00:05:04  * isaacsquit (Read error: Connection reset by peer)
00:05:18  * isaacsjoined
00:10:54  * pieternjoined
00:15:26  * isaacs_joined
00:15:26  * isaacsquit (Read error: Connection reset by peer)
00:15:27  * isaacs_changed nick to isaacs
00:18:39  * creationix|workquit (Remote host closed the connection)
00:19:23  * AvianFluquit (Ping timeout: 252 seconds)
00:21:51  * AvianFlujoined
00:23:52  <CIA-53>libuv: Ryan Dahl * r2c7e8bb / (include/uv.h src/unix/error.c src/win/error.c): Map EAI_NONAME to ENOENT - http://git.io/iSJXzA
00:35:19  <sh1mmer>https://github.com/joyent/node/issues/760 there seems like a legitimate API regression here from 0.4.1
00:36:04  <ryah>sh1mmer: we're matching API with v0.4.12
00:36:06  <sh1mmer>I can fix it if strings should be valid a format for "query" rather than an object
00:36:26  <ryah>that would have been a concern if we didn't ignore it
00:36:33  <sh1mmer>ryah: didn't work for me in 0.4.12 or HEAD
00:36:33  <ryah>now the new API is the correct API :)
00:36:40  <sh1mmer>ok
00:36:46  <sh1mmer>so string are now considered invalid
00:37:07  <sh1mmer>I'm asking because it was a regression from 0.4.1
00:44:29  * isaacsquit (Quit: isaacs)
00:44:51  * sh1mmerquit (Quit: sh1mmer)
00:51:03  * AvianFlu_joined
00:54:01  * AvianFluquit (Ping timeout: 260 seconds)
01:03:19  <CIA-53>node: Ryan Dahl * rfb3cccc / (3 files in 3 dirs): Upgrade libuv to 2c7e8bb - http://git.io/wUhvVg
01:03:20  <CIA-53>node: Ryan Dahl * rc83dda8 / (doc/api/dns.markdown lib/dns.js src/cares_wrap.cc): dns.lookup uses cares_wrap::GetAddrInfo - http://git.io/8VS5wQ
01:05:11  * AvianFlu_changed nick to AvianFlu
01:05:33  <CIA-53>node: Ben Noordhuis * r30deb31 / (3 files in 3 dirs): WIP - http://git.io/hrVjzw
01:05:33  <CIA-53>node: Ben Noordhuis * rfa00e09 / benchmark/http_pump.js : http pump benchmark - http://git.io/IxE2HA
01:05:34  <CIA-53>node: Ben Noordhuis * r27ee567 / benchmark/http_pump.js : squash - http://git.io/20kbfg
01:07:58  * ircretaryquit (Ping timeout: 260 seconds)
01:22:36  * isaacsjoined
01:32:59  * ircretaryjoined
01:34:18  * piscisaureusjoined
01:34:19  <ryah>bnoordhuis:
01:34:19  <ryah>% cat ~/local/bin/fix_macintosh_portrange
01:34:19  <ryah>#sudo sysctl -w net.inet.ip.portrange.first=32768
01:34:19  <ryah>sudo sysctl -w net.inet.ip.portrange.first=12000
01:34:19  <ryah>sudo sysctl -w net.inet.tcp.msl=1000
01:34:21  <ryah>sudo sysctl -w kern.maxfiles=1000000 kern.maxfilesperproc=1000000
01:35:35  * isaacsquit (Quit: isaacs)
01:45:41  * piscisaureusquit (Ping timeout: 256 seconds)
01:48:00  * brsonquit (Ping timeout: 252 seconds)
02:04:22  * piscisaureusjoined
02:04:24  <ryah>piscisaureus: https://github.com/joyent/node/blob/95d530f2b4faa57662d3b052e7637246b6a7025f/benchmark/http_simple.js#L32-34
02:09:55  * AvianFluquit (Quit: Leaving)
02:19:20  * pieternquit (Quit: pietern)
02:20:56  <CIA-53>node: Ben Noordhuis * r84d0b1b / (3 files in 3 dirs):
02:20:57  <CIA-53>node: http: improve http parser bindings
02:20:57  <CIA-53>node: Speeds up HTTP benchmarks by 10% on average. - http://git.io/iND-RA
02:32:07  <CIA-53>libuv: Bert Belder * rfb71386 / (include/uv-private/uv-win.h src/win/internal.h src/win/tty.c): Improve ansi escape code support on windows - http://git.io/j38xEA
02:35:08  <ryah>DrPizza: yt?
02:35:21  <DrPizza>hi
02:35:44  <DrPizza>ryah: 'sup?
02:36:11  <ryah>DrPizza: do you remember this problem with GYP in which is prints some random warning
02:36:14  <ryah>C:\Program Files\MSBuild\Microsoft.Cpp\v4.0\Microsoft.CppBuild.targets(1151 warning MSB8012: TargetPath(C:\Users\ryan\node\Debug\uv.lib) does not matc
02:36:18  <ryah>e Library's OutputFile property value (C:\Users\ryan\node\Debug\lib\uv.lib)
02:36:20  <ryah>is may cause your project to build incorrectly. To correct this, please mak
02:36:21  <DrPizza>yes
02:36:23  <ryah>re that $(OutDir), $(TargetName) and $(TargetExt) property values match the
02:36:24  <DrPizza>that's why I patched gyp
02:36:25  <ryah>ue specified in %(Lib.OutputFile). [C:\Users\ryan\node\deps\uv\uv.vcxproj]
02:36:27  <ryah>DrPizza: somehow your fix got clobbered
02:36:30  <DrPizza>if you take a look at the modification I made to gyp
02:36:32  <DrPizza>hmm
02:37:07  <DrPizza>ryah: let me find the fix
02:37:16  <ryah>DrPizza: we figured out how to finally submit patches to GYP so i'd like to give this to them
02:37:27  <ryah>(they don't check the issues list)
02:37:40  <DrPizza>ryah: they may not like the patch depending on if they depend on the behaviour without the patch
02:37:54  <DrPizza>for some reason, they hardcode gyp so that it builds libs into $(outputdir)/lib/
02:37:55  <ryah>DrPizza: have you ever used git-cl ?
02:38:03  <DrPizza>but they don't create the project files quite properly
02:38:05  <DrPizza>nope, what's it do?
02:38:14  <ryah>it's google's way of submitting patches
02:38:18  <ryah>it's really nice actually
02:40:39  <DrPizza>ryah: I mean, the way they have gyp is dumb, if they want libs in /lib/ then they should specify in the .gyp files, not in gyp itself
02:43:38  <DrPizza>ryah: but it's possible that they have some good but non-obvious reason for doing it, I don't know
02:49:41  <DrPizza>ryah: where's the best place to ask, their mailling list?
02:58:28  <ryah>DrPizza: they say the best place is gyp-dev list
02:58:33  <DrPizza>ok
02:58:42  <DrPizza>I will ask if they have a reason
02:58:52  <DrPizza>if they don't we can submit a patch
02:59:10  <DrPizza>if they do, I'll have to look more deeply into the project files
02:59:16  <DrPizza>to see how to set all the properties necessary
02:59:33  <ryah>DrPizza: https://groups.google.com/forum/#!forum/gyp-developer
02:59:41  <DrPizza>ok
03:33:13  * piscisaureusquit (Quit: ~ Trillian Astra - www.trillian.im ~)
03:34:51  * piscisaureus_joined
03:39:07  * piscisaureus_quit (Client Quit)
03:43:18  * bnoordhuisquit (Ping timeout: 244 seconds)
04:30:39  * ericktquit (Quit: erickt)
05:20:53  * ericktjoined
05:25:53  * sh1mmerjoined
06:03:39  * ericktquit (Quit: erickt)
06:05:38  * piscisaureusjoined
06:10:42  * bnoordhuisjoined
06:23:16  * AvianFlujoined
06:33:40  * piscisaureusquit (Quit: ~ Trillian Astra - www.trillian.im ~)
06:52:49  * bnoordhuisquit (Ping timeout: 252 seconds)
06:57:37  * piscisaureusjoined
07:06:45  * AvianFluquit (Ping timeout: 252 seconds)
07:19:37  * AvianFlujoined
07:36:07  * mralephjoined
08:10:33  * mralephquit (Quit: Leaving.)
09:18:42  * kuebkjoined
09:41:39  * sh1mmerquit (Quit: sh1mmer)
10:13:19  * AvianFluquit (Quit: Leaving)
13:05:54  * ericktjoined
14:05:28  * bnoordhuisjoined
14:19:10  * bnoordhuisquit (Remote host closed the connection)
14:36:00  * bnoordhuisjoined
14:41:24  * ericktquit (Quit: erickt)
15:33:46  * kuebkquit (Quit: HydraIRC -> http://www.hydrairc.com <- Organize your IRC)
15:38:09  * isaacsjoined
15:42:05  * bradleymeckjoined
15:42:48  <bradleymeck>node is statically linking to the openssl library for crypto right? (was digging through make file and couldnt find a line that explicitly statically linked)
15:47:05  <bnoordhuis>bradleymeck: only when we're building and linking against the in-tree version
15:47:27  <bnoordhuis>that's always the case on the windows, if that's what you're asking
15:49:24  <bradleymeck>but not always the case on *nix? i didnt see an option to set it to shared, so its just version to version how it links?
15:51:23  <rmustacc>On unix it defaults to using the system one that already exists.
15:51:37  <rmustacc>At least, it used to.
15:51:48  <bnoordhuis>yes, that's still the case
15:59:49  <isaacs>ryah: what's the max length of headers that node will let http_parser will keep parsing before quitting?
15:59:55  <isaacs>you know offhand?
16:05:45  * bnoordhuisquit (Ping timeout: 244 seconds)
16:11:22  * dapjoined
16:18:37  * sh1mmerjoined
16:26:04  * bnoordhuisjoined
16:34:55  * mralephjoined
16:37:03  * sh1mmerquit (Quit: sh1mmer)
16:41:32  <ryah>isaacs: 80kb?
16:41:47  <isaacs>oh, ok, yeah, that looks about right
16:42:18  * bnoordhuisquit (Remote host closed the connection)
16:42:33  <ryah>isaacs: https://github.com/joyent/node/blob/84d0b1bcc5154f67d9f200f3c852bb8c5b64c6ef/deps/http_parser/http_parser.h#L62-63
16:42:42  <isaacs>thanks
16:43:25  <isaacs>ryah: we should send back a "431 Request Header Fields Too Large" error in that case, maybe.
16:43:42  <isaacs>https://tools.ietf.org/html/draft-nottingham-http-new-status-02
16:43:46  <isaacs>from mnot and friends
16:44:07  <ryah>*shrug*
16:45:56  * piscisaureus_joined
16:56:48  * bnoordhuisjoined
17:09:54  <CIA-53>libuv: Ben Noordhuis * r9757a43 / src/unix/fs.c : unix: remove unused variable - http://git.io/l1mW4w
17:10:25  <ryah>mraleph: new v8 today?
17:10:28  <bnoordhuis>https://github.com/bnoordhuis/libuv/compare/remove-doubles <- review?
17:11:28  <ryah>lgtm
17:12:29  <rmustacc>Awesome!
17:15:17  * isaacsquit (Quit: isaacs)
17:15:24  * creationix|workjoined
17:15:28  <igorzi>ryah: https://github.com/igorzi/node/commit/7f04d5f105f2b78a190c2f191e4aa4c440868bc7
17:17:44  <CIA-53>libuv: Ben Noordhuis * rd396799 / (10 files in 4 dirs):
17:17:44  <CIA-53>libuv: Change return type of uv_get_*_memory() functions
17:17:44  <CIA-53>libuv: ... from double to uint64_t. Limit use of floating point in public API as much
17:17:44  <CIA-53>libuv: as possible. - http://git.io/6--hkA
17:21:06  <mraleph>ryah: new V8?
17:21:55  <CIA-53>node: Ryan Dahl * rfafb584 / test/simple/test-dgram-pingpong.js : Fix race condition in test-dgram-pingpong - http://git.io/CEX04Q
17:26:54  <ryah>mraleph: new release today?
17:27:18  <ryah>igorzi: thanks
17:27:18  <mraleph>I don't think so
17:27:34  <ryah>mraleph: :(
17:27:51  <mraleph>well it's Friday. we never push on Friday
17:29:04  * brsonjoined
17:29:42  <CIA-53>node: Igor Zinkovsky * r69b73f9 / (src/node.js test/simple/test-process-kill-null.js): make process.kill a no-op on windows - http://git.io/PVM7xg
17:51:36  * bradleymeckquit (Ping timeout: 240 seconds)
18:10:53  * AvianFlujoined
18:31:20  * isaacsjoined
18:41:17  <piscisaureus_>bnoordhuis: https://gist.github.com/1304600
18:44:27  * mikealjoined
18:44:44  * mikealquit (Client Quit)
18:47:45  <igorzi>ryah: test-child-process-fork2.js is currently broken on windows.. i'm investigating.. but it'd be good to have that fixed for the release
19:02:10  * bnoordhuisquit (Ping timeout: 260 seconds)
19:14:08  <indutny>mraleph: I've one js script that 2x slower on latest v8
19:14:16  <indutny>in comparison to node's 0.4.x v8
19:14:22  <indutny>raw js
19:14:24  <indutny>w/o any calls
19:14:32  <indutny>I mean native calls
19:15:20  <mraleph>show it to me.
19:20:51  <mraleph>indutny: ^^^ :-)
19:34:20  * bnoordhuisjoined
19:34:48  <ryah>win7: [15:45|% 100|+ 251|- 17]: Done
19:38:35  <indutny>mraleph: sorry
19:38:37  <indutny>one second
19:43:45  <indutny>mraleph: https://gist.github.com/08ec397dc4688fec0bab
19:43:46  <indutny>hehe
19:43:49  <indutny>try running test.js
19:43:59  <indutny>it'll be fast but 2x slower than on node 0.2.x
19:44:00  <indutny>oh
19:44:03  <indutny>0.4.x, sorry
19:44:26  <indutny>code looks like mess
19:44:28  <indutny>slightly
19:44:34  <indutny>because it's autogenerated
19:44:38  <indutny>mraleph: ^
19:48:09  <CIA-53>node: Ben Noordhuis * r0ae98df / wscript :
19:48:09  <CIA-53>node: build: fix wscript on solaris
19:48:09  <CIA-53>node: os.path.exists() in python 2.5.2 returns false for symlinks. - http://git.io/LccD0Q
19:49:45  <ryah>indutny: test/simple/test-debugger-client.js is hanging on Windows
19:49:51  <mraleph>indutny: omg
19:50:30  <mraleph>indutny: this is not going to be optimized by crankshaft (has switches with string cases). this is probably why it is slower
19:51:01  * bnoordhuisquit (Remote host closed the connection)
19:51:38  <ryah>mraleph: are you guys going to optimize that?
19:51:43  <indutny>mraleph: hehe :)
19:52:02  <ryah>mraleph: because i really want optimized switch over strings :)
19:52:05  <indutny>ryah: strange, I'm on mac right now. can I ask you to run 'node debug lib/fs.js'
19:52:15  <indutny>ryah: have you seen my code?
19:52:16  <indutny>:D
19:52:56  <indutny>mraleph: and even more - I'll be glad to see optimization for constant cases
19:53:03  <mraleph>ryah: someday maybe. we are also accepting patches (as usual) :-) it should be simple to implement switch over strings.
19:53:09  <indutny>switch { case C1: ... case Cn }
19:53:23  <indutny>mraleph: is it in hydrogen.cc ?
19:53:29  <indutny>mraleph: or in lithium.cc ?
19:53:31  <indutny>or in both
19:53:38  <mraleph>indutny: in hydrogen.cc
19:53:52  <mraleph>indutny: there is already support for switch with numeric cases.
19:53:54  <indutny>btw, I seen < 128 limitation
19:54:03  <mraleph>indutny: yes that is also.
19:54:53  <indutny>mraleph: why, because optimization lasts non-linear time?
19:56:07  <ryah>indutny: https://gist.github.com/1304788
19:57:29  <indutny>ryah: have we reverted isaacs commit?
19:57:35  <isaacs>no, aboutto land it
19:57:41  <indutny>ryah: ^
19:57:42  <indutny>;)
19:57:48  <isaacs>ryah: squash or would you prefer a revert first and then a better fix?
19:57:57  <ryah>isaacs: i guess two commits
19:58:17  <isaacs>k. lgtm? https://github.com/isaacs/node/commit/3f9db955297f9004abf176c7431d2186ab4cd1d1
19:58:46  <ryah>isaacs: yes
19:58:53  <isaacs>k, coming right up.
19:59:11  <mraleph>indutny: I think it just to avoid having graph that is too big.
19:59:52  <mraleph>and waste too much time optimizing it...
19:59:53  <indutny>mraleph: hm...
19:59:57  <indutny>yeah
19:59:59  <indutny>it's not O(n)
20:00:04  <indutny>that's what I'm talking about
20:02:56  <mraleph>building graph takes linear time
20:03:32  <indutny>so it's just a limit
20:03:33  <indutny>:)
20:03:34  <indutny>hehe
20:03:35  <indutny>ok
20:03:35  <mraleph>also I think all optimization passes are linear as well
20:05:39  * bnoordhuisjoined
20:10:32  <CIA-53>node: isaacs * r2d02e6a / (doc/api/repl.markdown lib/repl.js src/node.js): Add useGlobal flag for standard node repl - http://git.io/0mUycw
20:10:34  <CIA-53>node: isaacs * rcaf70f5 / (doc/api/repl.markdown lib/repl.js test/common.js):
20:10:34  <CIA-53>node: Revert "Don't use a separate context for the repl."
20:10:34  <CIA-53>node: This reverts commit b70fed48a7fb0ac884e6574253a89db6158b2f60. - http://git.io/I9B5Jw
20:17:02  <CIA-53>node: Ryan Dahl * rde09168 / lib/tls.js :
20:17:02  <CIA-53>node: Emit 'end' from crypto streams on close
20:17:02  <CIA-53>node: Fixes test/simple/test-tls-peer-certificate.js on Windows
20:17:02  <CIA-53>node: Patch from bnoordhuis.
20:17:02  <CIA-53>node: See also 75a0cf970fb48440a93a62796ab1f128fcbe7d76 - http://git.io/A_d2Gg
20:25:25  <ryah>win7: [04:13|% 100|+ 254|- 14]: Done
20:27:24  <ryah>^-- release
20:27:46  <bnoordhuis>piscisaureus: https://github.com/joyent/libuv/pull/224 <- don't forget
20:31:56  <creationix|work>ryah: :)
20:32:07  <creationix|work>I guess now would be a good time to pull in libuv
20:32:43  <igorzi>ryah: i have a fix for test-child-process-fork2.js (in libuv).. will commit shortly. can we include that?
20:33:02  <ryah>igorzi: yep
20:33:10  <ryah>=== debug test-eio-race ===
20:33:10  <ryah>Path: simple/test-eio-race
20:33:10  <ryah>trying to kill event loop ...
20:33:10  <ryah>C:\Users\ryan\node\test\simple\test-eio-race.js:62
20:33:13  <ryah> if (err) throw err;
20:33:15  <ryah> ^
20:33:18  <ryah>Error: ENOENT, The operation completed successfully.
20:33:20  <ryah> '/dev/zero'
20:33:23  <ryah>Command: C:\Users\ryan\node\Debug\node.exe C:\Users\ryan\node\test\simple\test-e
20:33:36  <ryah>i guess this is realtively easy to fix
20:35:09  * piscisaureus_quit (Read error: Connection reset by peer)
20:35:41  * piscisaureus_joined
20:38:29  <CIA-53>libuv: Igor Zinkovsky * r7738306 / (src/win/pipe.c src/win/tcp.c): windows: fixes queued_bytes for ipc pipes - http://git.io/RKoUpQ
20:38:48  <igorzi>ryah: ^--- pls update libuv in node
20:43:03  <ryah>igorzi: k
20:43:30  <creationix|work>piscisaureus: so, this ansi escape stuff on windows, how does that work. I just send normal ansi codes and they get translated automatically?
20:43:48  <piscisaureus_>creationix|work: when you write to uv_tty_t, yes
20:43:55  <creationix|work>nice
20:44:07  <creationix|work>luvit doesn't yet build on windows, but that will be nice when I get there
20:44:22  <creationix|work>and node v0.6.0 will build on windows ;)
20:44:33  <piscisaureus_>oh really ?
20:45:01  <creationix|work>well, there will be a binary on nodejs.org. Not sure hard building it actually is
20:46:58  <ryah>there's already a binary
20:47:20  <ryah>igorzi: --^
20:47:22  <ryah>er
20:47:23  <CIA-53>node: Ryan Dahl * r9b2335a / (16 files in 5 dirs): Upgrade libuv to 7738306 - http://git.io/PnqIKw
20:47:25  <ryah>igorzi: --^
20:49:11  <igorzi>ryah: thx.. can you rerun on win7 with that fix?
20:49:37  <igorzi>i seem to be getting more failiures than you: [06:09|% 100|+ 246|- 22]
20:50:24  <CIA-53>node: Ben Noordhuis * r3c3ec7b / (2 files in 2 dirs): gyp: upgrade to r1081 - http://git.io/CMqGSw
20:51:25  <igorzi>ryah: are you running in cmd.exe or something else? (do you have curl and cat)
20:52:01  <ryah>igorzi: cmd.exe - yes i have curl and cat
20:52:06  <ryah>let me rebuild
20:54:22  <igorzi>ryah: clean windows machines won't have curl nor cat.. do you want tests modified to not use those? or just require curl and cat in order to run tests?
21:14:02  * mikealjoined
21:24:28  <CIA-53>libuv: Ben Noordhuis * r9da53df / src/unix/core.c :
21:24:28  <CIA-53>libuv: unix: fix uv_freeaddrinfo() segfault on sunos
21:24:28  <CIA-53>libuv: The system freeaddrinfo() on SunOS does not handle NULL values graciously. - http://git.io/h3_pCQ
21:30:20  * AvianFluquit (Ping timeout: 260 seconds)
21:34:22  <igorzi>ryah: test-http-dns-error.js expects EBADNAME, and test-http-dns-fail.js expects ENOTFOUND.. is this due to switching from cares to getaddrinfo? seems like we should unify the error codes?
21:40:11  <ryah>igorzi: yeah - i messed up the error code
21:40:22  <ryah>igorzi: it's broken on unix too
21:40:25  <ryah>i have to fix it
21:42:09  <igorzi>ryah: do you want tests to continue depending on cat & curl when running on windows?
21:42:20  * AvianFlujoined
21:42:33  <ryah>igorzi: what would we use instead?
21:42:45  <ryah>i guess cat can be replaced with a windows thing
21:42:57  <ryah>but windows probably doesn't have any curl-like replacement installed by default
21:43:35  <igorzi>ryah: i've replace cat with more in some places.. but yeah, replacing curl will be more difficult.. probably some vbscript.
21:43:52  <ryah>igorzi: meh - i think depending on curl isn't bad
21:44:05  <igorzi>ryah: k
21:45:00  <igorzi>ryah: which curl are you using on windows? is it possible to include it in the tree?
21:49:18  <isaacs>igorzi/piscisaureus: can either of you verify the sanity of this suggestion? https://github.com/isaacs/npm/issues/1582
21:49:22  <isaacs>re: npm module paths.
21:49:26  <isaacs>install folders
21:49:54  <piscisaureus_>isaacs: looks sane to me
21:50:03  <isaacs>igorzi: you could use an XmlHttpRequest in a WSH script
21:50:06  <piscisaureus_>isaacs: although the install path should be configurable
21:50:07  <creationix|work>isaacs: isn't that what we were doing pre v0.4.x
21:50:26  <rmustacc>isaacs: Seems sane to me.
21:50:36  <isaacs>piscisaureus: this would just be how it defaults. if you set the `prefix` config to something else, it'd use that.
21:50:50  <piscisaureus_>isaacs: nah, don't use prefix
21:51:06  <isaacs>piscisaureus: the trick is that, on windows, the global prefix wouldn't add .. to the node install location, and won't add `lib` to the module location.
21:51:08  <piscisaureus_>isaacs: npm should just figure out where it's own executable is and derive the path from that
21:51:33  <isaacs>piscisaureus: how does that work for initial installation?
21:52:02  <isaacs>by the time it's "npm", it's just been called as something like: C:\node\node.exe C:\node\node_modules\npm\bin
21:52:06  <isaacs>by the time it's "npm", it's just been called as something like: C:\node\node.exe C:\node\node_modules\npm\bin\npm.js <args>
21:52:13  <creationix|work>isaacs: I see, I misread. Looks good (If my unsolicited opinion counts)
21:52:54  <isaacs>piscisaureus: unless i set some kind of env or --cli flag in the cmd file, which seems ugly.
21:53:11  <isaacs>creationix|work: your opinion counts, yes, thanks :)
21:55:44  <CIA-53>node: Ryan Dahl * refa95c2 / test/simple/test-eio-race.js : Fix test-eio-race on Windows - http://git.io/n68mvQ
21:55:57  <ryah>igorzi: just the curl.exe ? i think i installed it manually..
21:56:08  <ryah>igorzi: yeah including it in the tree might be okay..
21:56:18  <ryah>igorzi: but there are some other things we depend on like ab.exe
21:56:41  <creationix|work>isaacs: bte, node's require path semantics have spoiled me and I can't stand lua's require paths
21:56:59  <creationix|work>isaacs: so I'm re-implementing node style require in luvit
21:57:14  <ryah>commonjs ftw
21:57:41  <creationix|work>relative require and node_modules is commonjs?
21:59:04  <creationix|work>I just return the module instead of doing module.exports since evaluated scripts are always treated like functions in lua: https://github.com/creationix/luvit/blob/master/lib/fiber.lua
21:59:36  <creationix|work>hmm, but I guess that won't support recursive dependencies like exports.foo = bar does
21:59:38  <ryah>relative require is
21:59:51  <ryah>node_modules is isaac's work
22:00:03  <creationix|work>yeah, relative require is super nice. Everything in lua's package module is based on cwd
22:00:12  <ryah>yes
22:00:30  <creationix|work>if you try to require("/some/absolte/path") it will look in /my/cwd//some/absolute/path
22:00:30  <ryah>same with most programming languages - it's very annoying
22:00:48  <isaacs>creationix|work: node_modules is not commonjs
22:01:04  <isaacs>require.paths is commonjs
22:01:05  <creationix|work>isaacs: I figured as much it being called *node*_modules, but you never know
22:01:08  <isaacs>and the worse idea ever.
22:01:46  <creationix|work>I think I'll have absolute paths, relative paths to the script and node_module style paths in luvit
22:02:21  <creationix|work>not of this worrying about mutable paths arrays and the current cwd of the process mess
22:03:11  <ryah>okay. bnoordhuis is doing setNoDelay, im doing EBADNAME, and then we're releasing v0.5.10
22:03:23  <isaacs>badass.
22:03:26  <creationix|work>nodelay yay
22:03:59  <creationix|work>that will make dvvv happy (my super active luvit fan)
22:09:42  * dapquit (*.net *.split)
22:09:42  * CoverSlidequit (*.net *.split)
22:11:20  * dapjoined
22:11:20  * CoverSlidejoined
22:11:26  * DrPizzaquit (Excess Flood)
22:11:49  * DrPizzajoined
22:14:42  <ryah>test-signal-handler.js on windows acts very strange
22:15:27  <ryah>oh maybe not..
22:18:21  <igorzi>i guess it's the result of process.kill no-op
22:18:48  * creationix|workchanged nick to creationix
22:24:15  <CIA-53>node: Ryan Dahl * rd5a21a2 / (lib/dns.js test/simple/test-http-dns-error.js):
22:24:15  <CIA-53>node: getaddrinfo returns ENOTFOUND for invalid domain names
22:24:15  <CIA-53>node: change test-http-dns-error to reflect this. - http://git.io/pxPU6w
22:25:32  * AvianFlu_joined
22:25:41  <ryah>indutny:
22:25:41  <ryah>=== release test-debugger-repl ===
22:25:42  <ryah>Path: simple/test-debugger-repl
22:25:42  <ryah>Command: out/Release/node /Users/ryan/projects/node/test/simple/test-debugger-repl.js
22:25:47  <ryah>fail
22:25:53  <ryah>>:|
22:27:58  * AvianFluquit (Ping timeout: 260 seconds)
22:28:14  <indutny>haha
22:28:23  <indutny>not god
22:29:11  <indutny>not good
22:30:13  <indutny>ryah: strange
22:30:21  <indutny>ryah: it wfm
22:30:38  <indutny>on osx
22:31:46  * AvianFlu_changed nick to AvianFlu
22:36:05  <indutny>isaacs: ryah: https://github.com/joyent/node/pull/1921
22:37:30  <ryah>piscisaureus_: https://github.com/joyent/node/issues/1909
22:37:47  <ryah>indutny: i don't care. isaacs can land it.
22:37:55  <indutny>k
22:38:01  <indutny>debugger still wfm on osx
22:38:05  <indutny>can't test on win today
22:38:05  <ryah>indutny: hmm
22:38:10  <ryah>indutny: tht was on osx
22:38:16  <indutny>ryah: strange
22:38:24  <indutny>ryah: exactly same command works on osx for me
22:38:38  <ryah>% ./node test/simple/test-debugger-repl.js
22:38:38  <ryah>Error: timeout!
22:39:04  <indutny>[email protected]:~/Code/indutny/node > ./node test/simple/test-debugger-repl.js
22:39:04  <indutny>[email protected]:~/Code/indutny/node > ./node test/simple/test-debugger-repl.js
22:39:04  <indutny>[email protected]:~/Code/indutny/node > ./node test/simple/test-debugger-repl.js
22:39:04  <indutny>[email protected]:~/Code/indutny/node > ./node test/simple/test-debugger-repl.js
22:39:04  <indutny>[email protected]:~/Code/indutny/node > ./node test/simple/test-debugger-repl.js
22:39:05  <indutny>[email protected]:~/Code/indutny/node > ./node test/simple/test-debugger-repl.js
22:39:07  <indutny>:D
22:39:12  <indutny>6 times
22:39:28  <ryah>it's hanging for me each time
22:39:35  <ryah>no - not each time
22:39:37  <ryah>most times
22:40:17  <indutny>that's odd
22:40:27  <indutny>can you try running ./node debug lib/fs.js /cc mraleph
22:40:38  <indutny>mraleph: I seen that sometimes connection with debugger is broken
22:40:43  <indutny>s/is/was
22:40:45  <ryah>indutny: seems tow ork
22:41:08  <isaacs>indutny: why's that better this way?
22:41:08  <indutny>mraleph: it wasn't responding, while it was connected
22:41:24  <isaacs>indutny: don't have any strong feelings either way, really.
22:41:24  <indutny>isaacs: because noone should care about eval
22:41:34  <indutny>isaacs: ok
22:41:37  <indutny>isaacs: :)
22:41:56  <isaacs>k, that convinces me. :)
22:42:16  <indutny>haha
22:42:17  <indutny>:D
22:42:17  <indutny>great
22:42:22  <indutny>gtg
22:42:28  <indutny>almost 3am in moscow
22:42:51  <indutny>ttyl
22:43:00  <ryah>bnoordhuis: https://github.com/joyent/node/blob/ff942c6b39d06e81d82c30bc91a159a0440b6d9b/src/node_net.cc#L1297-1342
22:43:03  <ryah>indutny: by
22:43:04  <ryah>e
22:43:24  <bnoordhuis>sleep tight, indutny
22:45:00  <isaacs>running test-all with indutny's 9018906853895f07b149737b5f8c2775e551f549
22:45:10  <ryah>https://gist.github.com/1305185 <--- win7 fails igorzi, piscisaureus_
22:46:48  <igorzi>i'm looking at simple\test-fs-symlink.js and simple\test-fs-realpath.js
22:46:57  <igorzi>ryah & piscisaureus_ ---^
22:47:33  <ryah>we should be able to wrap fs.watch to create fs.watchFile
22:47:53  <ryah>bert has some concern about this - but it should work in most cases
22:49:35  * AvianFlu_joined
22:49:40  <igorzi>ryah: we'll do it for this release?
22:52:22  * mikealquit (Quit: Leaving.)
22:52:23  * AvianFluquit (Ping timeout: 252 seconds)
22:54:25  <CIA-53>node: Maciej MaƂecki * r8f1ba25 / (2 files in 2 dirs): test: #1899 - output in child process raises an exception - http://git.io/g1z3fg
22:56:09  <ryah>igorzi: no i think not
23:06:37  * AvianFlu_changed nick to AvianFlu
23:49:01  * mikealjoined
23:50:21  * mralephquit (Quit: Leaving.)