|
github.com/moarvm/moarvm | IRC logs at irclog.perlgeek.de/moarvm/today Set by moderator on 5 July 2013. |
|||
|
00:47
colomon joined
01:11
benabik joined
|
|||
| dalek | arVM: 63e9a82 | jimmy++ | / (3 files): small fix and cleanup |
01:48 | |
|
01:50
FROGGS_ joined
04:32
benabik joined
04:57
birdwindupbird joined
06:06
ssutch joined
|
|||
| FROGGS_ | JimmyZ++ | 06:24 | |
|
06:36
ssutch joined
06:54
cognominal joined
06:55
eternaleye joined
07:11
eternaleye joined
07:26
eternaleye joined
07:28
FROGGS joined
07:37
benabik_ joined
07:39
ssutch joined
07:40
patspam joined
|
|||
| dalek | arVM: e068ddd | (Tobias Leich)++ | nqp-cc/src/QASTOperationsMAST.nqp: added stubs for missing IO ops |
07:49 | |
|
08:17
dalek joined
08:45
PerlJam joined,
chipdude joined
|
|||
| FROGGS | question: a mast op's signature must be identical to the C-function, right? | 08:58 | |
| ohh, hmmm, maybe I got my answer | 08:59 | ||
| dalek | arVM: 4b51dfa | (Tobias Leich)++ | nqp-cc/src/QASTOperationsMAST.nqp: fix mentioned MVM_open-constants |
09:11 | |
| arVM: ee6e7b3 | (Tobias Leich)++ | / (2 files): mapped nqp::setencoding |
|||
| JimmyZ | FROGGS: you don't need change MVM_file_set_oshandle_encoding, or you need change interp.c too | 09:13 | |
| FROGGS | damnit | 09:14 | |
| I had grepped for set_oshandle_encoding and found nothing... | 09:15 | ||
| JimmyZ | amd fileops.h ... | ||
| *and | |||
| FROGGS | yes, sure | 09:16 | |
| JimmyZ | :P | ||
| dalek | arVM: 9aa3be3 | (Tobias Leich)++ | src/ (2 files): rename to setencoding, JimmyZ++ |
09:20 | |
|
10:16
not_gerd joined
|
|||
| not_gerd | o/ | 10:16 | |
| FROGGS | hi not_gerd | ||
| not_gerd | via HN: morepypy.blogspot.jp/2013/07/softwa...-lisp.html | 10:17 | |
| if it works out for the PyPy guys, might be interesting for MOarVM as well... | |||
| .tell jnthn GC with STM support the PyPy guys are building: bitbucket.org/pypy/stmgc/src/tip/c...at=default | 10:22 | ||
| yoleaux | not_gerd: I'll pass your message to jnthn. | ||
| jnthn is not entirely convinced by STM | 10:28 | ||
| Not for a language like Perl 6, anyway. | |||
|
10:34
itz_ joined
|
|||
| not_gerd | apparently, Perl6 already supports STM en.wikipedia.org/wiki/Software_tran...emory#Perl ;) | 10:36 | |
| pugs++ | |||
| FROGGS | jnthn: apparently there are no file io tests in the nqp testsuite, can I add tests for nqp::open, close, readall, ... ? | 10:45 | |
| yay, nqp::readallfh is working | 10:49 | ||
|
11:01
lizmat joined
|
|||
| jnthn | FROGGS: sure | 11:03 | |
|
11:29
not_gerd left
|
|||
| dalek | arVM: a06bfe5 | (Tobias Leich)++ | / (8 files): added nqp::readall_fh, and changed MVM_file_slurp to do open/readall/close |
12:43 | |
| arVM: 4ffb81d | (Tobias Leich)++ | / (8 files): map nqp::printfh and make it take only two args |
14:53 | ||
| JimmyZ thinks We can use github.com/antirez/linenoise for readline | 15:03 | ||
| hoelzro likes linenoise | 15:04 | ||
| JimmyZ | FROGGS: ^^ | ||
| FROGGS | I miss the rare "Windows Operating System" in the readme :/ | 15:06 | |
| JimmyZ | oh, yeah, I don't know how nodejs use linenoise in windows | 15:07 | |
| FROGGS | ohh, autocompletion... | ||
| pretty cool | |||
| JimmyZ | :D | 15:08 | |
| FROGGS | I'm going to work on simpler ops, and I dont think we need readline before rakudo hits moarvm | 15:16 | |
| JimmyZ | yes | 15:18 | |
| nwc10 | how far is NQP from bootstrapping on MoarVM? | 15:19 | |
| FROGGS | jnthn / JimmyZ: can you tell? | 15:20 | |
| JimmyZ | nwc10: see 6guts.wordpress.com/2013/05/31/moar...nd-rakudo/ | 15:22 | |
| "Wht's the rough timeline?" part | 15:23 | ||
| FROGGS | would be cool to have it bootstrapped within the next weeks... jnthn how far away are we from that point? | 15:28 | |
| JimmyZ thinks it wouldn't to be far after adding all of ops maps | 15:30 | ||
|
15:32
FROGGS[mobile] joined
15:50
birdwindupbird joined,
FROGGS joined
16:02
colomon joined
16:20
dalek joined
16:21
PerlJam joined,
Util joined
16:22
tadzik joined
17:00
lizmat_ joined
17:01
chipdude_ joined
|
|||
| dalek | arVM: 6a1d4a1 | (Tobias Leich)++ | / (8 files): fixed mkdir to take permissions as 2nd arg |
17:39 | |
| arVM: 8aef424 | (Tobias Leich)++ | src/io/dirops.c: typo |
19:59 | ||
| arVM: 9129002 | (Tobias Leich)++ | / (8 files): added tell_fh, also tweak readallfh readallfh used a mmap before, which doesnt tell about the position in the file. Switched to apr_file_read which does that. |
|||
|
20:06
ChanServ joined
|
|||
| dalek | arVM: 3f2908e | (Tobias Leich)++ | / (7 files): mapped nqp::eof |
20:42 | |
| FROGGS | meh, parrots readline is weird/broken | 21:15 | |
| timotimo | yeah | 21:22 | |
| it really is | |||
| timotimo remembers ;_; | |||
| FROGGS | but jvm's version is sane :o) | 21:23 | |
| I've got a test file like "line1\\rline2\\r\\nline3\\nline4", and parrots readline gives "line21" for the first readline call | 21:24 | ||
| timotimo | ;_; | ||
| oh i remember now. i wanted to receive a line of json data over a socket | 21:25 | ||
| but i got only a part of it | |||
| and no way to tell if i got a full line or not | 21:26 | ||
| FROGGS | now, since libapr has no readline... do I just call getc() a dozen of times? | ||
| timotimo | i had to write my own readline based on reading small chunks | ||
| FROGGS | yeah :/ | ||
| good that MoarVM is not such a code-monster, so one is able to find and fix these bugs | |||
|
21:39
ssutch joined
22:01
TimToady joined
22:07
colomon joined
22:28
colomon_ joined
22:42
_ilbot joined
|
|||
| moderator | github.com/moarvm/moarvm | IRC logs at irclog.perlgeek.de/moarvm/today | ||
|
22:53
benabik joined
|
|||
| diakopter | nwc10: eh, I'm still guessing aug/sept | 23:07 | |
| timotimo | oh, moarvm already has sockets? | 23:17 | |
| that's pretty cool | |||
| there's pretty many "ok"'s when running make nqptest | 23:18 | ||
| seems like the only thing missing is big integers; is there lots of fudging involved or something? ;) | |||
| diakopter | timotimo: I thought japhb added bigintegers | 23:20 | |
| timotimo | 12 days ago some bigint progress was merged by jnthn | 23:21 | |
| t/nqp/60-bigint.t ......... Failed 26/45 subtests - that's only partly missing | 23:22 | ||
| what's the magic incantation to create a .moarvm file from the nqp-moarvm-cc? | 23:23 | ||
| diakopter | <- has no clue | 23:25 | |
| diakopter looks | |||
| eh, --target=mbc --output=foo.moarvm ? | 23:28 | ||
| timotimo | thanks :) | 23:30 | |
| diakopter | did it work? | ||
| timotimo | didn't try | ||
| trying to get perl6-bench up again | 23:31 | ||
| and see what i can/haev to do to make it benchmark the tests | |||
| luckily perl6-bench removes the startup/compile time from the benchmark time | 23:32 | ||
| otherwise i suppose the compiler startup would be a bit bad :) | |||
| diakopter | which compiler? | ||
| timotimo | the crosscompiler that has to start a parrot vm first ;) | 23:33 | |
|
23:37
benabik joined
|
|||
| diakopter | ;) | 23:38 | |
| timotimo: when configuring moarvm, did you do --optimize? | 23:42 | ||
| timotimo | no, i did not. i should? | ||
| perl Configure.pl --optimize=3? or how? | 23:43 | ||
| don't leave me hanging, diakopter! :( | 23:46 | ||
| diakopter | just --optimize | ||
| timotimo | oh, ok | ||
| building ... building ... building ... | 23:51 | ||
| unfortunately the benchmark thing makes a completely clean build every time | |||
| so that takes a bit longer | |||
| i don't even see the processes show up in my process tree o_O | 23:56 | ||
| Internal error: invalid thread ID in GC work pass <- :( | |||
| i'm not sure the benchmark framework is recognizing the error as an unsuccessful run and will probably keep upping the workload infinitely now | 23:57 | ||