00:30:03  * jenkins-monitor1quit (Remote host closed the connection)
00:30:04  * jenkins-monitorquit (Remote host closed the connection)
00:30:12  * jenkins-monitorjoined
00:30:13  * jenkins-monitor1joined
00:54:24  * node-ghjoined
00:54:24  * node-ghpart
01:50:34  * Fishrock123quit (Quit: Leaving...)
04:00:43  * rmgquit (Remote host closed the connection)
04:01:18  * rmgjoined
04:06:12  * rmgquit (Ping timeout: 260 seconds)
04:14:47  <Trott>When I see a line like this in the Jenkins console, can I find the contents of the .sh file in a build repo somewhere?:
04:14:47  <Trott>[node-test-linter] $ /bin/sh -xe /tmp/hudson8963588080613176920.sh
04:34:44  <jbergstroem>No, its temporary
04:34:58  <jbergstroem>its usually how jenkins encapsulates bash we write in a job description
04:56:59  <Trott>So where I see that line in https://ci.nodejs.org/job/node-test-linter/2194/console ... I have no idea what happened there. And the behavior there is sooooo weird. mscdex and I are both puzzled as to why subsequent commands seem to be somehow using the wrong version of Node.js. So I was wondering maybe if it was something in that file? Because I have little
04:56:59  <Trott>else to go on...
05:00:50  <jbergstroem>I'll hvae a look
05:01:02  <jbergstroem>I'd have you debug but I don't trust you enough to give you permissions :-X
05:01:31  <jbergstroem>hm, that's indeed interesting
05:01:48  * rmgjoined
05:02:04  <jbergstroem>we're basically calling lint-ci
05:02:35  <jbergstroem>both machines were running nodejs 5.9.1.
05:02:59  <jbergstroem>updating to 5.11.0
05:03:29  <Trott>But shouldn't it be using the version in the Release directory for the linting job?
05:03:39  <jbergstroem>we don't build a node version in the linter
05:03:43  <jbergstroem>we call the host linter
05:03:48  <jbergstroem>that's why it finishes in <30s
05:03:55  <Trott>Ah!
05:04:10  <jbergstroem>Should we have to?
05:04:21  <Trott>OK, then. We either need to update to 6.0.0 or else I need to update a lint rule so that it will run on Node 5
05:04:31  <Trott>(It's one of our custom lint rules, so we can totally update it.)
05:04:45  <jbergstroem>for node 6 we'll have to call freebsd and say plez then wait for a day or two
05:05:01  <jbergstroem>I think @mscdex's tool needs to exit 1 on fail though
05:06:01  <Trott>From thealphanerd: "This is why we need to run linting and commit linting on separate boxes that have a Node.js environment." Disclaimer: He's been drinking. ("But I remember the conversation from last night. Or was it this morning. I don't know. I've been drinking.")
05:06:04  * rmgquit (Ping timeout: 240 seconds)
05:06:17  <Trott>Yes, there's a bug in mscdex's tool too. It's a double bug or something.
05:06:18  <jbergstroem>hehe
05:06:30  <jbergstroem>me/thealphanerd have been chit chatting too
05:06:34  <jbergstroem>think we're aligned
05:06:52  <jbergstroem>but saying that linters have to be bleeding edge (as in 0day) i'm not sure i 110% support
05:08:40  <jbergstroem>i'm not sure how 'separate' boxes would help herei f you get me; having to compile node $latest wouldn't really help linting time
06:04:32  <thealphanerd>jbergstroem if we always sourced the nightly for the linter we could do some good smoke testing :D
06:27:50  * node-ghjoined
06:27:50  * node-ghpart
06:28:01  * node-ghjoined
06:28:01  * node-ghpart
06:58:40  * othiym23quit (Ping timeout: 264 seconds)
06:59:16  * benglquit (Ping timeout: 264 seconds)
06:59:16  * joaocgreisquit (Ping timeout: 264 seconds)
06:59:16  * bretquit (Ping timeout: 264 seconds)
06:59:17  * phillipjquit (Ping timeout: 264 seconds)
07:01:07  * othiym23joined
07:01:11  * phillipjjoined
07:01:12  * joaocgreisjoined
07:02:31  * rmgjoined
07:03:50  * bengljoined
07:07:09  * rmgquit (Ping timeout: 246 seconds)
07:08:56  * bretjoined
07:30:22  * evanlucasquit (Ping timeout: 260 seconds)
07:52:50  <jbergstroem>thealphanerd: a nightly wouldn't guarantee that the commit we're testing would be in sync thouhg :(
09:03:18  * rmgjoined
09:07:54  * rmgquit (Ping timeout: 246 seconds)
10:35:48  * thealphanerdquit (Quit: farewell for now)
10:36:18  * thealphanerdjoined
13:04:51  * rmgjoined
13:09:31  * rmgquit (Ping timeout: 252 seconds)
13:13:53  * node-ghjoined
13:13:53  * node-ghpart
13:20:14  <joaocgreis>jbergstroem: I like the idea of using something known stable. An idea: there could be a daily job that runs on the lint machines that compiles one of our branches (`v6.x` for some stability?) and installs to a prefix that can be used by `node-test-lint`
13:21:48  <joaocgreis>when we choose where the jobs run we select labels, but there is also the option for individual machines and that job could run on all the linters
13:50:57  * Fishrock123joined
14:41:45  * evanlucasjoined
14:56:12  * stefanmbjoined
15:26:30  * rmgjoined
16:48:36  * sgimenoquit (Quit: Leaving)
18:19:10  <jbergstroem>joaocgreis: we could redeploy on linux and use nvm?
18:19:45  <jbergstroem>i personally am not a fan of the idea that linters should have to bleeding edge, but if thats where everyone wants to be.
18:24:00  <joaocgreis>I think it should be as stable as possible, I would suggest LTS if it can be done
18:25:12  <joaocgreis>I'm OK with nvm or daily job, whichever is less troublesome
20:28:54  <jbergstroem>joaocgreis: i'm happy to do LTS as well. Wouldn't using nvm contradict that though?
20:44:21  * Fishrock123quit (Read error: Connection reset by peer)
20:49:43  * node-ghjoined
20:49:43  * node-ghpart
20:52:24  * Fishrock123joined
21:00:13  * node-ghjoined
21:00:13  * node-ghpart
21:50:38  <joaocgreis>not sure, no experience with nvm. Makes sense to me to keep up with the latest version of LTS though
22:34:06  * Fishrock123quit (Quit: Leaving...)
22:49:36  * stefanmbquit (Quit: My Mac has gone to sleep. ZZZzzz…)
23:17:57  * stefanmbjoined