 23.3-65-g89f0f8978 would be the version | 14:00 |
| meena | you'd have to replace canonical with igalic in the makefile | 14:00 |
| dch | thanks you linked me a patch already very kindly | 14:00 |
| meena | i'll link you a usable patch with make makesum, too | 14:02 |
| dch | I finally decided I needed to write up the whole phoenix endpoint -> page render path for somebody | 14:04 |
| dch | like 5 years overdue | 14:04 |
| meena | dch: https://codeberg.org/meena/freebsd-ports/commit/dd907ac19af4472bc49178fa8e7a7e443b7e4fe7 | 14:13 |
| -ubottu:#cloud-init- Commit dd907ac in meena/freebsd-ports "net/cloud-init-devel: bump" | 14:13 | |
| meena | adding .patch on Codeberg works just like on github: https://codeberg.org/meena/freebsd-ports/commit/dd907ac19af4472bc49178fa8e7a7e443b7e4fe7.patch | 14:13 |
| meena | i need a nap before i go pickup my daughter | 14:14 |
| meena | dch: notice the += in PYDISTUTILS_INSTALLARGS+="--init-system=sysvinit_freebsd" | 14:17 |
| meena | the preexisting INSTALLARGS are the ones passing things like --prefix=${PREFIX} | 14:17 |
| dch | meena not much luck here atm on the dsidentify hack | 20:52 |
| meena | a shite | 20:52 |
| dch | https://git.sr.ht/~dch/ports/tree/main/item/net/cloud-init/files/patch-sysvinit_freebsd_dsidentify.tmpl | 20:52 |
| dch | on top of net/cloudinit though | 20:53 |
| dch | if that matters | 20:53 |
| meena | dch: what's ds-identify's log look like? | 20:55 |
| dch | https://www.irccloud.com/pastebin/QyLJfp5c/dsidentify.log | 20:56 |
| dch | started via `service dsidentify start` | 20:56 |
| dch | aah it doesnt pick up the env vars | 20:58 |
| dch | https://www.irccloud.com/pastebin/E4qVgPBO/env | 20:58 |
| dch | because cloudinit_env should be dsidentify_env I think | 20:58 |
| meena | do you wanna try? | 20:59 |
| dch | https://www.irccloud.com/pastebin/DfaQgtQY/boom | 20:59 |
| dch | that works | 20:59 |
| dch | I skipped the rest of the log | 20:59 |
| meena | dch: I'm mildly surprised, i think | 21:00 |
| dch | I would for sanity's sake make all of cloudinit_dsidentify_ ... just dsidentify_ | 21:01 |
| dch | I'm only running the dsidentify service, not the whole suits | 21:01 |
| dch | *suite | 21:01 |
| meena | hrmm | 21:01 |
| dch | https://git.sr.ht/~dch/ports/tree/main/item/net/cloud-init/files/patch-sysvinit_freebsd_dsidentify.tmpl | 21:02 |
| dch | but yes I share your general sense of surprise | 21:03 |
| meena | I'll add you as co-author | 21:03 |
| dch | share the blame, share the pain | 21:03 |
| meena | heh | 21:03 |
| dch | I reiterate, again, that sysvinit scripts should die in a fire, and like a phoenix from the ashes, freedesktop.org unit files should replace them | 21:04 |
| dch | just for ports | 21:04 |
| dch | leave my pid1 alone thanks | 21:04 |
| dch | so (1) I like this approach better than my *cough* dirty hack | 21:04 |
| dch | but (2) these vars, I'm not sure whether they all work like we expect | 21:05 |
| meena | what email do I use for you? | 21:05 |
| dch | dch@skunkwerks.at is fine | 21:05 |
| dch | they're all fine | 21:05 |
| dch | meena: let me give this 1 more whirl in a more-or-less fresh boot | 21:07 |
| dch | before we commit ourselves to immortalised infamy | 21:07 |
| meena | thank you for the help | 21:08 |
| meena | a proper fix would still be to template ds-identify during the installation process | 21:10 |
| dch | https://www.irccloud.com/pastebin/o9e5GO85/blargh | 21:12 |
| dch | https://www.irccloud.com/pastebin/5sX1DgyG/happy_dsidentify.log | 21:12 |
| dch | so, with that hack/patch, dsidentify runs correctly | 21:12 |
| dch | but the blargh paste above looks like network takes too long to come up before datasource bombs out | 21:13 |
| dch | that is more of a "me problem" than a cloudinit one | 21:13 |
| meena | can you give it a 👍 | 21:18 |
| meena | RFE https://github.com/canonical/cloud-init/issues/4497 | 21:24 |
| -ubottu:#cloud-init- Issue 4497 in canonical/cloud-init "[enhancement]: render ds-identify through template renderer" [Open] | 21:24 | |
| dch | ok, just finished a half-way decent test, this looks great | 21:29 |
| meena | dch: please give it an approving review: https://github.com/canonical/cloud-init/pull/4485 | 21:32 |
| -ubottu:#cloud-init- Pull 4485 in canonical/cloud-init "ds-identify: FreeBSD fixes" [Open] | 21:32 | |
| dch | already done | 21:37 |
| dch | its a lukewarm comment but definitely a +1 as WORKS ON MY MACHINE | 21:37 |
| * dch kicks a build and goes to sleep | 21:41 | |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!