01:24:05  * tilgoviquit (Ping timeout: 252 seconds)
01:31:38  * bnoordhuisquit (Ping timeout: 246 seconds)
02:32:41  * piscisaureus_quit (Read error: Connection reset by peer)
02:54:02  * alejandromgjoined
07:02:17  * alejandromgquit (Quit: leaving)
10:34:32  * bnoordhuisjoined
10:36:24  * piscisaureus_joined
11:16:20  * piscisaureus_quit (Ping timeout: 272 seconds)
12:00:46  * piscisaureus_joined
14:07:26  * hij1nxquit (Quit: hij1nx)
14:36:24  * bradleymeckjoined
15:07:07  * hij1nxjoined
15:12:04  * dvvjoined
15:12:17  <dvv>Hi, people!
15:12:25  <dvv>please consider https://github.com/indutny/candor/pull/19
15:12:26  <bradleymeck>jello
15:12:52  <dvv>same old fellows here :)
15:15:46  <bradleymeck>dvv probably til it takes off
15:15:53  <bradleymeck>err, off more
15:16:21  <bradleymeck>really want a tty module so I can make a repl though, not going to interfere with candor.io until i understand it more
15:17:58  <dvv>what ever is module system here?
15:18:05  <dvv>node-like?
15:18:48  <bradleymeck>dvv pretty much, you have a require function
15:19:20  <dvv>i see. about how files re organized?
15:19:29  <dvv>s/re/areee/
15:23:59  <indutny>hey
15:24:49  <indutny>dvv: are you sure your PR is up to date?
15:24:56  <indutny>dvv: can you rebase it over master?
15:25:21  <dvv>fresh. i'll do my best
15:25:33  <dvv>do you plan to keep clean history?
15:27:05  <indutny>yes
15:27:23  <dvv>right. virtue
15:27:48  <dvv>i do "git fetch origin ; git rebase origin/master", right?
15:28:04  <indutny>yes, something like that
15:29:30  <indutny>and please squash it
15:29:40  <indutny>at least fix up last merge commit
15:29:45  <dvv>into one commit?
15:30:04  <indutny>hm...
15:30:09  <indutny>no
15:30:13  <dvv>ok, rebased and pushed
15:30:16  <indutny>cool
15:30:19  <indutny>one more thing
15:30:30  <dvv>just teach me what you want the procedure to be
15:30:44  <indutny>can you take a look at candor's commit history and prefix your commits according to it?
15:30:54  <dvv>lemme snoop
15:36:02  <dvv>done (hopefully)
15:36:42  <indutny>one sec
15:37:20  <indutny>dvv: does it works fine on 32bit systems: PRIi64?
15:37:57  <indutny>though lgtm
15:39:20  <indutny>ok, pushing
15:41:10  <dvv>thanks
15:43:16  <indutny>dvv: no, thank you
15:43:25  <dvv>btw, libcandor.a is symlinked to build/out/Debug/libcandor.a, the latter lives deeper in out.target
15:44:01  <indutny>heh
15:44:16  <dvv>i wonder if it's my setup broken?
15:46:23  <indutny>build/out/Debug/libcandor.a is a file on my osx machine
15:46:29  <indutny>(not a symlink)
15:47:57  <dvv>then it is gyp's issue? i am on 64-bit ubuntu and Debug has out.target subdir
15:48:57  <indutny>I think gyp may create different makefiles for different platforms
15:49:05  <indutny>is that stuff a problem for you?
15:49:36  <dvv>nope. i just discovered candor and wanna make it build clean on my setup
15:50:21  <indutny>cool, btw join the google group if you haven't done this yet
15:50:32  <dvv>will do
15:50:50  <dvv>are there any "smoke" modules ready to stare at?
15:51:36  <indutny>"smoke"?
15:51:52  <indutny>there're only candor.io so far: https://github.com/creationix/candor.io
15:51:52  <dvv>i mean easy to explore
15:52:03  <indutny>ah, no native modules in VM itself
15:52:11  <indutny>dvv: https://github.com/creationix/candor.io/blob/master/lib/net.can
15:52:30  <creationix>yes, candor.io modules will be luvit-like
15:52:37  <indutny>but that's quite outdated
15:52:39  <dvv>sweet
15:52:48  <indutny>`new` keyword was replaced by `clone`
15:53:08  <creationix>dvv, it's not hard to add tty to candor.io. I just want to squash all bugs before moving on
15:53:19  <creationix>I moved too fast with luvit and it shows
15:53:30  <creationix>Timer.close isn't working
15:53:39  <creationix>and the uv loop is often stuck open
15:54:39  <dvv>right
15:56:02  * alejandromgjoined
15:58:56  * dvvquit (Ping timeout: 246 seconds)
16:00:31  <indutny>creationix: it's very nice to see more people interested in it :)
16:00:42  <indutny>creationix: btw, I'm 50% finished ia32 port
16:01:53  * dvvjoined
16:01:58  <creationix>nice
16:01:59  <creationix>that's fast
16:02:15  <creationix>indutny, any luck finding a good arm machine to aquire?
16:02:30  <indutny>creationix: haven't worked on that yet
16:02:41  <indutny>I've discussed JIT stuff with mraleph
16:02:43  <creationix>no problem, just curious
16:02:53  <indutny>after ia32 port, I'll try to do register allocation
16:02:59  <indutny>at non-optimized phase
16:03:09  <indutny>that may simplify both stubs and fullgen
16:03:35  <indutny>it's really pain-in-the-ass to handle all that stuff on ia32, because you've only 4-6 registers to work with
16:03:44  <indutny>2 of them are not really usefu
16:03:47  <indutny>s/usefu/useful
16:04:21  <creationix>crazy to think that v8 is faster on ia32
16:04:28  <creationix>though they spent a *lot* more effort on it
16:05:07  <indutny>yes, that's crazy
16:05:17  <creationix>luajit is mich faster on x64
16:05:19  <indutny>it's just hard to move from 16 registers to 6
16:05:20  <indutny>:D
16:05:31  <creationix>how many does arm have?
16:05:35  <indutny>oh
16:05:37  <indutny>maaany
16:05:45  <creationix>very different instruction set
16:05:46  <creationix>?
16:05:56  <indutny>ah, not so
16:06:01  <indutny>16
16:06:13  <creationix>not bad though
16:06:16  <indutny>and 16 xmm-like registers
16:06:17  <dvv>someone can provide a syntax highlighter for candor? :)
16:06:23  <creationix>I learned mips in school
16:06:29  <indutny>dvv: you can give it a try ;)
16:06:36  <creationix>I want to write one for ace
16:06:40  <dvv>need a list of tokens :)
16:06:47  <indutny>dvv: look at lexer
16:06:52  <indutny>dvv: src/lexer.h
16:06:58  <dvv>right
16:07:04  <creationix>dvv, if you write one for ace, I'll make sure it's accepted
16:07:07  <indutny>it's opensource, right? :)
16:07:12  <indutny>creationix: for c9?
16:07:15  <creationix>indutny, yep
16:07:22  <indutny>creationix: that would be neat! :)
16:07:31  <creationix>right now I'm working on a mini local editor that uses ace
16:07:41  <indutny>creationix: for what?
16:07:50  <creationix>eventually I hope to have c9 support for all the *uv variants
16:07:54  <indutny>aah
16:08:02  <indutny>well, more languages - better ;)
16:08:11  <creationix>ace already has lua and javascript
16:08:19  <creationix>obviously not candor, it's new
16:08:34  <creationix>but on top of language support, c9 has full integration for node
16:08:45  <creationix>interactive debugging and running of apps
16:09:17  <indutny>yeah, I know
16:09:19  <indutny>c9 is really powerful
16:19:51  * hij1nxquit (Quit: hij1nx)
16:21:32  <indutny>crap, xmm usage in ia32 is borked too
16:22:12  <creationix>are you sure ia32 is worth it ;)
16:22:35  <indutny>creationix: no :D
16:22:48  <creationix>seriously though, I do think it's a dying architecture
16:23:07  <indutny>may be you're right
16:23:08  <creationix>last time I tried to buy a netbook, most stores didn't even have them in stock anymore
16:23:10  <creationix>all discontinued
16:23:15  <creationix>replaced by tablets
16:23:18  <indutny>hm....
16:23:26  <indutny>lets put all that stuff into ia32 branch
16:23:33  <creationix>rackspace is switching all their vps boxes to 64 bit only
16:23:40  <indutny>and leave it for future :D
16:23:50  <creationix>someone will want ia32 for sure
16:23:56  <creationix>just wondering if it's worth the effort
16:24:05  <creationix>just like someone wants ppc on freebsd
16:25:14  <creationix>I can try to leave an arm machine plugged in for you if that helps
16:25:22  <creationix>but over internet it's going to be super slow
16:25:28  <creationix>even locally the lag is pretty bad
16:25:32  <indutny>yeah
16:25:57  <creationix>I wonder if anyone offers arm vps yet
16:26:00  <creationix>that would be cool
16:29:06  <indutny>I'm going to try finishing floating point stuff
16:29:09  * tilgovijoined
16:29:13  <indutny>I actually need a way to put xmm1 into memory
16:29:16  <indutny>and get it from it
16:29:23  <indutny>every other stuff should work fine
16:29:33  <creationix>xmml?
16:29:44  * tilgoviquit (Read error: Connection reset by peer)
16:29:44  <indutny>yeah, xmm registers, floating point registers
16:29:53  <indutny>who is tilgovi?
16:35:33  <creationix>Randall Leeds, he's really cool
16:35:45  <creationix>worked on couchdb, pouchdb
16:35:48  <creationix>helped with my leveldb
16:35:53  <creationix>(node bindings)
16:37:16  <indutny>aaah
16:37:18  <indutny>Randall
16:37:18  <creationix>I wonder how leveldb compares to your btree library
16:37:26  <indutny>creationix: well, btree is fast
16:37:31  <indutny>quite comprable
16:37:40  <creationix>I'm pretty sure leveldb uses a btree internally
16:37:40  <indutny>comparable*
16:37:47  <creationix>your's is smaller right?
16:37:52  <indutny>yes
16:37:55  <indutny>and it's in C
16:38:09  <creationix>btree bindings for candor would be the ultimate in bare-bones scriptable db
16:38:26  <creationix>but the I/O is blocking?
16:38:43  <indutny>haha
16:38:46  <indutny>yes it is
16:38:52  <indutny>but it may work in uv workers
16:38:56  <creationix>that's fine
16:38:59  <indutny>it's thread safe
16:39:01  <creationix>all fs in pretty much blocking
16:39:06  <creationix>filesystems suck
16:39:15  <creationix>the hardware is non-blocking
16:39:22  <creationix>all interrupt based and whatnot
16:40:10  <bradleymeck>well technically its vector table based on x86 based stuff… but thats a nightmare
16:40:47  <creationix>whatever the implementation, I'll bet it's true async as some level
16:40:52  <creationix>but what you see from C is blocking
16:41:11  <bradleymeck>creationix it is, if you really get down to it
16:41:41  <bradleymeck>creationix any way I could get a simple stdin and stdout raw stuff from candor.io as it is?
16:41:59  <creationix>blocking is easy
16:42:06  <creationix>bradleymeck, but if you want to make a repl, it needs to not block
16:42:13  <creationix>otherwise anything async will never fire
16:42:25  <creationix>bradleymeck, you can bind uv_tty_t
16:42:29  <creationix>I already did stream
16:42:31  <creationix>that's the hard part
16:42:37  <creationix>tty inherits from stream
16:42:42  <bradleymeck>ill take a crack at it at jsconf
16:42:49  <creationix>basically copy tcp, replace with ttf and bind to the right things
16:42:53  <creationix>*tty
17:08:01  * hij1nxjoined
17:09:16  <indutny>crap, any SSE/MMX experts here?
17:10:20  <dvv>(was doing molecular dynamic computer simulation using MMX looong ago) :)
17:10:22  <indutny>is there anyway to store xmm1 into memory address other than fst (on ia32)
17:10:44  <indutny>I mean is there any instruction, or only fst/fld
17:11:01  <indutny>movqd works fine with x64, but it move only half of quad word on ia32
17:13:49  <indutny>movdqa?
17:14:12  <dvv>sry, forgot everything... :)
17:14:21  <indutny>dvv: np
17:19:56  <indutny>yeah
17:19:59  <indutny>movdqa works!!!!!!!!
17:20:09  <indutny>one more step closer to the working ia32
17:30:16  <creationix>:)
17:33:34  <indutny>hm...
17:33:36  <indutny>or not :D
17:33:51  <indutny>strange, one test works, other don't
17:40:08  <indutny>ok, movdqu solved it
18:35:17  * alejandromgchanged nick to alejandro[away]
19:19:13  <indutny>oook
19:19:19  <indutny>I think I'll rewrite codegen
19:19:38  <indutny>it may generate well-optimized code
19:21:34  <indutny>ok
19:21:36  <indutny>time to sleep
19:21:36  <indutny>ttyl
19:36:45  * dvvquit (Ping timeout: 252 seconds)
19:42:04  * alejandro[away]changed nick to alejandromg[out]
20:41:09  * hij1nxquit (Quit: hij1nx)
20:49:17  * alejandromg[out]changed nick to alejandromg
21:03:54  * bradleymeckquit (Quit: bradleymeck)
22:36:50  * alejandromgchanged nick to alejandromg[out]
22:56:30  * alejandromg[out]changed nick to alejandromg
23:59:21  * piscisaureus_quit (Quit: ~ Trillian Astra - www.trillian.im ~)
23:59:29  * alejandromgquit (Quit: leaving)