|
github.com/moarvm/moarvm | IRC logs at irclog.perlgeek.de/moarvm/today Set by moderator on 16 July 2013. |
|||
| JimmyZ | FROGGS++, I was always patch nqp-moar-cc.nqp MASTTesting.nqp to win32 only for myself | 01:19 | |
| FROGGS: you can't patch to MVM_BIGINT_BINARY_OP, not all BINARAY_OP need that, i.e: MVM_BIGINT_BINARY_OP(sub) | 01:29 | ||
| FROGGS: only "or, xor, and" | 01:30 | ||
|
01:40
FROGGS_ joined
04:50
birdwindupbird joined
|
|||
| lizmat | diakopter: let me know and I will attend, not sure if I can contribute much on the Jenga end, though | 06:43 | |
| FROGGS_ | JimmyZ: k, will try then, thank you! | 06:48 | |
| JimmyZ | FROGGS_: :P | ||
|
07:11
FROGGS joined
|
|||
| FROGGS | JimmyZ: I use the two_complement now for and/or/xor only... gist.github.com/FROGGS/223f6f2933837d99315c | 07:29 | |
| if I use it for right/left shift, tests break | |||
| JimmyZ | FROGGS: you mean 18 and 20? | 07:50 | |
| FROGGS: 18 and 20 was failed before | 07:51 | ||
| FROGGS: do you mean use 'two_complement_shl' breaks test? | 07:52 | ||
| FROGGS | JimmyZ: I mean that if I uncomment gist.github.com/FROGGS/223f6f29338...-diff-L101 and comment L100, it breaks tests | 07:59 | |
| which is odd since nqp@parrot does it that way to actually pass tests... | 08:00 | ||
| JimmyZ | jnthn: ping | 08:02 | |
|
08:07
benabik joined
|
|||
| JimmyZ | FROGGS: two_complement_shl(ia, ib, n); shoulde be two_complement_shl(ib, ia, n); | 08:09 | |
| I think | |||
| dalek | Heuristic branch merge: pushed 80 commits to MoarVM/container by zhuomingliang | 08:50 | |
| Heuristic branch merge: pushed 20 commits to MoarVM by zhuomingliang | 08:51 | ||
| JimmyZ | ^^ this passed most 67-container.t test | 08:52 | |
| hmm, maybe not | 09:03 | ||
| jnthn: what's different between assignunchecked and assign? I can't see the differeent in the code s | 09:08 | ||
| *difference | 09:09 | ||
|
09:37
donaldh joined
09:54
eternaleye joined
|
|||
| dalek | arVM: 0ad0156 | jimmy++ | src/ (4 files): implement decont op |
10:05 | |
| FROGGS | JimmyZ++ # you are the man! :o) | ||
| JimmyZ decommutes | |||
| timotimo | don't you mean ... decontmute? | 10:07 | |
| FROGGS | hehe | 10:08 | |
| dalek | arVM: 5dc5893 | (Tobias Leich)++ | src/math/bigintops.c: added two_complement_bitop |
10:13 | |
|
11:26
cognominal joined
11:39
colomon joined
11:41
benabik joined
|
|||
| JimmyZ | Good evening, FROGGS | 11:42 | |
| FROGGS | hi JimmyZ | 11:43 | |
| JimmyZ | hello | 11:44 | |
| FROGGS | pro hint of the day: It is not recommended to run make/nmake in the same project directory from linux and windows at the same time. | 12:28 | |
| tadzik | *giggle* | ||
|
12:46
benabik joined
|
|||
| dalek | arVM: c5b1e85 | jimmy++ | / (6 files): added assign and assignunchecked ops |
12:54 | |
| arVM: 9a8bf18 | jimmy++ | src/ (3 files): small fixes |
13:03 | ||
|
13:07
colomon joined
13:58
benabik joined
|
|||
| dalek | arVM: 6a7db8b | jimmy++ | src/core/interp.c: fixed serious bug |
14:12 | |
| JimmyZ | ^^ you must do 'cur_op += 4;' before doing invoke | 14:14 | |
|
14:34
benabik joined
|
|||
| JimmyZ | jnthn: ping | 14:54 | |
| jnthn: how can I do 2-times invoke in a op? | 14:55 | ||
| jnthn | JimmyZ: You have to do it CPS style; see how load bytecode work.s | 14:56 | |
| JimmyZ | jnthn: I'm in assign op do fetch and store. fetch and store in the same op do 2-times invoke, I don't how to do cps | 15:05 | |
| jnthn: or I break assgin into two ops by use add_core_op? | 15:09 | ||
| jnthn | JimmyZ: No, it's handled entirley in C | ||
| JimmyZ: Leave it if you're not sure; the loadbytecode op is an example of doing this two-invoke thing. | 15:10 | ||
| dalek | arVM: 7063d72 | jimmy++ | src/core/interp.c: fixed decont |
15:27 | |
| JimmyZ | jnthn: ok :) | ||
|
15:40
colomon joined
|
|||
| JimmyZ | good night | 16:05 | |
| jnthn | 'night, JimmyZ | ||
| FROGGS | gnight JimmyZ | ||
| diakopter | how did I miss the loadbytecode op appearing | 16:11 | |
| jnthn | I dunno...how could so much that works now, work without it? ;-) | 16:15 | |
| diakopter | yeah.. | 16:17 | |
| makes me wonder what else I've missed :) | |||
| jnthn: my windows VM on the macbook retina is FAST with a capital F | 16:18 | ||
| like, cold boot in 4 seconds | |||
| with 8 GB ram and 8 cores | |||
| Parallels++ | 16:19 | ||
| took me a while to figure out how to get the resolution the way I wanted it when on the main hi-dpi screen and then also on a normal external monitor (27") | |||
| but once I figured that out, it's been awesome | |||
| jnthn | Taht's fast indeed :) | 16:21 | |
| diakopter | arnsholt: ping | 17:05 | |
|
17:08
FROGGS joined
|
|||
| dalek | arVM/dyncall: 13968dc | diakopter++ | build/Makefile.in: somehow I deleted the cat |
17:30 | |
| lizmat | .oO( wonder what Schrödinger would think of that ) |
17:33 | |
| dalek | arVM/dyncall: 98f5888 | diakopter++ | build/Makefile.in: goof |
17:37 | |
| arnsholt | diakopter: pong | ||
| diakopter | jnthn: have you heard from japhb? | 17:38 | |
| arnsholt: I wanted to answer your question from yesterday | 17:39 | ||
| arnsholt | Cool. I was pondering a bit of hacking right about now, so that's very good timing =) | 17:40 | |
| diakopter | arnsholt: well now that I fixed a couple of errors, the dyncall branch builds ok on linux | 17:41 | |
| arnsholt | Excellent | ||
| diakopter | jnthn indicated that CStruct repr on nqp-pvm does the right thing with the compose and type stuff | 17:43 | |
| so I'm starting there instead of with CArray | |||
| unless you think I should start another place.. | |||
| arnsholt | It's one place to start. On JVM I was actually going to start with the CPointer REPR | 17:44 | |
| diakopter | I guess CPointer would be simplest | ||
| heh. | |||
| arnsholt | Yeah, it's the bare minimum | ||
| You can even test without any of the REPRs too, mind | 17:45 | ||
| diakopter | yeah | ||
| arnsholt | Set up buildnativecall and nativecall, and bind printf with a single string argument for example | ||
| But TBH, I'd expect large swaths of the Parrot C code to be easily copy-pastable into MoarVM | 17:46 | ||
| diakopter | yeah I already did that with all of them :) | ||
| changing the boilerplate as I go | |||
| arnsholt | Cool. I think it'll mostly be a question of weeding out all the Parrot-specific stuff | 17:47 | |
|
18:01
colomon joined
|
|||
| dalek | arVM: 323e524 | (Tobias Leich)++ | src/6model/containers.h: newline at EOF, silences a warning |
18:02 | |
|
18:04
yoleaux joined
|
|||
| diakopter | .tell jnthn there's a bunch of places I'm finding that need the MVM_ASSIGN macro | 18:05 | |
| yoleaux | diakopter: I'll pass your message to jnthn. | ||
| diakopter | .tell jnthn need to do a full code audit one of these days | 18:06 | |
| yoleaux | diakopter: I'll pass your message to jnthn. | ||
| diakopter | .tell jnthn er, MVM_ASSIGN_REF | 18:10 | |
| yoleaux | diakopter: I'll pass your message to jnthn. | ||
|
19:36
cognominal joined
20:12
donaldh joined
20:51
cognominal joined
20:58
prammer joined
21:08
bronco_creek joined
|
|||
| FROGGS | jnthn: it seems to work | 22:05 | |
| jnthn | yays | 22:07 | |
|
22:42
cognominal joined
22:52
colomon joined
23:12
flussence joined
23:48
colomon joined
23:54
benabik joined
|
|||