|
github.com/moarvm/moarvm | IRC logs at irclog.perlgeek.de/moarvm/today Set by moderator on 4 June 2013. |
|||
| benabik | The abort seems to be tied to a stack checking of sprintf. If I change it from buf[20] to buf[21], it works. | 00:16 | |
| s/stack/buffer overflow/; I'll submit a pull request when I'm done putting the kid to bed. | |||
| There we go. Nothing like a fresh pull request on a new project. | 00:41 | ||
|
01:27
benabik joined
01:31
colomon joined
01:47
colomon joined
02:49
colomon joined
02:50
colomon_ joined
02:55
lizmat joined
|
|||
| dalek | arVM: a629202 | benabik++ | src/core/interp.c: Avoid minor buffer overflow in say_n t/moar/float.t fails on darwin/amd64 with an abort in __sprintf_chk. Google tells me that this is related to a libssp data protection library detecting a buffer overflow. Sure enough "-127243075316.468048" is 20 characters long which puts the null character one byte past the end of `char buf[20]`. Allocating a single extra byte for the buffer fixes the problem. |
03:05 | |
| arVM: cd0500b | benabik++ | nqp-cc/.gitignore: 77743cc | jimmy++ | build/Config/ (2 files): |
|||
| arVM: 62cb622 | jonathan++ | build/Config/ (2 files): Merge pull request #5 from zhuomingliang/patch-2 fixed build on win32 x86 platform |
|||
|
03:22
lizmat joined
|
|||
| dalek | arVM: bc92df7 | jonathan++ | src/core/coerce.c: Put buffer overrun fix in another place. |
03:27 | |
| benabik | This suggests that coerce could use testing so I would have noticed it failing. :-D | 03:29 | |
| jnthn | Yeah. | ||
| masak | time to write more tests! | ||
| benabik | \\o/ | ||
| jnthn | Well, the dev ops should also go away really and so the test should be updated to use the coerce op, I suspect. | ||
| nwc10 | shouldn't those 20s, er 21s, become a named constant defined once? | 03:30 | |
| benabik | Dev ops? | ||
| masak .oO( devops?! o.O ) | |||
| jnthn | benabik: There's a bank of ops which exist as "cheats" to ease the early development work on Moar. | ||
| Which we're kinda past now. | 03:31 | ||
| jnthn will have a look | |||
| nwc10: Probably, yes | |||
| benabik | say_* -> coerce_*_s, getstdout, write_fhs? | 03:34 | |
| jnthn | Well, a say_s should stay | 03:57 | |
| benabik | Ah, good. That would make things simpler. | ||
| jnthn | Maybe it needs to move | 03:58 | |
| jnthn looks | |||
| benabik | It is currently in dev bank. | 03:59 | |
| jnthn | yeah | ||
| grr, so we'd have say_s, but print. That can't be :) | |||
| So: say_s becomes say, print stays as it is, say_i/say_n go away, and both say and print move from dev into the io ops bank | 04:00 | ||
| benabik | A task! | 04:01 | |
| jnthn | If you want to work on it: after you update src/core/oplist, run perl6 tools/update_ops.p6 | 04:03 | |
| benabik | Was just tracking that down. But need to change interp.c manually? | 04:04 | |
| jnthn | yeah | ||
| That actually has to know what things mean :) | |||
|
04:08
FROGGS_ joined
|
|||
| benabik | Should the old ops be replaced with __INVALID_N__ or just vanish into the night? | 04:10 | |
| jnthn | Can vanish if you renumber to avoid gaps, which is probably easy for the dev bank | 04:11 | |
| At some point the op numbers get standardized and the ordering can be cleaned up. Not ready for that yet. | |||
|
04:26
colomon joined
04:49
birdwindupbird joined
|
|||
| benabik | Ah. Also have to modify MVM_opcounts_by_bank | 05:12 | |
| Wait. No. | 05:13 | ||
| jnthn | uh? | ||
| benabik | I just have to get the numbers correct. | ||
| jnthn | You shouldn't. | ||
| :) | |||
| benabik | C _D_ E F | ||
| jnthn | ;) | ||
| 'night | |||
| benabik | 'night | ||
|
05:35
tomyan joined
|
|||
| benabik | It is a little odd to keep writing 'coerce_is, str, int' | 05:38 | |
|
06:12
Guest1337 joined
|
|||
| benabik | Huh. threads.t fails when running 'make test' and succeeds when running alone. Stupid threading tests. (Fails with "Can't free STables in gen2 GC yet", so I imagine it's not a big deal.) | 06:30 | |
|
06:47
cognominal joined
|
|||
| benabik | I probably should have gone to bed instead of finishing changing those tests. Oh well. | 07:04 | |
| 'night | |||
|
08:14
flaviusb joined
08:33
flaviusb joined
08:47
lizmat joined
08:51
flaviusb joined
|
|||
| hoelzro | hi MoarVM folks | 09:16 | |
| moritz | hello | 09:17 | |
| hoelzro | why is it that the MoarVM has its own organization on GH? why not just put the repo under perl6/? | 09:18 | |
|
09:19
leont joined
|
|||
| moritz has no idea | 09:20 | ||
|
10:28
flaviusb joined
12:08
tgt joined
12:24
flaviusb joined
|
|||
| masak | hoelzro: one reason is that until MoarVM actually runs Rakudo Perl 6, whether it will succeed or not is still not 100% certain. so dissociating the Perl 6 project and the MoarVM project, at least in name, makes sense. | 12:41 | |
| hoelzro | hmm | ||
| I guess that makes sense. | |||
| masak | hoelzro: as I understood it, MoarVM could be moved into the 'perl6' organization in the near future, as more progress unfolds. | 12:42 | |
| but it may or may not do so. it might be that keeping the separation makes more sense. | |||
|
12:43
flaviusb joined
|
|||
| hoelzro | 3rdparty/apr/.libs/* is all pre-compiled, yes? | 12:47 | |
| nwc10 | no, built, but I think it's built by Configure.pl | 12:49 | |
| but there's been a bit of discussion about probing for installed useful system versions, and maybe not bundling, but downloading | 12:50 | ||
| hoelzro | ah ha | ||
| cognominal | masak, is the video and the slides of your presentations at YAPC::NA available somewhere (same for jnthn and pmichaud) | 12:52 | |
| hoelzro | hmm | ||
|
12:53
benabik joined
|
|||
| hoelzro | it's complaining about the files under .libs/ not existing (I removed them, due to linker errors) | 12:53 | |
| masak | cognominal: I have no idea. | 12:55 | |
| cognominal: I will put my slides online, as soon as I exit panic-slide-making mode or travel mode. | 12:56 | ||
| cognominal | oh, I thought you were done with your presentations. | ||
|
13:01
flaviusb joined
13:22
flaviusb joined
13:27
Ulti joined
13:32
lizmat_ joined
13:34
lizmat joined,
masak_ joined
13:37
kbenson1 joined
13:40
sorear_ joined
13:43
Guest1337 joined
13:45
FROGGS joined
13:47
ChanServ joined
|
|||
| FROGGS | o/ | 13:54 | |
| moritz | \\o | 13:55 | |
| lizmat | mornin' | 13:56 | |
|
14:00
Guest1337 joined
14:39
Guest1337 joined
14:48
sorear joined
14:59
FROGGS joined
15:01
bronco_creek joined
15:03
Guest1337 joined
15:04
birdwindupbird joined
15:06
benabik joined
15:38
pmichaud_ joined
15:41
moritz joined
15:43
japhb joined,
masak_ joined
15:54
pmichaud joined,
Guest1337 joined
16:08
kbenson joined
16:12
FROGGS joined
16:19
lizmat joined
17:01
Tene_ joined,
`patch`_ joined
17:08
Guest1337 joined,
masak joined,
diakopte1 joined,
dalek joined,
nwc10 joined,
prammer joined,
woosley joined
17:12
vmspb joined
17:40
benabik joined
18:13
FROGGS joined
18:22
lizmat joined
18:30
colomon joined
|
|||
| sorear | o/ | 18:31 | |
| colomon | japhb: A few days ago, Jen asked me "Is there a song about electric cars? Because Henry was just singing about them..." | ||
|
18:34
japhb_ joined
|
|||
| [Coke] | colomon; electric car, so good so far! | 18:40 | |
| colomon: www.youtube.com/watch?v=4BPU5mKipNo | |||
|
18:42
benabik joined
19:22
japhb__ joined
19:23
colomon joined
19:24
FROGGS joined
19:25
sorear joined
19:30
kbenson joined
19:32
Tene_ joined
20:06
lizmat joined
20:18
FROGGS joined
20:28
Guest1337 joined
20:31
lizmat joined
20:48
tgt joined
21:02
lizmat joined
21:09
donaldh joined
21:11
FROGGS joined
21:42
cognominal joined
22:19
FROGGS joined
22:34
colomon joined
23:00
tomyan joined
23:34
tgt joined
|
|||