00:03 Ven joined 00:23 Ven joined 00:44 Ven joined 00:48 pyrimidine joined 01:23 Ven joined 01:32 dalek joined 01:35 pyrimidine joined 01:44 Ven joined 01:50 leego joined 01:55 leego joined 01:58 pyrimidine joined 02:01 FROGGS_ joined 02:10 pyrimidine joined 02:23 Ven joined 02:46 Ven joined 02:49 ilbot3 joined 03:03 Ven joined 03:18 ggoebel joined 03:20 pyrimidine joined 03:23 Ven joined 03:42 Ven joined 04:02 Ven joined 04:22 Ven joined 04:43 Ven joined 05:03 Ven joined 05:05 pyrimidine joined 05:22 Ven joined 05:42 Ven joined 06:02 Ven joined 06:14 pyrimidine joined 06:22 Ven joined
nwc10 jnthn: spectest6 eventually gave up with a non-ASAN failure after just over 24 hours of looping: paste.scsys.co.uk/540693 06:22
so, I think that makes it a good moarning
(not perfect, but very much better)
06:48 pyrimidine joined 07:00 pyrimidine joined 07:08 domidumont joined 07:15 domidumont joined 07:29 JimmyZ joined 07:30 geekosaur joined 07:32 ilmari joined 07:35 leego joined 07:39 geekosaur joined 07:41 pyrimidine joined 07:57 geekosaur joined
samcv nwc10, do you usually build with clang or gcc? 07:57
08:04 leego joined 08:06 geekosaur joined 08:11 pyrimidine joined 09:01 zakharyas joined 09:08 pyrimidine joined 09:49 pyrimidine joined 10:07 pyrimidine joined 10:36 pyrimidine joined 10:51 pyrimidine joined 11:03 pyrimidine joined 11:07 pyrimidine joined 11:37 geekosaur joined 11:52 pyrimidine joined, Ven joined 12:11 Ven joined 12:31 pyrimidine joined, Ven joined 12:50 Ven joined
jnthn samcv: One comment but it's not very important... 13:23
samcv yeah i noticed that. it was already there I think before I changed things
jnthn Yeah, I noticed that too :) 13:24
samcv yeah so. gonna just leave it for now I think. just in case
jnthn Maybe I'll just merge it and clean up later? Or you want to do it now?
OK
samcv nah just leave it in now
jnthn Also noticed we call the struct NamedAlias but elsewhere it's name alias
Again, don't think I'm bothered enough about this to suggest we change it :) 13:25
Geth arVM: 816186484b | (Samantha McVey)++ | 4 files
Remove Unicode 1 names and add Unicode Name Aliases

This adds support in the codepointfromname op for all of the official Unicode name aliases. It removes the old Unicode 1 names in the name registry and replaces them with <control-xxxx> with the codepoint number.
There are only 5 codepoints which have changed their functionality aside from the 400+ Name Aliases added:
Correction:
   * BELL
... (18 more lines)		 13:29
arVM: fc12a60ec3 | (Jonathan Worthington)++ | 4 files
Merge pull request #497 from samcv/nou1names

Remove Unicode 1 names and add Unicode Name Aliases
13:50 pyrimidine joined 14:05 brrt joined 14:08 pyrimidine joined 15:05 pyrimidine joined 15:26 pyrimidine joined 15:46 pyrimidine joined
Geth arVM: 8442892451 | MasterDuke17++ | 8 files
Convert MVM_malloc+memset to MVM_calloc

In cases of `foo = MVM_malloc(...); memset(foo, 0, ...);`, convert them to `foo = MVM_calloc(...);		 15:52
arVM: d9482b9ba3 | (Jonathan Worthington)++ | 8 files
Merge pull request #498 from MasterDuke17/replace_malloc_plus_memset_zero_with_calloc

Convert MVM_malloc+memset to MVM_calloc
arVM/master: 5 commits pushed by jnthn++ 15:53
15:54 vendethiel joined
jnthn That's pretty :) 15:55
timotimo yup 15:58
17:36 domidumont joined
nwc10 ASAN is very excited by generate_codepoints_by_name 17:49
for example paste.scsys.co.uk/540698
I hope that the causeis obvious to someone. If not, I'll bisect 17:50
timotimo i'm glad we have you :) 17:55
lizmat nwc10: pretty sure samcv will have a look at it when she's awake again :-) 17:56
18:11 vendethiel- joined 18:35 vendethiel joined 19:50 vendethiel joined
timotimo samcv: does the backtrace in nwc10's paste ring a bell to you? 20:07
nwc10 I did bisect it. It's commit 816186484b5cc52f9ff1be6afa3b6f49264335bf 20:09
Remove Unicode 1 names and add Unicode Name Aliases 20:10
20:17 vendethiel joined
samcv yeah that struct didn't even exist before that commit 20:39
i am mostly awake let me look at it one sec
20:50 vendethiel joined
samcv how do I work with asan? I did ./Configure.pl --asan 21:08
not getting any messages
weird perl6 still works. but running moar i get that it's compiled with incompatible asan libs 21:09
==21995==Your application is linked against incompatible ASan runtimes.
i have clang 3.9.1-2 21:11
timotimo i haven't been able to build moarvm with asan last time i tried 21:13
jnthn samcv: If you have valgrind installed, it may be easier (doesn't need a special build) 21:15
It's slower, but should give at least as good diagnostics (sometimes better)
If you're on Linux valgrind should be a really easy install
samcv ok 21:21
i have valgrind
jnthn \o/ 21:22
There's a perl6-valgrind-m script btw
Which makes things super convenient, except on the occasion it doesn't :)
samcv yeah i have used it. i made my own perl6-callgrind script as well 21:23
jnthn ooh 21:25
I wondered a bit about making the perl6-valgrind-m respect a VALGRIND_TOOL env var or something so I could do VALGRIND_TOOL=massif perl6-valgrind-m ... 21:27
Which may be neater than a proliferation of scripts
timotimo right 21:29
samcv jnthn, how do i do leakcheck full using the valgrind script? 21:38
jnthn samcv: Lsast time I needed that I just edited it and put the arg after the valgrind call 21:45
samcv kk
jnthn (and was careful not to commit it :))
samcv heh
jnthn Maybe a generic VALGRIND_ARGS would be more useful 21:46
samcv m: say "\c[keycap: #, HUNDRED POINTS SYMBOL]" 21:47
camelia rakudo-moar 232641: OUTPUT«===SORRY!=== Error while compiling <tmp>␤Unrecognized character name keycap:␤at <tmp>:1␤------> say "\c[keycap:⏏ #, HUNDRED POINTS SYMBOL]"␤»
samcv m: say "\c[keycap: #]"
camelia rakudo-moar 232641: OUTPUT«===SORRY!=== Error while compiling <tmp>␤Unrecognized character name keycap:␤at <tmp>:1␤------> say "\c[keycap:⏏ #]"␤»
samcv m: say "\c[keycap: 3]"
camelia rakudo-moar 232641: OUTPUT«3️⃣␤»
samcv doesn't like the hash symbol hmm
m: say "\c[keycap: #]"
camelia rakudo-moar 232641: OUTPUT«===SORRY!=== Error while compiling <tmp>␤Unrecognized character name keycap:␤at <tmp>:1␤------> say "\c[keycap:⏏ #]"␤»
jnthn samcv: Might need to loosen stuff up here: github.com/perl6/nqp/blob/master/s...r.nqp#L157 21:51
ohh
Wow, we allow comments inside of \c[...] ??
samcv why…
jnthn I have no clue! 21:52
samcv it's basically already commented :P
jnthn Just reading the grammar and guessing :P
(Guessing that's why # is special-cased, I mean. But it calls <.ws> which will parse a comment. 21:53
timotimo we're mad! 21:55
samcv ok i got asan working
jnthn Nice 22:02
games &
samcv ok I fixed asan. will PR in a sec
Geth arVM: samcv++ created pull request #499:
Fix ASAN errors in generate_codepoints_by_name		 22:07
lizmat ah, the infamous off-by-one plot! 22:08
samcv yeah.
i had to change it to iterate down instead of up
well I maybe shouldn't have. i guess I _could_ have done it from 0 22:09
oh well it works now and still does every alias
lizmat I'll leave it to jnthn to merge, since it is not a blocker for you now, right ? 22:10
samcv no error results 22:11
if it was off by the opposite side it would fault
jnthn checks le PR 23:58

Powered by Cro and the Raku® Programming Language.

Please report any issues / comments / feature requests as an issue on App::Raku::Log.

Thank you!