00:02:56  <creationix>rphillips: ok, tiny build is done, the regular builds will be done soon assuming there are no problems
00:14:52  <creationix>rphillips: looks like all the windows builds finished. Thanks for the help today!
00:15:17  <creationix>https://github.com/luvit/luvi/releases/tag/v2.1.2
00:15:54  <rphillips>nice. thanks for building those
00:16:04  <rphillips>i'll try them out in a bit
00:35:38  <rphillips>yay
00:35:40  <rphillips>it works now
00:47:58  <rphillips>wow. 4.8 mb on osx
00:57:12  * DarkGodquit (Ping timeout: 272 seconds)
01:04:42  * travis-cijoined
01:04:43  <travis-ci>luvit/luvi#611 (fixes/windows_asm_fixes - d2b59f3 : Ryan Phillips): The build passed.
01:04:43  <travis-ci>Change view : https://github.com/luvit/luvi/compare/09676cdf0670...d2b59f3b8de8
01:04:43  <travis-ci>Build details : https://travis-ci.org/luvit/luvi/builds/68895476
01:04:43  * travis-cipart
01:04:43  * 18VAAFCR5joined
01:04:44  <18VAAFCR5>luvit/luvi#614 (release - a3b0d6e : Ryan Phillips): The build passed.
01:04:44  <18VAAFCR5>Change view : https://github.com/luvit/luvi/compare/43eca778a94e...a3b0d6e948af
01:04:44  <18VAAFCR5>Build details : https://travis-ci.org/luvit/luvi/builds/68901762
01:04:44  * 18VAAFCR5part
01:09:03  <rphillips>creationix: i think we need to tag a new release
01:09:22  <rphillips>the release branch is not pointing to a tag, which messes with the semver version within luvi
01:18:14  * travis-cijoined
01:18:15  <travis-ci>luvit/luvi#613 (master - a3b0d6e : Ryan Phillips): The build passed.
01:18:15  <travis-ci>Change view : https://github.com/luvit/luvi/compare/43eca778a94e...a3b0d6e948af
01:18:15  <travis-ci>Build details : https://travis-ci.org/luvit/luvi/builds/68896511
01:18:15  * travis-cipart
01:44:52  <rphillips>tagged 2.1.3
01:44:56  <rphillips>just needs the binaries
01:53:18  * travis-cijoined
01:53:19  <travis-ci>luvit/luvi#615 (master - b90d183 : Ryan Phillips): The build passed.
01:53:19  <travis-ci>Change view : https://github.com/luvit/luvi/compare/a3b0d6e948af...b90d183b11de
01:53:19  <travis-ci>Build details : https://travis-ci.org/luvit/luvi/builds/68906321
01:53:19  * travis-cipart
01:54:21  <rphillips>got darwin and source uploaded
01:57:08  <rphillips>building on pi1
02:06:35  * travis-cijoined
02:06:36  <travis-ci>luvit/luvi#617 (release - b90d183 : Ryan Phillips): The build was broken.
02:06:36  <travis-ci>Change view : https://github.com/luvit/luvi/compare/a3b0d6e948af...b90d183b11de
02:06:36  <travis-ci>Build details : https://travis-ci.org/luvit/luvi/builds/68906555
02:06:36  * travis-cipart
02:15:17  * travis-cijoined
02:15:18  <travis-ci>luvit/luvi#617 (release - b90d183 : Ryan Phillips): The build passed.
02:15:18  <travis-ci>Change view : https://github.com/luvit/luvi/compare/a3b0d6e948af...b90d183b11de
02:15:18  <travis-ci>Build details : https://travis-ci.org/luvit/luvi/builds/68906555
02:15:18  * travis-cipart
03:03:45  <creationix>rphillips: ok, started the builds
03:04:25  <rphillips>pi1 is almost done
03:04:26  <rphillips>thanks
03:43:38  * dan336joined
03:49:44  * dan336quit (Quit: Leaving.)
04:11:47  * SkyRocknRolljoined
05:20:56  * SkyRocknRollquit (Read error: Connection reset by peer)
05:50:41  * SkyRocknRolljoined
06:26:45  * SkyRocknRollquit (Ping timeout: 256 seconds)
06:38:50  * SkyRocknRolljoined
07:18:46  * DarkGodjoined
07:56:01  <sousoux>rphillips: yes. now exits properly. I sill have my http mods to check in and one problem in core where nil listeners can get called.
08:44:19  * sousouxquit (Ping timeout: 276 seconds)
09:09:13  * sousouxjoined
10:50:50  * torporjoined
11:09:40  * torporquit (Quit: Leaving.)
11:24:46  * torporjoined
12:13:41  * SkyRocknRollquit (Remote host closed the connection)
13:11:54  * SkyRocknRolljoined
13:20:57  <rphillips>bumped lit
13:40:34  * torporquit (Quit: Leaving.)
13:45:32  * torporjoined
13:52:23  <rphillips>good morning
14:03:50  <rphillips>we are having the worst luck bumping lua-openssl
14:22:47  * travis-cijoined
14:22:48  <travis-ci>luvit/luvi#620 (master - 68eeee1 : Ryan Phillips): The build has errored.
14:22:48  <travis-ci>Change view : https://github.com/luvit/luvi/compare/b90d183b11de...68eeee1e87f9
14:22:48  <travis-ci>Build details : https://travis-ci.org/luvit/luvi/builds/68980869
14:22:48  * travis-cipart
14:25:15  <creationix>rphillips: still ssl issues?
14:26:09  <rphillips>creationix: just that lua-openssl got reverted
14:26:15  <rphillips>or didn't get bumped
14:26:18  <rphillips>:/
14:26:31  <rphillips>we will get it on the next pass
14:27:15  * travis-cijoined
14:27:16  <travis-ci>luvit/luvit#2284 (2.2.3 - 4263001 : Ryan Phillips): The build passed.
14:27:16  <travis-ci>Change view : https://github.com/luvit/luvit/compare/2.2.3
14:27:16  <travis-ci>Build details : https://travis-ci.org/luvit/luvit/builds/68984464
14:27:16  * travis-cipart
14:27:49  <rphillips>got everything bumped
14:27:56  * dan336joined
14:28:22  <rphillips>the test case for lua-openssl seems to be hanging as well
14:28:47  <rphillips>tried a manually compiled luvi with the luvit test cases
15:03:49  * torporquit (Read error: Connection reset by peer)
15:04:00  * torporjoined
15:15:43  * torporquit (Quit: Leaving.)
15:27:52  * torporjoined
15:34:49  <rphillips>https://github.com/awslabs/s2n
15:44:40  <creationix>rphillips: s2n looks really nice
15:44:58  <rphillips>yeah, it does
15:49:53  <rphillips>sousoux: latest release should ease some of the ssl cpu usage
15:50:02  <sousoux>hi
15:50:12  <sousoux>I'm having a lot of issues with https
15:50:22  <sousoux>and keep-alive
15:50:35  <sousoux>the socket seems to stall
15:50:48  <sousoux>seems very timing dependend
15:50:51  <sousoux>t
15:51:53  <sousoux>I implemented keep alive in http which works great on http connections but when I move to https (on my embedded device not on my PC) the sockets from the browser stall after several requests.
15:52:04  <sousoux>I'm trying to track it down
15:52:28  <sousoux>rphillips: But any ideas welcome
15:52:40  <rphillips>luvit or luvi?
15:52:42  <rphillips>version?
15:52:44  <rphillips>test case?
15:56:04  <sousoux>luvit 2.2.2
15:57:14  <sousoux>To get a test case I have to rip the whole thing out of my application and specifically write one. I'm not going to do that for now. I thought that maybe you had experienced something similar.
15:59:57  <dan336>sousoux: are you making any HEAD requests to the webserver on the embedded device?
16:02:55  <rphillips>hmm
16:03:18  <dan336>I had an issue last week where making a HEAD request to a server which supported keepalive would cause the http-codec to be in the wrong state and hang indefinitely. creationix made the change in coro-http, but I don't see the change in luvit yet.
16:03:59  <rphillips>i wonder if that patch got ported over to luvit
16:04:27  <dan336>I checked and didn't see it.
16:06:23  <sousoux>dan336: no, no HEADs and it bugs out way earlier than that. I just do not see any more data from the socket. I'm also seeing some spurious stops ... the app exits with no message whatsoever.
16:06:50  <sousoux>I'm starting to suspect a build issue on the embedded platform. Probably openssl
16:06:55  <rphillips>which version of luvi?
16:07:51  <sousoux>677ed00dc861de3862b5ee4d6b1be572711f5f75
16:07:52  <rphillips>try using today's release
16:08:00  <sousoux>will do
16:08:03  <rphillips>luvit: 2.2.3, and luvi 2.1.3
16:08:49  <sousoux>I have to build bits separately to get everything working. I do libuv, luv, then luvi. luvit is pulled in by my app
16:09:02  <rphillips>hmm
16:09:14  <rphillips>we migrated to enabling asm optimizations in openssl
16:09:26  <sousoux>when?
16:09:45  <rphillips>today/yesterday
16:09:57  <rphillips>2.1.2 should probably not be used on windows
16:10:08  <sousoux>This is not windows!
16:10:20  <rphillips>yeah, but the other builds have asm optimizations now
16:10:22  <sousoux>linux on a freescale iMX28
16:10:57  <rphillips>make regular will still build without optimizations
16:13:38  <rphillips>sousoux: there was also a patch the other day gracefully shutdown a ssl socket
16:13:43  <rphillips>to gracefully*
16:17:08  <rphillips>siege -b -c5 -r200 -H "Connection: Keep-Alive"
16:17:41  <rphillips>seems to be working with luvit 2.2.3
16:34:57  <sousoux>hmm Unable to checkout '80896f0f3914dffe4228d8d8dd7e2c68e329c88c' in submodule path 'deps/lua-openssl'
16:39:01  <rphillips>hmm
16:39:08  <rphillips>master?
16:39:53  <creationix>dan336: right, as far as I know, luvit's http client still has the HEAD response bug
16:40:25  <rphillips>sousoux: hmm. got the same thing
16:40:36  <creationix>I don't know that code nearly as well and node style streams scare me.
16:41:01  <rphillips>sousoux: try again
16:41:11  <rphillips>i pushed to the luvit/lua-openssl.git repo
16:41:40  <creationix>(and the irony is I was one of the original authors of node streams back in summer 2010)
16:43:55  <rphillips>sousoux: that pin should be b5c7221cdde6263d4ecf8cba19a0b6eed511a9e2
16:43:57  <rphillips>fixed
16:45:32  <sousoux>rphillips: thanks. I pushed the 'rebuild all button' so I'm off for a tea!
16:49:20  <sousoux>rphillips: it made it past that
17:13:50  * travis-cijoined
17:13:51  <travis-ci>luvit/luvi#621 (master - c9db6bf : Ryan Phillips): The build passed.
17:13:51  <travis-ci>Change view : https://github.com/luvit/luvi/compare/68eeee1e87f9...c9db6bf9768c
17:13:51  <travis-ci>Build details : https://travis-ci.org/luvit/luvi/builds/69006427
17:13:51  * travis-cipart
17:14:37  * torporquit (Quit: Leaving.)
17:37:07  * sousouxquit (Ping timeout: 256 seconds)
17:49:27  * sousouxjoined
18:17:36  <creationix>rphillips: so do we need yet another luvi build today?
18:18:03  <rphillips>creationix: nope
18:18:05  <rphillips>we are good
18:18:27  <creationix>cool
18:20:51  <rphillips>creationix: have an arm9 box?
18:21:05  <creationix>no, why do you ask?
18:22:03  <creationix>well, I have several 32-bit microcontrollers, not sure which ones are arm
18:22:42  <creationix>hmm, the spark is Cortex M3
18:22:54  <creationix>The ESP is Expressif (not arm at all)
18:27:11  <rphillips>sousoux seems to be running on an arm9: iMX28
18:33:24  <creationix>rphillips: hmm, this is the cheapest and most accessable arm9 board I can find. Looks like a i.MX6 clone of the rPI http://www.newegg.com/Product/Product.aspx?Item=N82E16813455001&nm_mc=KNC-GoogleAdwords-PC&cm_mmc=KNC-GoogleAdwords-PC-_-pla-_-Embedded+Solutions-_-N82E16813455001&gclid=CInDp_WGuMYCFQiEaQodYb8Dvg&gclsrc=aw.ds
18:34:40  <rphillips>hmm. they have an i2
18:36:19  <rphillips>i2 is 64bit though
18:36:31  <rphillips>sousoux: are you running that board in 32 or 64bit mode?
18:37:19  <rphillips>appears to be 32bit
21:04:30  <rje>rphillips, creationix: https://github.com/virgo-agent-toolkit/rackspace-monitoring-agent/pull/762
21:05:25  <sousoux>rphilips: and creationix - I can get you both an iMX28 device
21:05:35  <creationix>sousoux: that would be awesome
21:05:36  <sousoux>cloudgate - www.option.com
21:05:48  <sousoux>Along with a dev build for it
21:05:56  <sousoux>of luvit
21:06:02  <rje>imx28 is pretty cool
21:06:18  <sousoux>I'm still having issues with https. I gave up for today. needed a clear mind
21:06:32  <sousoux>I'm starting to suspect the return path to the browser
21:06:46  <sousoux>I think it is hanging around waiting for something
21:07:13  <sousoux>Please send delivery address to m_dot_croome_at_option_dot_com
21:07:34  <sousoux>imx28 is very cheap!
21:07:59  <sousoux>much cheaper than the MX6
21:08:06  <creationix>sousoux: is that your email?
21:08:15  <sousoux>yes. professional one
21:09:52  <creationix>sent, thanks!
21:11:04  <rphillips>sent as well
21:11:06  <rphillips>thanks!
21:11:17  <sousoux>It definitely seems related to openssl. Unfortunately I'm using the build in luvi rather than our own since there were a lot of missing modules for lua_openssl. I plan on switching to ours. It may be that I have some compile options wrong.
21:11:36  <rphillips>sousoux: still building with `make regular`
21:11:38  <sousoux>But it also could be network related
21:11:38  <rphillips>?
21:12:05  <sousoux>I split the build up. We use buildroot and it is difficult to control subprojects with that
21:12:10  <rphillips>gotcha
21:12:25  <sousoux>so yes. more or less!
21:12:59  <sousoux>I also have some patches against libuv to allow interrupt polling but I don't think they are getting in the way
21:13:14  <sousoux>I'll track it down tomorrow
21:14:12  <sousoux>We are also using uclibc which generates its own series of issues.
21:14:39  <sousoux>It is a very lightweight platform which is why I chose luvit
21:16:04  <sousoux>creationix: I did a mod on lit to allow make to take an extra argument. Path to luvi. It is a trivial mod but would be useful for people doing cross platform builds.
21:16:30  <creationix>sousoux: you mean the luvi to embed?
21:16:35  <sousoux>yes
21:16:45  <creationix>sounds good. Send a PR if you want it included
21:16:55  <sousoux>I saw that it can download from url but that doesnt really work for me
21:17:14  <sousoux>ok
21:17:32  <creationix>we could overload the param. Just use uv.fs_access first to see if the path is a local file
21:17:35  <creationix>otherwise assume url
21:18:03  <sousoux>But that means modifying the meta data and I build for different platforms
21:18:18  <sousoux>I can do it but argument seemed simpler
21:18:25  <creationix>whichever
21:18:42  <sousoux>k
21:18:43  <creationix>using musl is a pain with buidtroot right?
21:18:51  <creationix>*buildroot
21:19:26  <sousoux>frankly we want to get to glibc ASAP
21:19:42  <sousoux>all the alternative libc's are a pain
21:20:17  <sousoux>I think our firmware guys looked at musl and rejected
21:20:35  <sousoux>But 4K automated tests to port is ...
21:21:35  <sousoux>not really port but suffer
21:23:35  <sousoux>One other patch I have outstanding is in core.lua. In the Emitter on handler there is a condition where a nil emitter can be run if a previous handler clears a listener.
21:24:04  <sousoux>needs a check to see if the handler is nil in the loop that runs all the handlers
21:25:59  <creationix>sousoux: I remember seeing that in my notifications, but I can’t find it now
21:26:23  <sousoux>It was mixed in the the http stuff so I deleted it
21:26:37  <sousoux>Its pretty obvious when you look at it
21:26:55  <sousoux>If you want to fix it before I get this solved
21:29:46  <creationix>sousoux: I see now. I clear out the empty slots after all handlers have run
21:29:55  <creationix>but that’s too late for inter-handler changes
21:31:40  <creationix>this should do it https://github.com/luvit/luvit/commit/3256bbf07985f25ea651cdfb5d764b1114b7fd11
21:35:08  * travis-cijoined
21:35:09  <travis-ci>luvit/luvit#2285 (master - 3256bbf : Tim Caswell): The build passed.
21:35:10  <travis-ci>Change view : https://github.com/luvit/luvit/compare/42630010a25c...3256bbf07985
21:35:10  <travis-ci>Build details : https://travis-ci.org/luvit/luvit/builds/69047443
21:35:10  * travis-cipart
22:02:01  * dan336quit (Read error: No route to host)
22:02:15  * dan336joined
22:58:26  * dan336quit (Quit: Leaving.)