00:01:53  * brsonquit (Quit: leaving)
00:23:01  * theColequit (Quit: theCole)
00:30:17  * txdvquit (Ping timeout: 244 seconds)
00:34:17  * txdvjoined
00:44:41  * igorziquit (Ping timeout: 245 seconds)
00:46:18  * brsonjoined
00:52:44  * dylukesjoined
01:28:48  * abraxasjoined
01:53:21  * k-s[AWAY]changed nick to k-s
01:57:13  * dylukesquit (Quit: Pipes are broken. Sending packets via Fedex.)
01:58:01  * k-schanged nick to k-s[AWAY]
01:59:31  * mmaleckichanged nick to mmalecki[zzz]
02:00:10  * bnoordhuisquit (Ping timeout: 276 seconds)
02:11:48  * orlandovftwjoined
02:12:00  * dshaw_joined
02:13:49  * mmalecki[zzz]changed nick to mmalecki
02:16:13  * dshaw_1joined
02:16:45  * dshaw_quit (Read error: Connection reset by peer)
02:19:07  * dshaw_1quit (Client Quit)
02:27:22  * brsonquit (Quit: leaving)
02:36:35  * orlandovftwquit (Ping timeout: 248 seconds)
03:02:07  * orlandovftwjoined
03:26:04  * pfox___quit (Ping timeout: 244 seconds)
03:26:55  * pfox___joined
03:54:28  * Ariaquit (Remote host closed the connection)
05:06:51  * sh1mmerquit (Quit: sh1mmer)
05:41:04  * TooTallNatequit (Quit: Linkinus - http://linkinus.com)
05:54:10  * eli-away_quit (Ping timeout: 276 seconds)
05:56:21  * eli-awayjoined
06:21:45  * mjr_joined
07:00:31  * isaacsjoined
07:05:01  * dshaw_joined
07:25:00  * milanijoined
07:25:45  <milani>hi can anyone help me to understand what `uv_ref (uv_default_loop ());` does?
07:26:19  <milani>does it force uv to wait for something to be complete? inorder to avoid program termination?
07:26:20  * rendarjoined
07:30:36  * pfox___quit (Ping timeout: 272 seconds)
07:45:08  <benvie>it adds a reference to the main loops refcount which will cause it to not quit
07:46:07  <milani>:) thx. I thought so.
07:47:06  <milani>I'm working on gui stuff. I can not get webkit to work. When ever I initialize webkit, segmentation fault occurs. I think it's about threads!
07:49:56  * mjr_quit (Ping timeout: 260 seconds)
07:51:43  <benvie>hah yeah that's likely to do with something about the V8
07:52:08  <benvie>getting them on the same V8 is tricky
07:52:59  <rendar>really? shouldn't be easy using threads with v8 and node?
07:53:52  <benvie>it's more about initializing webkit on top of the already running v8 in node
07:54:18  * paddybyersjoined
07:55:25  <milani>benvie: not that webkit! I mean webkitgtk+
07:55:31  <benvie>oooh
07:55:46  <milani>I think running webkit on top of v8 is a big fish for me!
08:09:26  * isaacsquit (Remote host closed the connection)
08:21:10  * dshaw_quit (Quit: Leaving.)
08:23:23  * dshaw_joined
09:08:38  * dshaw_quit (Quit: Leaving.)
09:34:05  * orlandovftwquit (Ping timeout: 272 seconds)
10:24:25  * piscisaureus_joined
10:47:54  * abraxasquit
10:49:29  * paddybyers_joined
10:52:21  * paddybyersquit (Ping timeout: 244 seconds)
10:52:21  * paddybyers_changed nick to paddybyers
11:54:25  * mmaleckichanged nick to mmalecki[out]
11:55:27  * bnoordhuisjoined
12:09:39  <piscisaureus_>bnoordhuis: hey
12:12:40  * benviequit
12:14:33  <bnoordhuis>piscisaureus_: ho
12:14:40  <piscisaureus_>bnoordhuis: ha
12:14:55  <bnoordhuis>piscisaureus_: hi
12:15:01  <piscisaureus_>bnoordhuis: hö
12:15:36  <piscisaureus_>bnoordhuis: uv_poll_t or uv_io_t or whatever the thing will be named. Does that have to be thread-safe, e.g. should users be able to create multiple poll watchers for the same fd?
12:15:51  <piscisaureus_>bnoordhuis: (and - on different threads / loops?)
12:16:27  <bnoordhuis>piscisaureus_: uhm... any particular reason to make that guarantee?
12:17:22  <piscisaureus_>bnoordhuis: Umm no I'd rather not
12:17:36  <piscisaureus_>bnoordhuis: but so if someone starts polling stdin stuff may break.
12:17:51  <bnoordhuis>how so?
12:18:05  <piscisaureus_>bnoordhuis: well if they poll stdin from 2 threads, I mean
12:18:09  <piscisaureus_>if we don't guarantee that
12:18:37  <bnoordhuis>people shouldn't do that
12:18:46  <bnoordhuis>just make a note of the gotchas in the README or uv.h
12:18:51  <piscisaureus_>People are idiots
12:18:52  <piscisaureus_>yeah
12:18:53  <piscisaureus_>ok
12:19:01  <bnoordhuis>btw, why did you merge that uv_is_closing patch?
12:19:17  <piscisaureus_>bnoordhuis: I agreed with it
12:19:21  <piscisaureus_>bnoordhuis: you don't
12:19:23  <piscisaureus_>?
12:19:34  <bnoordhuis>well... people need to track their own state
12:19:44  <bnoordhuis>i discussed that with dvv a while ago
12:19:44  <piscisaureus_>bnoordhuis: well yeah somewhat
12:19:55  <piscisaureus_>bnoordhuis: I told him that
12:20:10  <piscisaureus_>bnoordhuis: but I was thinking, what if you use uv_walk and you get the thing reported to you
12:20:13  <piscisaureus_>bnoordhuis: then i
12:20:15  <piscisaureus_>er
12:20:28  <piscisaureus_>then it'd be neat to be able to tell whether the handle is being closed or not
12:20:52  <bnoordhuis>embed the handle in another struct and maintain a state flag?
12:21:05  <piscisaureus_>bnoordhuis: yeah that is also possible ... but what's wrong with this?
12:21:13  <piscisaureus_>bnoordhuis: I mean, we also have uv_is_active
12:21:15  <bnoordhuis>it's not that that particular patch is horrendous but i don't like hand-holding
12:21:30  <bnoordhuis>yeah, which is a minor abomination unto me
12:21:30  <piscisaureus_>they can track that themselves can't they
12:21:49  <bnoordhuis>i could live with removing uv_is_active :)
12:21:49  <piscisaureus_>bnoordhuis: no, I think this is agreeable
12:22:23  <piscisaureus_>bnoordhuis: although what I don't like is that dvv discussed the thing with you and you said no
12:22:34  <piscisaureus_>bnoordhuis: then he proposes it to me
12:22:45  <piscisaureus_>bnoordhuis: that's kinda childish innit
12:23:04  <bnoordhuis>and then you didn't consult me :)
12:23:19  <piscisaureus_>bnoordhuis: luckily, I don't have to :-p
12:23:34  <bnoordhuis>no, but it always ends in tears if you don't
12:23:46  <piscisaureus_>bnoordhuis: but srsly, I think it's good to expose this kind of state to the user.
12:24:11  <bnoordhuis>but what's next? uv_is_readable / uv_is_writable?
12:24:12  <piscisaureus_>bnoordhuis: since the state is not an implementation detail, but it's part of the conceptual model that libuv uses
12:24:18  <bnoordhuis>don't tell me patches for that were landed
12:24:33  <piscisaureus_>bnoordhuis: uv_is_readable and writable have been in there for a while , yup
12:25:18  * milaniquit (Ping timeout: 252 seconds)
12:26:10  <piscisaureus_>bnoordhuis: we actually could not do without that
12:26:42  <piscisaureus_>bnoordhuis: since on windows you can have half-suplex pipes, you need them to determine the state of a pipe after connecting or accepting.
12:27:23  <piscisaureus_>(side note: half duplex is a really weird complex. Why is it not called simplex or monoplex?)
12:27:32  * benviejoined
12:27:36  <piscisaureus_>unplex
12:27:42  <bnoordhuis>noplex
12:27:49  <bnoordhuis>are we going to have uv_is_connected?
12:27:56  <piscisaureus_>hmm, good question
12:28:01  <piscisaureus_>I guess
12:28:39  <piscisaureus_>so what if a socket is sent to you over the ipc channel and you accept is
12:28:41  <piscisaureus_>*it
12:28:50  <piscisaureus_>how are you going to tell whether it is a server or a connection?
12:28:55  <piscisaureus_>^-- bnoordhuis
12:29:21  <piscisaureus_>bnoordhuis: maybe we should just export uv_handle_state() as a bitmask though
12:29:29  <piscisaureus_>so we can have the flags
12:31:19  <bnoordhuis>how are you going to tell whether it is a server or a connection? <- by having two ipc channels?
12:31:32  <piscisaureus_>ACTIVE <- if it's active
12:31:32  <piscisaureus_>REFFIN <- if it's ref'ing (under the new semantics)
12:31:32  <piscisaureus_>READABLE
12:31:32  <piscisaureus_>WRITABLE
12:31:32  <piscisaureus_>ACCEPTABLE <- This one is just to piss off bnoordhuis
12:31:33  <piscisaureus_>CONNECTION
12:31:33  <piscisaureus_>SERVER
12:31:34  <piscisaureus_>CLOSING
12:31:43  <piscisaureus_>bnoordhuis: two ipc channels?
12:31:48  <piscisaureus_>bnoordhuis: we're going to have that in node?
12:31:57  <bnoordhuis>i don't know
12:32:15  <piscisaureus_>bnoordhuis: it seems to me that uv_is_connected is the superior option right?
12:32:19  <bnoordhuis>my point is that you don't need a library function to determine the handle type
12:33:31  <piscisaureus_>bnoordhuis: not for the type, no, because the handle->type field is publiv
12:33:33  <piscisaureus_>*public
12:33:39  <piscisaureus_>bnoordhuis: but the flags field is not
12:34:30  <piscisaureus_>bnoordhuis: skype
12:35:12  <bnoordhuis>not now, i'll have to be afk in a minute
12:39:37  <piscisaureus_>bnoordhuis: slacking again?
12:40:18  <bnoordhuis>piscisaureus_: not quite. kinderdagverblijfintakegesprek
12:40:42  <bnoordhuis>i would love to hear americans pronounce that
12:41:04  <piscisaureus_>kinderdagverblijfintakegesprekwachtkamertijdschriftenstaperhouderrekje
12:41:30  <piscisaureus_>kinderdagverblijfintakegesprekwachtkamertijdschriftenstapelhouderrekje
12:46:04  * benviequit
12:49:04  * paddybyers_joined
12:51:32  * paddybyersquit (Ping timeout: 246 seconds)
12:51:33  * paddybyers_changed nick to paddybyers
13:01:16  <bnoordhuis>piscisaureus_: https://github.com/joyent/node/issues/3051 <- is that supposed to work?
13:06:40  * benviejoined
13:22:40  * mmalecki[out]changed nick to mmalecki
13:33:00  <piscisaureus_>bnoordhuis: ouch. Gigabyte-size files are broken in windows.
13:33:05  <piscisaureus_>bnoordhuis: do they work under unix?
13:33:15  <bnoordhuis>piscisaureus_: yes. what's broken?
13:33:28  <piscisaureus_>bnoordhuis: well the offset just wraps
13:33:40  <bnoordhuis>is it a 32 bit integer?
13:33:40  <piscisaureus_>bnoordhuis: so the max offset I can write is INT_MAX
13:33:43  <piscisaureus_>yes
13:33:47  <piscisaureus_>signed even
13:33:56  <bnoordhuis>hah, that's so not forward looking
13:34:02  <piscisaureus_>bnoordhuis: which is weird because I know windows supports 64bit offsets
13:34:05  <bnoordhuis>what syscall is that?
13:34:10  <piscisaureus_>bnoordhuis: no we're doing it wrong in node
13:34:14  <piscisaureus_>bnoordhuis: or libuv
13:34:18  <piscisaureus_>since windows itself can handle it
13:34:22  <bnoordhuis>node, probably
13:34:34  <bnoordhuis>we're using Int32Value() and Uint32Value() in most places
13:34:35  <piscisaureus_>bnoordhuis: then it would be broken as well under unix
13:35:45  <bnoordhuis>piscisaureus_: #define GET_TRUNCATE_LENGTH(a) ((a)->Uint32Value()) <- things like that
13:36:15  <bnoordhuis>#define GET_OFFSET(a) ((a)->IsNumber() ? (a)->Int32Value() : -1) <- or that
13:36:55  <bnoordhuis>should probably be something like static_cast<int64_t>((a)->NumberValue())
13:37:14  <bnoordhuis>or off_t
13:53:20  * pfox___joined
13:57:50  <piscisaureus_>bnoordhuis: NumberValue is double?
13:58:04  <piscisaureus_>bnoordhuis: probably we should also assert that it is in the precise range for doubles
13:58:24  <piscisaureus_>bnoordhuis: is there a way to get the mantissa size for platform doubles?
14:01:44  * pfox___quit (Ping timeout: 252 seconds)
14:03:38  * pfox___joined
14:08:10  * pfox___quit (Ping timeout: 244 seconds)
14:11:11  <piscisaureus_>bnoordhuis: I am pretty sure that it also does not work on unix, since we're using Int32Value
14:11:23  <piscisaureus_>in node_file.cc
14:51:42  * AvianFlujoined
14:53:39  * pfox___joined
14:56:41  <bnoordhuis>piscisaureus_: yes. i thought you were talking about libuv
14:56:46  <bnoordhuis>this being #libuv and all :)
14:57:11  <piscisaureus_>bnoordhuis: yeah well libuv-win should support 64 bit offsets
14:57:46  <piscisaureus_>bnoordhuis: well that is on 32 bit
14:57:58  <bnoordhuis>piscisaureus_: btw, re mantissa size, you probably want <float.h>
14:58:03  <piscisaureus_>bnoordhuis: er, *not* on 32 bit
14:58:26  <bnoordhuis>piscisaureus_: on 32 bit too, i assume?
14:58:45  <piscisaureus_>bnoordhuis: no :-(
14:58:55  <bnoordhuis>wut? windows :(
14:59:06  <piscisaureus_>bnoordhuis: uv_fs_write takes size_t and off_t :-(
14:59:10  <piscisaureus_>bnoordhuis: windows is not the problem here
14:59:20  <piscisaureus_>bnoordhuis: on x32 size_t is 32 bit
14:59:29  <piscisaureus_>bnoordhuis: so I think unix would be affected as well :-(
15:00:00  <bnoordhuis>oh, right
15:00:03  <bnoordhuis>we should fix that
15:00:07  <piscisaureus_>yes
15:00:15  <piscisaureus_>but we can't do that really on the 0.6 branch
15:00:19  <bnoordhuis>i'll open an issue
15:00:26  <piscisaureus_>I already did for node
15:00:30  <bnoordhuis>i just created a v0.8 milestone, this fits right in :)
15:00:37  <bnoordhuis>what's the issue #?
15:00:55  <piscisaureus_>https://github.com/joyent/node/issues/3053
15:02:57  * sh1mmerjoined
15:02:57  <bnoordhuis>btw, size_t is used for the buffer length, but the offset is off_t
15:03:09  <bnoordhuis>which is 64 bits on unices if _LARGEFILE_SOURCE is defined
15:03:14  <bnoordhuis>how does that work on windows?
15:04:05  <bnoordhuis>piscisaureus_: ^
15:04:37  <piscisaureus_>bnoordhuis: hmm
15:04:49  <piscisaureus_>my intuition says 32 bit
15:04:59  <bnoordhuis>okay. in that case it's a windows issue
15:05:46  <piscisaureus_>long
15:05:47  <bnoordhuis>quick googling returns some contradictory results
15:05:50  <piscisaureus_>yeah, that's always 32 bit
15:06:31  <bnoordhuis>right. windows issue it is then
15:07:15  <piscisaureus_>uv_off_t it is :-)
15:07:53  <piscisaureus_>bnoordhuis: we're not going to do 2135
15:08:18  <bnoordhuis>Stat fails with ENOENT on UNC paths containing spaces?
15:08:54  <piscisaureus_>*shrug* is that even allowed?
15:09:30  <piscisaureus_>bnoordhuis: well, works for me
15:09:49  <bnoordhuis>piscisaureus_: i don't even know what it means :)
15:10:17  <piscisaureus_>bnoordhuis: I said, 2135
15:10:54  <bnoordhuis>it's monday, you have to spoon feed it to me
15:11:06  <bnoordhuis>oh wait
15:11:11  <piscisaureus_>Dude. You should learn to type :-)
15:11:12  <piscisaureus_>or read :-)
15:11:18  <bnoordhuis>haha
15:11:29  <bnoordhuis>okay, duly noted
15:12:32  <piscisaureus_>bnoordhuis: I edited your issue
15:12:39  <piscisaureus_>bnoordhuis: since it is even the case on 64-bit windows
15:13:05  <bnoordhuis>windows :(
15:13:33  <piscisaureus_>bnoordhuis: but still, it doesn't work in node
15:13:41  <bnoordhuis>no
15:13:57  <bnoordhuis>we need to use NumberValue()
15:14:11  <bnoordhuis>with some checks for very small or large numbers
15:14:17  <piscisaureus_>bnoordhuis: and ASSERT that it is in the precise range
15:14:18  <piscisaureus_>yes
15:14:30  <piscisaureus_>or - rather check that it is in the precise range and throw
15:14:34  <bnoordhuis>yes
15:18:16  <piscisaureus_>bnoordhuis: so on unix you cannot write more than 2^31 bytes in a single write?
15:18:25  <piscisaureus_>bnoordhuis: (agreed that this is highly theoretical)
15:19:24  <bnoordhuis>piscisaureus_: size_t is unsigned
15:19:45  <bnoordhuis>so 2^32 on platforms where size_t is 32 bits
15:20:22  <piscisaureus_>bnoordhuis: ok. So this will always be 32 bits even windows, even on x64
15:20:28  <bnoordhuis>wut? why?
15:20:39  <piscisaureus_>bnoordhuis: because DWORD is always 32 bit
15:20:44  <piscisaureus_>and size is specified as DWORD
15:20:55  <bnoordhuis>there's not a 64 bits API?
15:20:59  <piscisaureus_>no
15:21:07  <bnoordhuis>...
15:21:08  <piscisaureus_>bnoordhuis: but *shrug* this is highly theoretical
15:21:11  * bnoordhuisis speechless
15:21:15  <bnoordhuis>oh, writes
15:21:18  <piscisaureus_>yes
15:21:22  <piscisaureus_>write sizes
15:21:23  <bnoordhuis>yeah, a max 4 GB write is fine
15:21:24  <piscisaureus_>I mean
15:21:39  <bnoordhuis>people won't be doing that much for the foreseeable future
15:23:23  <piscisaureus_>this is nice though
15:23:23  <piscisaureus_>To write to the end of file, specify both the Offset and OffsetHigh members of the OVERLAPPED structure as 0xFFFFFFFF. This is functionally equivalent to previously calling the CreateFile function to open hFile using FILE_APPEND_DATA access.
15:24:33  <bnoordhuis>yes, that's kind of nice
15:26:04  * isaacsjoined
15:29:31  <bnoordhuis>isaacs: morning
15:29:38  <isaacs>bnoordhuis: yo
15:29:43  <bnoordhuis>i have a question
15:29:54  <bnoordhuis>how was your weekend?
15:29:57  <bnoordhuis>that's not the question btw
15:30:06  <bnoordhuis>but i thought i'd start with the niceties :)
15:30:10  <isaacs>hahah
15:30:15  <isaacs>it's been lovely, thanks
15:30:18  <isaacs>what's up?
15:30:20  <bnoordhuis>great to hear :)
15:30:25  <bnoordhuis>what's the maintenance plan for v0.6?
15:30:26  <tjfontaine>now let him ruin it
15:30:28  <bnoordhuis>after v0.8 comes out that is
15:30:52  <isaacs>roughly the same as v0.4 after v0.6 came out.
15:31:00  <isaacs>ie, get people onto v0.8 as fast as we can
15:31:01  <bnoordhuis>declare it dead and move on?
15:31:16  <isaacs>we DO have a lot more users than we did with 0.4
15:31:20  <bnoordhuis>yes
15:31:26  <bnoordhuis>i would like to keep v0.6 alive for a while
15:31:29  <isaacs>yeah
15:31:37  <bnoordhuis>i'll maintain it if you don't want to
15:31:38  <isaacs>i'm sure joyent and voxer and yammer and others would as well
15:31:54  <isaacs>let's not declare it dead, but just declare it sleeping
15:32:14  <tjfontaine>not-quite-dead-yet
15:32:26  <isaacs>like, if there's a major bug that's easily fixed in a very low-risk way, sure, we'll do a build for it
15:32:48  <bnoordhuis>well, i would like to see minor bugs fixed as well
15:32:55  <isaacs>sure.
15:32:59  <bnoordhuis>like i said, i'll maintain it
15:33:01  <isaacs>but it must be low-risk
15:33:11  <bnoordhuis>or maybe someone else steps up
15:33:14  <bnoordhuis>one can hope :)
15:33:16  <isaacs>unless you're really THAT much of a masochist that you really enjoy doing maintenance builds, though, i don't want you spending a considerable amount of your time doing this.
15:33:27  <isaacs>we need you pushing on master.
15:33:45  <isaacs>i'm working on getting the maintenance builds to be virtually zero effort.
15:33:57  <bnoordhuis>in what way?
15:34:03  <isaacs>by automating a lot of it.
15:34:50  <bnoordhuis>good :)
15:35:30  <isaacs>once we get the CI stuff set up, it's going to be easier to track test results and performance. then the next step is a way to say "Make a build from commit cafebad08080", edit the commit message, and then it does everythig
15:35:47  <bnoordhuis>that sounds really nice
15:36:00  <isaacs>like, *everything*
15:36:12  <isaacs>the blog post, the tweet, the emails, the tag, etc, etc
15:37:03  <isaacs>if we have something like that, then a) weekly releases are not hellish, so we can move master faster, and b) 0.6 maintenance is a lot easier.
15:37:15  <bnoordhuis>yes. sounds good
15:37:27  <bnoordhuis>re v0.6, i would like to propose some kind of support schedule
15:37:36  <isaacs>k, what do you have in mind?
15:37:44  <bnoordhuis>nothing very specific yet
15:38:10  <bnoordhuis>but something along the lines of "bug fix support until september 30th, critical security fixes until december 31th"
15:38:33  <bnoordhuis>that should give people time enough to migrate to v0.8
15:38:38  <isaacs>yeah
15:38:47  <isaacs>i'd like to base the dates off of whenever 0.8.0 comes out.
15:38:53  <bnoordhuis>yes, certainly
15:39:11  <isaacs>like, 6 months for any bug fixes, 9 months for critical securyt
15:39:15  <isaacs>*security
15:39:22  <bnoordhuis>yes, exactly
15:40:00  * TooTallNatejoined
15:40:55  <bnoordhuis>isaacs: re v0.8, do you have a target date in mind?
15:41:13  <bnoordhuis>i want to have libuv v0.8 ready by the end of the month
15:41:40  <isaacs>I'd like to be release-worthy by June 1
15:42:02  <bnoordhuis>and what will the main features be? domains?
15:42:15  <piscisaureus_>umm that everything works well ?
15:42:15  <isaacs>domains and new stdio child_process api
15:42:20  <isaacs>and that :)
15:42:24  <isaacs>that's a big new feature for node :)
15:42:32  <bnoordhuis>i personally would like to see great cpu/heap profiler support go in
15:42:44  <bnoordhuis>but since that's still in the conceptual phase...
15:42:49  <isaacs>bnoordhuis: yeah
15:43:03  <isaacs>my feeling is, if it's a good idea, and it's done in time, sure, throw it in.
15:43:21  <isaacs>but the things that we're actually committed to are domains and stdio fixes
15:43:30  <bnoordhuis>right
15:43:33  <isaacs>the rest of what we need is pretty well finished and working at this point
15:43:38  <bnoordhuis>who is going to work on domains?
15:43:47  <bnoordhuis>piscisaureus_: ?
15:44:18  <isaacs>a first step should be to expose the open handles from libuv somewhere in node.
15:44:26  <isaacs>sorry, not open handles, open requests.
15:44:40  <isaacs>like, "why is the event loop not exiting"
15:45:01  <bnoordhuis>right
15:45:10  <bnoordhuis>that could be part of the refcount refactor
15:45:27  <isaacs>that will be very helpful for debugging.
15:45:32  <isaacs>it's a major pain point
15:45:49  <bnoordhuis>i know :)
15:46:05  <bnoordhuis>i did some work on that but it's pretty stale by now
15:46:13  <isaacs>domains is shaping up to be something that we can do largely at the JS layer with API. i'll take a swing at it.
15:46:20  <bnoordhuis>okay, cool
15:46:26  <isaacs>bnoordhuis: review, please? https://github.com/isaacs/node/commit/zlib-errors
15:47:42  <bnoordhuis>isaacs: is Object.keys better? i always use for in
15:47:52  <isaacs>bnoordhuis: Object.keys is faster.
15:47:58  <isaacs>bnoordhuis: doesn't look at __proto__
15:48:04  <bnoordhuis>ah okay
15:48:13  <isaacs>but it's not like that's really a bottleneck most of the time
15:48:17  <isaacs>just habit
15:48:25  <bnoordhuis>right, just curious
15:48:40  <bnoordhuis>the reason i use for in is because it's less typing :)
15:48:47  <isaacs>it's pretty common to do for(i in obj) if(obj.hasOwnProperty(i)) { for safety
15:49:08  <isaacs>and Object.keys(..).forEach(..) automatically implies that
15:49:23  <isaacs>also, no way to break
15:50:02  <isaacs>speaking of fast, and break, i'm off to breakfast.
15:50:52  <isaacs>i'll be jsconfing today, online sporadically.
15:59:41  <TooTallNate>lucky :p
16:00:32  * isaacsquit (Remote host closed the connection)
16:01:14  <piscisaureus_>ircretary: tell isaacs Can't we use substack to make a proposal for domains?
16:01:14  <ircretary>piscisaureus_: I'll be sure to tell isaacs
16:02:40  <piscisaureus_>bnoordhuis: can gcc deal with anonymous unions?
16:02:57  <piscisaureus_>bnoordhuis: and anonymous structs?
16:03:59  <tjfontaine>piscisaureus_: you mean http://gcc.gnu.org/onlinedocs/gcc/Unnamed-Fields.html#Unnamed-Fields ?
16:05:28  <piscisaureus_>yeah
16:05:59  * philipsquit (Excess Flood)
16:09:19  * philipsjoined
16:09:43  <piscisaureus_>bnoordhuis: is off_t signed or unsiged?
16:15:26  * k-s[AWAY]changed nick to k-s
16:17:01  <piscisaureus_>hmm
16:17:05  <piscisaureus_>more fuckups
16:17:07  * dapjoined
16:17:33  <mmalecki>spoiler: anything you try to get done today will eventually get fucked up
16:17:43  <mmalecki>it's that kind of day, piscisaureus_.
16:21:17  <AvianFlu>mmalecki is just mad cause he's on the wrong end of a strike for the first time XD
16:21:45  <mmalecki>OK FUCK
16:22:03  <mmalecki>so like, I'm going to have problems coming back home as well
16:22:18  <piscisaureus_>who'se striking?
16:22:18  <mmalecki>backage handlers are striking on friday/monday.
16:22:24  <mmalecki>air controllers
16:22:38  <piscisaureus_>maybe you should pay the guys
16:23:25  <mmalecki>right
16:23:46  <tjfontaine>no, you *personally* go out, and slip them a couple hundred to help you get home
16:23:53  <mmalecki>London -> Spain -> France.
16:24:00  <mmalecki>Spain -> France with a car.
16:24:06  <mmalecki>because fucking air contol
16:25:29  <piscisaureus_>lol
16:25:39  <tjfontaine>how long will the trip take? are we talking madrid or barcelona? cause one is significantly closer to france :P
16:25:47  <piscisaureus_>mmalecki: you have to get to charles de gaulle from spain?
16:25:55  <mmalecki>near Barcelona, I think
16:26:03  <piscisaureus_>oh, then it's not so bad
16:26:04  <mmalecki>Girona, actually
16:26:08  * dshaw_joined
16:26:24  <piscisaureus_>huh
16:26:30  <piscisaureus_>mmalecki: girona is also in spain
16:26:35  <mmalecki>piscisaureus_: I know
16:26:43  <tjfontaine>that's where he's starting his spain portion of the trip :)
16:26:54  <piscisaureus_>ah
16:26:57  <piscisaureus_>girona to paris is nice
16:27:02  <piscisaureus_>too bad it's about 1000 km
16:27:16  <AvianFlu>luckily mmalecki is only going to some little spot on the riviera
16:27:48  <AvianFlu>probably only a couple hundred km
16:27:54  <AvianFlu>which in real units is like, nothing
16:27:56  <AvianFlu>XD
16:27:56  <piscisaureus_>ah okay
16:28:00  <mmalecki>I hope we'll go by a van.
16:28:10  <mmalecki>we have 6 bottles of vodka.
16:28:24  <mmalecki>this is going to be a party van.
16:28:48  <piscisaureus_>You are living up to all expectations
16:29:05  <piscisaureus_>Drunk polish guys in a van that almost falls apart
16:29:14  <piscisaureus_>There's plenty of those here
16:30:01  <piscisaureus_>I sincerely dislike the synchronous uv_fs_t api
16:30:02  <AvianFlu>mmalecki, I once did this drive in a van with 8 guys and a bunch of whiskey http://g.co/maps/4bkpm
16:30:26  <tjfontaine>9that's a fun route
16:30:32  * orlandovftwjoined
16:30:42  <mmalecki>AvianFlu: neat. and we're doing it cross country!
16:31:07  * dshaw_quit (Quit: Leaving.)
16:31:33  <AvianFlu>the part on that map that goes through LA
16:31:34  <AvianFlu>took 5 hours
16:31:36  <mmalecki>also, I hope indexzero's flight will be all good
16:32:02  <mmalecki>or like, we could pick him up with that van.
16:32:05  <AvianFlu>I hate the 405 sooooooo much
16:32:12  <piscisaureus_>Is this the nodejitsu company trip?
16:32:18  <mmalecki>AvianFlu: is he around, per chance?
16:32:28  <AvianFlu>I'll be seeing him in like 20 minutes
16:32:29  <mmalecki>piscisaureus_: nah. me and him are speaking there
16:32:36  <piscisaureus_>ah, right
16:32:36  <AvianFlu>I'm waiting for marak to get dressed
16:32:44  <mmalecki>and we're in London with frontend girl who's speaking there too
16:33:00  <AvianFlu>the rest of us are at jsconf
16:33:18  <mmalecki>and my friend, who's just attending. it's complicated.
16:33:56  * pieternjoined
16:38:03  * AvianFluquit (Quit: Leaving)
16:42:15  <piscisaureus_>bnoordhuis: hey, yt?
16:42:28  <piscisaureus_>where are isaacs and igorzi when you need them?
16:49:37  <piscisaureus_>omg omg uv_fs_readdir is bad
16:54:25  <rendar>piscisaureus_: whats app? can i help?
16:56:33  <piscisaureus_>rendar: well I need to discuss with bnoordhuis and igorzi what has to be done about it
16:56:44  <rendar>oh, i see :-)
16:56:44  <piscisaureus_>rendar: but at the very least it should be a stream
16:57:34  <rendar>like a binary stream?
16:57:49  <piscisaureus_>no a stream of filenames
16:58:04  <rendar>oh yes, i agree
16:58:20  <piscisaureus_>maybe even a stream of dirent_t or something
16:58:28  <piscisaureus_>on windows and linux that can be optimized
16:59:36  <rendar>also a stream of strings would be cool like "dir1\0dir2\0dir3\0"
16:59:51  <rendar>but maybe unuseful at the v8 engine level..
17:06:27  * orlandovftwquit (Ping timeout: 250 seconds)
17:30:59  * mmaleckichanged nick to mmalecki[drinks]
17:46:59  * theCole_joined
17:51:24  * paddybyersquit (Read error: Network is unreachable)
17:51:34  * paddybyersjoined
17:53:40  * CoverSli1equit (Read error: Connection reset by peer)
17:57:19  * orlandovftwjoined
17:58:48  * brsonjoined
17:59:51  * igorzijoined
18:02:56  <TooTallNate>piscisaureus_: ping
18:02:59  <piscisaureus_>yo
18:03:15  <TooTallNate>so is there a way to spawn a cmd.exe subprocess that like... works?
18:03:40  <TooTallNate>customFds would probably work, but I need to catch the stdout data
18:03:45  <piscisaureus_>TooTallNate: huh. why?
18:03:52  <TooTallNate>piscisaureus_: codestre.am
18:04:06  <TooTallNate>would be awesome to get windows working
18:04:06  <piscisaureus_>TooTallNate: well, probably not
18:04:44  <piscisaureus_>TooTallNate: the only think I guess is to pump process.stdin to the child's stdin and process.stdout to the child's stdout
18:04:58  <piscisaureus_>well, vice versa for stdout
18:05:01  <piscisaureus_>but you'd not get colors
18:05:08  <TooTallNate>ya no colors is fine
18:05:19  <piscisaureus_>lemme try
18:05:20  <TooTallNate>the thing is, when i do that the subprocess appears unresponsive
18:05:30  <piscisaureus_>TooTallNate: what if you set stdin to raw mode?
18:05:38  <TooTallNate>it is
18:07:01  * theCole_quit (Quit: theCole_)
18:08:26  <piscisaureus_>TooTallNate: I think cmd just does not echo correctly
18:09:07  <piscisaureus_>TooTallNate: or maybe it takes control over the console or something
18:09:35  <piscisaureus_>TooTallNate: the weird thing is that as soon as I hit ^C suddenly my text pops up
18:09:59  <piscisaureus_>also, < cmd.stdin.write('dir\r\n') > works
18:10:10  <TooTallNate>can i see your script?
18:10:34  <piscisaureus_>sure
18:11:06  <piscisaureus_>TooTallNate: there
18:11:07  <piscisaureus_>https://gist.github.com/2285907
18:11:20  <piscisaureus_>TooTallNate: I think cmd.exe takes control over the console
18:11:33  <piscisaureus_>so we no longer receive keypresses
18:12:09  <TooTallNate>piscisaureus_: hmmm, maybe customFds: [0, -1, -1] would work?
18:12:21  <piscisaureus_>TooTallNate: maybe, lemme try
18:13:25  <piscisaureus_>yeah, that seems to work strange enough. The question is whether node actually sees the keys that you enter.
18:13:37  <piscisaureus_>TooTallNate: or that cmd is just echoing them to the console directly
18:13:44  <TooTallNate>piscisaureus_: i don't so much care about the keys, just the stdout of the child cmd.exe
18:13:59  <piscisaureus_>TooTallNate: so that means that you cannot record the text that the user types...
18:14:06  <piscisaureus_>not really could isit?
18:14:07  <piscisaureus_>*cool
18:14:24  <TooTallNate>piscisaureus_: oh, hmm, does cmd.exe not echo the input like bash?
18:14:52  <piscisaureus_>TooTallNate: well it does but it echoes straight to the console if you make it inherit stdin
18:15:07  <piscisaureus_>TooTallNate: basically the readline that cmd.exe uses is very low level
18:15:09  <TooTallNate>oh shoot
18:15:13  <piscisaureus_>TooTallNate: I have an idea though
18:15:29  <piscisaureus_>what if we create two consoles and we stitch them up over a pipe
18:15:30  <piscisaureus_>that would work
18:15:35  <piscisaureus_>TooTallNate: expect a demo soo
18:15:37  <piscisaureus_>*soon
18:15:48  * avalanche123joined
18:15:48  <TooTallNate>:) i like your spirit
18:16:12  * `3rdEdenjoined
18:18:51  * igorziquit (Ping timeout: 245 seconds)
18:25:27  * igorzijoined
18:26:27  <piscisaureus_>TooTallNate: hmm wait it actually works for me
18:26:46  <piscisaureus_>TooTallNate: it's just that the cmd sets stdin back to readline mode
18:27:19  <TooTallNate>piscisaureus_: so basically we need to capture what the user sees in their console. can we do that?
18:27:36  <piscisaureus_>TooTallNate: to a certain extent. hang on
18:27:47  <TooTallNate>which i guess includes keystrokes, so long as they're really printed to the console
18:29:55  * mjr_joined
18:33:50  <piscisaureus_>TooTallNate: hmm, almost works..
18:33:55  <TooTallNate>piscisaureus_: also, tty.setRawMode() takes a bool, not an fd
18:33:59  <TooTallNate>it's always fd 0
18:34:02  <piscisaureus_>ah, right
18:34:13  <TooTallNate>so you were entering readline mode there
18:34:47  <piscisaureus_>well in raw mode nothing works :-/
18:34:57  * mjr_quit (Quit: mjr_)
18:35:36  <piscisaureus_>ah right
18:35:46  <piscisaureus_>the key reported on enter is \r and not \n
18:36:36  <piscisaureus_>TooTallNate: I think I can make this work but I am afraid it's not going to be much use
18:37:02  <piscisaureus_>since this is probably going to break on all ends when you start something like vi
18:37:38  <TooTallNate>piscisaureus_: damn
18:38:02  <TooTallNate>how does like freeSSHd do it?
18:38:08  <piscisaureus_>I will try though
18:38:17  <piscisaureus_>does that work with vi?
18:38:26  <TooTallNate>dunno, i'll try it
18:41:18  <bnoordhuis>piscisaureus_: you rang?
18:42:06  <piscisaureus_>TooTallNate: try that https://gist.github.com/2285907
18:42:17  <piscisaureus_>bnoordhuis: what is the "flags" field for uv_fs_readdir for?
18:42:31  <piscisaureus_>bnoordhuis: I could not find anything in the header nor in man 3 readline
18:42:57  <bnoordhuis>piscisaureus_: it's passed to eio_readdir
18:42:58  <TooTallNate>piscisaureus_: what does ctrl+enter do?
18:43:38  <piscisaureus_>TooTallNate: nothing, it's just that <enter> is reported as \n to node and ctrl+enter is reported as \r
18:43:41  <bnoordhuis>piscisaureus_: you can give it some hints about sorting, i think
18:43:49  <bnoordhuis>that is, sorting the results
18:43:55  <piscisaureus_>TooTallNate: this is probably a libuv bug
18:44:17  <TooTallNate>piscisaureus_: cool
18:44:39  <piscisaureus_>TooTallNate: now try if vi still works... I guess not.
18:45:04  <piscisaureus_>TooTallNate: you would probably be better off with a screen recorder, or with a console capturing binary addon
18:45:05  <TooTallNate>nope :\
18:45:56  <piscisaureus_>TooTallNate: that would probably fix it anyway. Just spawn cmd.exe with all stdio inherited
18:46:09  <TooTallNate>piscisaureus_: but then i can't capture it?
18:46:18  <piscisaureus_>TooTallNate: you'd have to write a binary addon
18:46:25  <TooTallNate>oh
18:46:30  <piscisaureus_>it's not difficult though
18:46:34  <piscisaureus_>TooTallNate: http://msdn.microsoft.com/en-us/library/windows/desktop/ms684965%28v=vs.85%29.aspx
18:46:42  <piscisaureus_>TooTallNate: that lets you take a screen grab from the terminal
18:47:05  <piscisaureus_>so if you just run that on a set interval then you'd probably get there
18:48:22  <piscisaureus_>TooTallNate: I question the usefulness though... I don't think many people use vim in the terminal to edit files.
18:48:29  <piscisaureus_>but maybe you do ...
18:50:45  <TooTallNate>piscisaureus_: ya what you gave me will probably work for now
18:50:54  <piscisaureus_>TooTallNate: ok
18:50:56  <TooTallNate>piscisaureus_: not gonna go crazy over windows support :p
18:51:07  <piscisaureus_>TooTallNate: you are a wise man
18:51:10  <TooTallNate>piscisaureus_: i need to focus on getting 256 colors to work anyways :p
19:16:21  <piscisaureus_>bnoordhuis: does libev report anything if an error occurs while polling?
19:34:53  * dylukesjoined
19:43:53  * piscisaureus_quit (Quit: ~ Trillian Astra - www.trillian.im ~)
19:50:07  * pfox___quit (Ping timeout: 245 seconds)
20:10:44  * `3rdEdenquit (Quit: zZzz)
20:26:31  * piscisaureus_joined
20:28:24  * einaros_joined
20:28:37  * avalanch_joined
20:31:10  * mrb_bk_joined
20:31:28  * avalanche123quit (Ping timeout: 260 seconds)
20:31:28  * einarosquit (Ping timeout: 260 seconds)
20:31:28  * avsejquit (Ping timeout: 260 seconds)
20:31:28  * mrb_bkquit (Ping timeout: 260 seconds)
20:31:28  * rphillipsquit (Ping timeout: 260 seconds)
20:31:29  * avalanch_changed nick to avalanche123
20:31:31  * mrb_bk_changed nick to mrb_bk
20:31:38  * avsej_joined
20:32:57  * piscisaureus_quit (Quit: ~ Trillian Astra - www.trillian.im ~)
20:33:43  * rphillips_joined
20:34:00  <benvie>hmm are you trying to like tap into the repl or is it all console stuff
20:34:24  <benvie>how does this codestream work. Do I just put my colors all up on display
20:39:30  * isaacsjoined
20:45:55  * pfox___joined
20:48:17  * einaros_changed nick to einaros
20:50:18  <isaacs>bnoordhuis: Hey, pushed new changes to https://github.com/isaacs/node/commit/zlib-errors
20:50:27  <isaacs>bnoordhuis: incorporated your responses
20:58:07  * piscisaureus_joined
21:00:56  <piscisaureus_>bnoordhuis: is fixing o3 your new job?
21:02:56  * dylukesquit (Quit: Pipes are broken. Sending packets via Fedex.)
21:10:08  <bnoordhuis>piscisaureus_: the current state of affairs will not stand
21:10:10  <bnoordhuis>or do
21:10:37  <piscisaureus_>bnoordhuis: is o3 broken?
21:11:00  <bnoordhuis>piscisaureus_: somewhat
21:11:13  <bnoordhuis>i think it fails to handle EINTR when it calls write()
21:11:28  <piscisaureus_>huh
21:11:43  <piscisaureus_>I think o3 is obsolete
21:11:55  <bnoordhuis>yeah? it seems to be used in the c9 ide
21:12:08  <piscisaureus_>I think they only use the xml part
21:12:22  <piscisaureus_>not the crazy bindings generator
21:12:28  <bnoordhuis>oh. then it's probably something else
21:12:37  <bnoordhuis>i'm still not quite sure how everything ties together
21:12:41  * pfox___quit (Ping timeout: 252 seconds)
21:13:41  <TooTallNate>benvie: it's all about recording a terminal/console session
21:13:47  <TooTallNate>benvie: but only works on unix right now
21:13:59  <benvie>I'm working on it ;)
21:14:43  <TooTallNate>benvie: what approach are you thinking?
21:15:57  <benvie>well I already have stuff set up to work over a socket
21:19:26  * rendarquit
21:19:32  * pfox___joined
21:24:03  * pfox___quit (Ping timeout: 248 seconds)
21:26:49  * sh1mmerquit (Quit: sh1mmer)
21:28:02  <piscisaureus_>isaacs: that sounds an awful lot like mraleph
21:30:58  * pfox___joined
21:32:34  <isaacs>piscisaureus_: yep, that's the one
21:32:46  <bnoordhuis>isaacs: lgtm
21:33:26  <isaacs>bnoordhuis: thanks
21:33:28  <piscisaureus_>bnoordhuis: hey
21:33:30  <ryah>yo
21:33:45  <piscisaureus_>(I never seem to be able to talk to bnoordhuis these days)
21:34:16  <CIA-99>node: isaacs master * r01d46f3 / (3 files in 3 dirs): Fix #3052 Handle errors properly in zlib - http://git.io/CPxxLQ
21:34:27  <bnoordhuis>piscisaureus_: what's up?
21:34:28  <isaacs>ryah: hi
21:34:31  <bnoordhuis>sup ryah
21:34:41  <piscisaureus_>bnoordhuis: ev_io <- does that ever report an error?
21:34:52  <piscisaureus_>or does it just always silently stop the watcher?
21:34:55  <bnoordhuis>piscisaureus_: the type?
21:35:03  <bnoordhuis>or ev_io_start / ev_io_stop?
21:35:09  <piscisaureus_>bnoordhuis: no, I mean, when it calls the callbsck
21:35:12  <piscisaureus_>*callback
21:35:21  <piscisaureus_>I am also interested in ev_io_start
21:35:36  * pfox___quit (Ping timeout: 244 seconds)
21:36:02  <bnoordhuis>technically it can but i don't think it ever does
21:36:05  <bnoordhuis>piscisaureus_: ^
21:36:07  <ryah>marc lehmann wants the libev copyright notice displayed in the binary
21:36:31  <ryah>should also do that for the other deps
21:36:46  <isaacs>in the binary? like, as a string?
21:37:06  <ryah>i think in the installer is fine
21:37:07  <piscisaureus_>bnoordhuis: so in my proposal uv_poll_cb gets a status argument
21:37:22  <piscisaureus_>bnoordhuis: should we scrap that?
21:37:24  <isaacs>ryah: yeah, that's totally reasonable.
21:37:26  * pfox___joined
21:37:30  * mmalecki[drinks]changed nick to mmalecki
21:37:34  <piscisaureus_>time to get rid of libev :-p
21:37:37  <piscisaureus_>yeah
21:37:50  <bnoordhuis>piscisaureus_: what's the callback for?
21:38:02  <piscisaureus_>bnoordhuis: reports that a socket is readable and/or writable
21:38:25  <bnoordhuis>please no
21:38:37  <bnoordhuis>that's meaningless in edge-triggered mode
21:38:42  <isaacs>bbiab
21:38:43  * isaacsquit (Remote host closed the connection)
21:38:51  <bnoordhuis>80k users? congrats freenode!
21:39:10  <piscisaureus_>bnoordhuis: oh but I don't think uv_poll should be edge triggered.
21:39:29  <piscisaureus_>bnoordhuis: with epoll you can set that on a per-fd basis right?
21:39:33  <bnoordhuis>yes
21:40:00  <piscisaureus_>bnoordhuis: so ev_io should be level triggered and so should uv_poll
21:40:38  <piscisaureus_>ryah: I don't think that the binary distributions have any copyright notices whatsoever
21:41:22  <ryah>piscisaureus_: no, they dont
21:41:36  <ryah>i wanted to do it lazily
21:41:43  <piscisaureus_>heh
21:41:50  <benvie>http://codestre.am/4f7a0e2f9c58b6625701217a I did it
21:41:52  <benvie>ahaha
21:42:06  <piscisaureus_>benvie: empty here
21:42:11  <CIA-99>node: Ryan Dahl master * r0810936 / LICENSE : Fix links to libev and libeio licenses - http://git.io/nbrlQw
21:42:30  <ryah>first commit in 3 months \o/
21:42:33  <benvie>hmm
21:43:44  * pfox___quit (Ping timeout: 252 seconds)
21:43:56  <ryah>i think we should probably collect all the licenses into a single file
21:44:06  <ryah>so that it's easy to display with the installler
21:44:44  <piscisaureus_>ryah: so do we need these little stuffs from libuv as well - tree.h and nginx_queue?
21:45:28  <ryah>yeah
21:45:50  <piscisaureus_>ew
21:46:35  <piscisaureus_>do we also need to get gyp in there?
21:46:41  <ryah>yes
21:46:43  <piscisaureus_>and the npm and node-gyp licenses?
21:46:50  <ryah>*shrug* i guess
21:47:39  <piscisaureus_>ok
21:50:31  <benvie>ok there we go http://codestre.am/3b
21:50:56  * avsej_changed nick to avsej
21:52:13  <CIA-99>libuv: Ben Noordhuis master * r9c8f6dd / (config-unix.mk uv.gyp):
21:52:13  <CIA-99>libuv: build: define _DARWIN_USE_64_BIT_INODE=1 on OS X
21:52:13  <CIA-99>libuv: Fixes a segmentation fault on some OS X systems due to sizeof(struct stat)
21:52:13  <CIA-99>libuv: mismatches. - http://git.io/AqXy1A
21:52:13  <CIA-99>libuv: Ben Noordhuis master * r396d138 / (config-unix.mk uv.gyp): Merge branch 'v0.6' - http://git.io/XLs81w
21:52:14  <CIA-99>libuv: Ben Noordhuis v0.6 * r9c8f6dd / (config-unix.mk uv.gyp):
21:52:15  <CIA-99>libuv: build: define _DARWIN_USE_64_BIT_INODE=1 on OS X
21:52:15  <CIA-99>libuv: Fixes a segmentation fault on some OS X systems due to sizeof(struct stat)
21:52:15  <CIA-99>libuv: mismatches. - http://git.io/AqXy1A
21:52:53  <TooTallNate>benvie: hop in the chat
21:53:55  * travis-cijoined
21:53:55  <travis-ci>[travis-ci] joyent/libuv#164 (v0.6 - 9c8f6dd : Ben Noordhuis): The build is still failing.
21:53:55  <travis-ci>[travis-ci] Change view : https://github.com/joyent/libuv/compare/1795427...9c8f6dd
21:53:55  <travis-ci>[travis-ci] Build details : http://travis-ci.org/joyent/libuv/builds/1003136
21:53:55  * travis-cipart
21:54:15  <CIA-99>node: Ben Noordhuis v0.6 * rffee873 / (deps/uv/config-unix.mk wscript):
21:54:15  <CIA-99>node: build: define _DARWIN_USE_64_BIT_INODE=1 on OS X
21:54:15  <CIA-99>node: Fixes a segmentation fault on some OS X systems due to sizeof(struct stat)
21:54:15  <CIA-99>node: mismatches.
21:54:15  <CIA-99>node: Fixes #2061. - http://git.io/KNKj_w
21:54:16  <CIA-99>node: Ben Noordhuis master * r92c0c69 / common.gypi :
21:54:16  <CIA-99>node: build: define _DARWIN_USE_64_BIT_INODE=1 on OS X
21:54:16  <CIA-99>node: Fixes a segmentation fault on some OS X systems due to sizeof(struct stat)
21:54:17  <CIA-99>node: mismatches.
21:54:17  <CIA-99>node: Fixes #2061. - http://git.io/viF14Q
21:57:01  * travis-cijoined
21:57:02  <travis-ci>[travis-ci] joyent/node#684 (master - 0810936 : Ryan Dahl): The build is still failing.
21:57:02  <travis-ci>[travis-ci] Change view : https://github.com/joyent/node/compare/01d46f3...0810936
21:57:02  <travis-ci>[travis-ci] Build details : http://travis-ci.org/joyent/node/builds/1003071
21:57:02  * travis-cipart
22:00:31  <bnoordhuis>paddybyers: remember that one? ^
22:01:06  * coderaritychanged nick to coderarity-eatin
22:02:16  * travis-cijoined
22:02:16  <travis-ci>[travis-ci] joyent/node#685 (v0.6 - ffee873 : Ben Noordhuis): The build was fixed.
22:02:16  <travis-ci>[travis-ci] Change view : https://github.com/joyent/node/compare/0965d2d...ffee873
22:02:16  <travis-ci>[travis-ci] Build details : http://travis-ci.org/joyent/node/builds/1003151
22:02:16  * travis-cipart
22:09:12  * travis-cijoined
22:09:12  <travis-ci>[travis-ci] joyent/node#686 (master - 92c0c69 : Ben Noordhuis): The build is still failing.
22:09:12  <travis-ci>[travis-ci] Change view : https://github.com/joyent/node/compare/0810936...92c0c69
22:09:12  <travis-ci>[travis-ci] Build details : http://travis-ci.org/joyent/node/builds/1003154
22:09:12  * travis-cipart
22:09:48  * dylukesjoined
22:16:27  <ryah>what ever happened to that hash collision problem and snapshots?
22:16:33  <ryah>do we sill need to have snapshots off?
22:22:33  <piscisaureus_>yes
22:22:39  <piscisaureus_>ryah: ^
22:22:54  * coderarity-eatinchanged nick to coderarity
22:25:02  <ryah>sucks
22:25:36  <piscisaureus_>ryah: well there's bigger fish to fry
22:25:57  <ryah>sure
22:26:17  <piscisaureus_>ryah: also, I don't think the difference is noticable.
22:26:40  <ryah>how long until v0.8 ?
22:26:57  <piscisaureus_>*shrug* long
22:27:04  <piscisaureus_>or should I say 4 weeks
22:27:08  <piscisaureus_>there's a lot of stuff to be done
22:27:36  <ryah>will there be another v0.6 ?
22:27:49  <piscisaureus_>if we found big bugs, yes
22:27:56  <piscisaureus_>but we don't really want to
22:29:29  <ryah>this license issue might warrent a new release if v0.8 is 4 weeks out still
22:30:11  <ryah>would be stupid thought
22:30:14  <ryah>though
22:30:26  <piscisaureus_>ryah: oh so was Marc Lehman that anxious?
22:30:41  <piscisaureus_>ryah: I mean, we've done binary releases for a couple of months now
22:30:52  <piscisaureus_>(btw the 4 weeks is not really realisting)
22:30:56  <piscisaureus_>*realistic
22:34:50  * pfox___joined
22:39:18  * dylukesquit (Quit: Computer has gone to sleep.)
22:44:55  * dylukesjoined
22:53:59  * dylukesquit (Quit: Computer has gone to sleep.)
22:54:16  * igorziquit (Ping timeout: 245 seconds)
22:57:31  <paddybyers>bnoordhuis: so it was right all along?
23:00:05  * igorzijoined
23:01:25  <bnoordhuis>paddybyers: kind of
23:01:31  <piscisaureus_>So I am reading up on the libev mailing list
23:01:35  <bnoordhuis>it was a different define but the net result is that it no longer crashes
23:01:44  <piscisaureus_>seems that bnoordhuis has had it's own lehman moment
23:01:49  <bnoordhuis>oh yes
23:02:03  <bnoordhuis>i took it off-list after a while
23:02:11  <bnoordhuis>but he kept me sending these interminable emails
23:02:18  <bnoordhuis>so after a while i just stopped replying :)
23:02:35  <bnoordhuis>that guy really needs to learn to get to the point
23:02:47  <bnoordhuis>but he's german, he probably can't help it
23:04:27  <TooTallNate>piscisaureus_: link :p
23:05:03  <piscisaureus_>bnoordhuis: did you ever sign a cease fire ?
23:05:26  <piscisaureus_>TooTallNate: around 20th of december, about eio
23:05:54  <paddybyers>bnoordhuis: ok I see the difference
23:06:15  <paddybyers>only took 6 months :)
23:06:48  <bnoordhuis>piscisaureus_: not really
23:07:00  <bnoordhuis>at first i took the piss out of him
23:07:08  <bnoordhuis>but that really didn't get through
23:07:36  <piscisaureus_>bnoordhuis: yeah but the "when was your first contribution to the linux kernel" of yours was kinda lame
23:08:39  <bnoordhuis>piscisaureus_: that was to piss him off
23:08:59  <bnoordhuis>he got himself into a couple of big flamewars on the lkml
23:09:22  <pfox___>is there a libuv list?
23:09:40  <bnoordhuis>yes. but no one uses it
23:10:04  <bnoordhuis>re lkml, germans are quite overrepresented in lkml flamewars
23:10:09  <bnoordhuis>i don't know what's up with that
23:10:27  * dylukesjoined
23:12:24  <pfox___>IMO, a much bigger problem, in technology, than sexism is pedantic, humorless and over-confrontational nerds
23:12:36  * igorziquit (Ping timeout: 245 seconds)
23:12:44  <pfox___>the need for this feeling of mathematical correctness in all discussions is grating
23:13:49  <bnoordhuis>humorless... that's a quite apt description of your average german :)
23:14:06  <pfox___>i usually say "humorless and european" in my description of the problem
23:14:10  <pfox___>but i was feeling diplomatic today
23:14:37  <bnoordhuis>oh, no worries - dutchees don't feel affronted, we're the life of every party we go to
23:14:57  <bnoordhuis>for instance, there's this dutch joke: do you know what the smallest, thinnest book in human history is?
23:15:05  <bnoordhuis>three centuries of collected german humor!
23:15:09  <pfox___>a book of german jokes?
23:15:13  <pfox___>ah, heh
23:15:50  <bnoordhuis>we also have quite a few jokes on the flemish
23:15:54  <bnoordhuis>but then again, so do they
23:16:48  <bnoordhuis>piscisaureus_: are you at the office wednesday or thursday?
23:16:59  <piscisaureus_>bnoordhuis: both
23:17:10  <piscisaureus_>bnoordhuis: I thought you were only coming on friday?
23:17:33  <bnoordhuis>piscisaureus_: yeah, i'm changing that
23:17:35  * dapquit (Quit: Leaving.)
23:17:36  * igorzijoined
23:23:52  <dylukes>Hm, I worry I abuse C macros too much.
23:24:25  <dylukes>I end up using them to generate whole slews of functions w/ token pasting. Any sort of thing I can easily factor out :\.
23:26:28  * dapjoined
23:37:18  * k-schanged nick to k-s[AWAY]
23:48:14  * paddybyersquit (Quit: paddybyers)
23:49:08  <piscisaureus_>igorzi: hey
23:49:24  <igorzi>piscisaureus_: whatsup?
23:49:35  <piscisaureus_>igorzi: I was reading up on LSPs and it seems that providers can have their own implementation of GetOverlappedResult
23:49:54  <piscisaureus_>igorzi: do you think we should call WSAGetOverlappedResult instead of parsing the overlapped.Internal field directly?
23:49:55  <CIA-99>node: Ryan Dahl master * r55e971e / LICENSE :
23:49:55  <CIA-99>node: Include text of licenses in LICENSE file
23:49:55  <CIA-99>node: For easy inclusion in binary distributions. ref #3056 - http://git.io/Apu63A
23:49:55  <CIA-99>node: Ryan Dahl master * r30994aa / (tools/osx-pkg.pmdoc/02npm.xml tools/osx-pkg.pmdoc/index.xml):
23:49:55  <CIA-99>node: Mac installer shows license
23:49:55  <CIA-99>node: ref #3056 - http://git.io/P75mEg
23:51:26  <igorzi>piscisaureus_: hmm.. and that applies to ifs lsps?
23:52:17  <piscisaureus_>igorzi: well all isps can have their own but I think that IFS isps have not much option than to use the same semantics in the io status block.
23:52:35  <piscisaureus_>igorzi: but non-ifs lsp's can kind of do... anything really
23:52:48  <piscisaureus_>igorzi: they can also wrap io completion notifications for example
23:54:58  * isaacsjoined
23:56:06  <igorzi>piscisaureus_: so then yeah.. we should probably call WSAGetOverlappedResult.. but do you only want to do it for non-ifs lsps?
23:56:21  <piscisaureus_>igorzi: well yeah that would probably make sense
23:56:37  <piscisaureus_>igorzi: although... WSPGetOverlappedResult is pretty cheap by the look of it
23:56:45  <piscisaureus_>it doesn't involve a kernel mode transition
23:57:19  <isaacs>ryah, piscisaureus_: We should shoot for May 1 v0.8 feature complete, and try to release v0.8 by June 1
23:57:27  <isaacs>and yes, there will be a new 0.6
23:57:29  <piscisaureus_>igorzi: so there would only be the overhead of looking up the socket in the context, table, grabbing the function pointer out of the dispatch table, and calling the function
23:57:38  <isaacs>if nothing else, then just to update npm to 1.1.15
23:57:43  <piscisaureus_>igorzi: I have no idea how (in)efficient it is.
23:58:07  <piscisaureus_>isaacs: well maybe we should stop truncating fs.write() offset at 31 bits
23:58:16  <piscisaureus_>isaacs: yours truly got bit hard by that today :-)
23:59:12  <isaacs>Node has my permission to not show the npm license in the node installer.
23:59:18  <isaacs>:)
23:59:28  <piscisaureus_>we should add that to the installer
23:59:50  <ryah>isaacs: i updated the mac installer to show the license in master
23:59:57  <ryah>still need win master and both in v0.7
23:59:58  <ryah>er