github.com/moarvm/moarvm | IRC logs at colabti.org/irclogger/irclogger_logs/moarvm
Set by AlexDaniel on 12 June 2018.
Geth MoarVM: 1dc540f5f3 | (Ben Davies)++ | 3 files
Reimplement async socket introspection
 05:58
MoarVM: 40400fdcdd | (Aleks-Daniel Jakimenko-Aleksejev)++ (committed using GitHub Web editor) | 3 files
Merge pull request #991 from Kaiepi/async-introspection

Reimplement async socket introspection
travis-ci MoarVM build passed. Aleks-Daniel Jakimenko-Aleksejev 'Merge pull request #991 from Kaiepi/async-introspection 06:15
travis-ci.org/MoarVM/MoarVM/builds/466852047 github.com/MoarVM/MoarVM/compare/4...400fdcddc8
nine lizmat: no, the signedness issue is next 07:08
Geth MoarVM: lizmat assigned to niner Issue nqp::read-int16/32/64(oob,big-endian) allows out of buffer access github.com/MoarVM/MoarVM/issues/1021
b25de2702e | (Jonathan Worthington)++ | src/6model/reprs/MVMContext.c

The frame walker might vivify a lexical, and thus cause a GC run and so invalidate the object reference.		 09:06
lizmat hopes nine doesn't mind assigning the issue
lucasb hey, run MoarVM in NASA's static analyzer :) github.com/NASA-SW-VnV/ikos 10:42
lizmat looks interesting 10:49
jnthn Finally, I did some MoarVM stuff... 13:32
MoarVM: 7d3530308b | (Jonathan Worthington)++ | src/spesh/plugin.c
Set callsite in frame for JIT spesh plugin resolve

Otherwise, we won't properly mark the `args` buffer, which can lead to us processing outdated args.
timotimo oooh, very good
nwc10 the blocker being illness, $ork, lack-of-beer, $other, or $!private? 13:35
nwc10 hopes that he got the syntax correect in the joke part 13:35
jnthn Mostly $ork and a week and a bit's travel 13:36
nwc10 hopes that the travel was fun and generated lots of useful slide backgrounds 13:37
jnthn Well, I got a Swedish Christmas dinner :) 13:38
Geth MoarVM/master: 4 commits pushed by (Nick Logan)++, (Jonathan Worthington)++ 13:48
MoarVM: 8e078a2fed | (Jonathan Worthington)++ | src/core/threads.c
Improve handling of failure to join a thread

  * No need to panic, just throw
  * Include the error string, not just the code
Resolves #847.		 13:56
timotimo oh, jnthn, the ThreadPoolScheduler should probably catch failure to create a new thread, right? rather than the program dying? 14:01
jnthn I guess we could just note that in the scheduler debug log (if enabled), yes 14:02
timotimo and maybe not try to create a thread ~100 times a second :)
ah, there's an "exhausted" flag 14:04
travis-ci MoarVM build failed. Jonathan Worthington 'Improve handling of failure to join a thread 14:18
travis-ci.org/MoarVM/MoarVM/builds/467009775 github.com/MoarVM/MoarVM/compare/4...078a2fed52
timotimo hum. it's nqp-ified code, and there's many places where we try to add workers, and it'd have to return a different worker in those cases and ... ehhhh, maybe a task for another day 14:20
timotimo i wonder if nativerefs for lexicals could benefit from some spesh logging 16:26
so they could be turned into regular lexical lookups with a guard
Geth MoarVM: a8e63d2f59 | (Stefan Seifert)++ | 2 files
Sign extend values when reading less than 64 bit

readint returns a 64 bit signed integer. When reading smaller integers from a byte stream, we need to sign extend them to the full size.		 20:27
nine I hope someone knows a better way to achieve this ^^^ but it at least works 20:28
lizmat: ^^^
travis-ci MoarVM build passed. Stefan Seifert 'Sign extend values when reading less than 64 bit 20:45
travis-ci.org/MoarVM/MoarVM/builds/467183963 github.com/MoarVM/MoarVM/compare/8...e63d2f5936
Geth MoarVM: c32000ba8b | (Stefan Seifert)++ | src/core/interp.c
Use read_buf also when switching endianness

Calling read_buf once instead of multiple at_pos calls should be faster and also give the same error behavior whether we need to switch bytes afterwards or not.		 21:04
nine lizmat: ^^^ seems to fix the read-int.t tests 21:07

Powered by Cro and the Raku® Programming Language.

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

Thank you!