00:00:00  * ircretaryquit (Remote host closed the connection)
00:00:09  * ircretaryjoined
00:05:30  <trevnorris>okay. no one is ever fucking allowed to add stuff to lib/crypto.js in v0.10 again.
00:05:44  <srl295>jgi: as I understand from trevnorris "small icu" is the default for binary releases but not the default from ./configure
00:05:57  <trevnorris>merging that into v0.12 is a serious CF
00:06:42  <srl295>jgi: I've updated the wiki somewhat. I'll update the wiki again when it lands if that's OK
00:07:19  <srl295>^ trevnorris how did icu_config.gypi end up on .10 ?
00:07:48  <trevnorris>srl295: my fault. was hurriedly trying to get in a few things before EOD and it slipped in.
00:07:58  <trevnorris>the file sticks around after switching from v0.12 to v0.10
00:08:28  <srl295>trevnorris: right. No problem, just thought at first maybe someone back ported it.
00:08:56  <jgi>srl295: thanks for updating the Wiki, it’s more consistent now :) And yes, please update it again when it lands.
00:09:26  <srl295>jgi: welcome. I have made sure the configure --help and README are in sync with, you know, the code.
00:10:06  <srl295>jgi: I'll happily update the wiki as needed
00:12:21  <trevnorris>indutny: you're seriously going to have to double check my v0.10 to v0.12 merge. there are so many changes to crypto on both sides.
00:16:46  <jgi>srl295: cool thanks!
00:19:11  * toothrotjoined
00:21:33  * sblom_joined
00:22:00  * sblomquit (Ping timeout: 264 seconds)
00:31:22  * avalanche123quit (Remote host closed the connection)
00:32:31  * Fishrock123quit (Remote host closed the connection)
00:42:08  * avalanche123joined
00:43:54  * avalanche123quit (Remote host closed the connection)
00:44:19  * avalanche123joined
00:45:47  * avalanche123quit (Remote host closed the connection)
00:49:02  * quijotejoined
00:50:08  * c4miloquit (Remote host closed the connection)
00:50:25  * brsonquit (Quit: leaving)
00:51:54  * sblom_quit (Ping timeout: 245 seconds)
00:53:34  * quijotequit (Ping timeout: 245 seconds)
00:54:34  * sblom_joined
00:57:46  * fourqquit (Quit: leaving)
00:58:34  * avalanche123joined
00:59:00  * abraxas_joined
01:02:31  * jreyno40joined
01:02:51  <trevnorris>indutny: ping
01:02:56  <indutny>pong man
01:03:07  <indutny>trevnorris: wanna me review the merge?
01:03:21  <trevnorris>indutny: i'm trying to merge v0.10 to v0.12 and I don't know which tools/certdata.txt to use.
01:03:27  * jreyno40part
01:03:32  <trevnorris>indutny: they're different, but ben added them at the same time to v0.10 and v0.12.
01:03:43  <indutny>trevnorris: you could just scrap it out
01:03:46  <indutny>and get new certdata.txt
01:03:50  <indutny>do you know how to do it?
01:04:04  <trevnorris>nope. I think there are instructions from the commit message.
01:04:17  * abraxas_quit (Ping timeout: 260 seconds)
01:04:24  <trevnorris>yeah. it's there.
01:05:01  <indutny>trevnorris: ok, this should be the way to go
01:06:24  <trevnorris>okay. thanks
01:07:29  * sblom_quit (Read error: Connection reset by peer)
01:09:16  <trevnorris>jgi: I can't port caeb6773 to v0.12. you'll have to redo that PR.
01:12:05  <indutny>trevnorris: no problem :)
01:17:21  <jgi>trevnorris: alright, does it bock the merge?
01:25:12  * stagasquit (Ping timeout: 250 seconds)
01:37:11  * avalanche123quit (Remote host closed the connection)
01:43:37  * abraxas_joined
01:47:02  <srl295>jgi: will be out for a bit and/or till tomorrow PT. let me know if there are any next questions
01:47:34  <jgi>srl295: alright, I’m still working on setting up Jenkins jobs, I’ll let you know if I have any question, thanks!
01:50:01  * dshaw_joined
01:50:20  * quijotejoined
01:51:31  * avalanche123joined
01:51:55  * dshaw_quit (Client Quit)
01:53:22  * avalanche123quit (Remote host closed the connection)
01:54:39  <jgi>trevnorris: if that’s not a blocker (porting caeb6773), then I’ll create a GitHub issue
01:55:09  * quijotequit (Ping timeout: 260 seconds)
01:56:06  * avalanche123joined
02:04:56  * Fishrock123joined
02:08:13  <trevnorris>jgi: no. i'm just reverting the change.
02:08:27  <trevnorris>i'll have a note in the merge commit message that it was removed.
02:08:34  <trevnorris>and you can just resubmit a PR on the v0.12 branch
02:10:32  <jgi>trevnorris: alright, then I’ll create a GH issue with a proper milestone, otherwise I’ll forget about it
02:11:57  * dap_quit (Quit: Leaving.)
02:16:00  <jgi>trevnorris: I also created a 0.12.1 milestone since 0.11.15 should be the release candidate for 0.12.0
02:16:14  <jgi>trevnorris: and I assigned this isue to the 0.12.1 milestone
02:19:17  <jgi>trevnorris: see https://github.com/joyent/node/issues/8877
02:26:20  <indutny>trevnorris: btw
02:26:31  <indutny>trevnorris: you should probably consider this doing in a separate commit after merge
02:26:36  <indutny>and use v0.12's certdata.txt in the mrge
02:26:40  <indutny>jgi: heya
02:26:46  <indutny>jgi: have you seen my PushArgument patch?
02:26:55  * inolenquit (Quit: Leaving.)
02:30:48  * avalanche123quit (Remote host closed the connection)
02:34:18  <jgi>indutny: hey, yes I saw that, thanks!
02:34:28  <indutny>jgi: coolio
02:34:32  <indutny>jgi: what are your thoughts on this?
02:35:06  <jgi>indutny: I haven’t found the time to review it closely, but I saw that it’s been accepted by the V8 team, and that it fixes an actual issue
02:35:13  <indutny>wait
02:35:17  <indutny>this is a different thing
02:35:21  <indutny>it wasn't accepted by v8
02:35:33  <indutny>but it fixes the actual issue indeed
02:35:49  <indutny>I'm afraid that we won't be able to get any expertise from v8 team on this topic
02:35:57  <indutny>as they have never replied about the previous backport patch
02:36:00  <jgi>indutny: ah, let me take another look at the issue
02:36:02  <indutny>which is btw quite faulty
02:36:09  <indutny>this is another question
02:38:20  * iarnajoined
02:39:16  * c4milojoined
02:43:50  <jgi>indutny: I confused this issue with the one you mentioned on twitter recently. So I took another look at the fix for v0.10, and I’m afraid I’m not qualified to even have an educated opinion.
02:43:59  * c4miloquit (Ping timeout: 250 seconds)
02:45:27  <jgi>indutny: do we usually test these types of fixes (V8 fixes that cannot be reviewed by the V8 team) with a broader tests suite? I feel somewhat unconfortable with including this in the next stable release, but that comes mainly from my ignorance of V8’s internals.
02:45:49  <indutny>jgi: I tested it with v8's test suite
02:46:56  <indutny>jgi: can't say the same about the previous patch :)
02:47:09  <jgi>indutny: what previous patch?
02:47:19  <jgi>indutny: also, did we run benchmarks for this change?
02:47:39  <indutny>fd80a31e0697d6317ce8c2d289575399f4e06d21
02:47:43  <indutny>jgi: nah
02:47:48  <indutny>benchmarks won't matter here
02:47:51  <indutny>but you could run them if you want to
02:48:05  <indutny>so fd80a31e0697d6317ce8c2d289575399f4e06d21 is quite broken at the moment
02:48:19  <indutny>and I am not totally sure that I want to spend lots of time figuring it out :)
02:48:33  <indutny>actually, I spent the time on that PushArgument thing
02:48:43  <indutny>only because I thought it was relevant to ARM failures that I have seen with v0.12
02:48:50  <indutny>but it doesn't seem to be that related after all
02:49:35  <jgi>ok
02:50:11  <jgi>how many reports of the issue mentioned in https://github.com/joyent/node/issues/8730 have we had so far? Just 1?
02:50:50  * quijotejoined
02:53:09  <trevnorris>indutny: I imported the latest certdata.txt and it matches that in v0.12. so there was no change.
02:54:07  <indutny>jgi: I guess just one
02:54:21  <indutny>jgi: we could suggest issue reporter to patch v0.10 :)
02:54:41  <indutny>but the error doesn't manifest itself in general
02:54:49  <jgi>indutny: to build their own version of node and patch when they build it?
02:54:55  <indutny>it is very rare combination of "use strict"; arguments and new call
02:55:13  <indutny>jgi: hm... actually, I was a bit wrong here. this won't help them
02:55:29  <indutny>because the error was fixed in when.js
02:55:34  <indutny>they changed something in it :)
02:55:40  <indutny>and it has disappeared
02:55:44  <jgi>indutny: right, they have a workaround already
02:55:49  * quijotequit (Ping timeout: 260 seconds)
02:57:33  <jgi>indutny: I don’t feel super excited about landing a change so deep in V8 for a stable release, at the same time it clearly fixes an actual issue. However this issue seems to happen rarely, and has been reported only by one user. I would advise to not land it for now, but keep the change around in case it happens more regularly in the future.
02:57:52  <indutny>jgi: ok, could you please share your thoughts on that issue?
02:58:07  <jgi>indutny: yep
02:58:13  <indutny>thank you
02:58:24  <indutny>trevnorris: good!
03:00:19  <trevnorris>indutny: i'll definitely need your review on the crypto stuff. i'm already having problems getting everything in correctly. with all the vulnerabilities and such there were a lot of changes between the two. some of them that weren't ported immediately, and I might have lost them when bringing them over to the split files.
03:00:41  <indutny>trevnorris: ok, the rule of thumb here
03:00:55  <indutny>trevnorris: git remote add io git@github.com:iojs/io.js
03:00:58  <indutny>trevnorris: git fetch io
03:01:17  <indutny>trevnorris: git checkout io/v0.12 deps/openssl
03:01:28  <indutny>trevnorris: it has latest openssl with a proper gyp stuff
03:01:50  <indutny>trevnorris: I'm not totally sure about SSL3 things, though
03:02:03  <indutny>trevnorris: but I guess they need to be merged too
03:05:03  <trevnorris>indutny: eh? i'm merging joyent v0.10 to v0.12.
03:05:11  <indutny>trevnorris: I know
03:05:28  <indutny>trevnorris: I just suggested to use deps/openssl from iojs/v0.12
03:06:30  <jgi>indutny: done: https://github.com/joyent/node/pull/8871#issuecomment-67105993
03:13:35  <jgi>indutny: I’m taking off, please let me know on GitHub if you need anything from me
03:13:40  <jgi>have a good night all!
03:13:52  * jgiquit (Quit: jgi)
03:18:33  <indutny>thanks!
03:31:10  * avalanche123joined
03:36:17  * avalanche123quit (Ping timeout: 240 seconds)
03:41:25  * Left_Turnquit (Remote host closed the connection)
03:51:36  * quijotejoined
03:56:22  * quijotequit (Ping timeout: 255 seconds)
03:58:57  <MI6>joyent/node: trevnorris created branch merge-review2 - http://git.io/vITjOg
03:59:11  <trevnorris>indutny: there are 2 failing crypto tests ^
03:59:32  <indutny>need some help? :)
03:59:33  <trevnorris>i'm fixing something with child process then i'll be taking a look.
03:59:39  <trevnorris>if you'd like to offer it. :)
03:59:58  <indutny>i'll fetch it
04:00:00  <indutny>and run tests :)
04:00:07  <indutny>I guess I know one failure
04:01:15  <trevnorris>heh, already? that was fast. :P
04:03:38  <indutny>trevnorris: this is just a prediction
04:03:41  <indutny>let me build it and run testse
04:03:43  <indutny>tests*
04:04:04  <indutny>trevnorris: btw, on a selfish note :)
04:04:07  <indutny>trevnorris: have you seen https://blog.indutny.com/a.deoptimize-me-not ?
04:04:18  <indutny>trevnorris: I wonder how you find it?
04:04:58  <trevnorris>indutny: read it. good stuff!
04:05:07  <indutny>trevnorris: was it easy to read?
04:05:11  <indutny>informative?
04:05:33  <trevnorris>easy enough for the topic at hand, and very informative.
04:05:52  <trevnorris>I haven't thought much about deoptimization because it's usually something I just try not to do.
04:06:02  <trevnorris>but it was interesting to read how V8 does it.
04:06:21  * inolenjoined
04:06:40  * inolenquit (Client Quit)
04:07:44  <indutny>trevnorris: thanks
04:07:53  <indutny>trevnorris: ok, so the test-crypto-stream thing
04:08:24  <indutny>there should be assert(/Unsupported state/.test(err));
04:08:25  <indutny>assert
04:08:29  <indutny>not 00000
04:08:37  <indutny>or just Unsupported
04:08:40  <indutny>to make it compatible to io.js
04:10:57  * wolfeidau__changed nick to wolfeidau
04:11:06  <indutny>trevnorris: and the honor test
04:11:12  <indutny>trevnorris: you did a mistake in merge, AFAIK
04:11:22  <trevnorris>indutny: most likely. :P
04:11:24  <indutny>trevnorris: take a look at lib/tls.js in v0.10
04:11:31  <indutny>trevnorris: search for honorCipherOrder string
04:11:39  <indutny>trevnorris: then find the same place in _tls_wrap.js
04:11:59  <indutny>aaah
04:12:00  <indutny>you didn't
04:12:54  <indutny>trevnorris: ignore it for now
04:12:59  <indutny>I'll send a PR after the merge
04:12:59  <trevnorris>heh, okay.
04:13:08  <trevnorris>for what?
04:13:08  <indutny>trevnorris: please file a bug and assign it to me, though
04:13:12  <indutny>trevnorris: for that test
04:13:29  <trevnorris>for test-crypto-stream?
04:13:30  <indutny>test/simple/test-tls-honorcipherorder-secureOptions.js
04:13:32  <indutny>nope
04:13:33  <trevnorris>ah, okay.
04:13:37  <indutny>in test-crypto-stream
04:13:38  <indutny>replace 000000
04:13:42  <indutny>with Unexpected
04:13:46  <trevnorris>yeah. it's working now. thanks. :)
04:13:49  <indutny>the error message has changed
04:13:55  <indutny>s/has/has been/
04:13:56  <indutny>trevnorris: np
04:15:20  <trevnorris>is there a way to summarize the issue with test-tls-honorcipherorder-secureOptions.js?
04:16:20  <indutny>trevnorris: I have moved the creation of SecureContext to _tls_common.js
04:16:28  <indutny>and the secureOptions is maintained there now
04:16:40  * dshaw_joined
04:16:47  <indutny>so server.secureOptions is just the one that was set in tls.createServer({ secureOptions: 0 })
04:16:57  <indutny>but the test expects it to contain HONOR_CIPHER_ORDER flag
04:17:05  <indutny>if the honorCipherOrder: true was in createServer
04:17:10  <indutny>but this thing is set in _tls_common.js
04:17:15  <indutny>and is not visible to _tls_wrap.js
04:19:55  * dshaw_1joined
04:20:40  * iarnaquit (Remote host closed the connection)
04:23:48  * dshaw_quit (Ping timeout: 264 seconds)
04:28:16  * c4milojoined
04:32:50  * c4miloquit (Ping timeout: 250 seconds)
04:35:02  * kellabytequit (Ping timeout: 244 seconds)
04:36:04  * wolfeidauquit (Ping timeout: 244 seconds)
04:36:30  * iarnajoined
04:36:36  * yuskesh_quit (Ping timeout: 244 seconds)
04:37:20  <octetcloud>hey, I heard a rumour 0.10 doesn't support compiled addons on windows 8... I'm in the process of building a vm to verify, which will be slow. Anybody else heard this?
04:38:21  * wolfeidaujoined
04:39:03  * iarnaquit (Remote host closed the connection)
04:39:06  * kellabytejoined
04:39:26  * yuskesh_joined
04:43:49  * AlexisMochaquit (Ping timeout: 244 seconds)
04:49:27  * iarnajoined
04:52:07  * dshaw_1quit (Read error: Connection reset by peer)
04:52:26  * quijotejoined
04:52:30  * dshaw_joined
04:53:03  * iarnaquit (Remote host closed the connection)
04:56:57  * quijotequit (Ping timeout: 245 seconds)
04:57:35  * iarnajoined
04:57:58  * dshaw_quit (Quit: Leaving.)
05:05:53  * jgijoined
05:16:58  * jgiquit (Quit: jgi)
05:18:01  * jgijoined
05:18:05  * iarnaquit (Remote host closed the connection)
05:18:56  * jreyno40_joined
05:20:52  * jgiquit (Client Quit)
05:21:50  * jgijoined
05:22:02  * rmgquit (Read error: Connection reset by peer)
05:22:27  * rmgjoined
05:28:37  * felipealmeidaquit (Ping timeout: 245 seconds)
05:30:41  * avalanche123joined
05:34:08  * felipealmeidajoined
05:42:17  * iarnajoined
05:49:43  * seishunjoined
05:51:33  * Damn3dquit (Ping timeout: 258 seconds)
05:53:09  * quijotejoined
05:54:44  * petka_quit (Quit: Connection closed for inactivity)
05:55:31  * Damn3djoined
05:57:49  * quijotequit (Ping timeout: 260 seconds)
05:58:53  * bajtosjoined
06:02:08  * bajtosquit (Client Quit)
06:04:07  * bajtosjoined
06:05:59  * Fishrock123quit (Quit: Leaving...)
06:17:26  * c4milojoined
06:22:22  * c4miloquit (Ping timeout: 265 seconds)
06:26:45  * iarnaquit (Ping timeout: 260 seconds)
06:28:37  * octetcloudquit (Ping timeout: 260 seconds)
06:29:46  * octetcloudjoined
06:30:33  <MI6>joyent/node: Ben Burns v0.12 * 5b9e5bd : doc: clarify create{Read,Write}Stream fd option - http://git.io/kdcBgA
06:39:50  <chrisdickinson>we should probably merge https://github.com/joyent/node-website/pull/60 soon.
06:40:01  <chrisdickinson>does that repo follow the same guidelines as the node repo?
06:42:58  * janjongboomjoined
06:44:03  * wolfeidauquit (Read error: Permission denied)
06:44:44  * wolfeidaujoined
06:53:56  * quijotejoined
06:56:26  * avalanche123quit (Remote host closed the connection)
06:58:37  * quijotequit (Ping timeout: 265 seconds)
07:04:10  * janjongboomquit (Quit: My MacBook has gone to sleep. ZZZzzz…)
07:07:19  <jgi>chrisdickinson: I don’t know, we’ll have to ask tjfontaine tomorrow
07:16:43  * jgiquit (Quit: jgi)
07:26:57  * seishunquit (Ping timeout: 272 seconds)
07:33:09  * nickleeflyjoined
07:45:48  * janjongboomjoined
07:54:41  * quijotejoined
07:56:49  * avalanche123joined
07:59:48  * quijotequit (Ping timeout: 264 seconds)
08:01:27  * avalanche123quit (Ping timeout: 265 seconds)
08:06:29  * c4milojoined
08:08:07  * robertkowalskiquit (Ping timeout: 272 seconds)
08:08:16  * robertkowalskijoined
08:11:12  * c4miloquit (Ping timeout: 264 seconds)
08:12:14  * AvianFluquit (Ping timeout: 256 seconds)
08:14:09  * bajtosquit (Quit: bajtos)
08:14:20  * inolenjoined
08:15:23  * bajtosjoined
08:18:21  * quijotejoined
08:30:47  * rendarjoined
08:58:49  * hollandaisquit (Excess Flood)
09:06:07  * octetcloudquit (Ping timeout: 245 seconds)
09:10:54  * SergeiRNDjoined
09:13:09  * hollandaisjoined
09:14:48  * janjongboomquit (Ping timeout: 264 seconds)
09:16:19  * janjongboomjoined
09:17:13  * [spoiler]joined
09:23:58  * quijotequit (Ping timeout: 255 seconds)
09:27:11  * inolen1joined
09:27:12  * inolenquit (Read error: Connection reset by peer)
09:28:56  * stagasjoined
09:42:29  * chris_99joined
09:43:45  * iarnajoined
09:48:04  * iarnaquit (Ping timeout: 258 seconds)
09:53:09  * inolen1quit (Ping timeout: 245 seconds)
09:55:26  * c4milojoined
10:00:14  * c4miloquit (Ping timeout: 245 seconds)
10:14:52  * janjongboomquit (Ping timeout: 245 seconds)
10:15:57  * janjongboomjoined
10:19:13  * rmgquit (Remote host closed the connection)
10:21:19  * quijotejoined
10:25:41  * quijotequit (Ping timeout: 260 seconds)
10:25:51  * Left_Turnjoined
10:36:18  * AlexisMochajoined
10:36:20  * bajtosquit (Quit: bajtos)
10:39:42  * abraxas_quit (Remote host closed the connection)
10:41:47  * bajtosjoined
10:46:19  * bajtosquit (Client Quit)
10:46:55  * quijotejoined
10:47:17  * bajtosjoined
10:50:44  * bajtosquit (Client Quit)
11:01:27  * SergeiRNDquit (Quit: Leaving.)
11:02:36  * jreyno40_quit (Quit: jreyno40_)
11:13:17  * quijotequit (Ping timeout: 260 seconds)
11:19:44  * rmgjoined
11:22:34  * bajtosjoined
11:25:00  * rmgquit (Ping timeout: 264 seconds)
11:27:11  * stagasquit (Quit: Bye)
11:33:58  * hayesquit (Ping timeout: 256 seconds)
11:34:29  <bajtos>indutny: yt?
11:35:08  <bajtos>I am trying to run `make jslint` in io.js and the command fails with “cannot import name errorprinter"
11:35:48  * SergeiRNDjoined
11:35:57  <bajtos>the file `errorprinter` is on my filesystem: `tools/closure_linter/closure_linter/common/errorprinter.py`
11:37:30  <bajtos>never mind. Running `./setup.py install` inside `io.js/tools/closure_linter` fixed the problem
11:38:42  * hayesjoined
11:39:14  * quijotejoined
11:43:34  * quijotequit (Ping timeout: 250 seconds)
11:44:22  * c4milojoined
11:44:33  <indutny>yes
11:49:47  * c4miloquit (Ping timeout: 272 seconds)
11:54:32  * tarrudajoined
11:55:18  * tarrudachanged nick to Guest47277
12:06:06  * Guest47277quit (Quit: WeeChat 1.0.1)
12:06:26  * tarruda_joined
12:12:43  * AlexisMocha_joined
12:13:20  <bajtos>indutny: I wanted to ask you about the `make jslint` problem, but I have figured it out myself
12:14:24  * dembones_joined
12:14:25  * AlexisMochaquit (Ping timeout: 260 seconds)
12:14:27  * dembonesquit (Ping timeout: 260 seconds)
12:14:53  * janjongboomquit (Ping timeout: 260 seconds)
12:15:36  * janjongboomjoined
12:16:51  * trevnorrisquit (Ping timeout: 260 seconds)
12:17:47  * unprompt1djoined
12:18:01  * unpromptedquit (Ping timeout: 260 seconds)
12:18:59  * trevnorrisjoined
12:19:18  * Ralith_joined
12:20:11  * toothrotquit (Ping timeout: 260 seconds)
12:20:11  * Ralithquit (Ping timeout: 260 seconds)
12:20:11  * Raynosquit (Ping timeout: 260 seconds)
12:20:15  * \toothrotjoined
12:21:33  * Raynosjoined
12:22:21  * rvaggquit (Ping timeout: 260 seconds)
12:25:35  * rvaggjoined
12:28:30  * abraxas_joined
12:31:15  * yuskesh_quit (*.net *.split)
12:33:52  * abraxas_quit (Ping timeout: 255 seconds)
12:40:31  * nickleeflyquit (Quit: Connection closed for inactivity)
12:45:21  * [spoiler]quit (Quit: Leaving)
13:00:46  * bajtosquit (Quit: bajtos)
13:04:12  * bajtosjoined
13:06:34  * yuskesh_joined
13:23:43  * avalanche123joined
13:28:19  * avalanche123quit (Ping timeout: 255 seconds)
13:29:15  * [spoiler]joined
13:30:42  * janjongboomquit (Ping timeout: 245 seconds)
13:31:07  * \toothrotquit (Ping timeout: 272 seconds)
13:31:44  * janjongboomjoined
13:33:41  * c4milojoined
13:34:08  * bajtosquit (Quit: bajtos)
13:38:03  * c4miloquit (Ping timeout: 244 seconds)
13:57:11  <txdv>saghul2:
13:57:20  <txdv>did you kill the first saghul?
13:57:31  <txdv>https://github.com/kozross/awesome-c look libuv is in a list of awesome libs
14:05:07  * chris_99quit (Remote host closed the connection)
14:06:51  * chris_99joined
14:18:04  * abraxas_joined
14:21:42  * WalrusPonyjoined
14:22:22  * abraxas_quit (Ping timeout: 245 seconds)
14:22:24  * c4milojoined
14:30:13  * rendarquit (*.net *.split)
14:30:13  * WalrusPony1quit (*.net *.split)
14:30:13  * rphillipsquit (*.net *.split)
14:30:35  * rphillipsjoined
14:31:20  * rendarjoined
14:31:37  * janjongboomquit (Ping timeout: 264 seconds)
14:32:14  * janjongboomjoined
14:40:39  * lance|afkchanged nick to lanceball
14:56:19  * KennethWilkejoined
14:56:45  * Fishrock123joined
14:56:49  * SergeiRNDquit (Quit: Leaving.)
14:57:55  * rmgjoined
15:01:06  * piscisaureusjoined
15:02:17  * rmgquit (Ping timeout: 240 seconds)
15:04:06  <indutny>chrisdickinson: https://github.com/iojs/io.js/pull/172
15:04:07  <indutny>oops
15:16:07  * janjongboomquit (Ping timeout: 245 seconds)
15:17:17  * janjongboomjoined
15:31:35  * AvianFlujoined
15:51:14  * janjongboomquit (Read error: Connection reset by peer)
15:51:30  * janjongboomjoined
15:52:56  * seishunjoined
16:02:16  * janjongboomquit (Ping timeout: 250 seconds)
16:03:09  * janjongboomjoined
16:06:53  * abraxas_joined
16:08:07  * dshaw_1joined
16:11:17  * abraxas_quit (Ping timeout: 250 seconds)
16:14:00  * avalanche123joined
16:16:00  * dshaw_1quit (Quit: Leaving.)
16:26:03  * brsonjoined
16:41:07  * dshaw_joined
16:43:05  * tarruda_quit (Ping timeout: 264 seconds)
16:55:57  * rmgjoined
16:57:09  * iarnajoined
16:58:06  * jgijoined
16:58:10  <srl295>jgi: morning
16:58:51  <srl295>jgi: re: configure's return code.. maybe I'm the only one breaking .gyp enough to hit this? I figured if it was intentional, a PR wouldn't hurt
16:59:04  <srl295>jgi: is srl295/#15 ready for me to merge? it does look good
16:59:35  <jgi>srl295: I have a meeting right now, are you available around 10 (maybe before)?
16:59:52  <jgi>srl295: 10 AM PST that is :)
16:59:53  <srl295>jgi: think i have a meeting at 10, anyways i'm around today
17:00:01  <srl295>jgi: yep, PT
17:01:17  * iarnaquit (Read error: Connection reset by peer)
17:01:52  * iarnajoined
17:02:18  * c4miloquit (Remote host closed the connection)
17:02:45  <tjfontaine>trevnorris: ping?
17:02:48  <srl295>jgi: setting up SmartOS..
17:03:00  <tjfontaine>https://bluejeans.com/732994483/ meeting if others want to listen in (ideally silently)
17:03:26  * SergeiRNDjoined
17:03:58  <jgi>srl295: re: configure’s return code, I’m not sure I understand what you mean. I landed this PR: https://github.com/joyent/node/pull/8856, do you mean it shouldn’t have been landed?
17:04:37  <tjfontaine>indutny: are you still awake?
17:05:14  <jgi>srl295: I have SmartOS running mon my laptop, so I can (and I was) investigating that. So it’s cool if you have other stuff to do :)
17:05:55  <tjfontaine>AlexisMocha_: hey meeting?
17:07:12  * avalanche123quit (Remote host closed the connection)
17:08:20  <srl295>jgi: no its OK on both counts.
17:09:03  <srl295>Even without the return code fix, "configure && make" still fails, just at a different spot
17:10:13  * bajtosjoined
17:17:45  * SergeiRNDquit (Quit: Leaving.)
17:17:58  * inolenjoined
17:24:09  * janjongboomquit
17:25:34  * bajtosquit (Read error: Connection reset by peer)
17:25:42  * bajtos_joined
17:25:44  * yunong_joined
17:29:13  * yunongquit (Ping timeout: 258 seconds)
17:30:57  * dap_joined
17:31:04  * Ralith_quit (Ping timeout: 244 seconds)
17:33:07  * piscisaureusquit (Ping timeout: 255 seconds)
17:40:34  * quijotejoined
17:49:00  * SergeiRNDjoined
17:52:04  * a_lequit (Remote host closed the connection)
17:54:08  <creationix>is uv_loop_alive new? I don’t remember seeing that one
17:54:30  <creationix>I see the docs are 1.0.2, my bindings are 1.0.0. I need to update my bindings...
17:55:49  * abraxas_joined
17:59:16  <jgi>tjfontaine, chrisdickinson, AlexisMocha_, trevnorris, cjihrig: I’ll send the meeting minutes later today
18:00:18  * abraxas_quit (Ping timeout: 265 seconds)
18:02:04  <chrisdickinson>tjfontaine: I was mistaken, the queue patch touches all of the same parts of writable that the {un,}cork stuff touched :\
18:07:34  * octetcloudjoined
18:14:27  * bajtos_quit (Quit: bajtos_)
18:29:26  * iarnaquit (Remote host closed the connection)
18:31:05  * seldojoined
18:33:20  <tjfontaine>chrisdickinson: sad, well, if possible it's a really nice to have, but not necessary
18:34:50  * Ralithjoined
18:35:48  <chrisdickinson>tjfontaine: if the forward port of a commit has to add code to keep working, do we want to handle that in the merge commit, or as a separate commit?
18:38:13  <tjfontaine>chrisdickinson: normally I would suggest doing it in that commit if it's not that much, ideally again this is in a merge early merge often routine
18:38:37  <tjfontaine>chrisdickinson: in the case where you're merging multiple commits where the noise is high I would advocate a second follow up commit
18:39:03  <tjfontaine>merges are probably the one exception to the rule for cases where it's ok to introduce breakage
18:39:38  <chrisdickinson>has the v0.10->v0.12 merge still in progress? If that's going to be done soon (within the day) I can probably get the writable change to work
18:40:02  <chrisdickinson>but I'd hate to add a "gotcha" commit to that pile in advance
18:40:47  <tjfontaine>if trevor can't make progress on it, I will do it immediately after 0.10.34 release
18:41:17  <srl295>jgi: off calls, what's up?
18:43:55  * jgiquit (Quit: jgi)
18:45:44  * quijotequit (Ping timeout: 265 seconds)
18:49:47  <nathan7>hey tjfontaine
18:49:59  <nathan7>tjfontaine: I'm across from your office!
18:51:22  * c4milojoined
18:51:31  <nathan7>tjfontaine: Max (Bruning) said he'd be in the city today, but I can't reach him now
18:53:17  * jgijoined
18:53:24  <tjfontaine>nathan7: he's not in the office yet, I'm sure he will be though I saw him online earlier
18:53:32  <nathan7>tjfontaine: mhm
18:54:04  <nathan7>I didn't get an answer yesterday when I asked him what time he'd be there
18:54:17  <nathan7>so that's fine, my guess was off q=
18:54:37  <tjfontaine>hehe
18:54:50  <jgi>srl295: still spinning up jobs on Jenkins to get as many platforms as possible tested with —with-intl=small-icu
18:54:52  * nathan7chilling in the Peet's across
18:55:13  <jgi>srl295: are you investigating the SmartOS build error, or do you want me to do it?
18:56:14  * c4miloquit (Ping timeout: 258 seconds)
18:57:51  <MI6>joyent/node: misterdjules created branch v0.21 - http://git.io/mSFNqg
18:58:47  <jgi>arf...
18:59:16  <jgi>sorry, typo in the ‘branch’ list box on GitHub created that branch ^
18:59:21  * nathan7saw someone looking suspiciously much like Max get into the building
18:59:21  * dshaw_quit (Quit: Leaving.)
18:59:53  <tjfontaine>jgi: I hate that feature ...
18:59:55  <tjfontaine>"feature"
19:00:05  <jgi>tjfontaine: yep, I knew it would bite me some time, and here we are...
19:00:12  <jgi>deleting that branch now
19:00:38  <jgi>I will make sure I don’t make the mistake of deleting v0.12...
19:01:04  <srl295>jgi: that would be bad
19:01:16  <tjfontaine>recoverable though :)
19:01:39  <srl295>tjfontaine: right.
19:02:22  <jgi>alright, v0.21 deleted
19:02:25  <jgi>sorry again
19:02:37  * AlexisMocha_quit (Ping timeout: 240 seconds)
19:02:49  <srl295>jgi: I am going to try to integrate srl295#15 into the autoicu branch first. Did you try the setting here? https://github.com/srl295/node/issues/17#issuecomment-67189968
19:03:16  <jgi>srl295: I haven’t had the time to try it, but I will try it in a few minutes
19:03:30  <tjfontaine>jgi: I have done it a ton of times
19:03:44  <jgi>srl295: regarding tests, why don’t we assert if English is not supported: https://github.com/srl295/node/blob/srl-v0.12-autoicu/test/simple/test-intl.js#L77?
19:04:25  <srl295>jgi: while I wanted to try out smartos you might be quicker to test. I was very pleasantly surprised to see that imgadm supports tab complete
19:04:54  <jgi>srl295: re: https://github.com/srl295/node/pull/15, we don’t necessarily need to merge it into the autoicu branch
19:05:20  <srl295>jgi: havelocale tests whether english was chosen in the build options.
19:05:28  <jgi>srl295: since the first step is to have all builds built with —with-intl=small-icu
19:05:35  <srl295>jgi: OK.
19:06:16  <jgi>srl295: #15 can be landed later, when we made every platform (like SmartOS) work with --with-intl=small-icu
19:06:20  <srl295>jgi: we could assert in configure that you have to include English. Not sure we should though.
19:08:18  <jgi>srl295: basically, my concern is that this test could potentially become a no-op, and thus we would think that all tests pass, whereas actually this test would not run
19:09:43  <srl295>jgi: if you do --with-icu-locales=de,fr it still verifies that you have the Intl object and that it can format something, but it doesn't try to check the results against English. So it's not completely a no-op
19:09:45  <jgi>tjfontaine: it seems that our OS X agents are down: http://jenkins.nodejs.org/computer/osx-build/ and http://jenkins.nodejs.org/computer/osx-home/
19:10:26  <jgi>srl295: so that: https://github.com/srl295/node/blob/srl-v0.12-autoicu/test/simple/test-intl.js#L53 throws an error if it fails?
19:10:33  * qardquit (Quit: leaving)
19:12:13  <srl295>jgi: yes. I guess it could do dtf.format(date0) on line 77 and verify that format() can run without error.
19:13:22  <tjfontaine>jgi: I knew home was off, is build also off?
19:13:36  <jgi>tjfontaine: it seems to be back now
19:13:42  <jgi>tjfontaine: did you restart the agent?
19:13:44  <tjfontaine>nope
19:14:00  <jgi>ok, I clicked on the “Launch” button with the Java icon
19:14:16  <tjfontaine>that means you're now the slave :)
19:14:17  <jgi>I have no idea if it’s what restarted it
19:14:28  <jgi>hahaha
19:14:44  <tjfontaine>brb going into closet to check on laptop
19:16:23  * SergeiRNDquit (Quit: Leaving.)
19:16:29  * AlexisMochajoined
19:17:01  * qardjoined
19:17:56  * a_lejoined
19:20:18  * seldoquit (Remote host closed the connection)
19:20:51  * avalanche123joined
19:25:58  <tjfontaine>jgi: the box looks up right now?
19:26:18  <tjfontaine>home is what release builds are on, I should be able to bring that one to life
19:26:51  <jgi>tjfontaine: yes, the box seems to be up now, and my laptop didn’t become the OS X agent :)
19:27:41  <tjfontaine>:)
19:27:41  * quijotejoined
19:29:14  * seldojoined
19:29:21  <jgi>srl295: I think generally we want to make sure that when we do a release, the ICU support works for the locales that we say should work by default (which currently means the default locales root,en IIUC). Maybe we need a separate set of tests that run during the release process?
19:29:44  <srl295>jgi: the test.js could take an argument..
19:30:08  <srl295>jgi: it seems to me like, if you run configure --without-ssl you don't assert that ssl is there
19:31:46  <jgi>srl295: yes, an argument could work
19:32:03  <jgi>srl295: I don’t know what happens when we run tests against node built with —without-ssl
19:32:16  * quijotequit (Ping timeout: 258 seconds)
19:33:54  * jcrugzz_changed nick to jcrugzz
19:34:00  <srl295>jgi: console.error('Skipping because node compiled without OpenSSL CLI.'); ?
19:34:03  * seldoquit (Remote host closed the connection)
19:34:56  <srl295>jgi: if (!process.versions.openssl) { console.error('Skipping because node compiled without OpenSSL.');
19:35:18  <jgi>srl295: it does not seem to be present for all SSL tests though
19:35:45  <jgi>srl295: and it doesn’t mean that it’s always the right thing to do :)
19:37:48  <srl295>copy-paste: "now you have two bugs!"
19:38:09  * seldojoined
19:38:34  <jgi>srl295: We don’t have to find the most elegant solution right now, I just wanted to raise the question to know your thought about it. I can create an issue in joyent/node and we’ll have more time to think about it
19:39:31  <srl295>jgi: I think if we want to assert that a 'release binary' works a certain way , it would make sense to use an argument
19:39:57  <jgi>srl295: yes, I think it’s a good suggestion too, I’ll mention it in the issue
19:40:01  <jgi>srl295: thanks :)
19:40:13  <tjfontaine>jgi: looks like I have a broken libtool in that path?
19:40:15  <srl295>jgi: sure. Maybe you can mention what would actually call the test a certain way, etc.
19:40:39  <jgi>tjfontaine: what path?
19:40:52  <tjfontaine>for the builds on home
19:41:29  <jgi>tjfontaine: ah, I don’t know, has any recent build run on osx-home lately?
19:41:33  * iarnajoined
19:41:39  <tjfontaine>jgi: no, Im' fixing it up now
19:42:00  <tjfontaine>I didn't realize /usr/local/bin was before /usr/bin, not terribly surprising though
19:42:44  <tjfontaine>jgi: go ahead and retry that pkg build, ti should work
19:43:13  <jgi>tjfontaine: ah ok, you meant that build: http://jenkins.nodejs.org/job/nodejs-julien/DESTCPU=ia32,label=osx/lastBuild/consoleFull ran on osx-home
19:44:30  * abraxas_joined
19:44:44  * iarnaquit (Read error: Connection reset by peer)
19:44:54  <MI6>joyent/node: Julien Gilli merge-review * 934bfe2 : timers: Avoid linear scan in _unrefActive. - http://git.io/XioYXw
19:45:20  * iarnajoined
19:45:24  <jgi>tjfontaine: I’ll wait for the other one to finish first, but I’ll rerun it asap
19:45:27  <jgi>tjfontaine: thanks!
19:45:31  <tjfontaine>jgi: np
19:45:46  <tjfontaine>gonna try and grab some food before my 12pm call
19:47:33  <jgi>tjfontaine: see you later!
19:49:37  * abraxas_quit (Ping timeout: 264 seconds)
19:50:06  <jgi>srl295: yes, I’ll put as many details as possible in this issue
19:50:35  * SergeiRNDjoined
19:57:17  * a_lequit (Ping timeout: 244 seconds)
20:05:58  * piscisaureusjoined
20:06:04  * seldoquit (Remote host closed the connection)
20:07:52  * petka_joined
20:08:51  * a_lejoined
20:11:31  * avalanche123quit (Remote host closed the connection)
20:25:24  * SergeiRNDquit (Quit: Leaving.)
20:28:21  * quijotejoined
20:28:28  <trevnorris>tjfontaine: i've started on the v0.10 to v0.12 merge (branch merge-review2). working though an issue w/ child process, and indutny knows the issue on another test failure.
20:30:39  * a_lequit (Remote host closed the connection)
20:31:24  * a_lejoined
20:32:56  * quijotequit (Ping timeout: 244 seconds)
20:34:29  * AlexisMochaquit (Ping timeout: 245 seconds)
20:39:04  * tarrudajoined
20:39:28  * tarrudachanged nick to Guest49420
20:40:26  * c4milojoined
20:41:08  * quijotejoined
20:43:58  * Guest49420quit (Quit: WeeChat 1.0.1)
20:44:20  * tarruda_joined
20:45:14  * c4miloquit (Ping timeout: 245 seconds)
20:46:05  * quijotequit (Ping timeout: 264 seconds)
20:48:56  <chrisdickinson>tjfontaine: thoughts on landing https://github.com/joyent/node/pull/6442? it's got two LGTM's from folks, but you mentioned wanting to review in the PR.
20:49:03  * SergeiRNDjoined
20:55:31  * muriculajoined
20:55:42  * c4milojoined
20:55:55  * seldojoined
20:56:29  * avalanche123joined
20:58:51  * [spoiler]quit (Quit: Leaving)
21:01:47  * dshaw_joined
21:02:43  <jgi>srl295: is there any reason why we use ‘../../out’ instead of <(SHARED_INTERMEDIATE_DIR) sometimes in icu-generic.gyp?
21:03:18  <srl295>jgi: at some stages in the generation I could not get gyp to expand the variable
21:04:31  <jgi>srl295: interesting, is that a known bug with gyp, or is it normal?
21:06:35  <srl295>jgi: It's hard for me to tell what is normal with gyp!
21:06:42  <jgi>srl295: hehe :)
21:07:12  * davijoined
21:07:51  <srl295>jgi: exclamation marks aside, on a quick search of the file, those cases are inputs and outputs under windows. Yeah, it didn't work
21:08:35  <srl295>you are welcome to refine it if you think you can
21:08:45  <jgi>srl295: basically what happens on windows is that the tests runner looks if a directory ‘out’ exists, and if so runs the node binary from there, otherwise it runs it from .\Release or .\Debug depending on the type of build. Since we hardcode the creation of files in the ‘out’ directory in icu-generic, we always try to run tests with the node binary from .\out\Release or .\out\Debug on Windows, and that always fails. I’m trying to fi
21:08:46  <jgi>out the best way to fix that.
21:09:17  <srl295>jgi: hm.
21:10:13  <jgi>srl295: it’s easy to fix the tests runner, but hardcoding paths seems to be less than ideal, I’m concerned we may run into other issues.
21:16:00  <srl295>jgi: let me try it
21:16:15  <jgi>srl295: you’re doing all the work :)
21:16:31  <jgi>srl295: I’ll grab some lunch then, I’m too hungry to think right now
21:16:49  <srl295>jgi: I'll grab lunch while trhe build runs
21:17:03  <jgi>srl295: I’ll see you in a bit, feel free to post your findings here even if I disconnect, I’ll read the logs when I’m back
21:17:10  <jgi>srl295: enjoy your lunch :)
21:17:16  <srl295>jgi: likewise
21:19:06  * avalanche123quit (Remote host closed the connection)
21:21:53  * avalanche123joined
21:22:19  * daviquit (Ping timeout: 272 seconds)
21:31:26  * jgiquit (Quit: jgi)
21:31:58  * SergeiRNDquit (Quit: Leaving.)
21:33:12  * abraxas_joined
21:37:57  * abraxas_quit (Ping timeout: 240 seconds)
21:41:53  * quijotejoined
21:42:37  * lanceballchanged nick to lance|afk
21:46:26  * quijotequit (Ping timeout: 258 seconds)
21:50:19  * jgijoined
21:50:34  <jgi>srl295: I’m back
21:57:20  * seldoquit (Remote host closed the connection)
22:01:14  * tarruda_quit (Quit: WeeChat 1.0.1)
22:04:16  * seldojoined
22:11:40  * piscisaureusquit (Ping timeout: 255 seconds)
22:12:46  * seishunquit (Ping timeout: 250 seconds)
22:22:35  * seldoquit (Remote host closed the connection)
22:28:05  * Fishrock123quit (Remote host closed the connection)
22:32:29  * seldojoined
22:35:28  * sblomjoined
22:39:19  <muricula>How do I run a single test?
22:42:40  * quijotejoined
22:43:05  <trevnorris>muricula: python tools/test.py simple/test-<name>
22:44:14  <muricula> thanks
22:44:50  <muricula>Do you mind if I make a PR to add that to the README?
22:44:55  <muricula>trevnorris: ^^
22:45:06  <trevnorris>tjfontaine: ? ^
22:46:17  * Fishrock123joined
22:46:52  <muricula>Hold on there is no 'simple' directory in the repo that I see. This is a question about libuv, not io.js
22:47:13  * quijotequit (Ping timeout: 255 seconds)
22:47:15  <trevnorris>oh, heh. nm.
22:47:29  <trevnorris>then just put the name of the test
22:47:48  <trevnorris>e.g. ./run-tests <test-name>
22:49:06  <muricula>./out/Debug/run-tests test-spawn.c ? That doesn't work.
22:49:41  <tjfontaine>sorry was chatting with nathan7 in meatspace
22:49:52  <tjfontaine>python tools/test.py simple/*name-of-test*
22:50:14  <tjfontaine>oh libuv
22:50:26  <tjfontaine>muricula: ask for the list
22:50:38  <tjfontaine>muricula: there it's run-tests <inner test name>
22:50:51  <tjfontaine>muricula: it's not about the file, but the actual #define'd test name
22:51:08  <trevnorris>muricula: run ./run-tests --list
22:51:14  <trevnorris>that will dump a list of all tests.
22:51:23  <tjfontaine>muricula: there's also run-tests <test-name> <test-name> which runs it in a single process model such that you can invoke it more easily from a debugger
22:51:25  <muricula>ah thanks
22:52:08  <tjfontaine>and now for lunch
22:52:28  <muricula>thanks
22:52:46  <tjfontaine>chrisdickinson, jgi yes please land concatenated gzip fix in 0.12
22:55:31  <chrisdickinson>tjfontaine: will do
22:55:32  * c4miloquit (Remote host closed the connection)
22:55:52  * c4milojoined
23:01:30  * dshaw_quit (Quit: Leaving.)
23:07:07  <rendar>chrisdickinson, is that the patch for concatenating .gz files?
23:07:40  * c4miloquit (Remote host closed the connection)
23:08:14  <tjfontaine>it's for unzipping concatenated streams
23:08:28  <tjfontaine>concat during gzip is an exercise for the consumer
23:10:26  <rendar>tjfontaine, for concatenated stream you mean cat my1.gz my2.gz > all.gz -- so now 'all.gz' contains a concatenated stream of compressed data, right?
23:10:37  <tjfontaine>yup
23:11:05  <rendar>i see, so you can use some zlib APIs to recognize that, and decompressing those 2 streams in one operation
23:11:12  * sblomquit (Read error: Connection reset by peer)
23:11:20  <tjfontaine>gzip/bzip etc all handle this case already
23:11:50  <tjfontaine>basically they read until they hit another gzip/bzip header, and if the previous block uncompressed ok they barrell on
23:12:16  <chrisdickinson>cool, running tests now
23:12:25  <tjfontaine>it's how you're able to actually do parallel compression fairly easily (presuming you know how to chunk your data appropriately)
23:13:17  <muricula>So do I need to do something special to run tests under gdb/lldb?
23:13:27  <rendar>tjfontaine, i see
23:13:58  <muricula>https://pastebin.osuosl.org/13966/
23:14:14  <rendar>tjfontaine, hmm, parallel compression = each thread does one different chunk? what you mean exactly?
23:14:45  <tjfontaine>rendar: that's right one thread/process per chunk, and a supervisor to write the chunks out in order
23:15:04  <tjfontaine>[12-16] 22:51:23 <@tjfontaine> muricula: there's also run-tests <test-name> <test-name> which runs it in a single process model such that you can invoke it more easily from a debugger
23:15:08  <MI6>joyent/node: Luis Reis v0.12 * 6f6a979 : zlib: support concatenated gzip files - http://git.io/tKsqhQ
23:15:28  <muricula>ah thanks
23:15:32  <tjfontaine>muricula: that means you want to `r test-name test-name`
23:15:34  <rendar>tjfontaine, i see! and how node can do that? with multiple uv_loop?
23:16:10  <tjfontaine>rendar: in a binary module or something with ffi yes, easier in terms of child_process though
23:16:31  <tjfontaine>rendar: I did all this work, onceuponatime for python
23:16:54  <tjfontaine>rendar: warning ancient unmaintained code https://github.com/tjfontaine/zpyzpr and broken urls
23:17:13  <rendar>oh...very interesting
23:17:31  <rendar>iirc, isn't also git doing something that when you make a commit?
23:17:49  <tjfontaine>packfiles work in a similar format afaik
23:17:52  <tjfontaine>or aiui
23:18:37  <MI6>joyent/node: Julien Gilli v0.10 * 934bfe2 : timers: Avoid linear scan in _unrefActive. - http://git.io/jqpBGA
23:19:37  <trevnorris>tjfontaine: there's a problem with child process in the v0.10 merge that I cannot get right.
23:19:46  <nathan7>tjfontaine: our zlib is async anyway, no?
23:19:48  <tjfontaine>trevnorris: what's happening?
23:20:07  <nathan7>tjfontaine: or does that all go on the same thread
23:20:15  <rendar>nathan7, zlib is not async per se, but if you run it in another thread, it will be
23:20:28  <trevnorris>tjfontaine: v0.10 added a bunch of checks and v0.12 is failing them. but the code is so different I can't see how to make it compliant.
23:20:34  <tjfontaine>nathan7: it is, it's on the thread pool, but you can't really parallelize across multiple threads for the same context
23:20:37  <nathan7>rendar: the node core bindings
23:20:44  <tjfontaine>nathan7: so you need multiple contexts to do the compression
23:21:01  <nathan7>tjfontaine: a context being an instance of a zlib transform stream, right?
23:21:16  <rendar>tjfontaine, hmm, why you cannot use threads for the same context?
23:21:16  <tjfontaine>trevnorris: if it's too complex to do that, we can do it in a follow up commit, maybe it's a task that cjihrig could help with since he's done a lot of that work
23:22:19  <tjfontaine>nathan7: an actual zlib context, the thing that keeps the dictionary for the compressed stream
23:22:24  * abraxas_joined
23:22:32  <tjfontaine>rendar: because zlib isn't designed to do that
23:22:43  <rendar>tjfontaine, ok
23:22:45  <nathan7>tjfontaine: each zlib transform stream has a fresh one, right
23:22:50  <tjfontaine>yes
23:23:30  <rendar>tjfontaine, so if you want to parallelize, you'd go for compressing/decompressing N different context in N different threads, and not N different chunks in N different threads
23:24:07  <tjfontaine>rendar: yes
23:24:24  * AlexisMochajoined
23:25:28  <rendar>tjfontaine, for example, if you have a bunch of files, you can divide the compression work accross N threads, for *each* file, and then concatenate all data you get in one unique big stream, to create all.gz
23:25:48  <tjfontaine>rendar: yes
23:25:57  <cjihrig>trevnorris: glad to help if you need it
23:26:05  <rendar>but when you decompress it, you cannot use multiple theads
23:26:15  <tjfontaine>rendar: no, but gzip is ridiculously fast at decompression
23:26:20  <tjfontaine>so it's rarely a problem
23:26:31  <trevnorris>cjihrig: thanks. mind checking out the merge-review2 branch?
23:26:36  <trevnorris>heck, let me just make a PR for it.
23:26:54  * abraxas_quit (Ping timeout: 245 seconds)
23:27:26  <rendar>tjfontaine, i see, btw there was a lib different that zlib..cannot remember the name, like pz..or something...created as a multithreaded version of zlib
23:27:41  <tjfontaine>yes there's pbzip and pig or something for gzip
23:27:54  <tjfontaine>chrisdickinson: sad, we broke the windows build with zlib
23:28:15  * dshaw_joined
23:28:17  <rendar>tjfontaine, i see, so if they are multithreaded, can't them be used in node?
23:28:40  <trevnorris>cjihrig: https://github.com/joyent/node/pull/8886
23:28:41  <tjfontaine>rendar: depending on how they're implemented they can be, their use case is very limited though
23:28:53  <rendar>i see
23:30:48  * c4milojoined
23:31:18  <trevnorris>ugh. the deps/openssl upgrade in v0.10 is really noisy.
23:31:33  <trevnorris>tjfontaine: can we upgrade openssl in v0.12 just to simplify the v0.10 merge diff?
23:32:24  <tjfontaine>that seems more noisey, ideally that's how we handle upgrades where dependencies match
23:33:01  <trevnorris>more noisy in the commit log, but less noisy in the v0.10->v0.12 diff.
23:33:34  <tjfontaine>ya, but that's just part of the pain, ideally what happens is we merge early and merge often
23:33:51  <tjfontaine>we discussed this on the call this morning, if we can do that these merges hurt less
23:34:26  <trevnorris>yeah. i'm definitely for that.
23:35:46  <MI6>joyent/node: Shigeki Ohtsu merge-review * bddea12 : doc: Update LICENSE for zlib 1.2.8 (+2 more commits) - http://git.io/mFF6JA
23:36:37  <MI6>joyent/node: Trevor Norris merge-review2 * d28d392 : Merge branch 'v0.10' into merge-review2 - http://git.io/THRlkQ
23:39:27  * KennethWilkequit (Remote host closed the connection)
23:41:46  <cjihrig>trevnorris: just to confirm what you're seeing. 2 crypto failures and 1 child process failure
23:41:58  <cjihrig>looking at the child process failure
23:42:07  <tjfontaine>which crypto failures?
23:42:17  <trevnorris>cjihrig: hm. should only be one crypto failure now. but yeah. indutny knows what's going on there.
23:42:32  <trevnorris>the cp failure is the pain because of all the added tests in v0.10
23:42:40  <cjihrig>test-crypto-stream and test-tls-honorcipherorder-secureOptions
23:43:25  * quijotejoined
23:46:10  <trevnorris>the later is one indutny said he'd PR. the former is a simple test fix I thought I pushed up.
23:47:23  <MI6>joyent/node: Trevor Norris merge-review2 * 16ea117 : Merge branch 'v0.10' into merge-review2 - http://git.io/NjvTSw
23:47:48  <trevnorris>cjihrig: ^ test-crypto-stream shouldn't fail there. let me know if it does.
23:48:37  * quijotequit (Ping timeout: 272 seconds)
23:50:06  <tjfontaine>which subtest is failing in the honorcipherorder test?
23:52:33  * rendarquit (Quit: Leaving)
23:54:37  <trevnorris>indutny: ping
23:58:31  <cjihrig>trevnorris: have to run out for a little. i'll be able to look more when i get back
23:58:51  <MI6>joyent/node: Trevor Norris merge-review2 * 09b06be : Merge branch 'v0.10' into merge-review2 - http://git.io/tHfZsg
23:59:15  <trevnorris>cjihrig: thanks. i've cleaned things up a bit more and will continue to take a look.
23:59:37  <trevnorris>tjfontaine: will jenkins automatically test that branch?