02:07 MasterDuke joined
MasterDuke i haven't read it yet, but arxiv.org/abs/2411.11469 looks like it might be interesting 02:08
timo those are some bold claims in the abstract of that paper 02:27
03:18 MasterDuke left 07:27 [Coke] left 08:48 sena_kun joined 09:22 [Coke] joined 09:52 sena_kun left
lizmat hmmm looks like I git myself a relatively easy way of creating a heap corruption 11:05
*got
reliably
timo cool, share? :) 11:18
lizmat install raku.land/zef:lizmat/REPL?v=0.0.9 11:20
install Terminal::LineEditor
> repl
then enter: say ²³⁴⁴/₇;
I then get: 11:21
rakudo(59264,0x1ec19cc00) malloc: Heap corruption detected, free list is damaged at 0x6000036bc100
*** Incorrect guard value: 18433
rakudo(59264,0x1ec19cc00) malloc: *** set a breakpoint in malloc_error_break to debug
on a M1 fwiw 11:22
timo right
btw did you know about "fenced paste"? that may not be the real name
lizmat doesn't ring a bell 11:23
timo ok that is definitely not what it's called haha
bracketed paste
how do i know if the repl i just started from inside the repl is the REPL repl? do i need to rakudo -Msomething? 11:24
lizmat the module installs a "repl"script
timo oh, > was your shell 11:25
lizmat if you call "repl" as a CLI, you get the REPL version
yes, sorry if that wasn't clear
I meant: $ repl , duh :-)
timo that doesn't sound like a name it's fine to squat on :D
though of course it doesn't go into /usr/bin or anything 11:26
still ...
does it reproduce every time on your end?
lizmat yes, every single time
with the same 18433 and 0x1ec19cc00 values 11:27
could be an issue in Terminal::LineEditor
ah, yes, no problem with Linenoise as the editor 11:29
I'll make a Terminal::LineEditor issue
timo how can i tell which i'm in? 11:30
lizmat what does it say when you enter "ed" ? 11:31
timo Using the LineEditor editor
lizmat then that's the correct one
timo it's not explosive on my end, i'm running it with valgrind --trace-children=yes 11:33
lizmat ok, so it's M1 specific apparently
timo can you build a moar with asan? 11:34
ever tried that?
i just recently added some extra output to Configure.pl to help the user because there's some annoying additional steps necessary to be able to build rakudo when your moar has asan in it 11:35
lizmat no, I haven't recently: every time I did that, it messed up all of my workflows
timo i can imagine 11:36
lizmat it's just too disruptive for me, unless I'm really working on NQP / MoarVM related stuff directly
timo yeah, it's fair
lizmat anyways, looks like a Terminal::LineEditor issue, so japhb will know about it :-)
timo does japhb have a silicon mac to repro this? 11:37
lizmat I haz no idea: japhb is pretty busy moving house atm, so I doubt he'll have time to look at it soon anyway 11:38
perhaps patrickb might :-)
github.com/japhb/Terminal-LineEditor/issues/7 11:42
ok, enough excitement :-) 11:43
timo fair
patrickb Me no haz M1 either... 12:21
Actually I do. It's in a Mac stadium account. I use that machine for building the compiled releases. 12:25
timo: Want access?
13:17 finanalyst joined 13:25 patrickb left 13:26 patrickb joined 13:55 finanalyst left
timo not 100% sure i have the spare cycles to actually look deeply into this, but maybe it's not too hard to find once i point asan at it? 15:42
lizmat and yet another Rakudo Weekly News hits the Net: rakudoweekly.blog/2024/11/25/2024-...o-release/ 16:23
17:28 coleman left 17:30 coleman joined 18:12 lizmat_ joined 18:13 harrow left 18:14 harrow joined 18:15 lizmat left 18:21 lizmat_ left, lizmat joined 18:49 sena_kun joined 22:24 sena_kun left 22:36 sena_kun joined 23:14 sena_kun left