tbrowder .tell AlexDaniel i'm getting close to a useful nqp sub run-command based on Zoffix's code + jnthn's insights and help. it uses spawnprocasync and i would like to eventually submit it as a PR sub addition to nqp/src/core/IO.nqp. it will have tests, and it will be used for testing my previous nqp PR on the improved pod panic message. it runs successfully now, captures both stdout and stderr for use (but only if 00:09
yoleaux tbrowder: I'll pass your message to AlexDaniel.
tbrowder requested). i just need to clean it up a bit.
AlexDaniel k 00:10
yoleaux 00:09Z <tbrowder> AlexDaniel: i'm getting close to a useful nqp sub run-command based on Zoffix's code + jnthn's insights and help. it uses spawnprocasync and i would like to eventually submit it as a PR sub addition to nqp/src/core/IO.nqp. it will have tests, and it will be used for testing my previous nqp PR on the improved pod panic message. it runs successfully now, captures both stdout and stderr for use (but only if
tbrowder i’m getting segfault trying to build master branch nqp 01:37
timotimo yeah, known, there was a too-early version bump 01:39
AlexDaniel the segfault should go away with MVM_JIT_EXPR_DISABLE=1 01:43
MasterDuke or if you pull master MoarVM
AlexDaniel MasterDuke: oh, so we can rebump? 01:45
MasterDuke yeah. github.com/MoarVM/MoarVM/commit/a01cdb449c
Geth nqp: 03bb36a3b4 | (Aleks-Daniel Jakimenko-Aleksejev)++ | tools/build/MOAR_REVISION
[MoarVM Bump] a01cdb449 Disable getlexstatic_o […]

MoarVM bump brought: github.com/MoarVM/MoarVM/compare/2...ga01cdb449		 01:49
¦ nqp: version bump brought these changes: github.com/MoarVM/MoarVM/compare/2...ga01cdb449
rakudo: ad8823eedf | (Aleks-Daniel Jakimenko-Aleksejev)++ | tools/build/NQP_REVISION
[NQP Bump] Brings 8 commits

NQP bump brought: github.com/perl6/nqp/compare/2018....g03bb36a3b 03bb36a3b [MoarVM Bump] a01cdb449 Disable getlexstatic_o […] 4c9302764 correct return values for ```decoderempty```, they were reversed 51b9df0a8 [MoarVM Bump] Brings 6 commits 135154876 [js] eslint --fix 48dcb3f30 [js] Start taking grapheme boundaries into account more fc524076c [js] Make enumcharlist work better with graphemes 4c8ea814c [js] Strip marks more correctly in weird cases 9d664e6d6 [js] Scan in regexes by graphemes not by bytes
¦ rakudo: version bump brought these changes: github.com/perl6/nqp/compare/2018....g03bb36a3b
travis-ci NQP build passed. Aleks-Daniel Jakimenko-Aleksejev '[MoarVM Bump] a01cdb449 Disable getlexstatic_o […] 02:04
travis-ci.org/perl6/nqp/builds/340771205 github.com/perl6/nqp/compare/4c930...bb36a3b40e
AlexDaniel yay 02:05
travis-ci Rakudo build passed. Aleks-Daniel Jakimenko-Aleksejev '[NQP Bump] Brings 8 commits 02:37
travis-ci.org/rakudo/rakudo/builds/340771233 github.com/rakudo/rakudo/compare/c...8823eedf98
Geth nqp: tbrowder++ created pull request #411:
add a new built-in sub: run-command		 04:56
AlexDaniel tbrowder: don't have time to review right now but that looks very complicated for what it does 05:10
dogbert11 AlexDaniel: have you had time to test your bots against the latest fixes? 10:17
tbrowder .tell AlexDaniel maybe, but the scaffolding required to capture stderr or stderr is not very intuitive, is easy to get wrong, and IMHO, the wrapping is useful. it can certainly be simplified such as always returning both streams, and probably a better way to return the streams’ output. 11:30
yoleaux tbrowder: I'll pass your message to AlexDaniel.
Geth rakudo: 555db42a37 | (Zoffix Znet)++ | src/core/Backtrace.pm
Make Backtrace.gist more helpful

Fixes R#1521 github.com/rakudo/rakudo/issues/1521
Currently we go through Mu.gist, which is just Mu.perl, which just gives "Backtrace.new", leading users to believe it's an empty backtrace. ... (7 more lines)		 14:14
synopsebot R#1521 [open]: github.com/rakudo/rakudo/issues/1521 [LHF] Backtrace.gist could be better
AlexDaniel heh, whateverable has 14534 rakudo builds :D
yoleaux 11:30Z <tbrowder> AlexDaniel: maybe, but the scaffolding required to capture stderr or stderr is not very intuitive, is easy to get wrong, and IMHO, the wrapping is useful. it can certainly be simplified such as always returning both streams, and probably a better way to return the streams’ output.
Geth roast: ca8c23ae77 | (Zoffix Znet)++ | integration/error-reporting.t
Spec Backtrace.gist

R#1521: github.com/rakudo/rakudo/issues/1521 Rakudo fix: github.com/rakudo/rakudo/commit/555db42a37		 14:15
AlexDaniel all that taking 31G only, fun stuff
Zoffix ZOFVM: Files=1289, Tests=153214, 145 wallclock secs (20.14 usr 3.34 sys + 3079.15 cusr 154.43 csys = 3257.06 CPU) 14:17
dogbert2 AlexDaniel: have you tested jnthn's fixes against your bots? 14:19
AlexDaniel dogbert2: not yet, and that's a good idea I think 14:20
AlexDaniel looks at maxed out swap
Zoffix Alright. Totally burnt out. Time to retreat to the land of Fallout 4 and ignore all github and social media notifications. 14:21
AlexDaniel Zoffix++
Zoffix AlexDaniel: I'll be away for a few weeks, ignoring all notifications, so if I broke something I won't know about it, someone else will need to fix it
dogbert2 Zoffix--: get a good game and have a well deserved rest
Zoffix \o 14:22
jnthn Take care Zoffix o/
dogbert2 oops Zoffix+++
AlexDaniel Zoffix: that's ok, take care
dogbert2: we'll see how it goes :) 14:27
dogbert2 very interesting 14:28
Kaiepi can yall rec me some good advice on how to get into perl dev? 14:29
AlexDaniel Kaiepi: what are you interested in specifically?
Kaiepi i'm looking to fix some bugs that break callbacks in nativecall on openbsd and freebsd has issues with dns resolution 14:30
it's a fun project because it gives me a chance to learn more about how moar, nqp, and rakudo work 14:31
but i don't understand moarvm very well as i'm only just jumping 14:32
i think i should pay more attention to the github issues and pullreqs to find out where i should be looking for and learn what does what 14:33
dogbert2 Kaipei: a good first step is to fork github.com/rakudo/rakudo and then download and compile it on your system. nqp and moarvm will be pulled in as well 14:34
Kaiepi i already do that
dogbert2 but perhaps you've already done that 14:35
:-)
Kaiepi ./Configure.pl --prefix $HOME/.perl6 --debug=3 --optimize=0 14:36
dogbert2 and then 'make install ...'
Kaiepi yep
i've learned how to automate this 14:37
dogbert2 cool
have you tried running spectest's and stresstest's as well
Kaiepi yep 14:38
those are what i really want to fix
dogbert2 what kind of errors are you getting, testfails, SEGV's or panics
Kaiepi segfaults on the callback tests from rakudo 14:39
dogbert2 gdb time
Kaiepi damn right
i'm slowly getting the hang of it
dogbert2 ./perl6-gdb-m <failing test file> 14:40
jnthn Kaiepi: MoarVM can be built with two different libraries for handling native calling; it may be worth compiling it with the alternative one to at least try to rule out an issue with that
Kaiepi the alternative?
jnthn I think dyncall is the default everywhere, but a system libffi may also be used 14:41
It's a configure-time option
Kaiepi ah 14:42
i'll try that about
s/about/out/g
AlexDaniel oh gosh I might need to restart the bots again… 14:48
MasterDuke Kaiepi: have you seen github.com/edumentab/rakudo-and-nq...als-course ? it helped me get more familiar with how some of the stuff worked when i was just getting into it 14:52
Kaiepi OH 14:53
THAT'D BE REALLY HELPFUL
THANK YOU
MasterDuke sure, and jnthn++ for creating it 14:54
tadzik yep, that course is fantastic
Kaiepi bless you jnthn++
at somepoint i'll get those issues with dns on freebsd and the nativecall callbacks not working on openbsd 14:57
those are what i need to be able to get the lang to work proparly
MasterDuke dns? 14:58
Kaiepi dns resolution 14:59
on freebsd i couldn't get perl to resolve any
dogbert2 AlexDaniel: what happened to the bots? 15:12
AlexDaniel dogbert2: what do you mean? :)
dogbert2 :)
they seem to have been restarted 15:13
AlexDaniel well, if you're asking about the leak, then I think it's too early to talk about that
as for the second restart, I found a bug :)
if anybody is wondering why some bots are still not up, they will be in a few minutes once the test suite gets to them… 15:14
(and if anybody wonders why the test suite matters, it's because it creates precomp files which bots are not allowed to write… stuff is complicatedly weird) 15:15
Kaiepi dw i'm working on the freebsd and openbsd build bots 15:17
and finish the Net::DNS port 15:18
you've been a great help, thank you for the support 15:22
Geth rakudo: ash++ created pull request #1522:
Remove the Bridge method from the numeric classes		 15:36
AlexDaniel dogbert2: I think I'm seeing some extra segfaults :D 16:06
dogbert2 interesting
lizmat travelling& 16:08
AlexDaniel greppable6: Bridge 16:13
greppable6 AlexDaniel, 36 lines, 9 modules: gist.github.com/1160653db1dc2969dc...48b45cbd6c
Geth nqp: tbrowder++ created pull request #412:
add a test for stderr in spawnprocasync		 16:34
pmurias are there any benefits to fetching an str attribute with nqp::getattr instead of nqp::getattr_s? 16:46
jnthn No 16:47
pmurias Do we depend on nqp::getattr working (boxing) for native attributes? 16:49
Geth nqp/master: 5 commits pushed by pmurias++
Kaiepi anyway i'm not clear on how the tisket system works compared to github issues 17:14
jnthn pmurias: Not on purpose; the mechanism is actually for a longer-term feature (complex value types)
Which we don't yet support 17:15
jnthn bbl
AlexDaniel clickbaits news.perlfoundation.org/2018/02/gra...nd-im.html 17:20
I… am not sure how I feel about this… 17:21
at first I was so happy that someone is going to put so much work into the doc repo, but now I'm reading details
and it's so weird
“Contribute 1 entry-level tutorial per week, to be published preferably on non-perl publications (Medium, dev.to, hackernoon).” 17:22
so instead of contributing it to the doc repo it's going to be published somewhere else? That may be ok for marketing but not exactly what I'm expecting to see in terms of improving the docs 17:23
“Day 7: identify core of people working on the documentation and create an open channel of communication, for instance, a Telegram channel with the GitHub bot connected”
telegram channel?… seriously?
and so on and so forth 17:25
otherwise this is very promising 17:26
bisect: say rand 17:37
bisectable6 AlexDaniel, Bisecting by output (old=2015.12 new=555db42) because on both starting points the exit code is 0
AlexDaniel, bisect log: gist.github.com/61d21a6a522e8d5a9b...de0b8abcdc
AlexDaniel, (2015-12-25) github.com/rakudo/rakudo/commit/07...dc61f84053
AlexDaniel, The result looks a bit unrealistic, doesn't it? Most probably the output is different on every commit (e.g. ｢bisect: say rand｣)
AlexDaniel bisect: say rand
bisectable6 AlexDaniel, Bisecting by output (old=2015.12 new=555db42) because on both starting points the exit code is 0
AlexDaniel, bisect log: gist.github.com/bfe2507a92483bca89...9a37810b21 17:38
AlexDaniel, (2015-12-25) github.com/rakudo/rakudo/commit/07...dc61f84053
AlexDaniel, The result looks a bit unrealistic, doesn't it? Most probably the output is different on every commit (e.g. ｢bisect: say rand｣)
AlexDaniel dogbert2: do you feel adventurous? 17:39
[Coke] AlexDaniel: as someone on the GC and someone on the perl6 doc project, I have similar questions to you about that grant in particular; I'd appreciate it if you can add any of your questions you'd care to share as public comments on the proposal. 17:47
AlexDaniel [Coke]: ok 17:48
[Coke] Thank you, I appreciate it.
dogbert11 AlexDaniel: what kind of adventures are you referring to? 18:07
AlexDaniel dogbert11: I was thinking maybe you'll be able to play with that recent segfault, but now I'm realizing that it's probably a bit difficult to reproduce it locally 18:09
sooo nevermind
dogbert11 perhaps MasterDuke can investigate 18:16
btw, what about memory usage 18:17
AlexDaniel shareable6: uptime
shareable6 AlexDaniel, 3 hours and 18 minutes, 398.015625MiB maxrss. This is Rakudo version 2018.01-191-g555db42a3 built on MoarVM version 2018.01-89-ga01cdb449 implementing Perl 6.c.
AlexDaniel bisectable6: uptime 18:18
bisectable6 AlexDaniel, 3 hours, 18 minutes, and 43 seconds, 236.292969MiB maxrss. This is Rakudo version 2018.01-191-g555db42a3 built on MoarVM version 2018.01-89-ga01cdb449 implementing Perl 6.c.
dogbert11 is that good or bad
AlexDaniel dogbert11: so far it's great, but I think I'll be able to tell after like 24 hours
dogbert11 cool
Kaiepi where should i look if i want to implenet utf32? 18:27
lizmat Kaiepi: I'm not sure how jnthn's stance is on this 18:28
Kaiepi i'll wait to ask him
lizmat feels to me it should be as low level as possible for a "standard" encoding
having said that, it should be possible to implement it in pure userland with the new encoder/decoder API 18:29
Kaiepi yeah, i seems like something moarvm' doing
lizmat unfortunately, a. I'm still very foggy minded, and b. even if I weren't, I actually haven't looked at that API myself yet
Kaiepi what's the irc mailbot?
lizmat you mean, for asking a question? 18:30
Kaiepi yeah
lizmat ".ask"
".ask jnthn the question"
.ask lizmat where are you at ?
yoleaux lizmat: Talking to yourself is the first sign of madness.
Kaiepi sorry i cuolnd't find his nick in the convo
lizmat most irc clients have auto-complete 18:31
jn TAB is usually enough for me
Kaiepi it helps when servers has frooms with dadaist keyboar dsmashing for titles 18:34
|Tux| Rakudo version 2018.01-191-g555db42a3 - MoarVM version 2018.01-91-g1c79ca35d
csv-ip5xs0.793 - 0.795
csv-ip5xs-207.399 - 7.517
csv-parser11.449 - 11.514
csv-test-xs-200.530 - 0.557
test8.570 - 8.717
test-t2.531 - 2.550
test-t --race1.040 - 1.067
test-t-2045.016 - 45.907
test-t-20 --race15.897 - 16.551
 18:46
Kaiepi so for the github bots 19:02
githooks are what i'll rely on right?
seems simpla 19:03
lizmat Kaiepi: I think so, yes
Kaiepi perfect, i've tinkered with them before 19:04
Geth rakudo: e7af9a6a90 | (Elizabeth Mattijsen)++ | src/core/Cool.pm
Fix ($ = 42) vs ($ = "42).subst-mutate

As reported by lucasb: there should not be a difference in return value in case of a failed match.		 19:41
lucasb Can the line github.com/rakudo/rakudo/blob/mast...ol.pm#L184 be changed to "$value-to-subst-mutate = $str if $match" to avoid changing the type of the variable if the match failed? 20:11
lizmat lucasb: looking at that now 20:15
running spectest now 20:16
$match can be an empty string, right? so "if $match" wouldn't cut it 20:19
lucasb isn't it a match object?
lizmat ah... good point 20:21
Geth rakudo: 6101552658 | (Elizabeth Mattijsen)++ | src/core/Cool.pm
Cool.subst-mutate only mutates if there's a match

So the original Cool object remains unaltered if there was no match. Which is probably better for many reasons. lucasb++ for spotting/suggesting.		 20:34
lucasb yay! thank you, lizmat 20:36
Geth rakudo: 77794a28aa | (Elizabeth Mattijsen)++ | src/core/Str.pm
Make Str.subst-mutate always return a List on :g

Partial fix for GH #1523, tests needed		 21:24
synopsebot RAKUDO#1523 [open]: github.com/rakudo/rakudo/issues/1523 .subst/.subst-mutate should always return a List when :g is specified
lizmat situation with .subst is slightly more complicated
commuting&
Kaiepi what kind of detailed guides out theere for gdb? 21:28
i starting to work out how to use it, but there must be more 21:29
i'll chack the man page more lcosely but if you get any lemme know 21:32
MasterDuke heh, if you find a good one let me know 21:36
Kaiepi: timotimo knows a bunch about gdb, and he's created some helper scripts for doing rakudo/moarvm dev with it, might want to ping him 21:37
timotimo Kaiepi: o/
it me
lucasb to propose changes, can I open an issue on rakudo repo or on 6.d-prep ? 21:42
I want num vars to default to 0e0
m: say my num $
camelia NaN
lucasb is there any chances of this changing? 21:43
and, would it need to wait 6.d or it can change before it (if there is consensus)? 21:44
[Coke] in general, anything that is tested in the 6.c spec needs to go into 6.d if it's a change. 21:48
lucasb I don't think "my num" is tested in roast, but I'm not sure 21:50
[Coke] github.com/perl6/roast/blob/6.c-er....t#L12-L17
lucasb oh, right, it's there. thanks 21:51
timotimo ugh, i said i was here and then i disappeared 21:58
well, it's dinnertime now
jnthn lucasb: It's consistent with object variables being initialized to a type object: NaN is the undefined of num. It's like that by choice rather than by accident.
(Doesn't mean you can't request it be reconsidered, of course. But it was at least considered and intended.) 21:59
lucasb jnthn: I see, thanks 22:01
if there was at least a tiny chance, yes, I would request it to be reconsidered, but IDK what's the general feeling of other people 22:03
to me, it just feels strange. native int defaults to 0, so I would expect native num to default to 0e0
if I understand correctly, file static float/double variables in C also defaults to 0e0 22:04
jnthn No, it defaults to junk
oh, file static
Hmm...I always initilaize such things explicitly 22:05
jnthn bbl
geekosaur portability suggests you should initialize. security suggests it will be zeroed 22:06
Kaiepi there's something i noticed when working on the freebsd travis boto yesterday
geekosaur (but there's a stinger there as well: while modern CPUs will treat memory zeroes as value zeroes, there are possible value representations where that isn't true.)
Kaiepi my router has no ipv6 support without don't something like janking
buy anyway i tried to resolve an ipv6 domain name and failed 22:07
should there be a watning for that when that happens?
geekosaur resolving domain names shouldn't be affected by router ipv6 support. although if you are using the router's built-in DNS cache/forwarder then ti might not have AAAA record support 22:08
Kaiepi i am
port forwarding is also broken
lucasb geekosaur: I think the C standard says that file static vars *are* initialized. (dunno if all compilers are compliant) 22:09
Kaiepi might need to flash it with some custom firmware to make it work for me
geekosaur you can point directly to an external DNS source (e.g. google's) instead of the router. you can also use something like HE's ipv6 tunnels 22:10
Kaiepi using gooog'es dns sounds good 22:11
the ones my isp provides breal a lot 22:12
lucasb I submitted the proposal: Change the default value of native num variables from NaN to 0e0 22:29
github.com/perl6/6.d-prep/issues/4
Kaiepi i'm not a huge fan of nan
lucasb It's a little detail. I hope it changes, but it's ok if it gets rejected.
Please, leave you opinion there. Thanks people. 22:30
Kaiepi flashbacks to manually garabge collecting by making vals numb so its garbagec ollector can notice its dead
that gc is something else 22:31
aha 23:03
got more detailed crash info on the native callbacks thing 23:04

Powered by Cro and the Raku® Programming Language.

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

Thank you!