|
00:06
ssutch joined
|
|||
| [Coke] | I can't even load Test.pm | 00:12 | |
|
00:12
ssutch joined
|
|||
| FROGGS | hmmm, strange | 00:14 | |
| perl t/harness --fudge --moar --keep-exit-code --tests-from-file=t/spectest.data | 00:30 | ||
| t/spec/S01-perl-5-integration/basic.rakudo.moar ............... ok | |||
| t/spec/S02-lexical-conventions/begin_end_pod.t ................ ok | |||
| it still works for me | 00:31 | ||
|
01:17
tgt joined
|
|||
| [Coke] | FROGGS: do you have an account on host06 ? | 01:34 | |
| FROGGS | [Coke]: I'm not sure | 01:35 | |
| need to check (tomorrow) | |||
|
01:43
jnap joined
03:13
colomon joined
04:27
jnap joined
06:23
jnap joined
07:24
jnap joined
08:25
jnap joined
|
|||
| FROGGS | Building for 1805 at Mon Dec 23 08:43:31 2013 [...] | 08:43 | |
| seems this will go one for a while :o) | 08:44 | ||
| nwc10 | several days | 08:49 | |
| It will go faster with paste.scsys.co.uk/286834 | |||
| which I think should have been there in the first place | |||
| the Makefile defines a macro PROVE, but then never uses it | 08:50 | ||
|
09:12
brother joined
09:26
jnap joined
|
|||
| nwc10 | FROGGS: if you are running it, I think it's unlikely to be that good at telling you *why* there's a problem unless you apply the hacked-up mmap patches I sent to the list a few days ago, and possibly also that patch that aborts() in the write barrier if the new pointer is in fromspace | 09:33 | |
| FROGGS | k | 09:49 | |
| will do | |||
| JimmyZ | src/vm/moar/ops/perl6_ops.c:367: error: too many arguments to function ‘MVM_frame_find_invokee’ | 09:50 | |
| make: *** [dynext/libperl6_ops_moar.so] Error 1 | |||
| I got this | |||
| FROGGS | how on earth... | 09:56 | |
| JimmyZ | [root@li618-176 rakudo]# perl Configure.pl --backends=moar --prefix=../MoarVM/ | ||
| Use of uninitialized value in concatenation (.) or string at Configure.pl line 234. | |||
| FROGGS | nwc10: ahh, I need to copy the diffs out of the mails? | 09:59 | |
| ahh no, there are three patches in your mail from 15.12. | 10:01 | ||
|
10:07
retupmoca joined
10:13
[Coke] joined
|
|||
| timotimo | nwc10: what keeps you from commiting your local patches to a branch of moarvm so that it's easier to try it out locally? | 10:39 | |
|
10:50
cognominal joined
|
|||
| FROGGS | timotimo: I fear he has no commit bit | 10:50 | |
| and I dunno who can hand them out | 10:51 | ||
| t/nqp/56-role.t (Wstat: 139 Tests: 0 Failed: 0) | |||
| Non-zero wait status: 139 | |||
| timotimo | well, fortunately one can fork repositories on github :) | ||
| FROGGS | is succeeds under gdb and valgrind :o( | ||
| timotimo | i can make mokudo compile the test library when i add --stagestats >_> | 10:52 | |
| i'm confused. i run the same command make outputs with --stagestats and make wants to run it again and fails | 10:55 | ||
| FROGGS | timotimo: yeah, it wants to rebuild Test.moarvm almost every time | 10:57 | |
| I have no idea why though | |||
| timotimo | i remember there's a gcc plugin or script or something that can spot misuse of reference counting in cpython extensions | 11:11 | |
| maybe we can look at their code and do something similar for moarvm garbage collecting | |||
|
11:27
jnap joined
|
|||
| timotimo | gcc-python-plugin.readthedocs.org/...ecker.html | 11:34 | |
| the gcc python plugin developer seems to commit stuff every now and then, but not terribly much | 11:40 | ||
|
11:51
cognominal joined
12:28
jnap joined
12:30
tgt joined
|
|||
| timotimo | [1] 14978 segmentation fault (core dumped) ./perl6-m --target=mbc --output=blib/lib.moarvm --stagestats lib/lib.pm6 | 12:41 | |
| :( | |||
|
12:43
tgt joined
13:21
colomon joined
13:29
jnap joined
14:05
tgt joined,
jnap joined
14:10
jnap joined
14:42
colomon joined
14:45
lue joined
15:03
nwc10 joined
17:09
arnsholt joined
18:18
ssutch joined
19:03
colomon joined
20:23
FROGGS joined,
jnap joined
20:44
jnap joined
21:22
cognominal joined
|
|||