|
00:45
TimToady_ joined
00:46
bartolin_ joined
00:53
bloatable6 joined,
coverable6 joined
00:58
jpf joined
01:15
FROGGS_ joined
01:57
ilbot3 joined
|
|||
| samcv | not sure how to fix this. but just spent an hour trying to track down | 02:19 | |
| "undefined symbol: __asan_option_detect_stack_use_after_return" is what i got. wasted so much time compiling and recompiling and make clean and removing some stray .a and .o files | |||
| but the solution was deleting libmoar.so and moar binary and then make install... i'm not sure why that should change how it installs? does anyone know? | |||
| Geth | MoarVM: 677d4e9a89 | (Samantha McVey)++ | .editorconfig Use tabs for Makefile and Makefile.in in .editorconfig |
02:22 | |
| MoarVM: samcv++ created pull request #836: Ensure `make realclean` cleans cmp and libatomic_ops |
02:23 | ||
|
02:40
travis-ci joined
|
|||
| travis-ci | MoarVM build passed. Samantha McVey 'Use tabs for Makefile and Makefile.in in .editorconfig' | 02:40 | |
| travis-ci.org/MoarVM/MoarVM/builds/362431918 github.com/MoarVM/MoarVM/compare/b...7d4e9a8940 | |||
|
02:40
travis-ci left
02:44
travis-ci joined
|
|||
| travis-ci | MoarVM build passed. Samantha McVey 'Use tabs for Makefile and Makefile.in in .editorconfig' | 02:44 | |
| travis-ci.org/MoarVM/MoarVM/builds/362431918 github.com/MoarVM/MoarVM/compare/b...7d4e9a8940 | |||
|
02:44
travis-ci left
05:16
domidumont joined
05:23
domidumont joined
06:04
lizmat joined
06:43
robertle joined
|
|||
| nine | /win 10 | 06:48 | |
| buggable | nine, Thank you for entering Accidental /win Lottery! The next draw will happen in 3 weeks, 4 days, 17 hours, 11 minutes, and 51 seconds | ||
|
06:49
domidumont joined
|
|||
| timotimo | samcv: i've experienced moar not being installed at all if a moar proces is still running somewhere | 07:12 | |
|
07:14
brrt joined
|
|||
| brrt | good * | 07:14 | |
| yoleaux | 2 Apr 2018 12:48Z <timotimo> brrt: so how do i go about finding the right line number when printing a backtrace when there's jit involved? | ||
| brrt | .tell timotimo: you don't | ||
| yoleaux | brrt: What kind of a name is "timotimo:"?! | ||
| brrt | .tell timotimo that's not a resolved problem yet | ||
| yoleaux | brrt: I'll pass your message to timotimo. | ||
|
07:17
zakharyas joined
|
|||
| timotimo | i'm glad to hear i'm not just doing it wrong! :) | 07:18 | |
| yoleaux | 07:14Z <brrt> timotimo: that's not a resolved problem yet | ||
| timotimo | i'll go have breakfast :) | 07:23 | |
| samcv | nine: if you're around could you check out cry.nu/files/process.tar getting a "Unhandled exception: Probable version skew in pre-compiled 'gen/moar/CORE.setting' (cause: no object at index 140115913614632)" message. MasterDuke also i believe replicated it | 07:38 | |
| valgrind and asan are clean, and MVM_dump_backtrace doesn't seem to print anything since tc->cur_frame is NULL, not sure if that is expected or not. and here is a backtrace gist.github.com/samcv/73a6a156f577...2976b129b2 | 07:40 | ||
| i'm not sure if it's a rakudo or a MoarVM issue | |||
| nine | samcv: huh....I could reproduce it on the first run, but not since | 07:44 | |
| samcv | nine: are you using ASAN? | 07:46 | |
|
07:48
lizmat joined
|
|||
| timotimo recommends rr again | 07:48 | ||
| rr: reproducibly recommended | 07:50 | ||
| nine | samcv: no. Could repro it a couple of times now. I'd guess some MoarVM issue | ||
| timotimo | i'm a little late to the conf :| | 07:53 | |
| timotimo runs off | |||
|
08:13
avar joined
08:21
camelia joined
08:27
notable6 joined
08:29
camelia joined
|
|||
| samcv | if I disable spesh the message cannot be reproduced | 08:29 | |
| timotimo | does disabling inlining make thigns work? | 08:30 | |
| samcv | also i got .*** Error in `/home/samantha/perl6/bin/moar': free(): corrupted unsorted chunks: 0x00007f86c83626d0 *** but on/ly when i made a spesh log | 08:31 | |
| nine | Oddly enough though MVM_SPESH_BLOCKING=1 MVM_SPESH_NODELAY=1 does not make it more reproducible | 08:32 | |
| timotimo | perhaps dont-gc-in-spesh is a problem | 08:33 | |
| samcv | also with spesh_logit exits with 139 | 08:34 | |
| well after running it many times | |||
| timotimo | oh fun i'm not clearing out that one field there | ||
| oh, wrong, it's in a different struct altogether | 08:36 | ||
| oh, i was using the spesh alloc for this, ok, so i wasn't in trouble after all | 08:37 | ||
| but still, it's in spesh_candidate_add | 08:45 | ||
| it's probably my "retry after aborted inline" mechanism going awry | |||
| github.com/MoarVM/MoarVM/blob/mast...ine.c#L233 - samcv is there an obvious mistake here? perhaps i'm memcpy-ing wrong? | 08:47 | ||
| nine | It's definitely inline related | 09:03 | |
| timotimo | i can't reproduce the problem at all | ||
|
09:05
zakharyas joined
|
|||
| timotimo | nine: gist.github.com/timo/3203bb8b0d627...9e6af2a8bb - does this give you debug output? | 09:10 | |
|
09:10
lizmat joined
09:29
lizmat joined
09:32
AlexDaniel joined
09:33
lizmat joined
|
|||
| samcv | timotimo: i'm not that great with spesh stuff. i don't see anything but that may not mean much | 09:37 | |
| nine | timotimo: it....core dumps with that patch | 09:46 | |
| timotimo: in the fprintf in src/spesh/candidate.c:51 | 09:52 | ||
| __GI___libc_free (mem=0x7f6cf99ea000) at malloc.c:2956 | |||
| Geth | MoarVM: 300cf72062 | (Samantha McVey)++ | src/strings/ops.c Vectorize string flip operation for flat strings, ~25% speedup Improvement of 25% for 32 bit flat strings. Clang optimizes both 32bit and 8bit flat string flips, while GCC only vectorizes the 32bit flip. Regardless, we still see a 25% speedup for 8bit flat strings on GCC. |
10:01 | |
| MoarVM: efa8e16b91 | (Samantha McVey)++ | src/strings/ops.c Vect. turn_32bit_into_8bit_unchecked, 10% speedup for string_escape Vectorizes on both Clang 5 and GCC 6+7. Currently we only use this function in MVM_string_escape(), which gets a 10% speedup for strings that fit in 8 bits. |
|||
|
10:14
yoleaux joined
|
|||
| timotimo | nine: that's the one after MVM_spesh_dump(tc, g)? | 10:15 | |
| samcv | i'm headed to bed. night o/ | 10:16 | |
| timotimo | gnite samcv | ||
|
10:19
live_the_dream joined
10:27
lizmat joined
11:04
zakharyas joined
11:32
brrt joined
|
|||
| dogbert2 | timotimo: so how's the conference? | 11:32 | |
|
12:14
live_the_dream joined,
lizmat joined
|
|||
| timotimo | it's pleasant | 12:32 | |
|
12:36
Kaiepi joined
|
|||
| dogbert2 | cool | 12:43 | |
| timotimo | was just discussing a bit with jnthn, iryna, and liz | 13:05 | |
|
13:10
zakharyas joined
|
|||
| dogbert2 | aha | 13:26 | |
| brrt | say hi from me | 13:27 | |
| timotimo | i'll try to remember, but i assume jnthn would read irc before i see him tonight | 13:32 | |
| dogbert2 | he's been absent so far today | 13:33 | |
| timotimo | no | ||
|
13:33
Kaiepi joined
|
|||
| timotimo | oh, you mean from irc | 13:33 | |
| dogbert2 | yes | 13:34 | |
| .seen AlexDaniel | 13:40 | ||
| yoleaux | I saw AlexDaniel 04:14Z in #perl6: <AlexDaniel> sounds great | ||
|
14:04
zakharyas joined
14:10
lizmat joined
14:16
zakharyas joined
14:17
lizmat joined
|
|||
| AlexDaniel | dogbert2: yes? | 14:25 | |
| dogbert2 | wanna try a bisect | 14:26 | |
| github.com/MoarVM/MoarVM/issues/835 | |||
| AlexDaniel | ah yeah, I was about to do that yesterday | 14:31 | |
| but my brain hibernated | |||
| ok let's see… | |||
| dogbert2: what's _90in.csv ? | 14:37 | ||
| ah, nvm | 14:38 | ||
| ok, first test run: | 14:40 | ||
| bisect: old=2018.03 gist.githubusercontent.com/AlexDan...0/90_csv.t | |||
| bisectable6 | AlexDaniel, Successfully fetched the code from the provided URL. | ||
| AlexDaniel | will give a wrong result probably | ||
| bisectable6 | AlexDaniel, Bisecting by exit signal (old=2018.03 new=7faf431). Old exit signal: 0 (None) | ||
| AlexDaniel | we'll see | 14:41 | |
| bisectable6 | AlexDaniel, bisect log: gist.github.com/0a668681c83b1ed7be...0ddbfc393a | ||
| AlexDaniel, (2018-03-24) github.com/rakudo/rakudo/commit/b5...44f2de6c31 | |||
| AlexDaniel | c: b5685109322^,b5685109322 gist.githubusercontent.com/AlexDan...0/90_csv.t | 14:42 | |
| committable6 | AlexDaniel, Successfully fetched the code from the provided URL. | ||
| AlexDaniel, gist.github.com/25b892588172a7aa88...144d2a942b | |||
| AlexDaniel | that's a bit off | ||
| c: b5685109322^^^^ gist.githubusercontent.com/AlexDan...0/90_csv.t | |||
| committable6 | AlexDaniel, Successfully fetched the code from the provided URL. | ||
| AlexDaniel, gist.github.com/823e61f04204988025...f7b1853678 | |||
| AlexDaniel | or is it | ||
| c: b5685109322^^ gist.githubusercontent.com/AlexDan...0/90_csv.t | |||
| committable6 | AlexDaniel, Successfully fetched the code from the provided URL. | ||
| AlexDaniel, gist.github.com/9b78618d31094e963c...fb7c73df44 | 14:43 | ||
| AlexDaniel | c: b5685109322^ gist.githubusercontent.com/AlexDan...0/90_csv.t | ||
| committable6 | AlexDaniel, Successfully fetched the code from the provided URL. | ||
| AlexDaniel, ¦b5685109322^: «===SORRY!===Missing or wrong version of dependency 'gen/moar/Metamodel.nqp' (from '/home/bisectable/git/whateverable/data/all-modules/github/FROGGS/p6-Slang-Tuxic/lib/Slang/Tuxic.pm (Slang::Tuxic)') «exit code = 1»» | |||
| AlexDaniel | c: b5685109322 gist.githubusercontent.com/AlexDan...0/90_csv.t | ||
| committable6 | AlexDaniel, Successfully fetched the code from the provided URL. | ||
| AlexDaniel, gist.github.com/d84eb693ad9a73fb38...e91611fe3b | |||
| AlexDaniel | OK → weird SORRY → SEGV | ||
| dogbert2: what did you say you did to make it more reliable? | 14:44 | ||
| it looks like a 100% segfault on whateverable | |||
| but maybe not? | |||
| dogbert2 | I have a 70 script (is a bad golfer) and if run with MVM_SPESH_NODELAY=1 MVM_SPESH_BLOCKING=1 | ||
| sry, distracted at $work | |||
| AlexDaniel | c: MVM_SPESH_NODELAY=1 MVM_SPESH_BLOCKING=1 c2eae41,18252b6,b5685109322 gist.githubusercontent.com/AlexDan...0/90_csv.t | 14:46 | |
| committable6 | AlexDaniel, Successfully fetched the code from the provided URL. | ||
| dogbert2 | with those flags it's very reliable for me, might be enough for a bisect | ||
| committable6 | AlexDaniel, gist.github.com/934106605326e58662...b2fe380e11 | ||
| AlexDaniel | dogbert2: according to this ↑ it bisected just fine even without these flags | ||
| dogbert2 | cool, so you have it cornered then :) | 14:47 | |
| AlexDaniel | haha, cornered. It's a merge commit xD | 14:48 | |
| dogbert2: anyway, thanks for pinging, I'll write a comment | |||
| dogbert2 | coool | ||
| AlexDaniel | (although the result kinda doesn't make sense) | ||
| probably something else at play there | |||
| dogbert2 | should we closeit and then move it to rakudo? | ||
| AlexDaniel | probably. Will do | 14:49 | |
| dogbert2 | hopes that it's not brrt's iffy merge | 14:51 | |
| "Merge pull request #1642 from MasterDuke17/simplify_Telemetrys_new" hmm | 14:59 | ||
| c: MVM_SPESH_NODELAY=1 MVM_SPESH_BLOCKING=1 c2eae41cabdef70b86270 gist.githubusercontent.com/AlexDan...0/90_csv.t | 15:01 | ||
| committable6 | dogbert2, Successfully fetched the code from the provided URL. | ||
| dogbert2, ¦c2eae41: «===SORRY!===Missing or wrong version of dependency 'gen/moar/Metamodel.nqp' (from '/home/bisectable/git/whateverable/data/all-modules/github/FROGGS/p6-Slang-Tuxic/lib/Slang/Tuxic.pm (Slang::Tuxic)') «exit code = 1»» | |||
| dogbert2 | wat | ||
| c: MVM_SPESH_NODELAY=1 MVM_SPESH_BLOCKING=1 6a66b3593f606 gist.githubusercontent.com/AlexDan...0/90_csv.t | 15:03 | ||
| committable6 | dogbert2, Successfully fetched the code from the provided URL. | ||
| dogbert2, gist.github.com/cb9de4878c17674005...56278e868e | |||
|
15:03
lizmat joined
|
|||
| AlexDaniel | dogbert2: see github.com/rakudo/rakudo/issues/16...-378968016 | 15:04 | |
| brrt | AlexDaniel: this is a MoarVM issue most likely | ||
| AlexDaniel | brrt: yes, I agree | ||
| brrt | more importantly, I have the spesh bisect done :-) | 15:05 | |
| dogbert2 | the JIT is certainly innocent | ||
| brrt++ | |||
| it fails when doing an | 15:07 | ||
| grrr | |||
| so what's wrong with commit c2eae41 | |||
| AlexDaniel | probably nothing, but it revealed the problem | 15:08 | |
| dogbert2 | that commit c: MVM_SPESH_NODELAY=1 MVM_SPESH_BLOCKING=1 c2eae41cabdef70b86270 say "SEGV" | 15:09 | |
| c: MVM_SPESH_NODELAY=1 MVM_SPESH_BLOCKING=1 c2eae41cabdef70b86270 say "SEGV" | |||
| committable6 | dogbert2, ¦c2eae41: «SEGV» | ||
| AlexDaniel | dogbert2: probably nothing, it just shifts the moon a little | 15:12 | |
| ah, I said that already | 15:13 | ||
| ZofBot: I guess our conversation skills are identical | |||
|
15:14
lizmat joined
|
|||
| dogbert2 | so is it commit c2eae41 which manages to trigger a bug in spesh? | 15:16 | |
| it crashes when doing a if (IS_CONCRETE(obj)) call on a null pointer and the aforementioned commit is doing some stuff with that (I think) | 15:18 | ||
| c2eae41cabde: Replace .DEFINITE by nqp::isconcrete wherever possible: that is, when the value was only needed in a condition and not to be returned as a HLL Bool | 15:19 | ||
|
15:22
live_the1 joined
15:24
lizmat joined
15:39
zakharyas joined
15:45
Kaiepi joined
|
|||
| AlexDaniel | dogbert2: “so is it commit c2eae41 which manages to trigger a bug in spesh?” probably a combination of several commits | 15:47 | |
|
16:33
dogbert17 joined
16:46
robertle joined
16:47
live_the_dream joined
17:22
domidumont joined
17:29
zakharyas joined
18:08
Kaiepi joined
18:18
Kaiepi joined
18:22
Kaiepi joined
18:50
Kaiepi joined
19:54
zakharyas joined
20:03
Kaiepi joined
20:06
Kaiepi joined
20:30
lizmat joined
22:15
Kaiepi joined
|
|||
| Geth | MoarVM: 718512a30b | (Samantha McVey)++ | 108 files Add libatomic_ops as a submodule instead of in-tree The old in-tree 3rdparty/libatomic_ops files have been removed and they are replaced by a submodule. This submodule was created by merging only the commits and files from 3rdparty/libatomic_ops into the upstream libatomic_ops repository. |
22:31 | |
| MoarVM: cd8e407740 | (Samantha McVey)++ (committed using GitHub Web editor) | 110 files Merge branch 'master' into libato |
|||
| MoarVM: a342936cde | (Samantha McVey)++ (committed using GitHub Web editor) | 108 files Merge pull request #799 from samcv/libato Add libatomic_ops as a submodule instead of in-tree |
|||
| MoarVM: 6affa07a2c | (Zoffix Znet)++ (committed using GitHub Web editor) | README.markdown Add Appveyor badge |
22:33 | ||
| MoarVM: d115bf720f | (Zoffix Znet)++ | src/math/bigintops.c Use proper INFINITY constant Unbreaks Windows build |
22:53 | ||
|
22:59
travis-ci joined
|
|||
| travis-ci | MoarVM build failed. Zoffix Znet 'Add Appveyor badge' | 22:59 | |
| travis-ci.org/MoarVM/MoarVM/builds/362868302 github.com/MoarVM/MoarVM/compare/a...ffa07a2cc2 | |||
|
22:59
travis-ci left
|
|||
| samcv | well i opened an issue for this github.com/MoarVM/MoarVM/issues/837 | 23:09 | |
| the Probably version skew, no object found at index 140389328393323 | |||
|
23:11
travis-ci joined
|
|||
| travis-ci | MoarVM build failed. Zoffix Znet 'Use proper INFINITY constant | 23:11 | |
| travis-ci.org/MoarVM/MoarVM/builds/362874609 github.com/MoarVM/MoarVM/compare/6...15bf720f3a | |||
|
23:11
travis-ci left
|
|||