=== matsubara is now known as matsubara-lunch [12:11] * frankban lunches [12:12] ugh [12:12] if we could actually land Makyo's zoom-to-service branch that would be fab. uistage still has silly people [12:53] review request bribing time: https://codereview.appspot.com/10140047 please orangesquad or hatch who's not here yet or anyone else that loves tabs in the browser. === teknico1 is now known as tekNico [13:13] gary_poster: thanks for assigning that card to me, I got distracted [13:13] tekNico, np :-) [13:25] rick_h__: looking. [13:26] gary_poster: I'm assuming that adding the AGPL to the charm means adding the license file itself, and the header to all source files. Anything else? [13:27] tekNico, there's a Canonical-specific list to follow. that might be everything. Let me try to find it again... [13:29] tekNico, see https://sites.google.com/a/canonical.com/legal-services/software-ip [13:29] gary_poster: thanks [13:47] Makyo, hi. If you are around and can make it, I scheduled you on Tuesdays for the Juju core kanban review (check your calendar). If you can't make it today because of the short notice, np [13:50] frankban, hi. I saw that jcsackett's branch failed in jenkins only on the two test_charm_deploy tests. It doesn't look like the error tells us much, but do you have any deep suggestions? [13:51] non-deep suggestions: jcsackett runs the tests locally and sees how they do :-) [13:51] gary_poster: they pass. [13:51] jcsackett, that's the charm deploy tests too? [13:51] I figured the unit tests did :-) [13:52] gary_poster that's test-misc, right? [13:52] uh. I don't *think* so. looking. [13:52] jcsackett: no, that's bin/test-charm [13:53] frankban: ah. [13:53] yeah, that's what I thought :-) [13:54] jcsackett: instructions on how to run local ci tests are in docs/continuous-integration.rst [13:54] frankban, gary_poster: Invalid environment 'juju-gui-testing' [13:55] jcsackett: see docs/continuous-integration.rst ^^^ [13:56] jcsackett, yeah frankban gave you the doc to try. FWIW, this is entirely something that jenkins is supposed to catch, and that you didn't need to run beforehand. May I ask this? (1) you try to get the local ci tests running. Assuming you do, (2) verify that they pass for you. Assuming they do, (3) toss it to me and forget about it yourself. [13:56] jcsackett, you may skip to step 3 at any point if you get too fed up :-) === teknico1 is now known as tekNico [13:57] jcsackett, gary_poster: it seems selenium wants to click the deploy button while it's still not visible, taking a look locally seems reasonable. oh... is the charm panel still there? if not, we have to rewrite the deploy tests [13:58] gary_poster: sure, i can poke at them locally. [13:59] thanks jcsackett . frankban charm panel still there I think [13:59] upcoming branch removes it [13:59] rick_h__: still need that review? [13:59] and morning [14:00] hatch: yes please, and morning to you to [14:01] tekNico: sorry I didn't realize that the cookie code was given to you :) [14:02] hatch: yeah, I changed it anyway, can change it some more :-) [14:03] gary_poster: this is probably unrelated, but I am not able to deploy appflower (and other charms) in uistage [14:03] gary_poster: to make sure i am understanding this right, "local" is a bit of misnomer, as i need to run this out to canonistack or ec2? [14:04] * jcsackett is reading docs, as this is all new. [14:05] frankban, it works if you give it a unique name [14:05] hatch: what's the best time for a follow-up hangout for you? [14:05] frankban, some people think it is amusing to drag everything off to the right [14:06] I think they might even have scripted it [14:06] gary_poster: oh :-( [14:06] it's a weird kind of QA :-) [14:08] tekNico: pretty soon, I'm just reviewing ricks' branch first [14:08] hatch: ok, thanks [14:09] rick_h__: can you pull in the hash branch I landed last night and use the parsed hash? [14:09] hatch: it should have it since it landed in trunk [14:09] * rick_h__ goes to check diff... [14:09] I'm just asking because in browser.js you use window.location.hash [14:09] jujugui I just switch uistage to sandbox. Too annoyed at the script kiddies [14:10] yay [14:10] cool [14:10] hatch: heh, yea I've pulled your changes in the routing code, but in the browser subapp I just grab it from the browser. [14:11] :) [14:11] hatch: is the router adding the hash to the req in the app callables? [14:12] hatch: the router updates keep it from losing the hash and re-routing, but didn't see it provide it down into the app somewhere. [14:13] I thought that the hash was available in there [14:13] regardless [14:13] just booting one up to check [14:14] hatch: I don't think it's part of the yui parsed request object passed in [14:17] hmm looks like you are right, maybe you should use nsRouter.parse(url).hash [14:17] not entirely sure it's necessary however [14:18] hatch: yea, the goal was just to not have the router blow the hash away from window.location.hash. I don't see any harm in me reading it from there in the actual request handling bit. [14:18] yeah [14:18] hatch: it's r/o [14:18] r/o ? [14:18] read-only [14:18] orangesquad: could you please review https://code.launchpad.net/~abentley/charmworld/missing-content-type/+merge/168672 ? [14:19] abentley: looking [14:19] abentley: r=me [14:20] rick_h__: Thanks! [14:23] rick_h__: review done [14:24] hatch: reply inbound [14:25] jcsackett: if you get a second can you peek at my reply/update for the second +1 please [14:26] rick_h__: cool [14:30] tekNico: guichat? [14:30] hatch: I'm doing a review myself now :-) a few minutes and I'm done, ok [14:31] okee sure np [14:31] hatch: don't start another review yourself though ;-) [14:31] haha nope [14:32] hatch: do you have a second to chat? [14:32] lets do it [14:32] guichat [14:32] }"?cccccc [14:32] lol wth [14:32] dog knocked over my computer, yubikey hit couch. [14:32] haha [14:33] hey, this is better than it was a week ago; then i would just accidentally whack it myself. [14:35] orangesquad, hey. hazmat pointed out that if you go to http://uistage.jujucharms.com:8080/fullscreen/ and click on a category on right you get an error [14:35] known? [14:35] known fix? [14:36] gary_poster: looking. We had another bug with that fixed. Looks like a new one was hiding underneath [14:36] thx [14:37] gary_poster: filing a bug now and will get it on the board to look at. [14:37] rick_h__: The URL is wacky-- it has ? where it should have & [14:38] abentley: yea, and series is in there several times [14:38] abentley: will look into it right now as I've not gotten into the other card much yet, landing yesterday's stuff [14:41] hatch: hangout? https://plus.google.com/hangouts/_/02bb45411739e441fe107c9f66e2a8cc36ba4ba7?authuser=0&hl=en [14:41] tekNico: sorry in a hangout now :) [14:42] oh wow :-) [14:42] frankban: so, i've read through the docs, i'm unsure how to debug this test locally. how does one do that? [14:43] we need to update the CI docs with instructions on how to do it locally from start to finish [14:46] tekNico: guichat? [14:46] hatch: yep! [14:47] gary_poster: i think i may be jumping to option 3 here, as i cannot see how to do this locally on my machine, and if local actually means "running on ec2" i don't have any of that setup since having to reinstall and probably don't have the time/bandwidth to get that set up today. [14:48] jcsackett, ack but on call [14:48] dig. [14:48] jcsackett: I suggest configuring the juju-gui-testing env with the ec2 provider and then running NO_DESTROY=1 JUJU_GUI_TEST_BROWSERS="local-firefox" bin/test-charm. [14:49] jcsackett: later you can repeat the tests without boostrapping the env again with APP_URL="http://ec2-xxx-yyy.example.com" bin/test-charm [14:49] this way the tests will be run in your local firefox, and you can debug/pdb as usual [14:53] rick_h__: To properly fix bug #1176452, we think we need to change the JSON stored by mongo and elasticsearch. That will change the json emitted by the original API (e.g. http://staging.jujucharms.com/charms/precise/mysql/json). Do we know of anything still using that API? [14:53] <_mup_> Bug #1176452: ElasticHttpError:NumberFormatException[For input string: "yes" [14:53] abentley: looking [14:53] abentley: yes, until the branch that removes the old browser lands it's using and calling that and might throw errors since it auto loads results on init [14:54] abentley: we should chat with sinzui about getting the old browser code out of there soon now that we're landing browser by default. [14:54] sinzui, rick_h__: I'm up for a chat anytime. [14:55] gary_poster: you had mentioned there was a branch for removing the old browser already? Or did I mis-read that? [14:55] rick_h__, misread [14:55] gary_poster: ok, cool. [14:56] abentley, jcsackett and are about to have our our weekly call. We will be done about 30 minutes from now [14:56] sinzui: rgr [14:56] sinzui: Cool. [15:07] jcsackett, are you trying frankban's instructions or do you still want step 3? [15:08] I don't think he had juju environments and ec2 set up [15:08] and I wasn't sure how to run it all locally [15:25] jujugui ci is failing. could someone volunteer to investigate as critical task please? [15:27] gary_poster: sure [15:27] if the alert can wait :) [15:28] hatch alert is high priority too. frankban do you have half hour to investigate first? [15:28] or tekNico [15:28] gary_poster: I'm juggling two things already... [15:29] tekNico, then not you :-) [15:29] tekNico, what's your non-GPL thing? [15:30] gary_poster: the cookies warning, hatch just revolutionized it :-) [15:30] tekNico, heh oh ok [15:30] (is that a verb?) [15:30] it can be :) [15:30] uh, not sure [15:30] yeah [15:30] it is [15:30] ok, thanks :-) [15:31] the rules of English were made to be broken [15:31] hatch revolutionized the cookies industry! [15:31] lol [15:31] we're Americans. We like verbing things. [15:31] and forgetting u's [15:31] ie) colour [15:31] yeah. Canadians do that too, yeah? [15:31] nope we love our u's [15:32] oh. stinky u's. :-) [15:32] favour [15:32] lol [15:34] (actually three things, I'm also qa'ing a branch by frankban) [15:36] is there a UX bug for the width of the sidebar? [15:37] ^ luca__ [15:37] gary_poster: sorry, was OTP. looks like you've already implemented step 3? [15:37] gary_poster: taking a look with local firefox. [15:38] jcsackett, frankban is looking. maybe worth working with his instructions for the future though? [15:38] thank you frankban [15:38] rick_h__: is the browser sidebar set to min-width 350px on purpose? It should probably be max-width? [15:39] hatch: what is the bug with the side bar? [15:39] hatch: yea, it's on purpose [15:39] gary_poster: i've copied his instructions and when i next have slack i'll make sure i can follow them. [15:39] luca__: oh - well it looks 'wrong' on my monitor because it has a min-width of 350px not a max-width [15:39] ack thanks jcsackett [15:40] so there is just a bunch of whitespace beside everything that looks off [15:40] hatch: screenshot? [15:40] one moment [15:41] hatch: screenshot! :D [15:41] lol [15:41] https://www.evernote.com/shard/s219/sh/073fa338-306e-4fca-bb80-8f954e549add/f6d7b3c935bd8b66934bf7506262f081/res/e4ba1ae1-0155-4c09-af8e-5b15dff3b7a7/skitch.png [15:41] ^ rick_h__ luca__ [15:41] hatch: heh, well don't view it fullscreen on 1920px :P [15:42] hatch: there's talk about making things more responsive and dealing with 'grid units' but it's not in the immediate future [15:42] hah I'm nowhere near fullscreen on my monitor :) [15:42] rick_h__: can you explain what the issue is? o.O [15:42] it looks proper if we change the min-width to max-width [15:42] hatch: well I did say 1920px [15:43] rogpeppe, gentle reminder to reply to the two gui-related emails from william :-) [15:43] well...really just 'width' :) [15:43] luca__: so he's complaining about the wasted space to the right of the charm-tokens because it's seto to 25% and he's viewing it on a giant screen [15:43] that [15:44] rick_h__: I see, good spot hatch :P [15:45] gary_poster: will do, thanks for the poke [15:45] if you're ok with it being set to a fixed width of 350px instead of a min width it's a one line and 4character fix :) [15:45] thanks rogpeppe :-) [15:48] are failures in the simulator not supposed to trigger alerts? [15:57] jujugui call in 3 ;) [15:57] ugh thanks hatch [15:57] haha np [16:06] ---> https://codereview.appspot.com/10048043/ <--- [16:14] frankban: link? [16:14] jcsackett: https://codereview.appspot.com/9714047 [16:15] orangesquad: Could you please review https://code.launchpad.net/~abentley/charmworld/factory-uses-process-charm/+merge/168724 ? [16:15] * sinzui looks [16:17] Makyo: don't you have two reviews already for your "zoom to service" branch? [16:17] abentley, r=me [16:17] tekNico, no second LGTM, and Ben's out this week. [16:17] Makyo: ok, I'll do it [16:18] tekNico, thanks! [16:29] guihelp: does anyone have info about the "failed to load editorial content" notification? [16:30] rick_h__, ^^^ [16:35] frankban: I could be wrong with this - but I think that happens when the request to the jujucharms.com fails [16:36] hatch: ok thanks [16:39] frankban: looking [16:40] frankban: yea, I'll need the network info from the debug tools. Right now we know we've got a problem with the new routing code and searches which use query strings [16:40] frankban: I'm not aware of anything ootb right now on interesting loading [16:42] rick_h__: ack, just investigating CI failures, and that notification captured my attention [16:42] hatch: so yea, but the question would be what failed and why [16:43] frankban: definitely. If it's just in tests then it should be ok, but if you see it in use we need to figure out why it's failing. [16:43] rick_h__: there is no why, there is only do [16:43] *hatchisms* [16:43] hah [16:47] rick_h__: I am actually seeing those in a ec2 deployed GUI (used by CI tests) -> http://ec2-107-20-106-2.compute-1.amazonaws.com/ [16:47] frankban: ok, config issue. The url needs to end in a trailing / [16:48] frankban: both config-debug and config-prod have it, not sure where the config is from in the charm where it doesn't have it [16:48] rick_h__: you mean charmworldURL: "https://manage.jujucharms.com" ? [16:48] frankban: rgr [16:48] removing bootstrap from the page actually breaks damn near everything :/ [16:48] frankban: I mean 'roger' or yes [16:49] hatch: yea, why we left it for you :P [16:49] hatch: we took a peek at that back when we first started with design/huw in march [16:49] it's not just the alert button that huw had mentioned but everything [16:50] rick_h__: so the default value in the charm config is invalid. however, since this is something the user can change (providing a customized juju --config) I'd prefer the trailing slash to be handled by the GUI code... thoughts? [16:51] frankban: yes, we can add a bug to check for a /$ and add it if it's not there. [16:51] frankban: I'll add it now [16:51] rick_h__: thank you [16:55] rick_h__: in the meanwhile, I'll quickly fix the default value in the charm [16:58] frankban: thanks [17:44] gary_poster: so I have now extracted and replaced all of the bootstrap styles effecting the alerts button and made steps to gradually move away from bootstrap. But bootstrap still cannot be removed as other parts of the app rely on it [17:44] how far down this rabbit hole should I go? [17:46] frankban: was the CI failure caused by a missing trailing / ? [17:47] hatch: still investigating [17:47] ohh, I saw your commit and wondered if that was the fix [17:47] hatch: it can be related [17:55] hatch: jcsackett have any opinions on this tiny patch? https://codereview.appspot.com/10193043 [17:55] honestly about to self-review, but give you all a chance at it :) [17:55] hmm, actually yea I'm missing a line of removal still. doh! [17:56] combine now handles that, yes? [17:56] hatch: right [17:57] hatch: so it was double adding the qs and cauing issues later down the pipe [17:57] well then....guess you know what I'm going to ask for next :P [17:57] hatch: a pie? [17:57] a test!!! [17:57] weeeee tests! [17:57] hah, thought about it. but I'm testing that code that's no longer there no longer functions [17:58] :/ [17:58] hahaha [17:58] is there a test that checks to make sure the combine works properly? [17:58] hatch: correct, what was missing was a test that navigate had a query string on it, but now that navigate doesn't add a query string that test makes no sense [17:59] and by navigate I mean _navigate [17:59] sure okee [17:59] LGTM [17:59] abentley, I think search is broken/missing from staging? Any insights [17:59] ty [17:59] once you remove the line creating qes [17:59] hatch: yea, that's pushed up. Hit refresh [17:59] qa I mean [17:59] ARG [17:59] lol [17:59] qs [17:59] * sinzui can see that the staging logs are dominated by KeyError: 'branch_deleted' [17:59] log domination ftw! [17:59] ...or not [17:59] rick_h__: if i'm understanding this, you've found a path wherein the QS was preserved, and then parsed out and added again, yes? [18:00] sinzui: Oh, that sounds like a problem. [18:00] jcsackett: no, I found a path that had a url witha valid query string, and then...reappended a query string again [18:00] you just said what i said with different words. or at least what i was intending to say. [18:00] :-P [18:00] if it had removed it, it might have worked :) [18:01] "parsed out" or whatever [18:01] e.g. the nsRouter stuff now preserves QS, the code you deleted parsed the QS from the URL, then appended...but it was already there. [18:01] right. [18:01] anyway, kill it. [18:01] jcsackett: I follow ya, yea. [18:01] i can LGTM if you want, or you can self review. [18:01] jcsackett: LGTM please so I can say I was all official and people looked over my shoulders :) [18:05] sinzui: I believe abel's branch is broken and only works for new charms that have branch_deleted set. [18:05] sinzui: IOW a catch-22 [18:05] abentley, yep [18:06] I reported the bug. easy fix. [18:06] sinzui: Sorry I didn't notice this in the review. [18:06] sinzui: I can take it if you like. Just started on another bug. [18:06] abentley, okay. https://bugs.launchpad.net/charmworld/+bug/1190002 [18:06] <_mup_> Bug #1190002: KeyError branch_deleted [18:07] your choice of defensive coding or update and use the model. [18:07] gary_poster: so heads up, bug on the categories link that hazmat brought up is landing now. Small regression from last night to today so hopefully short lived and no pain. [18:08] cool thank you very much rick_h__ [18:08] next bug! /me is declaring it bug fixing day [18:08] :_) [18:09] gary_poster: did you see my q above? [18:09] hatch, no looking [18:10] hatch, only as far as we need today for header. thank you. if you've done any extra, great, but no need for more [18:10] honestly I'm not exactly sure what needs to be done - this is to change things for the grey header? [18:11] sinzui: Odd, branch_deleted gets set by the lp scan. I don't know how it's missing. [18:11] sorry hatch I meant the notifications thing [18:11] sinzui: Unless the branch *just* landed. [18:12] hatch, huw doesn't want to have to deal with bootstrap AIUI. I think he also wants to have YUI style css control over the notifications panel [18:12] abentley, ~zulcss/oneiric/nagios looks very old [18:12] well that's simply not going to happen without a new UI, the bootstrap styles controlling the notifications is ~150lns [18:13] gary_poster: retrying a CI run now. This is just to test if the problem was the error notification appearing due to an invalid config option in the charm defaults, and preventing the charm panel link to be clicked. If that's not the case, I'll continue tomorrow morning if not solved before. Created a critical card in Maintenance. [18:13] but I separated them out for him so they are easy to modify now at least [18:13] sinzui: No, I mean unless abel's branch *just* landed. Which I can see it didn't. [18:13] thank you frankban [18:14] abentley, right, it was merge by tarmac about 24 hours ago [18:14] abentley, the first error is 2013-06-11 08:58:07 [18:15] frankban I will ask someone else to run with it this afternoon if not fixed [18:15] hatch, uh. [18:15] gary_poster: cool, thanks [18:15] sinzui: But the last error is 2013-06-11 09:00:05 [18:15] hatch, I honestly have no idea what Huw's concern is tbh. I don't see why he can't change the top header without changing the notifications panel [18:16] ah [18:16] rick_h__, do you have any insight into above, maybe? [18:16] gary_poster: yeah - I'll propose my branch so you can see what I have done for him - see if it's what you had in mind [18:16] ok thanks hatch [18:16] sinzui: I think the problem was that the queued data didn't match the new code. [18:17] sinzui: And therefore it's no longer an issue. Which leaves the other problem mysterious. [18:17] hatch, how trivial is it to disable and remove the top search? [18:17] do you know off hand? [18:18] abentley, I see that ingest errors is current and nasty [18:18] umm - if I remember fairly trivial [18:18] not sure what the side effects would ne [18:18] be* [18:18] we could remove the html though really easy :) [18:19] sinzui: Where do you see it? I can't see a recent example in app.log or ingest-errors.log [18:19] abentley, ingest-lp.log [18:20] hatch, heh [18:20] sinzui: AttributeError: 'MongoQueue' object has no attribute 'charms'? Yes, that looks bad. [18:20] we would also need to remove the charms dropdown button [18:20] yes [18:21] one minute I'll look deeper [18:21] thank you [18:22] sinzui: bin/enqueue exhibits the error locally. [18:23] gary_poster: so I'd like to say that it would simply be removing things from the instance creation constructor but should probably allow......4h? JIC [18:23] hatch ok [18:24] hatch, are you up for this or would you like me to find someone else? This is more immediately pressing than the inspector work, but I really want that to proceed as well. If you'd rather not do this right now, lemme know [18:24] sure I can [18:24] abentley, shall we consider the script to remove ObjectId charms? Ot was run on staging 4 hours ago, and production in the last 15 minutes [18:24] this will definitely be a - diff :) [18:25] sinzui: What should we consider about the script? [18:25] :-) thanks hatch [18:26] abentley, I don't think the remove() call could cause the queue to loose charms, but maybe on second glance it is bad...http://pastebin.ubuntu.com/5755316/ [18:26] gary_poster: np! I'm going to move bens controller card into daily to free up space if that's alright? [18:26] abentley, this is the report from the production run...look good to me: https://pastebin.canonical.com/92568/ [18:27] sinzui: I don't think the script is implicated. [18:27] sinzui: I think that ingest is broken, and all the current charms are in a bad state because of the last ingest. [18:27] fab, then production probably wont go blank [18:27] hatch, how 'bout this [18:28] sinzui: For example, http://staging.jujucharms.com/charms/precise/apache2/json has an 'errors' member. I bet all branches do. [18:28] (1) propose your/ben's current branch. use Ben's card for that [18:28] s/branches/charms [18:28] abentley, yes, that is what I thought at first when I saw the branch_deleted, but the [18:28] AttributeError: 'MongoQueue' object has no attribute 'charms' [18:28] (2) hopefully you 'll have reviews for your alerts dropdown by then and can land [18:28] confuses me [18:28] (3) then you'll have room :-) [18:29] sinzui: I get that locally when I run bin/enqueue. I've never run your script. I think abel never tested locally. [18:29] hah sure - `recess` is throwing an error but doesn't tell me what the error is :/ [18:30] abentley, we have two branches found by interfaces...and one is ~chilicuil that we discussed before [18:32] sinzui: Also, the API lists only two charms. [18:32] http://staging.jujucharms.com/api/2/charms [18:33] abentley, yep, those two. We can added a defensive check to the code. [18:34] sinzui: Sorry, I forget what context they were relevant to. Are they potentially deleted? [18:34] hmm [18:34] sinzui: bzr certainly thinks they aren't branches. [18:35] abentley, ~chilicuil is my Cthulhu example form a few days about charms without branches that dominate the error logs [18:36] the recent change inverts the case where good charms error and dominate the logs [18:36] sinzui: It [18:36] sinzui: It is the charm that had no contentType earlier, suggesting a deleted branch. [18:37] yeah [18:38] sinzui: Anyhow, once we fix this MongoQueue error, I expect the non-missing branches to be fixed. [18:41] * hatch wants to remove recess from our build [18:43] sinzui: It appears to be a one-line fix: http://pastebin.ubuntu.com/5755825/ [18:43] :) [18:49] sinzui: Could you please review https://code.launchpad.net/~abentley/charmworld/enqueue-broken/+merge/168788 ? [18:50] sinzui: I've successfully run enque and ingest-queued. [18:50] r=me [18:53] sinzui: I wonder whether we should drain the queues as part of a source code upgrade? [19:00] abentley, I think we should. [19:00] abentley, do you have time for https://code.launchpad.net/~sinzui/charmworld/enable-migration/+merge/168791 [19:00] sinzui: looking. [19:01] sinzui: I think production thinks it's already on version 7, so you need to rename version 7 to version 8. [19:05] abentley, production is not on 7, it failed to migrate [19:05] sinzui: Even after you made the 7 migration a no-op? Are you sure? [19:06] oh! [19:06] abentley, you are correct. [19:06] * sinzui copies and renames [19:09] benji hijacked you into call tomorrow of potential juju core task. let me know if you'd rather have a tooth pulled :-) [19:09] heh [19:09] what time? [19:10] benji 1500 utc. on your calendar [19:10] * benji takes out his UTC to Faux-Eastern slide rule. [19:10] heh, 11 for me benji [19:13] hatch: can you give me a ping when you get a sec? [19:14] I need someone to second my build issue isn't just me [19:14] what's up? just eating lunch [19:14] hatch: so can you check out trunk, make clean-all && make [19:14] and try to run the tests? [19:14] in a browser, test-server, not test-debug [19:14] sure, [19:16]