01:44:51  * DarkGodquit (Quit: Leaving)
02:14:24  * joconnorquit (Ping timeout: 245 seconds)
04:37:48  * imzyxwvujoined
05:04:44  * imzyxwvuquit (Ping timeout: 244 seconds)
05:12:11  * a__quit (Ping timeout: 245 seconds)
05:25:28  * a__joined
08:19:08  * hdmsquit (Quit: hdms)
08:25:22  * DarkGodjoined
09:16:11  * a_lequit (Ping timeout: 250 seconds)
09:54:05  * a_lejoined
11:29:06  * imzyxwvujoined
11:46:24  <imzyxwvu>https://ci.appveyor.com/project/creationix/luvi/build/artifacts(the url found in luvit's README.md) shows "Project not found or access denied."
13:24:48  <rphillips>imzyxwvu: thanks... fixed
13:26:15  * travis-cijoined
13:26:15  <travis-ci>luvit/luvit#1532 (luvi-up - 36de919 : Ryan Phillips): The build passed.
13:26:15  <travis-ci>Change view : https://github.com/luvit/luvit/compare/9af210e56a61...36de919c43f1
13:26:15  <travis-ci>Build details : http://travis-ci.org/luvit/luvit/builds/49596678
13:26:15  * travis-cipart
14:04:22  * hdmsjoined
15:14:31  <rphillips>https://github.com/luvit/luvit/pull/604
15:14:36  <rphillips>just a cleanup
15:15:45  * travis-cijoined
15:15:45  <travis-ci>luvit/luvit#1533 (cleanup/child_process - 62aa3f2 : Ryan Phillips): The build failed.
15:15:45  <travis-ci>Change view : https://github.com/luvit/luvit/commit/62aa3f296f82
15:15:45  <travis-ci>Build details : http://travis-ci.org/luvit/luvit/builds/49611518
15:15:45  * travis-cipart
15:20:06  <rphillips>hmm
15:22:09  <rphillips>fixed
15:22:51  <imzyxwvu>what are the differences between luvi-tiny and luvi?
15:23:42  * travis-cijoined
15:23:42  <travis-ci>luvit/luvit#1536 (cleanup/child_process - 3bf71ff : Ryan Phillips): The build was fixed.
15:23:42  <travis-ci>Change view : https://github.com/luvit/luvit/compare/62aa3f296f82...3bf71ff5b1dd
15:23:42  <travis-ci>Build details : http://travis-ci.org/luvit/luvit/builds/49612602
15:23:42  * travis-cipart
15:24:17  <rphillips>imzyxwvu: luvi-tiny doesn't have openssl and probably zlib
15:24:44  <imzyxwvu>rphillips: got it. thanks.
15:27:50  * dan336joined
16:01:35  <imzyxwvu>lua c libraries on windows need to link to lua apis. now luajit is static linked into luvi and the lua apis are not exported.
16:01:39  <imzyxwvu>so why not try to add /DLUA_BUILD_AS_DLL when compiling luajit static library? then i can simply link the lua c libraries against the exe file.
16:09:41  * imzyxwvuquit (Ping timeout: 245 seconds)
16:12:44  * DarkGodquit (Ping timeout: 252 seconds)
16:26:03  * DarkGodjoined
16:28:44  <rphillips>That might work
16:29:30  <rphillips>creationix: would be slick to use the ace editor on the Luvi and Luvit playground to editor and run code
16:32:56  * UniOnjoined
17:04:44  <creationix>yeah, making a playground would be fun
17:05:06  <creationix>tricky to host though
17:05:22  <creationix>it would be easier to add luvit to something like c9.io or one of the other alternativesa
17:08:02  * cledevjoined
17:29:25  <creationix>alright, I’m moved back into my office now
17:29:27  <creationix>it’s cold out here
17:34:56  * joconnorjoined
17:41:44  * joconnorquit (Read error: Connection reset by peer)
17:43:59  * joconnorjoined
17:48:27  <rje>creationix: have time for a g+/vidyo?
17:49:34  <creationix>rje: sure
17:56:13  * cledevquit (Ping timeout: 264 seconds)
18:01:50  * DarkGodquit (Ping timeout: 265 seconds)
18:03:33  * travis-cijoined
18:03:33  <travis-ci>luvit/luvit#1537 (cleanup/child_process - 9d404b1 : Ryan Phillips): The build passed.
18:03:33  <travis-ci>Change view : https://github.com/luvit/luvit/compare/3bf71ff5b1dd...9d404b1c624f
18:03:33  <travis-ci>Build details : http://travis-ci.org/luvit/luvit/builds/49634073
18:03:33  * travis-cipart
18:11:17  * hdmsquit (Quit: hdms)
18:35:46  <creationix>rphillips: what if lit defaulted to installing packages without the maintainer’s username
18:35:55  <creationix>so installing creationix/git would install to modules/git
18:36:09  <rphillips>i think that would be preferrable
18:36:16  <rphillips>would ease the pain on name conflicts
18:37:05  <creationix>and if there were conflicts of the same name by different maintainers?
18:37:21  <rphillips>perhaps fix upstream
18:37:46  <creationix>well, I mean they might be entirely different packages
18:37:58  <rphillips>true... maybe alias to a local directory
18:38:01  <creationix>creationix/git is git core, but foo/git might be a wrapper around cli git
18:38:27  <creationix>aliasing is easy enough to implement. it’s the deep dependencies I’m worried about
18:44:10  <rch>i really like how npm does it
18:56:15  * a_lequit (Remote host closed the connection)
18:56:55  * a_lejoined
18:58:56  <creationix>rch: I don’t, it gets out of hand real quick in medium to large projects
18:59:08  <rch>hmm i guess i haven't seen that happen
18:59:16  <creationix>I kinda like how npm used to do it. It would only install deep if there were conflicts at the top level
19:05:50  * DarkGodjoined
19:09:03  <creationix>rch: rphillips: what if the require could read directly from the lit database?
19:09:43  <rphillips>hmm. how would that get packaged into the binary?
19:09:56  <creationix>that’s my concern.
19:10:17  <rch>heh
19:10:39  <creationix>there are two use cases. They are roughly analgous to static and dynamic linking
19:10:41  <rch>crazy idea: optionally structure the bundle as a litlike cas
19:11:00  <creationix>rch: great idea
19:11:14  <rch>heh
19:11:17  <creationix>so store a subset of the db internally in the binary
19:11:26  * rchwondered if that was tim's kind of crazy
19:11:34  <creationix>lol
19:12:57  <creationix>I guess we should only worry about the static case for now
19:13:02  <creationix>optimize later
19:13:16  <creationix>(optimize for space later, optimize for simplicity now)
19:14:29  <rphillips>+1
19:15:08  <creationix>how black-box are third-party dependencies
19:15:16  <creationix>storing them in a cas makes them much more opaque
19:17:27  <creationix>ok, so by default I think dependencies will be aliased to the package name only in code so “git” for “creationix/git@1.2.3”
19:17:48  <creationix>and require will be embedded with the alias list in addition to the caller’s path
19:17:53  * travis-cijoined
19:17:53  <travis-ci>luvit/luvit#1538 (luvi-up - f9ef3f0 : Ryan Phillips): The build passed.
19:17:53  <travis-ci>Change view : https://github.com/luvit/luvit/compare/36de919c43f1...f9ef3f069e4c
19:17:53  <travis-ci>Build details : http://travis-ci.org/luvit/luvit/builds/49643162
19:17:53  * travis-cipart
19:18:43  <creationix>so when I require(‘git’) with a dependency to “creationix/git@1.2.3” it will search in lit (local and global just like the “modules” paths) for such a package
19:19:48  <creationix>though there will probably be only two databases. The user one defined in the user’s lit config and the local one embedded in the bundle at “bundle:lit.db” or something
19:20:13  <rphillips>gotcha
19:20:26  <creationix>name conflicts never happen in the local files that are part of a package. It’s deep dependencies that have the issue
19:20:47  <creationix>and deep dependencies are 99% 3rd part libs and should be pulled from the lit cas
19:43:39  <rphillips>creationix: still seeing the lua-openssl crash?
19:44:16  <creationix>haven’t tried today
19:44:54  <creationix>that first link I sent you a while back seems to reproduce it pretty reliably. Let me see if I can find what revision that was
19:45:34  <creationix>I think this revision https://github.com/creationix/voodoo-spark-lit/tree/5ecc208ff3ea1e6b8ee1d359e79a1dc7a3342bea
19:47:56  <rphillips>hmm. bad request
19:48:09  <creationix>bad request is expected, it’s invalid according to the API
19:48:17  <creationix>but on my machine is triggers the issue a lot
19:48:21  <creationix>make sure to checkout that revision
19:48:22  <rphillips>hmm. k
19:49:10  <rphillips>creationix: linux?
19:49:14  <creationix>osx
19:49:20  <rphillips>thanks
19:49:48  <creationix>in lldb it acts slightly different, but I seem to get a lot of BAD_ACCESS errors in x509 stuff
19:50:21  <creationix>https://gist.github.com/creationix/f467e77c3f228f59e2d6
19:55:02  <rphillips>i saw the same thing the other day, but now running vsl every 5 seconds it doesn't crash
19:55:19  <rphillips>is the luvi binary the latest within luvi-binaries?
19:55:30  <creationix>probably not, what should it be
19:55:53  <creationix>oh, luvi-binaries, let me try that
19:56:06  <rphillips>$ md5 Darwin_x86_64/luvi
19:56:13  <rphillips>MD5 (Darwin_x86_64/luvi) = 57d72006e51b8f088fca87240b3005e5
19:56:18  <rphillips>that is the binary i'm using
19:57:11  <creationix>ok, using that binary I still see lots of crashes
19:57:18  <rphillips>hmm
19:57:50  <creationix>rphillips: https://gist.github.com/creationix/d0ae6dc513bdcec9cade
19:58:58  <creationix>a pretty even mix of segfaults and invalid frees running direct
20:00:16  <rphillips>and I saw the same thing a few days ago... weird
20:00:50  <creationix>I think lldb can capture both types too https://gist.github.com/creationix/d0ae6dc513bdcec9cade#file-lldb-txt
20:00:52  <rphillips>https://www.evernote.com/shard/s9/sh/183c9829-9fd9-4dc9-b1c9-e5490363232a/1a6dee3a991374d5b4a6d148e911f84f
20:01:44  <creationix>same git hash, same binary md5, same platform
20:01:45  <creationix>weird
20:02:09  <rphillips>i did reboot... i wonder if you have any pending osx updates
20:02:49  <creationix>I installed a couple updates on the 2nd, not sure if I’ve rebooted since then. I thought it made me reboot
20:02:52  <creationix>I can reboot again to see
20:03:05  <rphillips>just for sanity
20:04:58  <rphillips>so weird
20:05:04  <rphillips>i verified the tree hash
20:07:50  <creationix>yep, rebooting doesn’t change anything
20:08:44  <rphillips>and I take it that `make test` works in the luvit tree?
20:09:23  <creationix>yep, even including the test-https.lua you wrote
20:09:31  <rphillips>man
20:09:40  <rphillips>i'll try reproducing it some more
20:10:09  <rphillips>i did see it at one point, and it was consistent
20:10:30  <creationix>my computer is using 5.68Gb according to activity monitor
20:10:38  <creationix>(out of 8Gb total)
20:11:32  <rphillips>6.2 Gb here
20:12:56  <creationix>➜ luvit git:(luvi-up) ✗ uname -a
20:12:57  <creationix>Darwin Tims-MacBook-Pro.local 14.1.0 Darwin Kernel Version 14.1.0: Mon Dec 22 23:10:38 PST 2014; root:xnu-2782.10.72~2/RELEASE_X86_64 x86_64
20:13:02  * joconnorquit (Remote host closed the connection)
20:13:55  <creationix>rphillips: try collecting manually in the test file maybe?
20:14:01  * joconnorjoined
20:14:19  <creationix>it’s during gc that triggers the issues
20:14:32  * joconnorquit (Read error: Connection reset by peer)
20:15:02  * joconnorjoined
20:15:23  * joconnorquit (Read error: Connection reset by peer)
20:15:54  * joconnorjoined
20:16:25  * joconnorquit (Read error: Connection reset by peer)
20:16:50  * joconnorjoined
20:17:16  * joconnorquit (Read error: Connection reset by peer)
20:17:44  * joconnorjoined
20:18:10  * joconnorquit (Read error: Connection reset by peer)
20:18:39  * joconnorjoined
20:19:05  * joconnorquit (Read error: Connection reset by peer)
20:19:34  * joconnorjoined
20:20:02  * joconnorquit (Read error: Connection reset by peer)
20:20:31  * joconnorjoined
20:20:58  * joconnorquit (Read error: Connection reset by peer)
20:21:28  * joconnorjoined
20:22:05  * joconnorquit (Read error: Connection reset by peer)
20:22:23  * joconnorjoined
20:22:46  * joconnorquit (Read error: Connection reset by peer)
20:23:17  * joconnorjoined
20:23:48  * joconnorquit (Read error: Connection reset by peer)
20:24:12  * joconnorjoined
20:24:39  * joconnorquit (Read error: Connection reset by peer)
20:25:07  * joconnorjoined
20:25:35  * joconnorquit (Read error: Connection reset by peer)
20:25:57  <creationix>rphillips: what timing do you get with time curl -I https://api.spark.io/
20:26:04  * joconnorjoined
20:26:18  <creationix>I can’t seem to ping it, they must have disabled ping
20:26:31  * joconnorquit (Read error: Connection reset by peer)
20:26:59  * joconnorjoined
20:27:26  * joconnorquit (Read error: Connection reset by peer)
20:27:53  * joconnorjoined
20:28:22  * joconnorquit (Read error: Connection reset by peer)
20:28:49  * joconnorjoined
20:29:15  * joconnorquit (Read error: Connection reset by peer)
20:30:34  * joconnorjoined
20:31:04  * joconnorquit (Read error: Connection reset by peer)
20:31:30  * joconnorjoined
20:31:57  * joconnorquit (Read error: Connection reset by peer)
20:33:43  * joconnorjoined
20:34:11  * joconnorquit (Read error: Connection reset by peer)
20:34:39  * joconnorjoined
20:35:04  * joconnorquit (Read error: Connection reset by peer)
20:35:36  * joconnorjoined
20:36:00  * joconnorquit (Read error: Connection reset by peer)
20:37:44  * joconnorjoined
20:38:12  * joconnorquit (Read error: Connection reset by peer)
20:39:42  <rphillips>curl -I https://api.spark.io/ 0.04s user 0.01s system 7% cpu 0.637 total
20:40:47  <creationix>curl -I https://api.spark.io/ 0.03s user 0.01s system 10% cpu 0.334 total
20:41:14  <creationix>I know when I added more code before it made the crashes go away. Maybe it has to be super fast
21:04:31  * joconnorjoined
21:04:55  * joconnorquit (Read error: Connection reset by peer)
21:05:22  * hdmsjoined
21:30:45  * joconnorjoined
21:31:11  * joconnorquit (Read error: Connection reset by peer)
21:31:41  * joconnorjoined
21:32:20  * joconnorquit (Read error: Connection reset by peer)
21:32:33  * joconnorjoined
21:33:07  * joconnorquit (Read error: Connection reset by peer)
21:34:26  * joconnorjoined
21:34:47  * joconnorquit (Read error: Connection reset by peer)
21:36:32  * joconnorjoined
21:36:57  * joconnorquit (Read error: Connection reset by peer)
21:37:23  * joconnorjoined
21:38:17  * joconnorquit (Read error: Connection reset by peer)
21:39:09  * joconnorjoined
21:39:40  * joconnorquit (Read error: Connection reset by peer)
21:39:42  <rphillips>hmm
21:39:57  <rphillips>seems like a localhost server would be fast enough to show the crash
21:40:05  * joconnorjoined
21:40:36  * joconnorquit (Read error: Connection reset by peer)
21:41:52  * joconnorjoined
21:42:25  * joconnorquit (Read error: Connection reset by peer)
21:42:45  <creationix>that’s a good idea
21:42:52  * joconnorjoined
21:43:21  * joconnorquit (Read error: Connection reset by peer)
21:44:38  * joconnorjoined
21:45:06  * joconnorquit (Read error: Connection reset by peer)
21:45:33  * joconnorjoined
21:45:57  * joconnorquit (Read error: Connection reset by peer)
21:46:19  <rphillips>reproduced it on linux
21:46:25  <creationix>yay
21:47:14  * joconnorjoined
21:47:50  * joconnorquit (Read error: Connection reset by peer)
21:48:11  * joconnorjoined
21:48:45  * joconnorquit (Read error: Connection reset by peer)
21:49:55  * joconnorjoined
21:50:21  * joconnorquit (Read error: Connection reset by peer)
21:50:52  * joconnorjoined
21:51:20  * joconnorquit (Read error: Connection reset by peer)
21:52:36  * joconnorjoined
21:53:07  * joconnorquit (Read error: Connection reset by peer)
21:55:39  * joconnorjoined
21:56:04  * joconnorquit (Read error: Connection reset by peer)
21:56:24  * joconnorjoined
21:56:39  * cledevjoined
22:18:24  <rphillips>https://github.com/zhaozg/lua-openssl/pull/44
22:18:30  <rphillips>i think this makes sense
22:18:48  <rphillips>the x509 store is managed by lua, so we always want the lua gc to cleanup after it
22:19:20  <creationix>that would explain the double frees
22:19:33  <rphillips>the reference count is sometimes '1' and that is why it's sorta intermiddent on different platforms
22:20:05  <rphillips>the internal openssl ssl state doesn't check the reference count, it just frees it
22:21:28  <creationix>so it fixed the issue for you best you can tell right?
22:22:52  <rphillips>damn. well it doesn't crash
22:23:09  <rphillips>i just checked the openssl code and it creates a cert_store by default
22:26:38  * a_lequit (Remote host closed the connection)
22:27:11  * a_lejoined
22:34:57  * cledevquit (Ping timeout: 264 seconds)
22:34:59  <rphillips>ok. updated that PR
22:35:05  <rphillips>we will see what george says
22:36:25  <creationix>interesting
22:36:42  <creationix>I’m almost done with adding alias support to lit
22:36:55  <creationix>also the default alias will be the name minus the first segment (the maintainer’s name)
22:37:37  <rphillips>cool
22:38:43  <rphillips>luvit tests pass as well
22:39:19  <rphillips>the memory transfers ownership to the ssl context... meh
22:50:28  <creationix>yeah, memory ownership is the hardest part of C
22:50:35  <creationix>otherwise, it’s not that bad a language
23:13:41  <creationix>rphillips: ok, lit now uses flatter namespaces for packages
23:13:58  <creationix>this is a breaking change since internal requires all need to change to “foo” instead of “author/foo”
23:14:21  <creationix>and any lit packages that depend on other lit packages need to be modified and bumped and re-published
23:14:29  <creationix>I republished the ones used internally to lit
23:15:01  <creationix>also you can alias things in your dependencies by using a string key instead of the default numerical index
23:15:09  <creationix>but that’s not a breaking change since nobody did that yet
23:17:08  * a_lequit (Read error: Connection reset by peer)
23:17:35  * a_lejoined
23:48:12  * UniOnquit (Quit: Leaving)