00:26:02  * node-ghjoined
00:26:02  * node-ghpart
00:26:31  * node-ghjoined
00:26:31  * node-ghpart
00:26:46  * node-ghjoined
00:26:47  * node-ghpart
00:53:04  * node-ghjoined
00:53:04  * node-ghpart
02:01:42  * Steve_p_joined
02:01:45  * Steve_pquit (Ping timeout: 268 seconds)
02:04:02  <joyee>test-requireio-osx1010-x64-1 is having trouble fetching origin `Connection to github.com closed by remote host.`
02:05:53  <joyee>I see it's offline now
02:06:48  <joyee>https://github.com/nodejs/build/issues/1563
02:07:43  <joyee>Rackspace windows machines are having trouble `warning: failed to remove Release/: Directory not empty`
02:09:25  <joyee>I will wipe the workspace of `test-rackspace-win2008r2-x64-3` (and 4,5,6)
02:10:42  <joyee>(Oh wait looks like if you wipe they all get wiped..)
02:15:54  <joyee>`test-packetnet-centos7-arm64-1` is having trouble with `which: no node`, `test-scaleway-ubuntu1604-armv7l-3` is `mkdir: cannot create directory ‘out/junit’: No such file or directory`: both taken offline in https://github.com/nodejs/node/pull/24103/#issuecomment-435942156
02:29:43  <refack>joyee: that a post mortem?
02:30:17  <refack>Since at least the scaleway arm and the macOS1010 we're handled
02:30:24  <joyee>refack: they are marked offline but I am not sure if the issues are fixed?
02:30:30  <joyee>Oh, good, thanks
02:31:39  <joyee>Do the workers depend on an global node readily available?
02:32:06  <refack>No, on the contrary, if they find node they fail
02:32:41  <joyee>Then this looks a bit odd..https://ci.nodejs.org/job/node-test-commit-arm/nodes=centos7-arm64-gcc6/19703/console
02:32:51  <refack>I think the packetnet arm had truble yesterday
02:33:23  <joyee>on a side note, wow `make run-ci -j 50`
02:33:45  <refack>ERROR: --without-ssl is incompatible with --openssl-no-asm
02:34:02  <refack>I run a special test job without ssl
02:34:26  <refack>It's an actual ARM clustet
02:34:34  <refack>It actually has 64 cores
02:34:36  <joyee>But this was triggerd from node-test-pull-request?
02:36:15  <refack>Jenkins is making assumptions. I use `rebuild` to start https://ci.nodejs.org/job/node-test-commit/22899/
02:36:23  <joyee>um, wait, it's running the node -p "process.versions.openssl"
02:36:33  <refack>Yeah
02:37:22  <refack>I do still need to look at the windows rackspaces... But in a few...
02:37:28  <joyee>But there is no guarantee that "process.versions.openssl" matches what you are going to build..
02:37:48  * Steve_p_quit (Ping timeout: 252 seconds)
02:38:33  <refack>Only for testing, it is output by the binary you are testing
02:38:52  <joyee>Wasn't that for testing npm install?
02:39:29  <refack>No, npm install doesn't work if your binary doesn't have crypto
02:39:45  <joyee>The testing does not really need to build docs in the first place
02:39:48  <refack>So you need to skip any npm install
02:39:57  <joyee>So the shell call should probably not defined
02:40:05  <refack>Yeah, it's a sanity test
02:40:30  <joyee>No I mean..it should be "run", not defined in the global scope
02:41:08  <joyee>Otherwise whenever you run `make whatever` it gets run
02:41:59  <refack>Well we use it for two things. Sanity test the doc making
02:42:09  <refack> And actually build the docs
02:42:30  <refack>For (1) we should use $(NODE)
02:42:30  <joyee>But if you are not even doing anything related to doc..
02:42:40  <joyee>why would you need node?
02:42:59  <joyee>like you don't need node to make clean
02:43:09  <refack>It's a tool, like the benchmarks we don't want it to rot
02:43:36  <refack>The docs are built with an npm package
02:44:09  <refack>So for sanity you want the fresh binary
02:44:25  <refack>To actually build the docs, you don't care
02:45:29  <refack>Digfooding/selfhosting/bootdtrating is not trivial
02:48:21  <refack>Want a real headache. Think about how you built node with gyp.js?
02:48:52  <joyee>https://www.irccloud.com/pastebin/o8LG0C6a/
02:49:08  <joyee>Not I am talking about this
02:49:58  <refack>Yeah
02:50:05  * node-ghjoined
02:50:05  * node-ghpart
02:50:56  <refack>If you think of it as part of the test, you skip is no crypto
02:51:16  <refack>*if
02:51:55  <refack>Jest test that is works, and throw away the output
02:52:14  <joyee>the workers don't run doc tests right? only the linter job
02:52:26  <refack>They all do
02:52:34  <refack>The linter just lints
02:52:43  <refack>It doesn't compile
02:52:59  <refack>(we compile the MD to HTML)
02:53:36  <joyee>right..test-ci: doc-only
02:53:39  <refack>And do some YAML processing, and do cross reference between the docs and the code
02:54:14  <refack>So we test that those scripts don't brack
02:54:46  <refack>Sort of like CITGM
03:07:56  * Steve_pjoined
03:28:16  <joyee>Anyways I opened https://github.com/nodejs/node/pull/24115
03:28:53  <joyee>So that if you don't need to build doc locally, you don't attempt to the bad build of node
03:29:30  <joyee>Which, come to think of it, is why we need to run that to test for openssl in the first place - because you need to build the docs and that depends on openssl in a strange way
03:35:52  <joyee>Right...I don't think gyp.js is going to work for building node, but not a bad idea to bundle it in node-gyp?
03:36:39  * Steve_pquit (Ping timeout: 252 seconds)
03:38:25  <refack>Fador made gyp.js embeded with duc-type, so you get a single .c file that any compiler can compile trivialy.
03:39:02  <refack>(I love stuff that comes in a single file, and just work)
03:42:21  <refack>Ahhhhh, I see what you mean about always running it.
03:42:42  <refack>Isn't that what := is for?
03:42:51  <refack>Or is that the opposite
03:43:09  <refack>NINJA ALL THE THINGS
03:53:07  <joyee>Does gyp.js only support ninja?
03:53:37  <refack>ATM
03:54:01  * Steve_pjoined
03:54:17  <refack>It also has a ninja compatible engine in js
03:54:28  <joyee>I've never actually figured out where ninja is available or how to install it, I just use the one from depot_tools
03:55:52  <refack>It's available in the GitHub repo and via package managers like apt and yum
03:56:27  <joyee>Yeah that's what I imagined..
03:56:39  <refack>It's also small and can bootstrap itself. Building is 1 minute
03:57:51  <refack>It built on linuxONE no problems
03:57:55  <joyee>oh, you mean it can built iself?
03:58:08  <refack>Yes
03:58:20  * Steve_pquit (Ping timeout: 252 seconds)
03:58:23  <refack>Even on Windows
03:59:03  <refack>Just a little python
03:59:20  <refack>To start things
04:00:34  <refack>I actually wanted to Port it to python, and embed it into GYP, so we could stop using make
04:01:04  <joyee>You mean core?
04:01:16  <refack>Yes
04:01:20  <joyee>The Makefile we have is actually sort of a script
04:01:50  <refack>The generated makefiles are horible
04:02:54  <joyee>Oh you mean, instead of `make -C out`, we `ninja -C out`?
04:03:10  <joyee>But the top level Makefile is probably still going to be there
04:03:26  <joyee>To call `ninja -C out`
04:03:47  <refack>Yeah, like you said, it's just a bunch of scripts with names
04:04:10  <refack>Like a big bash switch
04:04:28  <refack>Almost all the targets are PHONEY
04:04:54  <joyee>Well it does do dependency detection well enough, that's the part pure script does not provide
04:05:23  <joyee>Although some of the dependency is tracked by git
04:05:38  <refack>Just a little bit. I think < 10%
04:06:04  <joyee>Yeah most of them are still PHONY
04:06:22  <refack>It's ok as a frontend...
04:06:58  * Steve_pjoined
04:07:12  <joyee>It's probably not terribly bad to convert the PHONY ones into tools/something.py or tools/something.sh
04:07:47  <joyee>tools/something.sh should be trivial to port anyway
04:10:22  <refack>I saw this pattern somewhere to actually do a bash script with a big switch($1)
04:10:42  <refack>Looks nice, will work in Windows too
04:11:19  <refack>Anyway I'm off, big day tomorrow. America is voting and I'm turning 40
04:11:38  <refack>🦕
05:10:09  * node-slack-botpart
05:10:33  * node-slack-botjoined
05:18:24  * Steve_pquit (Ping timeout: 252 seconds)
05:25:21  * Steve_pjoined
07:52:24  * Steve_pquit (Ping timeout: 252 seconds)
08:08:22  * Steve_pjoined
08:12:45  * Steve_pquit (Ping timeout: 252 seconds)
08:29:01  * Steve_pjoined
08:33:39  * Steve_pquit (Ping timeout: 252 seconds)
08:49:46  * Steve_pjoined
08:54:55  * Steve_pquit (Ping timeout: 268 seconds)
09:10:17  * Steve_pjoined
09:14:38  * Steve_pquit (Ping timeout: 245 seconds)
09:30:58  * Steve_pjoined
09:35:07  * Steve_pquit (Ping timeout: 240 seconds)
09:51:31  * Steve_pjoined
09:56:40  * Steve_pquit (Ping timeout: 272 seconds)
10:12:11  * Steve_pjoined
10:16:30  * Steve_pquit (Ping timeout: 252 seconds)
10:32:51  * Steve_pjoined
10:37:24  * Steve_pquit (Ping timeout: 252 seconds)
10:43:24  * sxajoined
10:44:14  * sxaquit (Client Quit)
10:53:26  * Steve_pjoined
10:58:06  * Steve_pquit (Ping timeout: 272 seconds)
11:05:56  * Steve_pjoined
11:24:09  * Steve_pquit (Ping timeout: 268 seconds)
11:53:56  * Steve_pjoined
12:20:53  * Steve_pquit (Ping timeout: 245 seconds)
12:22:21  * Steve_pjoined
12:35:04  * Steve_pquit (Ping timeout: 268 seconds)
12:58:59  * lucalanziani_quit (Quit: ZNC 1.7.0 - https://znc.in)
13:03:33  * lucalanzianijoined
22:50:03  * node-ghjoined
22:50:03  * node-ghpart
22:50:41  * Steve_pjoined
23:03:22  * Steve_pquit (Read error: Connection reset by peer)
23:03:48  * Steve_pjoined
23:38:41  * node-ghjoined
23:38:41  * node-ghpart
23:49:11  * Steve_pquit (Ping timeout: 260 seconds)