01:03 lizmat joined 03:57 vendethiel joined 04:45 TimToady joined 06:08 vendethiel joined 06:21 FROGGS joined 06:34 lizmat joined 06:40 vendethiel joined 07:33 Ven joined 07:55 Ven_ joined 08:30 FROGGS joined 09:43 brrt joined
brrt \o 09:43
brrt is clearly missing sufficient brainpower to deal with moarvm this week
09:44 zakharyas joined
brrt also: laziness. i knew i should've taken ref of .o the moment i closed my laptop 09:44
sorry :-( 09:45
FROGGS brrt: eat a snickers :o)
brrt yeah, i suppose i'll do something like that 09:46
10:11 Ven joined
jnthn This morning so far: "Why is MSVC giving me an insane stacktrace?" "Hmm...gdb gives me the same impossible stack trace." "Gah, valgrind agrees and whines loudly." "Oh, it actually happens." 10:28
dalek arVM: c14339c | jnthn++ | src/core/frame. (2 files):
Function to return without running exit handlers.

Will be used from a Perl 6 ext-op that fakes up a frame in order to run constraint checks under it, and then wants to tear it down. If we do try to run the exit handlers in this case, we've got a frmae that is not set up for it, and so get all manner of odd behavior.		 10:45
timotimo interesting thing that happened 11:24
that one there, i mean 11:25
arnsholt jnthn: So much fun, one might think today's a monday! =) 11:30
nwc10 nooooooooo - I was hoping that the week is all downhill from here 11:31
and this week has 5 working days in it
(and so do all the following weeks, until the APW)
(unless I burn up holiday)
11:38 FROGGS joined 12:01 Ven joined 12:05 brrt joined
brrt \o 12:05
a really evil stacktrace may be the result of being interrupted in a JIT frame 12:06
timotimo src/strings/ops.c:26:19: warning: return type defaults to ‘int’ [-Wimplicit-int] 12:28
MVM_STATIC_INLINE MVM_string_check_arg(MVMThreadContext *tc, MVMString *s, const char *operation) {
i can fix this
dalek arVM: f092e80 | timotimo++ | src/strings/ops.c:
MVM_string_check_arg is void, not (implicitly) int
 12:29
12:36 Ven joined 13:25 Ven joined
brrt oh, timotimo, i've another pretty one 13:51
for the sp_get_* series
there's a difference between getting a signed and an unsigned byte/word/dword from memory
because you have to alternate between zero-extension (unsigned) and sign-extension moves 13:52
note to self 13:54
we need a conversion node in the expression tree
13:59 JimmyZ_ joined
brrt maybe the whole sp_get_* stuff isn't really worth it 14:13
14:23 Ven joined 14:39 Ven joined 14:48 Ven joined
jnthn is closing in on which optimization busts the Test::CSV test suite 14:49
Relax, it wasn't the exception ops :P
darn, narrowed it down to findmeth... 14:53
It seems that we spesh-time select the incorrect method. 15:05
*an
It's incredibly odd
15:06 lizmat joined 15:07 Ven joined 16:19 lizmat joined
timotimo right, i was thinking we might want signed and unsigned 16:24
i'm not building all those for the jit, btw :)
japhb jnthn: With the bugs you're finding, I'm starting to have "how did it ever work?" feelings. That's a bit scary. 16:26
timotimo: p6weekly likely this week? 16:27
jnthn japhb: Mostly they're rarely tried things or things that only happen in sufficiently rare circumstances that you go a long way without noticing them. 16:28
Given the sorts of language features we're dealing with along with the optimizations we're performing, we do pretty well. 16:30
16:30 mj41 joined
jnthn Especially given the "budget" :) 16:30
japhb Well sure.
jnthn (Yes, I'm funded now, but wasn't for MoarVM work until now. :)) 16:31
(And others ain't.)
japhb I know. :-(
Wish I could change that.
jnthn But yeah, this stuff's hard in places, and the analysis to really understand bugs and do proper fixes is time-consuming and needs a good deal of patience. 16:32
mj41 timotimo: p6weekly - my two "tips": github.com/mj41/Perl6-Analytics-re...m-prev2.md and www.slideshare.net/michaljurosz1/pe...and-perl-5 16:33
ohh, afk 16:37
timotimo thanks, mj41 :) 16:48
japhb: i was going to! but i'm procrastinating rather effectively :P
japhb wonders if there's a good song for procrastination ... maybe something riffing on Stevie Wonder's song "Superstition" 16:53
timotimo a single song isn't nearly enough to procrastinate effectively 16:55
japhb I guess I should have s/for/about/, since that's what I actually meant. 16:56
Songs *for* procrastination? Comedy ones where you have to pay attention to lyrics. 16:57
timotimo OK :)
arnsholt Once again, I need to point out how nice the 6model code is to work with 16:59
Really, a gem among C APIs
timotimo argh, i turned on the screen reader on my gnome3 and can't turn it back off 17:00
it's woefully useless with weechat over ssh m)
17:23 FROGGS joined
jnthn timotimo: So if I type stuff, it'll say it? :) 17:34
jnthn goes for some dinner & 17:35
lizmat m: my $a = "🍺"; say ++$a; say --$a
camelia rakudo-moar 14458a: OUTPUT«🍻␤🍺␤»
[Coke] nice. 17:36
timotimo more like: whenever i switch tabs or someone says something, it starts reading from the top 17:37
lizmat m: my $a = "🐪"; say ++$a; say --$a 17:38
camelia rakudo-moar 14458a: OUTPUT«🐫␤🐪␤»
lizmat just a bit of -Ofun :-)
17:56 lizmat joined
FROGGS m: say uninames "🐫🐪" 17:59
camelia rakudo-moar 4f3687: OUTPUT«BACTRIAN CAMEL DROMEDARY CAMEL␤»
FROGGS can't see them :o(
lizmat :-( 18:00
timotimo me neiter :( 18:04
they look like whitespace to me ;\
lizmat ah... they look fine on my MBP :) 18:06
japhb And here on Ubuntu LTS.
lizmat .u 🍻🍺
m: say uninames "🍻🍺"
camelia rakudo-moar 4f3687: OUTPUT«CLINKING BEER MUGS BEER MUG␤»
japhb Though a lot of these want to be fullwidth, and they get stuffed in narrow columns, thus overlapping
jnthn I don't see them well on IRC, but my iPhone renders 'em nice :) 18:55
19:04 dagurval joined 19:29 Peter_R joined 19:51 lizmat joined 19:56 lizmat_ joined 20:33 colomon joined 21:06 sivoais joined 21:15 Ven joined 21:37 colomon joined 21:46 lizmat joined

Powered by Cro and the Raku® Programming Language.

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

Thank you!