00:00:04  <indutny>piscisaureus_: so create a folder, install 'redis' module here
00:00:16  <indutny>grab the example from this https://github.com/joyent/node/pull/3661
00:00:19  <piscisaureus_>indutny: are you serious
00:00:22  <indutny>and fixtures from node's project
00:00:24  <indutny>yeah, I am :D
00:00:28  <tjfontaine>heh
00:00:50  <piscisaureus_>indutny: ok, let's do that at some other time, okay
00:00:56  <piscisaureus_>indutny: like, monday or something
00:01:02  <indutny>oook
00:01:14  <indutny>hahaha
00:01:23  <piscisaureus_>indutny: are you in the us monday, or are you in russia?
00:01:49  <indutny>piscisaureus_: I'm leaving US at 4pm
00:01:51  <indutny>at monday
00:01:55  <piscisaureus_>indutny: oh, right
00:02:00  <indutny>piscisaureus_: and will be at 10am at Tuesday in russia
00:02:02  <piscisaureus_>indutny: I will try to test this on monday
00:02:06  <tjfontaine>is there even a usable redis-windows thing?
00:02:12  <piscisaureus_>good question
00:02:14  <indutny>oooh
00:02:20  <piscisaureus_>I think igorzi wrote that
00:02:21  <indutny>right, there're no such thing
00:02:29  <piscisaureus_>there is
00:02:29  <isaacs>indutny: i'll try it out
00:02:33  <piscisaureus_>but I don't know what the state of it is
00:02:36  <indutny>isaacs: cool, thanks man
00:02:45  <isaacs>just gotta figure out how to run redis on windows :)
00:03:23  <piscisaureus_>isaacs: the windows drone sits on a 0.8 branch with fedor's patches applied
00:03:29  <piscisaureus_>isaacs: so if you want to you can continue there
00:03:39  <isaacs>oh, sweet
00:03:41  <isaacs>thanks
00:04:01  <indutny>isaacs: well, you can run redis on other drone
00:04:03  <piscisaureus_>indutny: sorry, I don't feel like working now and I also think I am entitled to not-work on saturday night at 2am
00:04:10  <indutny>hahaha
00:04:19  <indutny>piscisaureus_: understood, I'm going to grab some beer soon
00:04:40  <isaacs>piscisaureus_: yes, definitely
00:04:53  <isaacs>it's kind of novel for me to play around with this ssl stuff on windows :)
00:05:09  <isaacs>but i'm sure i'll move on to other stuff soon
00:05:34  <isaacs>oh !@#~
00:05:39  <isaacs>no msysutils on there
00:05:46  <isaacs>i don't even know how to use a computer without the unix tools.
00:05:53  <tjfontaine>:)
00:06:26  <isaacs>oh, wait, no, it has them
00:06:29  <isaacs>just not "which" for some reason
00:07:11  * Ariajoined
00:07:37  <isaacs>hm... no official support for redis on windows
00:07:42  <isaacs>sounds like someone should be using libuv ):
00:08:31  <piscisaureus_>isaacs: https://github.com/MSOpenTech/Redis
00:08:45  <isaacs>yah, i found https://github.com/dmajkic/redis.git also
00:08:56  <isaacs>that's the one linked from the main redis site
00:09:04  <piscisaureus_>isaacs: that one is older
00:09:12  <isaacs>makes sense
00:09:15  <piscisaureus_>but
00:10:35  <piscisaureus_>isaacs: https://github.com/MSOpenTech/redis/blob/bksavecow/msvs/bin/release/redisbin.zip
00:10:50  * piscisaureus_should stop
00:11:07  <isaacs>nice
00:11:50  <isaacs>have you noticed that a lot of windows/msft folks seem to refer to all unixes as "Linux"?
00:12:07  <Aria>Huh! Do they?
00:12:14  <isaacs>eah
00:12:25  <isaacs>in subtle ways, in the middle of a README or conversation.
00:13:10  <Aria>Huh. I'm so going to notice that now.
00:13:19  <isaacs>even when specifically corrected, like "It's linux-specific? Or is it the same on Mac and Sunos and BSD?" "Yeah, those too. Anyway, the linux version..."
00:14:15  <isaacs>as a unix lover who only uses linux begrudgingly, it's sort of irksome in a petty way.
00:15:55  <Aria>Heh. Isn't it?
00:16:17  <Aria>I cut my teeth on Solaris, and really have a soft spot for some parts of BSD. This whole linux thing is just eh!
00:21:18  <isaacs>Aria: i find that linux is kind of the second worst at everything
00:21:26  <isaacs>but that's juuusst good enough.
00:21:29  <Aria>Yep!
00:21:53  <Aria>And a strange obsession with virtual filesystems.
00:21:55  <indutny>what the hell is linux?
00:22:08  <isaacs>cuz the minute you say, "It's crap with ui and peripherals and wifi and displays, and literally every single thing you care about on a laptop", then people are like, "Yeah, but you wouldn't run a server on os x, would you?"
00:23:16  <isaacs>and if you point out that the maintenance tools all lie and that process management sucks compared to smf, and the lack of dtrace and stuff, they're like, "So, I guess you run smartos on your laptop, then, right?"
00:23:24  <Aria>Heh. Yep.
00:23:56  <indutny>isaacs: actually, I think my example could be rewritten to use process.send
00:24:03  <indutny>isaacs: I don't really care and it's not important :)
00:24:14  <isaacs>indutny: yeah, but having a redis example as well is fucking awesome
00:24:21  <isaacs>indutny: because almost everyone uses redis for sessions
00:24:26  * mikealjoined
00:24:29  <isaacs>mongo is pretty good for sessions, too
00:24:39  <indutny>isaacs: well, redis is better
00:24:42  <indutny>because of expire functionality
00:24:42  <isaacs>depends
00:24:55  <indutny>yeah, it depends
00:24:59  <isaacs>if your sessions are longer-lived, and there are more of them, slightly more agressive persistence and sharding are nice features.
00:25:07  <isaacs>but redis is fast and super simple
00:25:16  <indutny>they've 5 minutes timeout
00:25:32  <isaacs>right
00:25:37  <isaacs>so redis is better here, clearly :)
00:25:37  <indutny>in OpenSSL
00:25:43  <indutny>indeed
00:25:46  <indutny>it's for security reasons
00:26:01  * EhevuTovjoined
00:27:38  * mikealquit (Client Quit)
00:29:30  <isaacs>is there a command in vcbuild to actually install node?
00:29:35  <isaacs>the msi is failing because it can't sign
00:29:52  <piscisaureus_>isaacs: that won't cause it to fail
00:30:00  <piscisaureus_>isaacs: it will just build an unsigned
00:30:13  <piscisaureus_>isaacs: if that doesn't work, run "vcbuild nosign"
00:30:23  <isaacs>ah, yes, nosign
00:30:33  <isaacs>that's the key
00:31:58  <indutny>isaacs: very interesting experience, isn't it? :)
00:41:15  * mikealjoined
00:41:16  <indutny>ok, time to grab so beer
00:49:11  <piscisaureus_>yeah
00:49:14  <piscisaureus_>time to get outta here
00:49:17  * piscisaureus_quit (Quit: ~ Trillian Astra - www.trillian.im ~)
00:53:54  * ericktjoined
00:56:10  * indexzerojoined
01:05:40  <isaacs>indutny: didn't get to it, sorry. got distracted, and now it's time for dinner with mikeal and yusuke and malte
01:05:57  <isaacs>i'll try it out tonight maybe
01:05:58  <mikeal>does yusuke know how to get here?
01:06:04  <isaacs>mikeal: not sure
01:06:27  <isaacs>mikeal: do you have a phone number for him or anything?
01:06:36  <mikeal>yes
01:06:40  <mikeal>i'll put it in the email
01:16:29  <davisp>Anyone ever seen the first uv_fs_stat call return ENOENT where a retry fixes it?
01:16:40  <davisp>Ie, the file exists, but it seems like the first call always fails for some reason
01:17:49  <davisp>Also, it looks like req->errorno is alwasy set to ENOENT even though result swaps back to 0x0
01:20:06  * chobi_e_changed nick to chobi_e
01:23:34  * chobi_echanged nick to chobi_e_
01:41:44  * brsonquit (Ping timeout: 264 seconds)
01:43:24  * brsonjoined
01:51:52  <davisp>Hrm. It looks like its the first uv_fs_* request that fails no matter what.
02:07:44  * EhevuTovquit (Quit: This computer has gone to sleep)
02:10:17  <indutny>mikeal: so you're still hanging around with node people? ;)
02:15:07  * AlbireoX_changed nick to AlbireoX
02:21:42  * chobi_e_changed nick to chobi_e
02:23:11  * loladiroquit (Ping timeout: 246 seconds)
02:23:47  <davisp>And I sure can't reproduce in a small test case. I wonder if it requires being started in a therad.
02:23:50  <davisp>thread even.
02:28:19  <indutny>davisp: are you doing access to one loop from different threads?
02:28:27  <indutny>davisp: if so - that's not going to work
02:29:26  <davisp>indutny: no, I'm following the rules there.
02:29:57  <davisp>indutny: Near as I can tell it works fine except for the very first call to a uv_fs_* function. Perhaps its the way I use the async thing.
02:30:10  <indutny>hm...
02:30:25  <indutny>I'm not really in FS stuff, but if you can write a test case - I should be able to fix it
02:30:32  <indutny>s/FS/FSEvents/
02:32:02  <indutny>davisp: ^
02:32:13  * chobi_echanged nick to chobi_e_
02:32:20  <davisp>indutny: its not fsevents. just plain uv_fs_stat is doing this
02:32:33  <indutny>oh, right
02:32:34  <davisp>indutny: I'm trying to work up that test case now. I'll ping you if I find it
02:32:39  <indutny>ok
02:41:15  <davisp>indutny: I sure can't find it with a test case. I'll try and debug it in situ for awhile longer
02:41:52  <indutny>isaacs: basically, after testing for quite a long time this feature looks working to me
02:42:07  * c4miloquit (Remote host closed the connection)
02:42:09  <indutny>isaacs: so I propose to do a code review of it and merge it in a master
02:42:17  <indutny>isaacs: I mean asking ben to review it
02:42:39  <indutny>ircretary: tell bnoordhuis to take a look (review) this https://github.com/joyent/node/pull/3661 once he'll have time ;)
02:42:39  <ircretary>indutny: I'll be sure to tell bnoordhuis
02:42:46  <indutny>ircretary: thanks, I owe you a beer
02:42:46  <ircretary>indutny: I'm not sure what to do with that command. Ask for help in PM.
02:50:27  * EhevuTovjoined
02:56:39  * AlbireoXchanged nick to Guest51731
02:56:59  * Guest51731changed nick to AlbireoX
02:58:21  * AlbireoXchanged nick to Guest38405
02:59:08  * Guest38405changed nick to AlbireoX
03:02:09  <davisp>And I'm a moron
03:02:46  <davisp>indutny: I wasn't passing null-terminated strings to libuv. I'm gonna go hide in the corner now.
03:02:54  <indutny>oh
03:03:05  <indutny>yeah, that happens time to time :)
03:04:43  <indutny>that's why I really like passing string length along with the string's value
03:04:49  <indutny>it always works and easy to check
03:07:02  * brsonquit (Quit: leaving)
03:07:48  <davisp>indutny: Yeah, and it'd save me a memcpy in this case
03:08:22  <davisp>Granted I don't expect the stat function needs to be too terribly efficient, its just one of those whiney dev moments when I have data X and need to fit it into API Y
03:08:34  <deoxxa>does libuv play nice with anything other than glibc?
03:08:36  <davisp>brb, celebratory smoke after finding that.
03:09:18  <davisp>deoxxa: Anything other? I'm not aware of anything it doesn't play well with.
03:09:39  <davisp>brb thoug
03:09:44  <deoxxa>tried to compile with musl, for arm, and never actually got it to the end
03:10:04  <indutny>deoxxa: I don't think we have ever tested it for anything else
03:10:10  <indutny>s/for/with
03:10:24  <indutny>deoxxa: but what do you get, segfaults, test failures, what?
03:10:36  <deoxxa>couldn't even get it to compile all the way
03:10:45  <indutny>deoxxa: well, log would be helpful
03:11:12  <deoxxa>a log from before or after the first hacks i made to get it part of the way?
03:12:32  <deoxxa>i had to remove the stuff for subsecond comparison in fs-stat.c, change -std=c89 to -std=c99 in config-unix.mk, and stop libuv defining HAVE_IFADDRS_H in unix/linux/core.c
03:12:58  <deoxxa>oh and the `typedef intptr_t ssize_t' line in uv.h
03:13:57  <deoxxa>the real problems are in libev, it looks like
03:14:04  <indutny>aah
03:14:44  <deoxxa>libev relies on features.h
03:14:49  <deoxxa>which musl tells me is "bogus"
03:15:07  <deoxxa>(i believe musl here, it's compliant to its own detriment)
03:15:30  <indutny>system programming is odd
03:15:53  <deoxxa>indeed it is, which is why i use libuv so i don't have to do it :P
03:18:31  <deoxxa>`#if LINUX_VERSION_CODE >= 0x020609 && __GLIBC_PREREQ(2, 4)' << from ev/config_linux.h
03:18:48  <deoxxa>might be able to avoid this by reconfiguring libev
03:19:15  <deoxxa>is there anything i should know about how libuv configures libev before i go poking around there?
03:19:44  <indutny>I don't think I can be helpful in this question
03:19:49  * EhevuTovquit (Quit: Leaving)
03:20:00  <indutny>you may try asking bnoordhuis or piscisaureus once they'll be online
03:21:13  <deoxxa>looks like it's just plain old autotools annoyances
03:21:56  <deoxxa>basically because musl is such a niche libc, most projects using autoconf don't work very well with it and usually require manual patches to get running
03:49:51  * ericktquit (Quit: erickt)
04:15:58  * brsonjoined
05:25:43  * Ariaquit (Remote host closed the connection)
05:36:18  * ericktjoined
06:09:02  * ericktquit (Quit: erickt)
07:19:44  * sh1mmerquit (Quit: sh1mmer)
07:33:11  * AlbireoXchanged nick to AlbireoX`Away
07:48:07  * rendarjoined
08:17:43  * hzjoined
08:22:40  * paddybyersjoined
08:26:31  * paddybyersquit (Client Quit)
08:35:24  * saghulquit (Remote host closed the connection)
08:38:46  * AndreasMadsenjoined
08:46:32  <davisp>Anyone have an idea on common causes of triggering this assertion? https://gist.github.com/3f80e0dbcda1665ff3ef
08:47:06  <davisp>I'm basically in a "add_new_things, run_loop_once" type of loop where add_new_things may or may not add new things.
08:47:40  <davisp>Its a very rare error so not super easy to capture with gdb
08:49:34  <davisp>oh, there it might be. Looks like I might be calling uv_fs_req_cleanup more than once per uv_fs_* call
08:51:48  * brsonquit (Ping timeout: 252 seconds)
08:59:23  * paddybyersjoined
09:09:58  * paddybyersquit (Quit: paddybyers)
09:19:31  * txdvquit (Remote host closed the connection)
09:30:12  * paddybyersjoined
09:37:39  * brsonjoined
09:39:13  * paddybyersquit (Quit: paddybyers)
09:42:25  * brsonquit (Ping timeout: 272 seconds)
09:48:25  * brsonjoined
09:54:38  * brsonquit (Ping timeout: 246 seconds)
10:07:26  * loladirojoined
10:08:28  * mmaleckijoined
10:09:04  * txdvjoined
10:18:28  * theColejoined
10:18:40  * indexzeroquit (Quit: indexzero)
12:33:19  * bnoordhuisjoined
13:04:17  * chobi_e_changed nick to chobi_e
13:39:08  <CIA-108>node: Ben Noordhuis v0.8 * r26f1bc8 / src/node_crypto.cc : crypto: fix DecipherUpdate() memory leak - http://git.io/8NCHmQ
13:42:12  * `3rdEdenjoined
14:02:45  * chobi_echanged nick to chobi_e_
14:08:51  <indutny>bnoordhuis: hey man
14:08:59  <indutny>how is it going?
14:14:34  <bnoordhuis>indutny: great. you?
14:14:39  <indutny>great too
14:14:44  <bnoordhuis>you're going to ask me to review that PR, aren't you? :)
14:14:48  <indutny>nooo
14:14:54  <indutny>just talking
14:14:59  <indutny>:)
14:15:02  <bnoordhuis>hah, okay
14:15:08  <indutny>I'm sure you've received that message
14:15:11  <indutny>so that's up to you
14:15:20  <bnoordhuis>i'll look at it today
14:15:29  <indutny>basically, I wanted to ask if you know anything I can work on today
14:15:32  <bnoordhuis>i skimmed it yesterday actually and it mostly lgtm
14:15:46  <bnoordhuis>but i need to review it a little more in-depth
14:15:48  <indutny>cool
14:15:55  <indutny>yeah, it's really security sensitive
14:16:06  <indutny>it should just pass data to OpenSSL in case of any trouble with packets
14:16:12  <bnoordhuis>re things you can work on... is the bug tracker empty?
14:16:17  <indutny>hahaha
14:16:22  <bnoordhuis>:)
14:16:25  <indutny>well, it contains a lot of shit
14:16:32  <indutny>which should be filtered out
14:16:56  <bnoordhuis>that in itself is, if not glamorous work, necessary work
14:17:29  <bnoordhuis>closing invalid bug reports is a great way to practice snippy english
14:18:19  <indutny>thanks man
14:18:21  <bnoordhuis>i try to cram in as much condescending passive aggressiveness as possible without actually crossing the threshold, you know?
14:18:57  <bnoordhuis>i half-kid but it sometimes takes constraint to not go all out
14:19:28  <bnoordhuis>anyway, there must be interesting bugs in there :)
14:20:01  * `3rdEdenquit (Quit: Leaving...)
14:20:08  <bnoordhuis>indutny: https://github.com/joyent/node/issues/3651 <- this one might interest you
14:20:29  <bnoordhuis>i did some quick checks already but i'm not sure if it's a bad test, bad expectations of the user or a bug in node
14:20:55  <indutny>oh, my gosh
14:21:12  <indutny>yeah, will check
14:21:50  * `3rdEdenjoined
14:27:14  * japjjoined
14:28:44  * theColequit (Quit: theCole)
14:33:35  <indutny>must be 'binary' encoded strings
14:33:48  <indutny>I think we should remove that stuff eventually
14:33:50  * paddybyersjoined
14:35:19  <mmalecki>indutny: ++, binary strings are just wrong
14:35:21  <kohai>indutny has 28 beers
14:36:48  <indutny> v8::V8::AddGCPrologueCallback((GCPrologueCallback)dtrace_gc_start);
14:36:48  <indutny> v8::V8::AddGCEpilogueCallback((GCEpilogueCallback)dtrace_gc_done);
14:37:11  <mmalecki>indutny: right, these are events i'm tracing
14:37:15  <mmalecki>oh, gchat, right
14:37:39  <indutny>oh :)
14:39:45  <japj>does anyone know how long until the nodeup cast happening today (or did I miss it already?)
14:41:33  <bnoordhuis>son of a... wtf is up with github lately?
14:41:54  <indutny>bnoordhuis: ?
14:42:34  <bnoordhuis>indutny: the issue tracker is very flaky. things like when i try to post a comment, the 'comment on this issue' button is grayed out
14:42:38  <bnoordhuis>or simply doesn't work
14:42:47  <indutny>ah, that's known
14:43:02  <indutny>issue tracker is really flacky last days
14:43:13  <indutny>you should try clearing your browser's cache
14:43:22  <indutny>may be they fixed some issue, but you hadn't got them yet
14:43:25  <bnoordhuis>hah, that's so internet explorer 6
14:47:12  <indutny>bnoordhuis: btw, you know why I stopped working on shared-memory version of tls storage?
14:47:22  <indutny>s/you/do you/
14:47:36  <bnoordhuis>indutny: no. i have this feeling you're going to tell me though :)
14:47:41  <indutny>haha
14:47:42  <indutny>indeed
14:47:47  <indutny>because we're using execve
14:48:05  <indutny>oh, execvp, actually
14:48:12  <indutny>and it's detaching any shared memory
14:48:16  <bnoordhuis>oh right, the shared mapping won't survive that
14:48:20  <indutny>yes
14:48:27  <bnoordhuis>sorry, i should've thought of that
14:48:34  <indutny>I thought about mmaping files, but there're no way to use mutexes with that
14:48:42  * `3rdEdenquit (Quit: Leaving...)
14:48:46  <indutny>bnoordhuis: np, it's good to learn such things by practice
14:48:47  <indutny>:)
14:49:16  <bnoordhuis>indutny: sysv and posix shm are simple memory segments
14:49:29  <indutny>bnoordhuis: and?
14:49:31  <bnoordhuis>i.e. you can store cross-process mutexes in them
14:49:44  <indutny>you mean named shms?
14:50:00  <indutny>I don't think there's a way to share anonymous shms after execvp
14:50:17  <bnoordhuis>indutny: yes, named shm
14:50:38  <indutny>well, it's good, but it not really closs-platform, I guess
14:50:55  <indutny>well, by saying not-cross-platform I mean that it won't work on some platforms at all
14:51:10  <bnoordhuis>mutexes / semaphores may not, no
14:51:29  <indutny>ah, so share mutexes and use mmap for memory
14:51:40  <bnoordhuis>yes
14:51:43  <indutny>that may work
14:51:52  <indutny>but I really think we should try parsing TLS frames
14:51:56  <bnoordhuis>in theory, sem_*() should work on any unix
14:52:03  <indutny>not on osx
14:52:03  <bnoordhuis>the exception being os x, i think
14:52:06  <bnoordhuis>right :)
14:52:07  <indutny>yes
14:52:28  <bnoordhuis>that's an outstanding bug in libuv btw
14:52:32  <indutny>so, basically, good thing about ClientHello parser is that we can parse TLS extensions here too
14:52:46  <indutny>em... do we have semaphores in libuv?
14:52:52  <bnoordhuis>yes
14:52:55  <indutny>or do you mean semaphore abstraction in general
14:52:56  <indutny>ah
14:53:00  <indutny>ok
14:53:01  <bnoordhuis>uv_sem*
14:53:13  <bnoordhuis>they don't work on os x, though
14:53:29  <bnoordhuis>because i foolishly assumed that apple would've taken the effort to emulate them on os x
14:53:31  <deoxxa>there was a writeup by some postgres guy the other day about this, i think
14:53:33  <bnoordhuis>silly me, i know
14:53:43  <AndreasMadsen>hi, I'm trying to run a v8 profiling using node --prof, but when I run mac-tick-processor on the log nothing happens.
14:53:59  <AndreasMadsen>Any idea why
14:54:05  <bnoordhuis>deoxxa: link?
14:54:08  <deoxxa>yesss, here we are: http://rhaas.blogspot.com.au/2012/06/absurd-shared-memory-limits.html
14:54:17  <deoxxa>(might be relevant)
14:54:36  <indutny>AndreasMadsen: where is v8.log located?
14:54:41  <indutny>AndreasMadsen: how are you running it?
14:54:48  <bnoordhuis>deoxxa: thanks
14:55:13  <deoxxa>actually, it looks to be dealing with a different issue. still might be useful though?
14:55:52  * loladiro_joined
14:55:57  <AndreasMadsen>indutny: I run node --prof test/benchmark/parser.js && ~/GitHub/node/deps/v8/tools/mac-tick-processor
14:56:13  <AndreasMadsen>and the v8 log is located in the cwd
14:56:16  <indutny>ok, you should try moving your v8.log to the directory with v8
14:57:02  <bnoordhuis>deoxxa: interesting read
14:57:19  <bnoordhuis>deoxxa: yes, separate issue but still interesting :)
14:57:42  <AndreasMadsen>indutny: hmm, didn't help
14:57:55  <indutny>AndreasMadsen: interesting, had you built d8?
14:57:55  <AndreasMadsen>I executed tools/mac-tick-processor
14:58:37  <indutny>AndreasMadsen: you should build d8 first
14:58:42  <indutny>AndreasMadsen: scons -j4 d8
14:58:45  <indutny>in v8's folder
14:58:53  <indutny>and better do it in other v8 folder
14:59:17  <indutny>because otherwise you'll endup borked :)
14:59:21  <AndreasMadsen>indutny: I did build d8, but that was in a separate directory, the version might be another
14:59:23  <indutny>s/endup/end up
14:59:40  <indutny>AndreasMadsen: why are you running mac-tick-processor in another directory then?
15:00:02  <indutny>it's executing ../../d8 for processing stuff
15:00:20  * loladiroquit (Ping timeout: 264 seconds)
15:00:20  * loladiro_changed nick to loladiro
15:00:26  <AndreasMadsen>indutny: because it didn't work, and I had to try something
15:00:46  <indutny>well, yo should copy v8.log to v8 folder
15:00:51  <indutny>be sure that this folder contains d8
15:00:56  <indutny>and only then run mac-tick-processor
15:01:22  <AndreasMadsen>indutny: Already tried that
15:01:35  <indutny>AndreasMadsen: oh, interesting
15:01:45  <AndreasMadsen>indutny: will make ia32.release work, I don't have scons
15:02:07  <indutny>AndreasMadsen: you may try inserting some logging into mac-tick-processor script
15:02:15  <indutny>the same pattern works for me for about one year
15:02:33  <tjfontaine>i use the nprof from node-profiler instead of the shipped tcik processor stuff
15:03:03  <tjfontaine>https://github.com/bnoordhuis/node-profiler I dunno if should trust the developer or not
15:03:35  <bnoordhuis>please do. i'm putting your cpu's spare cycles to good use
15:04:13  <indutny>tjfontaine: he has not a lot of followers on github https://github.com/bnoordhuis
15:04:17  <indutny>not sure if he's trustable
15:04:22  <indutny>:D
15:04:41  <bnoordhuis>aww, i try :(
15:06:05  <indutny>omg, it's so hot in NY
15:06:10  <tjfontaine>heh
15:06:13  <indutny>Y IT IS SO HOT HERE
15:06:17  <tjfontaine>welcome to merica!
15:06:18  <LOUDBOT>DAV3XOR UNDER SIEGE, DAY 153
15:06:22  <AndreasMadsen>tjfontaine: bnoordhuis: thanks will try that
15:06:41  <tjfontaine>AndreasMadsen: fwiw I don't use the actual native module just the script that gets built
15:07:13  <bnoordhuis>fwiw, i've been considering splitting it off into a separate module
15:07:20  <tjfontaine>that'd be glorious. :)
15:08:46  * loladiro_joined
15:09:59  * loladiroquit (Ping timeout: 248 seconds)
15:09:59  * loladiro_changed nick to loladiro
15:12:32  <AndreasMadsen>omg, installed profiler, setup profiler.resume(); and profiler.resume(); then executed "node --prof --prof_lazy test/benchmark/parser.js" it runs but there is no v8.log
15:12:52  <AndreasMadsen>setup profiler.resume(); and profiler.pause();
15:14:17  <bnoordhuis>AndreasMadsen: the v8 profiler api has been broken for a while now
15:15:03  <AndreasMadsen>bnoordhuis: hmm
15:15:13  <AndreasMadsen>bnoordhuis: that makes life quite difficult
15:15:37  <bnoordhuis>AndreasMadsen: yeah, it kind of sucks
15:15:49  <bnoordhuis>biab, dinner
15:15:50  <tjfontaine>you can still use normal --prof just not the focused profiling you might want to do :)
15:16:15  <AndreasMadsen>tjfontaine: okay, thanks that will do
15:16:20  <indutny>bnoordhuis: do we have a define for darwin platform?
15:16:26  <indutny>bnoordhuis: like #ifdef PLATFORM_DARWIN
15:16:54  <bnoordhuis>indutny: i cannot reply to your inquiry as i'm having dinner right now
15:17:04  <bnoordhuis>also, no
15:17:04  <indutny>is this autoreply?
15:17:06  <indutny>:)
15:17:07  <bnoordhuis>yes
15:17:08  <indutny>_DARWIN_C_SOURCE
15:36:43  <indutny>ok, osx has no semaphore_trywait
15:36:47  <indutny>unfortunatelly :(
15:37:27  <indutny>oh, there's a hack possible
15:41:21  * mmaleckichanged nick to mmalecki[brb]
15:43:38  <indutny>oh, this function exists in header, but not in library
15:43:44  <indutny>really nice, Apple!
15:46:23  <deoxxa>lol
15:46:33  <indutny>bnoordhuis: https://github.com/joyent/libuv/pull/488/
15:46:42  <indutny>no sem_trywait
15:46:44  <indutny>:(
15:54:35  <indutny>aaah
15:54:40  <indutny>that was a typo in function name
15:54:57  <indutny>omg, I really scared that it was compiling
15:56:20  <indutny>bnoordhuis: ok, implemented trywait for osx too
16:19:16  <indutny>bnoordhuis: btw, we ain't using semaphores anywhere, I guess
16:19:19  <indutny>why do we need them?
16:20:43  * AlbireoX`Awaychanged nick to AlbireoX
16:42:16  * chobi_e_changed nick to chobi_e
16:44:30  * c4milojoined
16:52:11  * mmalecki[brb]changed nick to mmalecki
16:56:56  * mikealquit (Quit: Leaving.)
17:22:34  * `3rdEdenjoined
17:40:21  * V1joined
17:41:50  * c4miloquit (Remote host closed the connection)
17:42:18  * `3rdEdenquit (Disconnected by services)
17:42:23  * V1changed nick to `3rdEden
17:51:04  * mikealjoined
17:58:18  * japjquit (Remote host closed the connection)
18:02:41  * c4milojoined
18:04:50  * paddybyersquit (Quit: paddybyers)
18:05:44  * dshaw_joined
18:10:20  * `3rdEdenquit (Quit: Leaving...)
18:40:14  * piscisaureus_joined
18:40:50  <piscisaureus_>unix, why u no have frename / funlink ?
18:54:03  * EhevuTovjoined
18:54:18  <piscisaureus_>indutny: https://www.pidder.com/pidcrypt/
18:54:42  <piscisaureus_>indutny: but I bet it will be dog slow.
19:08:35  * c4miloquit (Remote host closed the connection)
19:25:43  * piscisaureus_changed nick to piscisaureus
19:55:01  * c4milojoined
20:00:55  * perezdquit (Quit: perezd)
20:06:59  * c4miloquit (Remote host closed the connection)
20:18:15  * voodootikigodquit
20:33:20  * loladiroquit (Ping timeout: 264 seconds)
20:33:34  * paddybyersjoined
20:35:37  * loladirojoined
20:42:36  * philipsquit (Excess Flood)
20:44:15  * philipsjoined
20:52:45  * `3rdEdenjoined
20:53:32  * brsonjoined
20:59:30  * felixgejoined
20:59:30  * felixgequit (Changing host)
20:59:30  * felixgejoined
21:05:22  * ericktjoined
21:07:23  <isaacs>piscisaureus: srsly
21:07:29  <isaacs>piscisaureus: and why not flink?
21:07:38  <isaacs>seems pretty obvious. i have an fd, i want to give it a filename.
21:07:45  <isaacs>it MUST have the ability to do this *somehow*, right?
21:07:57  <piscisaureus>yeah, but apparently it doesn't :-(
21:08:03  <piscisaureus>I have more use cases for funlink
21:14:19  <bnoordhuis>piscisaureus: unlinkat() ?
21:14:38  <piscisaureus>bnoordhuis: No I don't want to unlinkat, I want to unlink the open file itself
21:14:45  <piscisaureus>not "some file in an open dir"
21:25:19  * `3rdEdenquit (Quit: Linkinus - http://linkinus.com)
21:37:50  <CIA-108>libuv: Fedor Indutny v0.8 * r3a54805 / (include/uv-private/uv-unix.h src/unix/thread.c): unix: thread: use mach semaphores on osx - http://git.io/X6y6DQ
21:38:27  <indutny>bnoordhuis: hi
21:38:28  <indutny>paddybyers: hi
21:38:30  <indutny>piscisaureus: hi
21:38:39  <indutny>bnoordhuis: so we have mach semaphores in uv now
21:38:42  <indutny>but why? :)
21:38:57  <piscisaureus>yes, explain that
21:39:59  * travis-cijoined
21:39:59  <travis-ci>[travis-ci] joyent/libuv#487 (v0.8 - 3a54805 : Fedor Indutny): The build is still failing.
21:39:59  <travis-ci>[travis-ci] Change view : https://github.com/joyent/libuv/compare/be09be7f3efd...3a548053b8d7
21:39:59  <travis-ci>[travis-ci] Build details : http://travis-ci.org/joyent/libuv/builds/1812362
21:39:59  * travis-cipart
21:40:11  <bnoordhuis>indutny, piscisaureus: to help out e.g. the rust guys
21:40:20  <indutny>bnoordhuis: em... with what?
21:40:26  <indutny>I mean why "mach"?
21:41:25  <bnoordhuis>rust is an example but they have a lot of platform specific mutex / condvar / semaphore code
21:41:52  <bnoordhuis>indutny: what do you mean 'why mach'? because os x doesn't support anything else?
21:42:02  <indutny>well, it was working
21:42:07  <indutny>at least tests were passing
21:42:15  <indutny>but I don't really know why v8 and others are using mach
21:42:18  <indutny>instead of posix
21:43:39  <bnoordhuis>indutny: because sem_init() and friends don't work
21:43:49  <bnoordhuis>we discussed that this afternoon, didn't we?
21:43:56  <indutny>yes, we've discussed that
21:44:02  <indutny>and there was something about shared stuff
21:44:16  <indutny>and I can't remember how, but we get to the point where we decided that they ain't working :)
21:44:39  <bnoordhuis>oh, that's because anon shared memory doesn't survive execve
21:44:52  <bnoordhuis>i guess we'll have to use named memory instead
21:44:57  <indutny>eys
21:45:04  <indutny>bnoordhuis: ok, nevermind
21:45:11  <indutny>we just have mach semaphores on mach platform now
21:45:12  <indutny>:D
21:45:15  <indutny>for some reason
21:45:19  <bnoordhuis>good for us, eh?
21:46:24  <indutny>I hope so
21:46:34  <indutny>:D
21:47:46  <mmalecki>indutny: o hai
21:47:55  <mmalecki>indutny: hij1nx wants to know if you're at ace hotel in ny
21:51:31  <indutny>mmalecki: o hai
21:51:39  <indutny>well, I'm walking all around
21:51:52  <indutny>leaving it now
21:51:59  <indutny>mmalecki: is he near you?
21:52:56  <mmalecki>indutny: no, I'm talking to him on skype
21:54:00  <indutny>mmalecki: oh, say hi to him then ;)
22:06:16  * hzquit
22:08:18  * AndreasMadsenquit (Remote host closed the connection)
22:25:14  * philipsquit (Excess Flood)
22:27:45  * philips_joined
22:30:59  * mjr_quit (Quit: mjr_)
22:39:55  * mmaleckiquit (Quit: Lost terminal)
22:40:15  * mmaleckijoined
22:50:22  <mmalecki>I love it how node v0.8.1 noticably reduced our laod balancer response time
22:50:45  <mmalecki>in some cases it's half of what we've been seeing on 0.6
23:09:25  * rendarquit
23:16:20  * paddybyersquit (Quit: paddybyers)
23:16:58  * EhevuTovquit (Quit: This computer has gone to sleep)
23:30:39  * EhevuTovjoined
23:31:19  * felixgequit (Quit: felixge)
23:43:18  * piscisaureuschanged nick to piscisaureus_
23:46:27  * piscisaureus_quit (Read error: Connection reset by peer)
23:48:08  * piscisaureus_joined
23:53:49  * piscisaureus_quit (Ping timeout: 251 seconds)
23:57:39  * loladiroquit (Read error: Connection reset by peer)
23:57:45  * loladiro_joined