jnthn samcv++ xx * 00:01
samcv jnthn: releases page is updated now :) 00:03
homepage too
cleaning up the script I wrote to generate the releases page at least
changed the home page manually
jnthn :-) 00:04
That's still a huge reduction in the work needed to keep it updated 00:05
samcv yes!
Geth MoarVM: 65273a403e | (Samantha McVey)++ | docs/moarvm.org_releases.pl
Add script that generates moarvm.org/releases.html

This reads docs/ChangeLog and requires the Text::Markdown Perl 5 module.		 00:25
MoarVM: ac9f24145c | (Samantha McVey)++ | docs/release_guide.md
Update release_guide.md w/ info on generating releases.html
MasterDuke now that we've had the release any objection to someone merging github.com/MoarVM/MoarVM/pull/785 ? 01:27
yoleaux 25 Jan 2018 07:17Z <brrt> MasterDuke: seeing it, I will check it out as well. I was hoping you'd find bugs like these :-)
MasterDuke .tell brrt cool. what do you think about github.com/MoarVM/MoarVM/pull/787 (sp_findmeth template)? should it be merged as is, or are you working on a different fix (than marking the op invokish)? 01:28
Geth MoarVM: d3ce626b49 | MasterDuke17++ (committed by timo) | src/jit/core_templates.expr
Add JIT template for atkey_o
 02:30
MasterDuke timotimo++ 02:31
i have another 6 or 7 templates so far in a branch, but i'm waiting to see if the box_i fix will be simple before i create the PR 02:32
timotimo: want to look at a segv inducing getdynlex template? 02:42
timotimo probably going to sleep too soon for that :| 02:54
but if you paste it i'll have a look tomorrow
MasterDuke gist.github.com/MasterDuke17/a8993...fb66ab9727
timotimo the lexical_registry in the cur_frame->static_info->body.lexical_names seems broken? 03:00
hm, ok, so 03:06
the template is supposed to call MVM_frame_getdynlex with cur_frame's caller, right? 03:07
>>> print cur_frame 03:08
$33 = (MVMFrame *) 0x7ffff7fd4068
>>> print tc->cur_frame->caller
$34 = (MVMFrame *) 0x7ffff6593c28
however!
>>> print tc->cur_frame
$35 = (MVMFrame *) 0x7ffff7fd4030
see the cur_frame that got passed from the template is just a couple bytes offset from the cur_frame
*maybe* that's broken. or maybe that's just because it's on the stack memory region
m: say 0x7ffff7fd4030 - 0x7ffff7fd4068 03:09
camelia -56
timotimo an MVMFrame is 184 bytes long
MasterDuke: i think the ^caller macro requires a (load ...) 03:10
or maybe instead of addr it should have a getf
^caller is only used in callercode, and i'm not sure if that's currently exprjitted anywhere 03:11
cool 03:12
(macro: ^caller () (^getf (^frame) MVMFrame caller))
with this replacing the ^caller macro it seems to get much further in the nqp making
MasterDuke: ^- go forth and make things awesome :)
timotimo off to bed
also, it's not entirely unlikely that the ^params macro wants to be changed the same way 03:13
MasterDuke cool, thanks! i'll give that a shot 04:18
Geth MoarVM: MasterDuke17++ created pull request #788:
Add more jit templates		 04:57
brrt good * #moarvm 07:10
samcv good * brrt 23:12
oh i guess you're offline

Powered by Cro and the Raku® Programming Language.

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

Thank you!