[00:53] emulation is slooow [02:11] /me goes eod real soon now [02:14] axw, babbageclunk or anastasiamac_: https://github.com/juju/juju/pull/6794 [02:15] * anastasiamac_ looking [02:15] looking [02:17] not looking [02:18] thumper: what's the gcwriter thing? doesn't seem related [02:18] * anastasiamac_ has a feelling that babbageclunk will miss out on magic \o/ [02:18] huge reduction in test runs? magic! [02:18] axw: it isn't entirly related [02:18] but I did it while tracing the timings [02:18] seemed useful, [02:18] oh man, is it like a fun surprise PR? almost tempted to click. [02:18] so left it in [02:19] thumper: it's going to route all command output to test output? I think I'd prefer that was opt-in, might lead to a lot of spam [02:19] i'd say anything that imporves readibility of our logs and speeds up diagnosis is useful - happy for u to leave it. +1 from me [02:19] hmm... [02:20] if it's not that spammy, leave it [02:20] when we get the test output, I think it is because something is broke [02:20] the more detail when that happens IMO is useful [02:20] fair enough [02:21] * anastasiamac_ wonders if we can abandon 1.25.x since we can't land [02:21] yeah - nobody uses that! :-P [02:22] thanks [02:22] blahdeblah: :D [02:22] install base, schminstall base [02:23] * thumper hopes this fixes some of the arm64 issues [02:23] I have a feeling though that we'll need to target other tests too [02:24] I almost feel like running the entire test suite with a modified cert.NewCA function that writes to a local file with the call stack of every call site. [02:24] Just to see how many times we call it [02:24] to identify the test suites... [02:24] * thumper wonders how hard that would be to coble together quickly [02:25] probably not too hard... [02:25] * thumper attacks [02:29] * thumper looks for a quick file append call [02:32] thumper: os.Open(filename, os.O_WRONLY | O_APPEND | O_CREATE, 0600) -ish ? [02:32] yeah, had got to that [02:32] natefinch: but thanks [02:32] :) no helper AFAIK [03:05] still building:| I'll see if it works tomorrow [03:05] later #juju-dev [03:05] redir: o/ [07:08] Bug #1656188 opened: metric-collect error when deploying charm === mup_ is now known as mup === jamespag` is now known as jamespage [13:33] Bug #1656304 opened: juju2/maas2: nodes allocation fails if any bridge created in MAAS

[16:47] katco: do you still that data point? the deployment last night blew my network and then I was absorbed by a hangout so I did not dare try again but I have some time now (sorry for not mailing this yesterday) [16:47] perrito666: hey thanks for checking in; no i'm all good [16:47] perrito666: tyvm for the attempt [16:48] katco: I am curious, do you think the bootstrap is what killed my net or was just a coincidence? [16:49] perrito666: frobware could better answer that question; i don't know =/ [16:49] lol [16:49] perrito666: i am only just touching this branch [16:49] perrito666: but i can say i haven't had network problems [16:49] katco: perhaps just an unhappy coincidence [16:50] perrito666: sorry if i caused you any issues! [16:51] nah, just restarted the connection [16:51] I start every juju test with the assumption my machine is going down, something that stayed in my head since the times of local provider :p [16:52] * katco cries [16:52] katco: did I just bring back bad juju memories? [16:52] perrito666: unfortunately not memories. i just hate that our test suite does what it does [16:53] katco: oh, I was talking about testing by hand, I dont think our test suite ever killed my computer [16:53] it killed mine a couple times. [16:53] perrito666: i've had our test suite muck with nics [16:53] natefinch: oh yeah :) [16:53] perrito666: and it compiles things, and other weird weird stuff [16:54] hey folks [16:54] mattyw: hey, long time! [16:54] it starts mongo. I mean, geez [16:54] I am pretty sure that running the suite with sudo would do some interesting wrecking [16:54] katco, hey hey, happy new year [16:54] mattyw: hehey [16:54] natefinch: and further just starts it by running the binary =| [16:54] perrito666, hello there, happy new year to you too [16:55] mattyw: happy new year [16:55] so I'm juju bootstrapping, and hoping my local jujud to be uploaded, and it seems to not be "just doing the right thing" [16:56] mattyw: you can use --build-agent flag [16:56] katco, I can do, do we still need to do that? [16:56] if you want to be sure it uses your local binary, yes [16:56] mattyw: if it's not building the agent and you want it to, yes ;) [16:57] I was looking for that flag but didn't see it in the help [16:57] mattyw: generally it does the right thing, but the rules can be a bit opaque, and the outcome can change depending on what is in streams [16:57] is it secret, or have I forgotten how to read? [16:57] based on state far outside your control or knowledge, sometimes it'll do that anyway (based on your local client version and the versions in streams)... but if you want to be sure, then use --build-agent [16:58] mattyw: juju bootstrap -h |grep build-agent |wc -l [16:58] yeah, it's in my help [16:59] katco, not sure how I feel about there being a command to prove I'm wrong :) [16:59] natefinch, I see it now thanks [16:59] natefinch, I didn't know what I was looking for last time I looked :) [16:59] mattyw: rofl; few things in life are so black and white. i enjoy such certainty! ;) [17:02] katco, natefinch thanks for the help [17:02] mattyw: thanks for stopping by. good to hear from you [17:03] katco, I'm always, here, just normally like hiding in the corner [17:03] :) === akhavr1 is now known as akhavr === akhavr1 is now known as akhavr [22:33] bbiab [23:37] Woohoo ARM64 bootstrapped!