|
00:18
daxim joined
00:29
__sri joined
00:34
tokuhirom joined
00:39
brother joined
05:03
woosley joined
|
|||
| dalek | arVM: 5ef6691 | jimmy++ | / (8 files): Add CArray REPR |
05:23 | |
| arVM: 3ef8643 | jimmy++ | / (8 files): Add CStruct REPR |
05:24 | ||
| arVM: b0aa29f | jimmy++ | src/6model/reprs/CStruct.c: small fixes |
06:02 | ||
| arVM: 87bacef | jimmy++ | src/6model/ (3 files): small refactor |
06:36 | ||
|
06:41
vincent22 joined
|
|||
| dalek | arVM: 840ab82 | jimmy++ | src/io/ (3 files): change MVM_stat_* macro to uppercase |
07:03 | |
|
08:06
brrt joined
08:13
zakharyas joined
08:32
FROGGS joined
08:39
lizmat joined
08:44
lizmat joined
08:54
woolfy joined
08:55
lizmat joined
09:27
lizmat joined
09:34
colors joined
09:43
woolfy joined
09:47
lizmat joined
09:49
woolfy joined
|
|||
| jnthn | hmm, a tcg todo is "- Add float and vector support." :) | 09:59 | |
| nwc10 | tcg? A JITting type thing? | 10:00 | |
| jnthn | The float support is quite a missing thing :) | ||
| nwc10: Yeah. | |||
| nwc10: Got some interest from somebody to work on it in GSoC, so pondering options a little :) | |||
| brrt | here's the interest :-) | 10:07 | |
| although i've learned that i only have a week! left to apply | 10:08 | ||
| :-o | |||
|
10:09
woolfy left
|
|||
| jnthn | :-O | 10:20 | |
| Well, we can make that. | |||
| I'm willing to mentor. :) | |||
| brrt | fortunately so :-) | 10:22 | |
| i hope to have some time for it this weekend | |||
| many things to be planned, worked out | 10:24 | ||
| nwc10 | I thought that the real "skill" in proposal writing is to figure out a description which sounds good, is actually viable, but is sufficiently flexible it doesn't commit to any kind of hard pass/fail deliverable | 10:25 | |
| ie, as long as the student works reasonably hard and diligently, something will be delivered which is useful | 10:26 | ||
| brrt | …. i have slightly higher hopes than 'something' personally | 10:27 | |
| :-) | |||
| but not really the brain space left to think too deep about it | 10:28 | ||
| jnthn | So long as you have moar brain space in the summer :P | 10:31 | |
| brrt is finally going to look up the word 'moar' | 10:32 | ||
| ah i see | |||
| jnthn | Stands for Metamodel On A Runtime :) | 10:33 | |
| Sicne the VM came out of taking 6model and sticking a runtime around it :) | |||
| brrt | that i knew | ||
| timotimo | is there something that must be done on moarvm to flush stdout when we're breaking off execution with a stack trace? | 10:34 | |
| or is that something for nqp? | |||
| brrt | … one of my ideas for now is that you might need / want to represent a JIT-ed call frame / stack somewhat different from a non-jitted one | 10:45 | |
| and that you might want to use a REPR for that so that the gc can trace it | 10:46 | ||
| .. probably call frames are already REPRS | |||
| timotimo | MVMStaticFrame is a repr at least | 10:51 | |
| well, and there's MVMCode, which seems to be something different, but is contained in MVMStaticFrame as a pointer | 10:54 | ||
| brrt | uhuh | 10:55 | |
| i'll need to check that out, haven't really got time for that | |||
| but thanks for the pointer :-) | |||
| timotimo | i didn't invest much time in the analysis either >.< | ||
| brrt off for lunch | 10:57 | ||
| jnthn | We should keep JITted ones the same at first. | 11:24 | |
| 'cus it makes your deopt a lot simpler. | |||
| JimmyZ | jnthn: how hard to add float support? | 11:44 | |
| timotimo | probably depends on the number of supported backends? :\ | 11:47 | |
|
12:04
brrt joined
|
|||
| brrt back from lunch | 12:12 | ||
| jnthn | JimmyZ: No clue. :) | 12:29 | |
| JimmyZ: But my question is, "if that's missing what else is" | |||
| JimmyZ: Not to mentin that it has the extra abstraction layer problem, which I think might help us at first, but hinder us in the long run. | 12:30 | ||
|
12:30
lizmat joined
|
|||
| JimmyZ | jnthn: agreed, I like the dynasm one if the generator is written by perl 6, but by lua :P | 12:32 | |
| btw | 12:33 | ||
| not serious | |||
| jnthn | It's not a runtime dependency, though | 12:34 | |
| It's not even a buidl dependency | |||
| JimmyZ | yeah | ||
| jnthn | In so far as we can check in the .c file | ||
| JimmyZ | just developer needs it :P | ||
| jnthn | So then it's only a tiny dependency for anybody who is going to hack on the JIT :) | ||
| A full Lua is 200KB or so :) | |||
| JimmyZ | yeah, to me, dynasm is much easy than tcg | 12:35 | |
| because it has less abstraction | 12:36 | ||
|
12:55
lizmat joined
13:01
lizmat joined
13:32
woolfy joined
14:05
btyler joined
14:12
jnap joined
|
|||
| brrt | not that it is super-important, but lightning doesn't seem that hard, even if a bit RISC-y | 14:51 | |
|
15:27
colomon joined
|
|||
| jnthn | brrt: lightning gives us licensing issues though :( | 15:51 | |
| brrt | oh | ||
| i see | |||
| fair enough | 15:52 | ||
| actually, no, that is not fair though | |||
| one free software project not being able to use another free software project because of licensing is just madness | |||
|
15:52
lizmat joined
|
|||
| brrt | we're working together, people! | 15:53 | |
| jnthn | Apart from when we aren't. | ||
| If you're Stallman, LLVM is a "huge setback". | |||
| 'cus it's not GPL'd. | 15:54 | ||
| If your interest is in "enforcing freedom" then the GPL makes sense. | |||
| brrt | …. i'm sorry but i had to laugh about that | ||
| jnthn | I'm more in this because I like making nice things and sharing them. :) | ||
| brrt | 'i thought i could organise freedom' | ||
| didn't know stallman was scandinavian | |||
| jnthn | So my concerns/motivations for doing open soruce are entirely different. | ||
| brrt | well, yeah, i guess mine too | 15:55 | |
| japhb__ | jnthn: That would presumably be why you call it 'open source' instead of 'free software'. :-/ | ||
| jnthn | japhb__: Right. | ||
| japhb__ | (That :-/ was about the war of words, not you, of course.) | 15:56 | |
| jnthn | japhb__: Well, I know I choose licenses that folks who want to own the meaning of "free software" would get upset about, so it's easier to just say "open source" :) | ||
| japhb__ | nodnod. I sometimes end up just using 'FOSS' as a "Look, I don't plan to grind your axe either way. That's irrelevant to the current discussion." | 15:57 | |
| jnthn | *nod* | 15:58 | |
| brrt | it is hardly ever relevant | 15:59 | |
| but anyway | 16:00 | ||
| :-) | |||
| japhb__ | brrt: Yep. But some otherwise good to work with people will get their hackles up if you "wear the other gang's colors" | 16:02 | |
| brrt | people and their tribes | 16:03 | |
| japhb__ | I wonder how many metaphors I can mix there. | ||
| brrt | some people will get angry when i telll them i have - gasp - perl5 programs | ||
| can't make everybody happy :-) | |||
| jnthn | THEY'RE SCRIPTS!! :P | ||
| ;) | |||
| japhb__ | LOL | ||
| japhb__ has a plan to make everyone happy -- which involves nitrous oxide | 16:04 | ||
| jnthn | Well, it's getting towards dinner time here... :) | 16:09 | |
.oO( nitrous...that's slang for nutritious, right? ) |
16:10 | ||
| brrt | right :-D | 16:13 | |
| jnthn | Time to find a drink/nom and head home... bbott * | 16:23 | |
| & | |||
|
16:25
lizmat joined
|
|||
| brrt | yeah, me too | 16:27 | |
|
16:27
brrt left
|
|||
| FROGGS | be back old time traveller? | 16:32 | |
|
16:40
lizmat joined
17:06
oakleaf joined
|
|||
| TimToady | yes, back when he's a trainee again | 17:42 | |
| FROGGS | I hope not, I like him as he is :o) | 17:43 | |
|
17:52
lizmat joined
18:01
woolfy left
|
|||
| jnthn | "on the train" :) | 18:21 | |
| TimToady | as I said, when you're a trainee :P | 18:31 | |
| jnthn | :D | 18:32 | |
| dalek | arVM: c1b60eb | jonathan++ | build/Makefile.in: Install missing dyncall headers. |
18:35 | |
|
18:36
FROGGS[mobile] joined
18:46
FROGGS[mobile] joined
18:51
FROGGS[mobile] joined
18:59
FROGGS[mobile] joined
|
|||
| nwc10 | jnthn: All tests successful. | 19:18 | |
| jnthn | nwc10: For...Rakudo? | 19:27 | |
| nwc10: Assume this means "above patch fixes the Rakudo build on latest Moar"? | |||
| nwc10 | yes. | 19:30 | |
| jnthn | Grr...looks like Moar's shell is blindly turning /s to \s... | 19:53 | |
| cognominal | a moar shell? what is that? like a unix shell? I thought that moarvm was the meat not the shell. | 20:16 | |
| jnthn | cognominal: the shell op | 20:18 | |
| cognominal | jnthn++ | 20:21 | |
|
20:25
tgt joined
|
|||
| dalek | arVM: f3599cb | jonathan++ | src/io/procops.c: Don't \ every / on nqp::shell on Windows. Otherwise we screw up args like /nologo and so forth. |
20:41 | |
|
21:02
rurban joined
|
|||
| nwc10 | ./moar --version | 22:40 | |
| This is MoarVM version 2014.02-77-gc1b60eb | |||
| the key part being: Linux raspberrypi 3.10.25+ #622 PREEMPT Fri Jan 3 18:41:00 GMT 2014 armv6l GNU/Linux | 22:42 | ||
| and it's old - only 256M of RAM | |||
| (no swapping. onto NQP...) | 22:43 | ||
| jnthn | yes, but can it build NQP? :P | ||
| nwc10 | no. | ||
| MAST::Local index out of range | |||
| jnthn | ugh | 22:45 | |
| nwc10 | at gen/moar/stage2/QAST.nqp:4818 (src/vm/moar/stage0/QAST.moarvm:assemble_to_file:19) | ||
| and a lot more ugh | |||
| I should go to bed | |||
| about an hour ago | |||
| jnthn | 'night | 22:47 | |
|
23:00
FROGGS joined
23:15
timotimo joined
23:36
ggoebel11112 joined,
colomon_ joined
23:37
jnap1 joined
23:38
cognominal__ joined
|
|||
| dalek | arVM: f8fdc70 | jnthn++ | src/6model/reprs.c: Register CStruct. |
23:41 | |
|
23:44
jnap2 joined
23:50
brother joined
|
|||