00:08:36  <refack>We try to, but sometimes it gets overridden. The ubuntu17 is just slow
00:42:52  * node-ghjoined
00:42:52  * node-ghpart
00:44:54  <Trott>tniessen: You can use Rebuild on the arm fanned job rather than Resume Build on the larger job.
01:29:51  * Danibal19joined
01:32:46  * Danibal19quit (Remote host closed the connection)
05:34:15  <Trott>Anyone around who can help with this? https://ci.nodejs.org/job/node-test-commit-linuxone/5199/nodes=rhel72-s390x/console
05:34:30  <Trott>https://www.irccloud.com/pastebin/SgM2ng28/
05:34:56  <Trott>Uh, let's try that pastebin again....
05:34:59  <Trott>https://www.irccloud.com/pastebin/XGKppyLJ/
05:35:03  <Trott>There we go...
05:54:12  * jferrerojoined
05:58:55  * jferreroquit (Ping timeout: 246 seconds)
06:20:34  * m00c0w_joined
06:21:21  * m00c0w_quit (Remote host closed the connection)
06:46:18  <refack>Trott: wrong box https://usercontent.irccloud-cdn.com/file/h487drEJ/Screenshot_20180918-024531.png
07:33:03  * seishunjoined
08:56:18  * seishunquit (Ping timeout: 245 seconds)
09:15:10  * seishunjoined
09:25:28  * seishunquit (Ping timeout: 245 seconds)
09:34:53  * seishunjoined
09:35:26  * sxajoined
09:45:06  * seishunquit (Ping timeout: 252 seconds)
10:04:14  * Guest7087joined
10:07:06  * Guest7087quit (Remote host closed the connection)
10:41:41  * qbitquit (Ping timeout: 244 seconds)
10:58:18  <tniessen>Can I resume node-test-windows-fanned?
10:58:53  <tniessen>And ubuntu1404-32 is still failing really weird:
10:58:55  <tniessen>git: malloc.c:2374: sysmalloc: Assertion `(old_top == (((mbinptr) (((char *) &((av)->bins[((1) - 1) * 2])) - __builtin_offsetof (struct malloc_chunk, fd)))) && old_size == 0) || ((unsigned long) (old_size) >= (unsigned long)((((__builtin_offsetof (struct malloc_chunk, fd_nextsize))+((2 *(sizeof(size_t))) - 1)) & ~((2 *(sizeof(size_t))) - 1))) && ((
10:58:55  <tniessen>old_top)->size & 0x1) && ((unsigned long) old_end & pagemask) == 0)' failed.
10:59:03  <tniessen>Not the first time I see this
11:02:59  * node-ghjoined
11:02:59  * node-ghpart
11:03:06  <tniessen>I opened https://github.com/nodejs/build/issues/1499
11:08:32  * node-ghjoined
11:08:33  * node-ghpart
11:08:55  * qbitjoined
11:09:12  * qbitchanged nick to Guest70267
11:19:37  * silverwhitefishjoined
11:24:23  * silverwhitefishquit (Remote host closed the connection)
11:25:37  * axsuul-joined
11:30:23  * axsuul-quit (Remote host closed the connection)
11:49:15  * jackcjoined
11:50:43  * jackcquit (Remote host closed the connection)
12:30:12  * Guest70267changed nick to qbit
12:33:41  <Trott>refack: Heh, well *that's* embarrassing.
12:59:32  * node-ghjoined
12:59:32  * node-ghpart
13:11:53  <refack>Trott: since we are responsible software engineers, I'd call that a bug in the UX
13:37:03  * node-ghjoined
13:37:03  * node-ghpart
13:44:53  <Trott>Although I wonder how the linter job succeeded....
13:45:38  <refack>linter does not rebase, and uses `master` as default...
13:46:49  <node-slack-bot>[trott] thumbsup emoji
14:03:49  * node-ghjoined
14:03:49  * node-ghpart
14:07:45  * node-ghjoined
14:07:45  * node-ghpart
14:08:00  * node-ghjoined
14:08:00  * node-ghpart
14:17:30  * node-ghjoined
14:17:30  * node-ghpart
14:20:35  * node-ghjoined
14:20:35  * node-ghpart
14:21:07  * node-ghjoined
14:21:07  * node-ghpart
14:24:50  * node-ghjoined
14:24:50  * node-ghpart
14:27:44  * node-ghjoined
14:27:44  * node-ghpart
14:33:07  * node-ghjoined
14:33:07  * node-ghpart
15:08:48  <joyee>Is it me or are the blinking ball icons in the CI job pages new? (I am guessing that means "waiting for workers")
15:11:54  <refack>Yes. We had a plugin called "New status Icons", but I removed it during the CPUpocalipse
15:49:43  * Fishrock123joined
16:41:46  * ProAliasjoined
16:43:52  * ravndal11joined
16:44:33  * ProAliasquit (Remote host closed the connection)
16:49:09  * ravndal11quit (Ping timeout: 252 seconds)
17:18:27  * MrFantastik2joined
17:23:12  * MrFantastik2quit (Remote host closed the connection)
18:01:29  <ryzokuken>refack: https://ci.nodejs.org/job/node-test-commit-linux-containered/nodes=ubuntu1604_sharedlibs_openssl110_x64/7177/console
18:01:32  <ryzokuken>infra fail?
18:01:44  * Fishrock123quit (Remote host closed the connection)
18:02:13  <refack>No, it the `13:57:41 modified: deps/v8/third_party/jinja2/LICENSE` you need to rebase
18:02:41  * Fishrock123joined
18:10:48  <ryzokuken>refack: cool!
18:14:13  <joyee>https://ci.nodejs.org/job/node-cross-compile/18938/nodes=cross-compiler-armv6-gcc-4.9.4/console
18:14:29  <joyee>https://www.irccloud.com/pastebin/mjwgVaj3/
18:15:07  <joyee>hmm interesting...why is this a syntax error?
18:16:10  <joyee>Where can I find the script run by node-cross-compile ?
18:18:22  <refack>https://ci.nodejs.org/job/node-cross-compile/configure
18:18:52  <joyee>thanks
18:19:04  <refack>joyee: my guess `CONFIG_FLAGS` causes a syntex error in your changed Makefile
18:21:54  <joyee>ah, that's a Makefile error? I thought it was a node error..
18:23:08  <refack>BTW: that's going to be a painful change... I don't think rerunning configure and `make all` twice is going to work.
18:23:47  <refack>We need to find a way to just pass the new file and rebuild just `./node`
18:26:12  <joyee>hmm, why?
18:26:43  <refack>`./configure` can't be part of a make target
18:26:58  <refack>The pattern of `./configure && make` should stay valid
18:27:17  <joyee>it's still valid, just without code cache
18:27:55  <joyee>Maybe the change should happen in the Jenkins
18:28:21  <refack>I'll try to read the PR and find a way
18:29:15  <joyee>With the current build files, to build with code cache we'll need to do that `make with-code-cache` does
18:29:52  <joyee>It may be more viable to just get the CI do that 2-pass build step
18:30:51  <refack>I'm now trying to solve https://github.com/nodejs/node/pull/22920 & https://github.com/nodejs/node/issues/22006
18:31:47  <refack>Our build is not stable. running `./configure && make` causes wide rebuild, also `make && make`
18:32:09  <joyee>What did you mean by wide rebuild?
18:32:21  <joyee>Rebuilding from scratch?
18:32:32  <refack>Rebuilding of a few 100 files
18:32:51  <joyee>Does that happen on every platform?
18:32:54  <refack>and relinking of several v8 libs, nodelib, and node
18:33:03  <joyee>I noticed the relinking too
18:33:03  <refack>I think so
18:33:49  <joyee>Is that getting in the way of gyp fixes for embedded builtins?
18:34:01  <refack>I think I can solve the `make && make` issue, but `./configure && make && ./configure && make` is still hard
18:35:11  <refack>It's just if we run `make run-ci -j2` it will fail because we do `./configure && make & make test` so the two parallel makes break each other
18:35:24  <joyee>Maybe just do that in the release CI? (although then we will have a risk of disparity between release builds and testing builds)
18:35:42  * sxaquit (Ping timeout: 272 seconds)
18:36:43  <refack>I'm close to solving it... And if we can do the code-cache in `GYP` like mksnapshot, that will be best, so we only need to run `./configure` once
18:36:46  <joyee>Why do we need to run `make && make test` though? `make test` is already running all?
18:37:10  <refack>BTW that very similar to the embedded builts procedure
18:37:20  <joyee>I've attempted to do `mksnapshot` but to no avail. That needs some refactoring in node.cc
18:37:41  <joyee>(which is still quite messy despite a lot of effort of refactor)
18:38:23  <joyee>in essense, mksnapshot, code cache and embedded builtins all need the same kind of procedures
18:39:09  <joyee>build a binary once, generate something, then embed that into another binary - the first and the second binary may or may not be using the same main code
18:39:59  <joyee>It would be easier if we can refactor node.cc so that we can build a slimmed down version of node compiled with a different main function dedicated to generate those stuff
18:40:49  * ebeyer_joined
18:40:58  <joyee>But I am still not sure how to do the refactoring properly..:/
18:43:45  * ebeyer_quit (Remote host closed the connection)
18:45:18  <refack>I'll look at that after I finish with the embeds. I believe we can do it a bit more simple.
18:47:20  <refack>Then I should really write a doc about our build procedure (v8_libbase -> torque -> v8_base -> mksnapshot -> v8 -> js2c -> nodelib -> node)
18:47:25  <joyee>Speaking of which, should we notify the electron people about that change?
18:47:57  <joyee>or nodejs/deliever-channels in general
18:48:56  <refack>about the embeds, maybe not. About the code-cache, I think so.
18:49:26  <refack>Actually electron are now building with GN, so definatly
18:49:57  <refack>Maybe I'll help them, I'm changing a few gyp files
18:50:07  <richardlau>I left a comment in the PR, but re. https://ci.nodejs.org/job/node-cross-compile/18938/nodes=cross-compiler-armv6-gcc-4.9.4/console is node-msft-cross-compiler-1 an ARM host? Otherwise I suspect that the error is trying to run an ARM binary on the wrong architecture.
18:51:02  <refack>richardlau: also makes sense. target/host toolchains are painful
18:51:37  <richardlau>(The way to tell would be to look at host_arch and target_arch in config.gypi. We used to get configure output in the job output but that was recently suppressed).
18:52:21  <refack>Ohh, thank you I need to fix that.
18:53:29  <refack>but we could still look at `config.gypi` in the workspace browser
18:53:59  <refack>https://www.irccloud.com/pastebin/XOW2PgKb/
18:54:17  <refack>yep
18:55:19  <richardlau>That (workspace browser) only works as long as there hasn't been another run of the job.
18:55:33  <refack>yeah, we were lucky
18:56:08  <richardlau>but yes, host_arch: ia32 suggests that running out/Release/node compiled for armv6 isn't going to work :)
19:04:52  <refack>richardlau: see https://github.com/nodejs/node/pull/22935
19:05:07  <richardlau>refack: already commented on it :)
19:08:14  * Fishrock123quit (Remote host closed the connection)
19:08:32  * Fishrock123joined
19:18:00  * snadge23joined
19:22:45  * snadge23quit (Remote host closed the connection)
20:07:09  <joyee>hmm so how did we manage to do the mksnapshot on those jobs?
20:08:07  <refack>we run them as "host". I'm not sure what comes out but it works
20:08:37  <refack>I'll investigate
20:42:39  <joyee>Looks like ` test-softlayer-ubuntu1404-x64-1` has been keeping a bunch of jobs running over 6hrs?
20:42:52  <joyee>https://ci.nodejs.org/job/node-test-commit-linux/nodes=ubuntu1404-64/
20:43:04  <joyee>ooh yeah...a bunch of jobs waiting on the worker
20:45:39  <joyee>umm, https://ci.nodejs.org/computer/test-digitalocean-ubuntu1404-x64-1/ is idle, while https://ci.nodejs.org/computer/test-azure--msft-ubuntu1404-x64-1 is 404 but it's ssh-able
20:46:21  <joyee>the digitalocean one is waiting on https://ci.nodejs.org/job/node-test-napi-modules-citgm/nodes=ubuntu1404-64/444/ which is also 404..
20:47:44  <joyee>https://ci.nodejs.org/job/node-test-napi-modules-citgm/444/ this has been running for 3 days. I'll kill it first.
20:49:14  <refack>Just what I was about to say
20:49:24  <joyee>Should I restart the azure one?
20:49:35  <joyee>Not sure what happened to that one
20:49:53  * Natchjoined
20:50:07  <refack>we don't have a worker for it
20:50:17  * Natchchanged nick to Guest86307
20:50:27  <joyee>so there are only 2 executor for ubuntu-1404-x64?
20:50:56  <refack>yes
20:51:01  <refack>https://ci.nodejs.org/label/ubuntu1404-64/
20:51:34  <joyee>OK...let's wait until these 2 finish the queue
20:54:31  <refack>the msft one in https://ci.nodejs.org/computer/node-msft-cross-compiler-1/
20:54:39  <refack>*is
20:54:39  * Guest86307quit (Remote host closed the connection)
20:55:18  <joyee>ah, I only grepped my .ssh/config
20:55:25  <joyee>I'll use the labels next time
20:57:49  <refack>I kinda want to rename the node...
20:58:23  <joyee>https://ci.nodejs.org/job/node-test-binary-windows/COMPILED_BY=vs2017-x86,RUNNER=win2012r2,RUN_SUBSET=1/20072/console
20:58:45  <joyee>Looks like https://github.com/nodejs/build/issues/1496 isn't working...?
20:59:45  <refack>yes
21:00:05  <joyee>I'll paste it there
21:00:09  <refack>But maybe it's just a resume of a job from before #1496
21:00:31  <joyee>It was started 6hrs ago
21:01:08  <refack>The resume. The original was https://ci.nodejs.org/job/node-test-pull-request/17261/
21:01:47  <refack>https://ci.nodejs.org/job/node-test-commit-windows-fanned/20792/console
21:01:51  <refack>it did delete
21:01:59  * node-ghjoined
21:01:59  * node-ghpart
21:02:20  <joyee>delete?
21:03:19  <refack>delete the temp branch with the binaries
21:07:10  <joyee>Its jenkins config got left out?
21:09:22  <refack>No, I'm wrong https://ci.nodejs.org/job/git-delete-branches/38608/parameters/ is the got that would have deleted, but it didn't
21:10:28  <refack>ahhh, but the branch is named after the job `jenkins-node-test-commit-windows-fanned-20792` and the job number changes
21:17:00  <joyee>| 2018-09-18 21:00 | 10 | 8 | 19 | 7 | 56 | 9.64% |
21:17:04  <joyee>oh no
21:17:11  <joyee>https://www.irccloud.com/pastebin/gW6o4Rbo/
21:17:31  <refack>Because of the ubstable
21:17:38  <refack>We fixed that
21:20:26  * goofie12joined
21:22:42  <joyee>https://github.com/nodejs/reliability/issues/16
21:22:57  <joyee>Trying to work through other failures..
21:23:20  * goofie12quit (Remote host closed the connection)
21:23:52  * node-ghjoined
21:23:52  * node-ghpart
21:29:17  * node-ghjoined
21:29:17  * node-ghpart
21:53:37  <refack>GYP has a bug/limitation with `make` output. It creates non-compilation actions as `FORCE_DO_CMD`, so if they are part of a `type: none` target, they must run every time
21:54:04  <refack>I can fix the `.gyp` files and put the actions in a "real" target. Or fix GYP?
21:54:09  <refack>What should I do
22:31:49  * Fishrock123quit (Quit: Leaving...)
22:42:42  * qbitquit (Ping timeout: 272 seconds)
23:09:21  * qbitjoined
23:09:40  * qbitchanged nick to Guest30740
23:26:53  * Guest30740changed nick to qbit
23:28:12  * RabbiBob19joined
23:31:06  * RabbiBob19quit (Remote host closed the connection)