github.com/moarvm/moarvm | IRC logs at colabti.org/irclogger/irclogger_logs/moarvm
Set by AlexDaniel on 12 June 2018.
Geth MoarVM: Kaiepi++ created pull request #1001:
Include system library paths in Configure.pl if configured to use them		 01:20
samcv Kaiepi, reviewed your PR 05:40
brrt ohai #moarvm 17:43
yoleaux 23 Nov 2018 19:18Z <nine> brrt: I sure hope I can help :)
brrt I can't seem to find the concentration to actually get to work on it :-( 17:44
(it being, the implementation of nativecallinvoke in the expr jit, especially testing that it is correct)
lucasb is there any way to segfault moarvm on purpose from Perl 6 code? 17:49
lucasb and if there isn't, can't one exist? 17:50
I remember seeing this "feature" in some environment/language 17:51
japhb lucasb: Do you want to *actually* segfault, or do you just want to send it sig 11 or so? 17:52
lucasb just for the purposes of learning what a legit segfault looks like in a debubging tool
japhb Oh, I see.
brrt write a C program, you'll get one
lucasb oh, then bind it with Nativecall?
brrt I'm not sure why you want to hurt moarvm so much
japhb lucasb: Yeah, it shouldn't be hard to do it with nativecall
brrt :-P 17:53
japhb brrt: He wants to learn how to debug a segfault.
lucasb I don't want a standalone C program, I want to segfault from within Perl 6 code
japhb So cause one and look at the crash in a debugger
brrt hehe
well, i'll make one for you, no problem
lucasb brrt: you will? :-)
brrt just wait, they'll pop up soon enough 17:55
but we try not to have segfaults. I'm not sure we're right about that, but that's what we do
lucasb you are not sure moarvm should have *no* segfaults? (Just trying to learn/understand :-) 17:57
brrt well, something like java will not emit checks for null pointers, but translate a segmentation vialotion to a null pointer to a NullPointerException 18:00
that makes the code quite a bit smaller for the common case that a pointer is indeed not null
there's another school of thought that says all UB is a safety problem. I'm not so sure what my position on that is 18:02
lucasb but a NullPointerException is "catchable", right? (dunno Java)
brrt yes
but that's not a problem.
when a SEGV is raised, and it happens to be caused by access of a null pointer, it is really not all that hard to write a signal handler to catch it and translate to a catchable exception 18:03
I forgot which signals could not be caught, but SIGSEGV is not one of them 18:04
lucasb ah, I remembered where I've seen it: Redis! 18:05
redis.io/commands/debug-segfault
DEBUG SEGFAULT performs an invalid memory access that crashes Redis. It is used to simulate bugs during the development.
I think this could be useful, can I haz it in moarvm?
brrt 'no' 18:07
timotimo easy to build a module that does it 18:08
brrt :-P
lucasb hehe, ok. I'll push this project into my todo list :) 18:09
timotimo it'll be two lines 18:11
Geth MoarVM: c0799ec146 | (Ben Davies)++ | Configure.pl
Include system library paths in Configure.pl if configured to use them

Fixes #1000		 19:48
MoarVM: 6307866af8 | (Samantha McVey)++ (committed using GitHub Web editor) | Configure.pl
Merge pull request #1001 from Kaiepi/configure2

Include system library paths in Configure.pl if configured to use them
lucasb What was "s/\n/ /g" before, became "s/\n//g", but maybe pkg-config never outputs more than one line, so probably harmless, otherwise things could get concatenated. 20:02
travis-ci MoarVM build passed. Samantha McVey 'Merge pull request #1001 from Kaiepi/configure2 20:10
travis-ci.org/MoarVM/MoarVM/builds/459918511 github.com/MoarVM/MoarVM/compare/6...07866af856
brrt .tell jnthn you were right, .net is much superior to java / jvm 21:37
yoleaux brrt: I'll pass your message to jnthn.
brrt as I am learning the hard way 21:40
timotimo i don't see where exactly spesh logs are freed 22:07
oh 22:09
maybe i see where it happens
lizmat And another Perl 6 weekly hits the Net: p6weekly.wordpress.com/2018/11/26/...onga-grep/ 22:57
jnthn Quite a few things for a slow week :) 23:06
yoleaux 21:37Z <brrt> jnthn: you were right, .net is much superior to java / jvm
jnthn lizmat++
timotimo jnthn: the spesh worker and threads are supposed to recycle spesh logs, right? 23:38
because i'm seeing MVM_spesh_log_create be called over and over 23:42
hum. my idea that maybe it's not running gc was not supported by evidence 23:47
bleh. the watched pot never boils 23:55

Powered by Cro and the Raku® Programming Language.

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

Thank you!