00:01:36  * mikealquit (Quit: Leaving.)
00:13:23  * mikealjoined
00:21:25  * dshaw_quit (Read error: Operation timed out)
00:30:42  * mralephquit (Quit: Leaving.)
00:34:54  * rendarquit
00:35:21  * xaqquit (Remote host closed the connection)
00:45:23  * travis-cijoined
00:45:23  <travis-ci>[travis-ci] joyent/libuv#103 (v0.6 - fca18c3 : Brandon Philips): The build is still failing.
00:45:23  <travis-ci>[travis-ci] Change view : https://github.com/joyent/libuv/compare/c0e7044...fca18c3
00:45:23  <travis-ci>[travis-ci] Build details : http://travis-ci.org/joyent/libuv/builds/731340
00:45:23  * travis-cipart
00:59:56  * mikeal1joined
01:03:32  * mikealquit (Ping timeout: 240 seconds)
01:08:49  <bnoordhuis>TooTallNate: https://github.com/joyent/node/commit/ba0892ba910bb6c60ac4e9ed2df2c37f67efdea7#commitcomment-1022157 <- where did our comments go?
01:09:14  <TooTallNate>??? weird
01:09:15  <TooTallNate>haha
01:09:18  <TooTallNate>damn github
01:09:24  <TooTallNate>been acting weird today I think
01:10:43  <CIA-99>node: Joshua Holbrook v0.6 * r4065b24 / (doc/api/child_processes.markdown lib/child_process.js): child_process: remove dummy "setsid" option setting - http://git.io/8xDmwQ
01:12:08  <CIA-99>node: Joshua Holbrook master * r56cfcea / (doc/api/child_process.markdown lib/child_process.js): child_process: remove dummy "setsid" option setting - http://git.io/XaDc5g
01:18:57  * travis-cijoined
01:18:57  <travis-ci>[travis-ci] joyent/node#521 (v0.6 - 4065b24 : Joshua Holbrook): The build passed.
01:18:57  <travis-ci>[travis-ci] Change view : https://github.com/joyent/node/compare/6343179...4065b24
01:18:57  <travis-ci>[travis-ci] Build details : http://travis-ci.org/joyent/node/builds/762383
01:18:57  * travis-cipart
01:23:15  <CIA-99>node: Blake Miner master * r7343f8e / (doc/api/crypto.markdown doc/api/tls.markdown lib/tls.js):
01:23:16  <CIA-99>node: tls: add `honorCipherOrder` option to tls.createServer()
01:23:16  <CIA-99>node: Documented how to mitigate BEAST attacks. - http://git.io/-z1SyQ
01:28:07  * travis-cijoined
01:28:07  <travis-ci>[travis-ci] joyent/node#522 (master - 56cfcea : Joshua Holbrook): The build is still failing.
01:28:07  <travis-ci>[travis-ci] Change view : https://github.com/joyent/node/compare/af7960b...56cfcea
01:28:07  <travis-ci>[travis-ci] Build details : http://travis-ci.org/joyent/node/builds/762405
01:28:07  * travis-cipart
01:32:17  * AvianFluquit (Quit: Leaving)
01:34:16  * mikeal1quit (Quit: Leaving.)
01:38:13  * travis-cijoined
01:38:13  <travis-ci>[travis-ci] joyent/node#523 (master - 7343f8e : Blake Miner): The build is still failing.
01:38:13  <travis-ci>[travis-ci] Change view : https://github.com/joyent/node/compare/56cfcea...7343f8e
01:38:13  <travis-ci>[travis-ci] Build details : http://travis-ci.org/joyent/node/builds/762447
01:38:13  * travis-cipart
01:59:03  <isaacs>bnoordhuis: yeah, i've noticed that i'm getting emails about comments, but they're gone from the site.
02:04:05  * dshaw_joined
02:08:21  * dshaw_quit (Client Quit)
02:14:44  <bnoordhuis>odd, innit?
02:19:37  * pfox___quit (Ping timeout: 245 seconds)
02:21:51  * TooTallNatequit (Quit: Leaving...)
02:24:52  * pfox___joined
02:34:23  * mikealjoined
02:35:12  * TooTallNatejoined
02:36:09  * brsonquit (Quit: leaving)
02:38:59  * AvianFlujoined
02:45:54  * xaqjoined
03:01:05  * pieternquit (Quit: pietern)
03:06:04  * TooTallNatequit (Quit: Leaving...)
03:09:39  * philipsquit (Excess Flood)
03:12:04  * philipsjoined
03:24:07  * isaacsquit (Remote host closed the connection)
03:29:34  * mikealquit (Ping timeout: 276 seconds)
03:54:22  * bnoordhuisquit (Ping timeout: 244 seconds)
04:05:03  * xaqquit (Remote host closed the connection)
04:12:21  * TooTallNatejoined
04:27:06  * indutny_sleepingchanged nick to indutny
04:32:45  * TooTallNatequit (Quit: Linkinus - http://linkinus.com)
05:12:00  * xaqjoined
05:36:36  * dap1joined
05:39:04  * dapquit (Ping timeout: 252 seconds)
06:07:00  * benviequit
06:12:02  * benviejoined
06:12:21  * benviequit (Client Quit)
06:13:53  * benviejoined
06:15:00  * pfox___quit (Ping timeout: 260 seconds)
06:15:47  * felixgejoined
06:46:08  * isaacsjoined
07:09:27  * AvianFluchanged nick to AvianusAsleepus
07:58:08  * stephankquit (Quit: *Poof!*)
08:01:57  * xaqquit (Remote host closed the connection)
08:03:19  * rendarjoined
08:07:56  * mikealjoined
08:12:46  * mralephjoined
08:28:39  * felixge_joined
08:28:39  * felixge_quit (Changing host)
08:28:39  * felixge_joined
08:32:02  * felixgequit (Ping timeout: 240 seconds)
08:32:03  * felixge_changed nick to felixge
08:49:33  * mralephquit (Quit: Leaving.)
09:17:00  * mralephjoined
09:22:28  * abraxasjoined
09:23:54  * trondn_joined
09:24:29  <trondn_>anyone tried using libuv on solaris? make test the current master gives me quite a few coredumps...
09:40:25  * dshaw_joined
09:57:00  * mralephquit (Quit: Leaving.)
09:58:44  * isaacsquit (Remote host closed the connection)
09:59:27  * mralephjoined
10:06:50  * mralephquit (Quit: Leaving.)
10:36:27  * abraxasquit
10:38:39  <trondn_>Hmm.. is make test supposed to work on the current master branch? It fails for me on Solaris, macosx and debian…
10:38:54  <trondn_>Am I supposed to use another branch perhaps?
12:01:59  * mmalecki[zzz]changed nick to mmalecki
12:39:44  * felixgequit (Read error: Connection reset by peer)
12:40:15  * felixgejoined
12:42:03  * felixge_joined
12:42:03  * felixgequit (Read error: Connection reset by peer)
12:42:04  * felixge_changed nick to felixge
12:43:36  * felixge_joined
12:43:36  * felixge_quit (Changing host)
12:43:36  * felixge_joined
12:43:37  * felixgequit (Read error: Connection reset by peer)
12:43:37  * felixge_changed nick to felixge
12:47:40  * felixgequit (Client Quit)
12:51:16  * felixgejoined
12:51:16  * felixgequit (Changing host)
12:51:16  * felixgejoined
12:54:38  * felixgequit (Read error: Connection reset by peer)
12:54:54  * felixgejoined
13:12:09  * felixgequit (Read error: Connection reset by peer)
13:13:10  * felixgejoined
13:13:17  * pfox___joined
13:18:26  * bnoordhuisjoined
13:20:11  * AvianusAsleepuschanged nick to AvianFlu
13:36:33  * mmaleckichanged nick to mmalecki[away[
13:36:35  * mmalecki[away[changed nick to mmalecki[away]
13:45:44  * philipsquit (Excess Flood)
13:46:35  * philipsjoined
14:01:19  * piscisaureus_joined
14:26:42  * dshaw_quit (Quit: Leaving.)
15:06:26  <piscisaureus_>bnoordhuis: http://codereview.chromium.org/6539006
15:16:11  <indutny>piscisaureus_: ++
15:16:11  <kohai>piscisaureus_ has 6 beers
15:16:29  <indutny>ah, I thought that is a strings patch
15:16:35  <indutny>-=0.5 then :D
15:18:18  <CIA-99>libuv: Ben Noordhuis master * r0459097 / src/unix/kqueue.c :
15:18:18  <CIA-99>libuv: unix: fix fs_event refcount bug on darwin
15:18:18  <CIA-99>libuv: The event loop got unref'd twice when the handle was closed. - http://git.io/qDrWOg
15:18:28  <CIA-99>node: Ben Noordhuis master * rebb79cd / test/simple/test-fs-watch.js :
15:18:28  <CIA-99>node: Revert "test-fs-watch: Add faster failure, and link to #2813"
15:18:28  <CIA-99>node: This reverts commit 9a6f936c8a62c1a6edad3bda0faa8d8112b7cf92. - http://git.io/Yx0X_w
15:18:28  <CIA-99>node: Ben Noordhuis master * rbab59f3 / deps/uv/src/unix/kqueue.c : uv: upgrade to 0459097 - http://git.io/mxRnYg
15:20:08  * travis-cijoined
15:20:08  <travis-ci>[travis-ci] joyent/libuv#110 (master - 0459097 : Ben Noordhuis): The build is still failing.
15:20:08  <travis-ci>[travis-ci] Change view : https://github.com/joyent/libuv/compare/2f886c8...0459097
15:20:08  <travis-ci>[travis-ci] Build details : http://travis-ci.org/joyent/libuv/builds/766036
15:20:08  * travis-cipart
15:30:07  <piscisaureus_>hehe
15:30:20  <piscisaureus_>no I suspect that's why v8 doesn't work on solaris
15:32:21  * mikealquit (Quit: Leaving.)
15:33:39  * travis-cijoined
15:33:39  <travis-ci>[travis-ci] joyent/node#524 (master - bab59f3 : Ben Noordhuis): The build is still failing.
15:33:39  <travis-ci>[travis-ci] Change view : https://github.com/joyent/node/compare/7343f8e...bab59f3
15:33:39  <travis-ci>[travis-ci] Build details : http://travis-ci.org/joyent/node/builds/766048
15:33:39  * travis-cipart
15:55:22  * AvianFluquit (Quit: Leaving)
15:56:13  * mikealjoined
16:36:13  * mikealquit (Quit: Leaving.)
16:41:50  * mmalecki[away]changed nick to mmalecki
16:47:46  * stephankjoined
16:53:28  * dshaw_joined
16:54:54  * benviequit
17:02:41  * isaacsjoined
17:03:43  * isaacsquit (Read error: Connection reset by peer)
17:04:47  * isaacsjoined
17:05:36  <tjfontaine>isaacs: because you were just recently talking about it "[02-29] 04:24:05 < trondn_> anyone tried using libuv on solaris? make test the current master gives me quite a few coredumps..."
17:05:44  * benviejoined
17:07:13  * AvianFlujoined
17:11:06  <isaacs>tjfontaine: not sure.
17:11:15  <isaacs>tjfontaine: but yes, that looks like a problem.
17:28:51  * txdvjoined
17:36:24  <CIA-99>node: Bert Belder v8string * ra8600ea / (7 files in 4 dirs): Utf8 encoder - http://git.io/yvzvFQ
17:36:24  <CIA-99>node: Bert Belder v8string * r2257925 / (benchmark_utf8.js deps/v8/src/api.cc src/node_unicode.cc): Optimizations and fixes - http://git.io/zzqeHg
17:36:25  <CIA-99>node: Bert Belder v8string * radcac38 / common.gypi : Compile with NDEBUG - http://git.io/SC_oDA
17:36:25  <CIA-99>node: Bert Belder v8string * rcc16602 / (3 files): Erik's optimization of the v8 encoder - http://git.io/ekSi-Q
17:37:03  * travis-cijoined
17:37:03  <travis-ci>[travis-ci] joyent/node#525 (v8string - cc16602 : Bert Belder): The build failed.
17:37:03  <travis-ci>[travis-ci] Change view : https://github.com/joyent/node/compare/a8600ea^...cc16602
17:37:03  <travis-ci>[travis-ci] Build details : http://travis-ci.org/joyent/node/builds/767406
17:37:03  * travis-cipart
17:38:59  <piscisaureus_>isaacs: is npm back up?
17:39:28  <tjfontaine>it was not long ago
17:39:58  <mmalecki>I believe it is, not seeing many failed deploys :D
17:41:08  * trondnjoined
17:42:34  * dshaw_quit (Quit: Leaving.)
17:43:56  <trondn>anyone awake?? should I be able to successfully run make test on git://github.com/joyent/libuv master branch? (it fails on solaris, debian and macosx for me)..
17:44:45  <trondn>I'm happy to look into failures, but I'd like to start from the branch that is supposed to work ;)
17:45:35  <AvianFlu>trondn, on debian, I usually see ~7 test failures
17:47:54  * felixgequit (Quit: felixge)
17:48:48  <trondn>AvianFlu: ok.. so this is expected?? Reason I ask is because I'm planning to use it in another project and as a requirement there is to have it working without gcc ;) I'm done with my patch, but I wanted to verify that it still works as expected but that's hard to tell given all of the test failures ;)
17:49:20  <bnoordhuis>non-gcc? icc?
17:50:12  <isaacs>piscisaureus_: yes
17:51:29  * AvianFluquit (Ping timeout: 244 seconds)
17:51:33  <trondn>bnoordhuis: my requirement is that a C project shouldn't require anything else than a C compiler ;) not a given one ;)
17:52:23  * AvianFlujoined
17:53:17  <bnoordhuis>trondn: agreed, but we only actively support gcc and msvc
17:53:36  <bnoordhuis>i don't mind taking a non-intrusive patch for icc but we can't maintain it
17:53:47  <trondn>bnoordhuis: why not: We actively support C89 ;)
17:54:43  <trondn>bnoordhuis: This is all needed for building with Sun Studio compilers: https://github.com/trondn/libuv/commit/f6caddf945e6081ae836ef4732816236b091c031
17:54:44  <bnoordhuis>trondn: well, i don't have icc installed for one :)
17:54:49  <bnoordhuis>oh, sun studio
17:55:56  <bnoordhuis>like i said, i don't mind taking the patch but actively supporting it is another thing
17:56:22  * mikealjoined
17:57:00  <trondn>sure. but I need to figure out which tests are bogus and if I have real failures first :S
17:57:31  <bnoordhuis>can you gist the output of make test?
17:57:54  <trondn>2 sec
18:00:05  <trondn>http://pastebin.com/PZnw0H2J
18:00:34  <trondn>(I've already fixed one test with: https://github.com/trondn/libuv/commit/0afada3baf513e3d7f37c6e7b234db8f8fa4a0ee
18:02:34  <bnoordhuis>trondn: pipe_bind_error_addrinuse is known bad, the *_ref tests are part of a refactoring, process_title probably can't be implemented on sunos, spawn_and_kill_with_std should pass but is probably broken, the fs_event tests are broken because event ports doesn't support that particular feature
18:04:04  <bnoordhuis>trondn: with what version of solaris are you testing it? udp-options passes for me on smartos
18:04:22  <trondn>I tried it on illumos..
18:04:38  <trondn>SunOS storm 5.11 oi_151a2 i86pc i386 i86pc
18:04:44  <bnoordhuis>hmm okay, that needs further investigation
18:04:49  <trondn>(ehh. openindiana)
18:05:11  <trondn>I can try on my solaris 10 boxes as well (x86 and sparc)
18:05:25  <bnoordhuis>yeah, might be a good idea
18:05:28  <trondn>when will the refactor be done?
18:05:59  <bnoordhuis>that's a good question :)
18:06:13  <bnoordhuis>it kind of stalled as of late
18:09:20  <trondn>ok.. I'll look into the solaris files and see what I can do about the stuff there to see if it's all implemented ;)
18:11:03  <trondn>whats the current status on win32/64? is it working as good as on unix? or is there any limitations there?
18:12:32  <bnoordhuis>it's on par with unix
18:12:59  <trondn>great
18:15:39  <piscisaureus_>isaacs: http://codereview.chromium.org/6539006 <-- port this to gyp and try if it fixes solaris.
18:16:00  <piscisaureus_>isaacs: (told bnoordhuis about it too)
18:16:04  * benviequit
18:16:26  <isaacs>piscisaureus_: port to gyp, and also revert the -O3 thing?
18:16:48  <piscisaureus_>isaacs: keep O3
18:16:55  <isaacs>right
18:17:04  <piscisaureus_>isaacs: instead just pass -fno-strict-aliasing if the offending gcc version is detected
18:17:13  <isaacs>i see
18:17:20  <isaacs>piscisaureus_: i'll see if that fixes it.
18:17:36  <isaacs>because -O3 is probably nice. 3 optimizations are more than 1
18:17:37  <isaacs>:)
18:18:02  <bnoordhuis>actually...
18:18:09  <bnoordhuis>i did a test a while ago
18:18:24  <bnoordhuis>and i got better performance with -Os or -O2 than with -O3
18:18:32  <isaacs>crazy
18:18:46  <isaacs>well, maybe we test anduse -O2 (which also breaks now)
18:18:47  <piscisaureus_>well, -O2 is better than -O0 right :-)
18:18:57  <isaacs>really, th big difference is between -O1 and -O0
18:19:01  <piscisaureus_>but yes the benefit of O3 over O2 is dubious
18:19:19  <piscisaureus_>(at least that's what I saw in mingw)
18:19:25  <bnoordhuis>yeah
18:19:35  <bnoordhuis>-O3 does aggressive inlining and loop unrolling
18:19:54  <bnoordhuis>that probably doesn't work too well on cpus with small L1 caches
18:20:34  <bnoordhuis>(caveat emptor: that is conjecture on my part)
18:22:50  * benviejoined
18:23:55  * mralephjoined
18:24:48  * isaacsquit (Remote host closed the connection)
18:29:24  * piscisaureus_quit (Quit: ~ Trillian Astra - www.trillian.im ~)
18:31:22  * TooTallNatejoined
18:31:23  * bnoordhuisquit (Ping timeout: 245 seconds)
18:32:56  * `3rdEdenjoined
18:36:44  * perezdjoined
18:41:25  * xaqjoined
18:45:36  * sh1mmerquit (Quit: sh1mmer)
18:56:11  * mikealquit (Quit: Leaving.)
19:11:35  * dshaw_joined
19:13:09  * brsonjoined
19:15:54  * mralephquit (Ping timeout: 260 seconds)
19:16:11  * AndreasMadsenjoined
19:17:10  * mralephjoined
19:17:16  <AndreasMadsen>Can I send socket handlers on windows, only server handlers seems to work.
19:17:37  * dshaw_quit (Quit: Leaving.)
19:17:40  <AndreasMadsen>I get a ENOTSUP error.
19:18:18  * dshaw_joined
19:22:32  <AndreasMadsen>indutny: I'm almost done with fork-socket patch but I need a windows expert who should I write to?
19:29:35  * brsonquit (Quit: leaving)
19:29:49  * brsonjoined
19:32:24  <indutny>AndreasMadsen: probably peiscisaureus_
19:32:44  <AndreasMadsen>indutny: okay thanks
19:48:55  * mikealjoined
20:13:48  * mikealquit (Quit: Leaving.)
20:17:34  <igorzi>AndreasMadsen: on windows only server sockets can be sent over IPC
20:18:01  <igorzi>AndreasMadsen: what's fork-socket patch? i can help with that
20:18:27  <AndreasMadsen>igorzi: https://github.com/joyent/node/pull/2772
20:18:44  <AndreasMadsen>a highly requested feature
20:19:16  <AndreasMadsen>igorzi: Will sockets be supported on windows?
20:19:28  <igorzi>AndreasMadsen: yeah, we can add that
20:20:02  <igorzi>AndreasMadsen: what's the use case? one process accepts a connection, and then sends it to a worker?
20:20:14  <AndreasMadsen>igorzi: yeah
20:20:52  <AndreasMadsen>igorzi: sticky-session is a use case
20:21:11  <igorzi>AndreasMadsen: _setSimultaneousAccepts is not needed for a connection socket
20:21:40  <AndreasMadsen>igorzi: you just said that sockets aren't supported
20:21:54  <AndreasMadsen>igorzi: and setSimultaneousAccepts is only used on windows
20:22:03  <igorzi>AndreasMadsen: right.. when they will be supported, setSimultaneousAccepts will not be needed
20:22:57  <AndreasMadsen>igorzi: Okay, actually I have a fix ready
20:24:23  <igorzi>AndreasMadsen: i can add support for sending connection sockets to libuv later this week or next week
20:25:02  <AndreasMadsen>igorzi: that would be awesome, thanks
20:26:19  <indutny>AndreasMadsen: I told ya ;)
20:26:25  <indutny>about setSimultaneousAccepts
20:27:30  <AndreasMadsen>indutny: yeah, but except for your awesomeness it wasn't very trustworthy.
20:29:48  * CoverSli1equit (Read error: Connection reset by peer)
20:34:39  * xaqquit (Remote host closed the connection)
20:35:49  <indutny>AndreasMadsen: man, you're quite offensive in a last days
20:37:58  <AndreasMadsen>indutny: I have been quite offline the last days
20:38:16  <AndreasMadsen>indutny: But I'm sorry if I have offended you
20:38:23  <AndreasMadsen>^ That wasn't offensive
20:39:41  <AndreasMadsen>indutny: I try not to be drunk while coding, if you do that to I'm sure we will get along nicely
20:39:49  <AndreasMadsen>that too
20:39:51  <AndreasMadsen>*
20:39:59  <mmalecki>wtf.
20:40:28  <AndreasMadsen>mmalecki:http://piscisaureus.no.de/log/2012-02-23#13:58:23.707
20:40:51  <indutny>mmalecki: wtf is a right word :)
20:41:27  * mralephquit (Quit: Leaving.)
20:41:38  * mmaleckigoes back to hacking on kohai
20:42:19  <indutny>(138 commits, 10391 additions, 2825 deletions)
20:42:28  <indutny>love C++ ^ a lot of text, not a lot of work :)
20:42:28  <mmalecki>indutny: woot, where?
20:42:32  <indutny>mmalecki: dotlang
20:42:36  <mmalecki>indutny: ah, nice
20:43:02  <mmalecki>"10 additions and 81 deletions." <- kohai :D
20:43:08  <mmalecki>my last commit
20:43:13  <AndreasMadsen>indutny: autch
20:43:24  <mmalecki>but yeah, well, back to work
20:43:41  <indutny>just FYI I'm going to be away for 5-7 days
20:43:47  <indutny>I mean totally offline
20:44:02  <mmalecki>indutny: oh? where are you going to be?
20:44:09  <mmalecki>indutny: (if that's not a secret)
20:44:19  <mmalecki>indutny: and make sure to send team@ e-mail
20:44:23  <indutny>unfortunately it's exactly that case
20:44:26  <indutny>yeah, going to do that too
20:44:45  <mmalecki>indutny: oh, you're going to work for KGB as a spy?
20:44:56  <mmalecki>OH DAMN IT WAS A SECRET.
20:45:04  <AvianFlu>isn't it the GRU these days?
20:45:13  <AndreasMadsen>wee all know indutny is a russian spy
20:45:16  <AndreasMadsen>they all are
20:46:01  <indutny>it's GRU
20:46:06  <indutny>and I know where you're living
20:46:42  <indutny>lol :D
20:47:06  <mmalecki>oh well. I'll go and cancel all my appointments
20:47:18  <mmalecki>I'm happy I got to meet you guys
20:47:57  <AvianFlu>lol
20:49:31  * AndreasMadsenquit (Ping timeout: 246 seconds)
20:50:06  <mmalecki>AvianFlu: tell people I'm going to miss some standups
20:50:08  <mmalecki>lot of them
20:50:23  <mmalecki>unless there's some next step
20:50:30  <mmalecki>and they have skype there
20:50:41  <indutny>mmalecki: you'll attend on a lot of standups here too
20:50:48  <indutny>mmalecki: hahahahaha
20:51:15  <AvianFlu>hahahahahaha
20:52:26  <mmalecki>lol indutny
20:54:00  <indutny>ok, time to go
20:54:03  <indutny>ttyl guys
20:54:08  * indutnychanged nick to indutny_7_days
20:54:27  <mmalecki>"7 days" is pretty, uhm, suggestive
20:55:38  <mmalecki>damn, I forgot movie title
20:55:42  <mmalecki>but yeah, see you indutny_7_days
21:05:15  * leifquit (Read error: Operation timed out)
21:12:53  * TooTallNatequit (Read error: Operation timed out)
21:14:11  * TooTallNatejoined
21:16:51  * sh1mmer_joined
21:17:56  <pquerna>so, we are setting SSL_CTX_set_session_cache_mode to server
21:18:02  * dap1quit (Ping timeout: 244 seconds)
21:18:09  <pquerna>and as long as we use the same SSL_CTX, we should be using the ineternal openssl cache
21:19:23  * isaacsjoined
21:22:42  <pquerna>isaacs: so, from the code side, we are doing everything fine afaics.
21:23:01  <pquerna>isaacs: and in manual testing, i get a session id back from the server in node...
21:23:13  <pquerna>openssl s_client -connect 127.0.0.1:12122 -reconnect
21:24:15  <pquerna>i can reproduce the crappy performance though, on my desktop i only get ~620 req/second
21:24:33  <pquerna>which, if i run `openssl speed rsa2048`
21:24:37  <pquerna>is about what I get
21:24:52  * dapjoined
21:24:55  <pquerna>i'll profile to confirm, but sure seems to me is we are just running all the rsa ops in the event loop
21:24:56  <isaacs>pquerna: hm, so it's more interesting than session id caching?
21:25:07  <pquerna>well, it just means as a single threaded server
21:25:14  <pquerna>unless you do $Magic to run your rsa multi-threaded
21:25:20  <isaacs>i see
21:25:20  <pquerna>this is the expected performance
21:25:31  <isaacs>hm... how come apache22 does it way faster than us, then?
21:25:38  * trondnquit (Quit: Leaving.)
21:25:44  <pquerna>because it'll run its RSA on all his 4 cores
21:25:49  <pquerna>and get about 4x the performance
21:25:57  <isaacs>ahh, i see
21:26:06  <isaacs>so, it's doing the $Magic?
21:26:30  <pquerna>openssl speed rsa2048 -multi 8
21:26:35  <pquerna>its doing the equiv of that ^^
21:26:48  <pquerna>because its just running the rsa op inside the thread/process pool
21:26:51  <isaacs>i see
21:27:34  * trondnjoined
21:27:43  <pquerna>i'm unsure about the session id report; openssl says we are setting it, and the code all looks correct, it could be something with the size of the test/number of clients though
21:27:52  <pquerna>i'll mess around with that a bit to make sure we know how big the cache is
21:28:14  <isaacs>awesome.
21:29:44  * trondnquit (Client Quit)
21:30:15  <pquerna>also, our bundle openssl is quite old
21:30:22  <pquerna>it would be ideal to be on the newest 1.x
21:30:46  <pquerna>there are a bunch of performance improvements; "The easiest way to improve RSA performance is to upgrade to OpenSSL 1.0.0, which increases performance by 25%."
21:32:50  <isaacs>oh, yeah, i think i saw that a while back
21:38:02  * ljacksonjoined
21:42:17  <pquerna>so, we don't muck with it
21:42:18  <pquerna>#define SSL_SESSION_CACHE_MAX_SIZE_DEFAULT(1024*20)
21:42:35  <pquerna>so, ~20k is the default number of cached sessions per-SecureContext
21:42:55  <pquerna>it is possible to use SSL_CTX_sess_set_cache_size to tweak it
21:43:04  <pquerna>seems fine
21:55:24  <isaacs>pquerna: would you be interested in dropping some of this knowledge on the mailing list thread/github issue?
21:56:36  <CIA-99>node: isaacs smartos-compile-fno-strict-aliasing * rc98adce / (common.gypi configure): Use -fno-strict-aliasing instead of -O1 for SmartOS - http://git.io/rAKcQQ
21:59:22  * isaacsquit (Read error: Connection reset by peer)
21:59:42  * isaacsjoined
22:00:29  * isaacstopic: librarian: umbrella victim
22:05:07  * piscisaureus_joined
22:11:02  <piscisaureus_>isaacs: did -fno-strict-aliasing fix it?
22:11:50  * travis-cijoined
22:11:50  <travis-ci>[travis-ci] joyent/node#526 (smartos-compile-fno-strict-aliasing - c98adce : isaacs): The build failed.
22:11:50  <travis-ci>[travis-ci] Change view : https://github.com/joyent/node/commit/c98adce
22:11:50  <travis-ci>[travis-ci] Build details : http://travis-ci.org/joyent/node/builds/769240
22:11:50  * travis-cipart
22:12:05  <piscisaureus_>isaacs: if so, please consider always using this fix when gcc_old is detected
22:12:17  <piscisaureus_>isaacs: since afaik it affects not only solaris
22:12:39  <mmalecki>anyone here in close relationship with http-parser and willing to take a look at https://github.com/joyent/node/issues/2849#issuecomment-4247707 to make sure I'm sane?
22:13:53  <isaacs>piscisaureus_: can you review that patch?
22:14:01  <isaacs>piscisaureus_: i didn't merge into master
22:14:09  <isaacs>piscisaureus_: but yes, it does fix it.
22:14:31  <isaacs>piscisaureus_: i have no problem doing this for all unixes, but it seems like Linux and OS X are fine without it, and i'm running gcc_old on OS X
22:20:04  <piscisaureus_>isaacs: I can confirm that this used to happen on linux and cygwin as well when we previously fixed it. But we treated only gcc <= 4.5.x as old.
22:20:16  <piscisaureus_>isaacs: also, are you sure you're not compiling with clang?
22:20:26  <isaacs>hm.... no, i'm not sure of that.
22:20:30  <isaacs>how would i tell?
22:20:40  <TooTallNate>gcc --version
22:20:40  <kohai>gcc has -1 beer
22:21:02  * sh1mmer_quit (Quit: sh1mmer_)
22:21:10  <TooTallNate>isaacs: on osx (lion at least), gcc is llvm, and gcc-4.2 is gnu gcc
22:21:21  <isaacs>$ gcc --version
22:21:22  <isaacs>i686-apple-darwin11-llvm-gcc-4.2 (GCC) 4.2.1 (Based on Apple Inc. build 5658) (LLVM build 2336.1.00)
22:21:22  <isaacs>Copyright (C) 2007 Free Software Foundation, Inc.
22:21:22  <isaacs>This is free software; see the source for copying conditions. There is NO
22:21:24  <isaacs>warranty; not even for MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.
22:21:40  <piscisaureus_>isaacs: ah that is gcc-llvm
22:21:45  <isaacs>i see.
22:21:48  <piscisaureus_>isaacs: that would probably not be affected
22:21:49  <isaacs>sothat's why it works?
22:22:22  <isaacs>but that'd still register as "gcc_old" here, because gcc -dumpversion is
22:22:24  <isaacs>$ gcc -dumpversion
22:22:25  <isaacs>4.2.1
22:23:12  <isaacs>so we should perhaps test gcc<4.6.0 and non-llvm as "gcc_old"?
22:23:37  <piscisaureus_>isaacs: well, that is *really* old if it were normal gcc :-)
22:23:42  <piscisaureus_>isaacs: that's 2008 or so
22:24:09  <isaacs>heh
22:24:34  <isaacs>honestly, i could not care less, as long as it works.
22:24:36  <isaacs>:)
22:25:01  <isaacs>gotta run, i'll be back in a minute or 15
22:25:03  * isaacsquit (Remote host closed the connection)
22:25:41  <piscisaureus_>isaacs: right. yes, detect gcc_old on any unix but not if llvm
22:33:32  * dshaw_quit (Quit: Leaving.)
22:40:45  * dapquit (Quit: Leaving.)
22:41:10  * dapjoined
22:43:03  * dshaw_joined
22:51:05  * sh1mmerjoined
22:53:59  <pquerna>i tried to impart some ssl knowledge.
22:55:59  * AvianFlu_joined
22:57:21  * AvianFluquit (Ping timeout: 244 seconds)
23:03:57  * isaacsjoined
23:06:43  * AvianFlu_changed nick to AvianFlu
23:11:45  * rendarquit
23:31:48  * `3rdEdenquit (Quit: Zzzzz gnite)
23:34:48  * bnoordhuisjoined
23:39:32  * Ariajoined