»ö« Welcome to Perl 6! | perl6.org/ | evalbot usage: 'p6: say 3;' or rakudo:, or /msg camelia p6: ... | irclog: irc.perl6.org or colabti.org/irclogger/irclogger_logs/perl6 | UTF-8 is our friend!
Set by moritz on 22 December 2015.
dalek c: 4d09760 | (Zoffix Znet)++ | doc/Type/Mix.pod6:
Correct that Mix.Bag/BagHash .Ints not .rounds

Per today's changes[^1], the coercion now .Ints the weights instead of .round'ing
  [1] github.com/rakudo/rakudo/commit/a09c8dc99e		 00:07
p3rln00b And my internet went out before I could fix MixHash too :( 00:12
p3rln00b sacrifices a goat to computer gods
dalek c: fe570ad | (Zoffix Znet)++ | doc/Type/MixHash.pod6:
Correct that MixHash.Bag/BagHash .Ints not .rounds

Per today's changes[^1], the coercion now .Ints the weights instead of .round'ing
  [1] github.com/rakudo/rakudo/commit/a09c8dc99e		 00:13
p3rln00b haw...
llfourn it seems that "require" doesn't work if it's in an exported sub 00:16
require Foo; => "no such symbol 'Foo'"
ah it only fails if there isn't a symbol 'Foo' defined in Foo.pm6 00:18
llfourn I think require should return a handled failure. Just because you require something doesn't mean you expect there to be symbol defined in required compunit. 00:23
Herby__ question: if i have a multiline file, is there a slick way to split the file into a mutlidimension array? 01:03
[['this','is','line','one'], ['this,'is','line','two']]
[Coke] m: slurp.say 01:04
camelia rakudo-moar a09c8d: OUTPUT«Céad slán ag sléibhte maorga Chontae Dhún na nGall␤Agus dhá chéad slán ag an Eireagal ard ina stua os cionn caor is coll;␤Nuair a ghluais mise thart le Loch Dhún Lúich’ go ciúin sa ghleann ina luí␤I mo dhiaidh bhí gleanntáin ghlas’ G…»
[Coke] m: say slurp.lines>>.words 01:06
camelia rakudo-moar a09c8d: OUTPUT«((Céad slán ag sléibhte maorga Chontae Dhún na nGall) (Agus dhá chéad slán ag an Eireagal ard ina stua os cionn caor is coll;) (Nuair a ghluais mise thart le Loch Dhún Lúich’ go ciúin sa ghleann ina luí) (I mo dhiaidh bhí gleanntáin ghlas’…»
[Coke] am: say slurp.lines>>.words.perl
m: say slurp.lines>>.words.perl
camelia rakudo-moar a09c8d: OUTPUT«(("Céad", "slán", "ag", "sléibhte", "maorga", "Chontae", "Dhún", "na", "nGall").Seq, ("Agus", "dhá", "chéad", "slán", "ag", "an", "Eireagal", "ard", "ina", "stua", "os", "cionn", "caor", "is", "coll;").Seq, ("Nuair", "a", "ghluais", "mise", "thart",…»
[Coke] there you go
Herby__ sweet. thanks! 01:07
AlexDaniel m: say lines».words.perl 01:09
camelia rakudo-moar a09c8d: OUTPUT«(("Céad", "slán", "ag", "sléibhte", "maorga", "Chontae", "Dhún", "na", "nGall").Seq, ("Agus", "dhá", "chéad", "slán", "ag", "an", "Eireagal", "ard", "ina", "stua", "os", "cionn", "caor", "is", "coll;").Seq, ("Nuair", "a", "ghluais", "mise", "thart",…»
AlexDaniel Herby__: perhaps try github.com/tony-o/perl6-csv-parser or github.com/Tux/CSV ? 01:12
AlexDaniel Herby__: not sure if that is your case, but it sounds like you actually have a csv file… ;) 01:13
Herby__ I do :)
I tried CSV::Parser and my script was pretty slow. although its likely due to the algorithm i was using
dylanwh_ Feicim focail Ghaeilge... Tá mé iontas. 01:15
perlawhirl yeah, CSV parsing tends to be a little slow. one of my scripts i use to search for an entry does a `if $line.contains($query)` before doing a `$csv.parse($line)` which helps speed things up 01:16
CSV parsing is one of those things that's hard to do yourself. if you're absolutely positive all your cells are single-word entries with no commas, you could just do `$line.split(',')` 01:17
Herby__ luckily my file fits that criteria so thats how i ended up handling it. then i started wandering how to split a file into mutli-d array 01:20
perlawhirl another option i've been known to employ is open your CSV in Excel, go into `Control Panel > Regional & Language > Additional settings`, change `List separator` to something other than ","... something you know isn't in your file... even a weird char like "§", then save your "CSV" file in Excel 01:25
now you have a §-separated file :D
don't forget to change the regional settings back as soon as you're done
then you can more confidently do $line.split('§') 01:26
Herby__ nice idea, never thought about that 01:39
BenGoldberg Or you could install a C library for parsing csv, like libcsv, then use NativeCall. 01:52
Another option would be 'use Text::CSV_XS:from<Perl5>;' 01:56
Herby__ anyone heard any news on the Perl 6 book that is in the works? 03:26
TEttinger are the features of perl 6 stable enough for there to be a physical book? 03:35
it seems like stuff is still changing week-to-week
geekosaur the features are stable. bugs are being fixed, but still have to pass the spec tests 03:46
and yes, the vast majority of changes are bugfixes 03:49
geekosaur rarely you will see something added to 6.c-errata (bugs in the spec, more or less), even more rarely added to 6.d (the next release-to-be) 03:50
Herby__ speaking of bugs... my simple script failing is bugging me: pastebin.com/b1kDhuaf
i'm trying to download an image. i'm sure its something simple
but that fails with "Cannot resolve caller spurt(IO::Path: HTTP::Response, Str, :b); none of these signatures match:" 03:51
any ideas? 03:52
perlawhirl $img is still a HTTP::Response object, and spurt is expecting Blob or Cool 03:58
Herby__ just saw my mistake :) 03:59
spurt "testfile.png", $img.content does the trick
if I wanted to download all of the xkcd comics, would this be a good scenario to try to do in parallel? 04:00
i've never really messed with parallel/concurrent scripts 04:01
i realize thats a pretty generic question
perlawhirl Theoretically... yes. In practice, don't hammer Randall's server with too many requests :D 04:03
Herby__ yeah, i wouldnt go crazy and hammer the server non-stop. 04:04
been trying to think of a simple project to learn a little about concurrency and parallelism
perlawhirl also... just back to your script, the `:b` in `$user-agent.get(%content<img>)` is not neccessary 04:05
Herby__ just changed it, thanks 04:07
perlawhirl Herby__: one more thing 04:56
perlawhirl is at work and getting distracted
Herby__ do tell
perlawhirl you passed :b to spurt, but the actualy option is :bin. now in this case, it worked, because the multi that is invoked when you call spurt with a Blob enforces the :bin option... but just FYI 04:57
Herby__ hmm k 04:58
i need do some more reading on Blob, Any, Cool, Mu etc...
perlawhirl if you don't pass the right named argument it could cause headache's in the future... methods will generally ignore unknown named parameters
Herby__ cant say I have a great grasp on that stuff
perlawhirl well this is more a question of multiple-dispatch. given a particular function or method, different code may run depending on what args you provide 05:01
when you call the Function `spurt()`, it takes your args and passes them to the Method `.spurt()` 05:02
there are 2 `.spurt` methods, and they will ignore any unknown named parameter, like `:bin` or `:Herby`. one of those methods gets run when you passed binaray data (aka, a Blob), and a different `.spurt` method gets called for something... non-Blobby :) 05:03
Herby__ that makes sense 05:05
perlawhirl The Blobby spurt automatically turns on the `:bin` option for you... so it worked. Perl strives to DWIM (Do What I Mean)
Herby__ does it do the same thing for :a, :append?
perlawhirl no 05:06
the only reason it turned on `bin` is because that makes sense when passing binary data (Blob) to the method
it will just ignore ':a' because it doesn't have a named parameter called `:a` 05:07
ie, it will silently clobber your file... hence why i'm telling you all this 05:08
you need to be sure that you read the docs and pass the appropriate options
perlawhirl named parameters are not like command line args... close enough is not good enough :D 05:11
AlexDaniel m: say ‘hello world’ ~~ m:g/$<abc>=\w/ 05:31
camelia rakudo-moar a09c8d: OUTPUT«(｢h｣␤ abc => ｢h｣ ｢e｣␤ abc => ｢e｣ ｢l｣␤ abc => ｢l｣ ｢l｣␤ abc => ｢l｣ ｢o｣␤ abc => ｢o｣ ｢w｣␤ abc => ｢w｣ ｢o｣␤ abc => ｢o｣ ｢r｣␤ abc => ｢r｣ ｢l｣␤ abc => ｢l｣ ｢d｣␤ abc => ｢…»
AlexDaniel scratches his head 05:32
AlexDaniel m: say ‘hello world’ ~~ m:g/(\w)/ 05:33
camelia rakudo-moar a09c8d: OUTPUT«(｢h｣␤ 0 => ｢h｣ ｢e｣␤ 0 => ｢e｣ ｢l｣␤ 0 => ｢l｣ ｢l｣␤ 0 => ｢l｣ ｢o｣␤ 0 => ｢o｣ ｢w｣␤ 0 => ｢w｣ ｢o｣␤ 0 => ｢o｣ ｢r｣␤ 0 => ｢r｣ ｢l｣␤ 0 => ｢l｣ ｢d｣␤ 0 => ｢d｣)␤»
AlexDaniel I should be getting more sleep o_o 05:35
there is probably some good explanation, but the way I see it is… it's just crazy… 05:36
bombittable6: releases say ‘hello world’ ~~ m:g/(\w)/
committable6 AlexDaniel, ¦«2015.10,2015.11,2015.12,2016.02,2016.03,2016.04,2016.05,2016.06,2016.07.1,2016.08.1,HEAD»: (｢h｣␤ 0 => ｢h｣ ｢e｣␤ 0 => ｢e｣ ｢l｣␤ 0 => ｢l｣ ｢l｣␤ 0 => ｢l｣ ｢o｣␤ 0 => ｢o｣ ｢w｣␤ 0 => ｢w｣ ｢o｣␤ 0 => ｢o｣ ｢r｣␤ 0 => ｢r｣ ｢l｣␤ 0 => ｢l｣ ｢d｣␤ 0 => ｢d｣)
geekosaur did you mean \w+? 05:37
AlexDaniel no
AlexDaniel in fact, looks like it works fine. It's just that gist that I'm trying to make sense of 05:38
AlexDaniel bombittable6: releases put ‘hello world’ ~~ m:g/(\w)/ 05:39
committable6 AlexDaniel, ¦«2015.10»: ===SORRY!=== Error while compiling /tmp/Rnk6t7SOHK␤Undeclared routine:␤ put used at line 1␤ «exit code = 1»␤¦«2015.11,2015.12,2016.02,2016.03,2016.04,2016.05,2016.06,2016.07.1,2016.08.1,HEAD»: h e l l o w o r l d 05:40
AlexDaniel oops
but yeah
ShimmerFairy AlexDaniel: I think the gist is hurt by the fact that camelia can't output newlines where they should be, along with the fact that a space-separated list for a list of Match objects isn't the best idea. 05:41
geekosaur seems like it, yeh
AlexDaniel committable6: HEAD say ‘hello world’ ~~ m:g/(\w)/; .say for ^1000
committable6 AlexDaniel, gist.github.com/2d537f325d03368ef2...7473820e7c
geekosaur needs a reindenter for dd output, the dd version of that is incomprehensible as is
AlexDaniel ShimmerFairy: here is the output with newlines 05:42
AlexDaniel ShimmerFairy: and it's not any better. What is “0 => ｢h｣ ｢e｣” ? 05:42
ShimmerFairy AlexDaniel: yeah, that's an issue I've run into before with printing stuff involving Match objects; the gist of the list is doing space-separated values, but the Match objects have inherently multi-line gists 05:43
AlexDaniel ahhhhhhhhhhhhh
committable6: HEAD .say for ‘hello world’ ~~ m:g/(\w)/; .say for ^1000
committable6 AlexDaniel, gist.github.com/e8ce78af3f8921343b...a15d728be3
AlexDaniel aaaaaaaaaaaaahhhhhhhhh. 05:44
ShimmerFairy: thank you very much
ShimmerFairy np, it's definitely confusing when you're not aware of the issue :)
geekosaur right, that was why I was trying to clean the dd output
ShimmerFairy (and Match.gist can't return a string with a "\n" at the end because then that could mess up who-knows-what expecting not-\n-terminated gists, e.g. say $thing.gist) 05:45
I think the issue here is ultimately "we need a smarter gist perhaps" :P
geekosaur or at least Match needs a smarter .gist 05:46
ShimmerFairy I recall some parts of the spec speculating on various named options to .gist for controlling various ways the string is returned. I wonder if that idea would be useful for cases like "list .gist works best with elements' .gist following a certain pattern". 05:48
ufobat i've got a new computer and .. i am doing rakudobrew build zef for the first time.. and it fails :( 07:38
ufobat hm maybe i should take the 2016.09 instead of moar-nom 07:43
El_Che ufobat: what OS?
ufobat linux ubuntu 16.06
gist.github.com/ufobat/225f6f2af63...558159ff7f 07:44
El_Che github.com/nxadm/rakudo-pkg/releases
for in the mean time :)
ufobat yeah.. why not :-) 07:45
ufobat okay i am taking the deb package now.. 2016.09 also segfaults :/ 07:48
and it works 07:52
nope it segfaults too :/ it just takes longer
*sigh
ufobat any advice? 07:54
RabidGravy booM! 08:22
DrForr !Moob
El_Che ufobat_: segfaults? do you mean rakudo or when installing zef? 08:25
ufobat_ perl6 segfaults when i want to install zef 08:26
RabidGravy eugh 08:30
running it with perl6-gdb-m might help find where it is doing that 08:31
but it shouldn't segfault with no native call involved 08:32
DrForr Today's jam: www.youtube.com/watch?v=CuVd9oK16V0
El_Che ufobat_: installing
ufobat_ i have no c/debugging skills, does this help: gist.github.com/ufobat/73a23653a6f...b1304b74f6 08:34
stmuk_ ufobat_: see github.com/MoarVM/MoarVM/issues/410 and related tickets
turning off HLE on linux may help 08:35
github.com/tadzik/panda/issues/324...-250345236 08:38
ufobat_ stmuk_, i am not sure .. do i need a glibc without HLE or can i just compile a perl6 without HLE? 08:41
stmuk_ I think you need a glibc w/o it but I tried that myself 08:44
not sure if there is one in any distro repos either 08:45
rakudo 2016-06 may still work too )see github.com/MoarVM/MoarVM/issues/41...-249211127 08:47
ufobat_ that is probably the most easy way for me to get a perl6, i will try 08:50
stmuk_ actually if you could check 06 works and 07 doesn't it would help in confirming a possible connection between the BSD and Linux/HLE issue 08:53
ufobat_ stmuk_, i will build a 2016.07 and 2016.06 and check if there are differences 08:58
El_Che ugexe: ping me if around. The build-panda scripts runs fine: github.com/nxadm/rakudo-pkg/blob/m...as_user.sh . The build zef script (github.com/nxadm/rakudo-pkg/blob/m...s_user.sh) fails: gist.github.com/nxadm/6350f63e5bc9...c03bfaf43b . It tries to write to /opt/rakudo/share/perl6/site instead of ~/.perl6 (I even added a 09:00
Iinst#$HOME/.perl6 to be sure)
RabidGravy I'm having a hubristic moment - looking at ladonize.org/index.php/Main_About and thinking "how difficult could that be" 09:07
El_Che RabidGravy: yes, you can 09:31
:)
RabidGravy :) 09:37
RabidGravy but I don't like the way it multiplexes the different web service protocols 09:54
RabidGravy yeah, it should be able to use the Content-Type to distinguish between XML and JSON for example and then examine the content to refine the selection of protocol 10:01
*not* use different paths for each one
jnthn Indeed, using differnet paths for different representations of the same resource should be laid to REST... :P 10:02
RabidGravy I did an interview yesterday and had a strange moment when I realised they didn't quite "get" REST and had to explain like I might a junior 10:05
jnthn 'tis rather common 10:06
orangeblock why is [1e30, -1e30, 1, 3].sum == 4, but [1e30, 1, 3, -1e30].sum == 0? 10:07
psch m: say [1e30, -1e30, 1, 3].sum == 4; say [1e30, 1, 3, -1e30].sum == 0
camelia rakudo-moar a09c8d: OUTPUT«True␤True␤»
jnthn Floating point
e implies Num 10:08
orangeblock but they are the same numbers
just in different order
psch not in ieee 754
the order of operations matters
jnthn Yes, but floating point arithmetic doesn't commute
Thus why things like 2.5 default to Rat, not Num, in Perl 6. 10:09
orangeblock oh i see
so what would I have to do to make the right example print 4?
psch m: say [10 ** 30, 1, 3, -(10**30)].sum 10:10
camelia rakudo-moar a09c8d: OUTPUT«4␤»
psch . o O ( floating point lives near its work place, so it doesn't commute )
RabidGravy anyhoo, just off out to the supermarket, play nicely 10:11
orangeblock I get it now thanks a ton!
timotimo could this god damned HLE segfault bug please go die in a fire? 10:19
ufobat_ stmuk_, with 2016.07 it doesn't work but with 2016.06 it does not segfault anymore, hope that helps 10:37
timotimo ufobat_: does your system have a /lib64/noelision folder perhaps? 10:39
ufobat_ timotimo, nope 10:41
timotimo the debian bug that tracks the HLE problem claims things have been fixed on the first of january this year
timotimo so why hasn't this trickled down into ubuntu? 10:42
timotimo seems like on ubuntu a patch was suggested but rejected upstream and then no patch has been put in for ubuntu itself 10:48
timotimo should rakudo officially recommend against using ubuntu on newer intel cpus? 10:59
the distro wars live again! 11:00
DrForr Because of the sky...er, neural network instructions?
timotimo skylake, not skynet :)
moritz timotimo: you could add a note to the README, but usually nobody reads the README 11:01
timotimo we can after-the-fact point at the README when someone comes in with that crash
it'd be nicer if we could point out something an ubuntu user could just run to get rid of the bug
like, there must be a hundred different PPAs that give you glibc with hardware lock elision disabled
DrForr silently corrupts the FPU division tables while he's there. 11:02
timotimo DrForr: at least in that situation everybody will have seen that before ;) 11:03
though many new programmers may not have seen the FPU div bug in the wild in their lifetime
DrForr We are Pentium of Borg. Division is futile. You will be approximated.
El_Che Rakudo could supply its own libc 11:04
:)
en emacs and systemd
timotimo rakudobrew could output a warning about newer intel cpus and ubuntu before building zef, or if building zef failed 11:25
stmuk_ ufobat_: thanks 11:31
timotimo: I don't think its just the processor bug, the same (or similar) bug exists on all BSD systems
timotimo: maybe the HLE is triggering something else 11:32
stmuk_ is the string "HLE" present in /proc/cpuinfo on these systems? 12:02
ufobat_ on mine, yeah 12:05
> cat /proc/cpuinfo| grep -i hle
flags: fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush dts acpi mmx fxsr sse sse2 ss ht tm pbe syscall nx pdpe1gb rdtscp lm constant_tsc art arch_perfmon pebs bts rep_good nopl xtopology nonstop_tsc aperfmperf eagerfpu pni pclmulqdq dtes64 monitor ds_cpl vmx smx est tm2 ssse3 sdbg fma cx16 xtpr pdcm pcid sse4_1 sse4_2 x2apic movbe popcnt tsc_deadline_timer aes xsave avx f16c rdrand lahf_lm abm 3dnowprefetch ep
b intel_pt tpr_shadow vnmi flexpriority ept vpid fsgsbase tsc_adjust bmi1 hle avx2 smep bmi2 erms invpcid rtm mpx rdseed adx smap clflushopt xsaveopt xsavec xgetbv1 dtherm ida arat pln pts hwp hwp_notify hwp_act_window hwp_epp
stmuk_ ah good 12:07
El_Che github.com/nxadm/rakudo-pkg/releas....10Support 12:38
stmuk_ its the release this w/e isn't it? 12:41
El_Che it was release yesterday/today depending on your location 12:43
[Coke] pretty sure the rakudo compiler release is this weekend, yes 12:59
p3rln00b Yes
p3rln00b NeuralAnomaly: status 13:00
NeuralAnomaly p3rln00b, [✘] Next release is today. Since last release, there are 48 new still-open tickets (6 unreviewed and 0 blockers) and 37 unreviewed commits. See perl6.fail/release/stats for details
p3rln00b Well... the bot is off by +/- 24 hours :) 13:00
dalek osystem: 0034e08 | (Brian Duggan)++ | META.list:
add Log::Async
 13:03
osystem: 793b7a8 | (Zoffix Znet)++ | META.list:
Merge pull request #263 from bduggan/p6-log-async

add Log::Async
RabidGravy woo! 13:16
RabidGravy So it appears I have a new contract starting on Tuesday 13:34
what larks
DrForr That would be with a certain company at Westfield? 13:35
Xliff__ Guys, any word on what is causing this problem? 13:37
gist.github.com/Xliff/0f537c12b415...9484a25b1a
I've been fighting this one for months.
RabidGravy DrForr, indeed 13:39
ugexe El_Che: you still have to tell zef where to install to; `perl6 -Iinst#$HOME/.perl6 zef ...` is only telling perl6 where to look for libraries and tells zef nothing
`zef --install-to=inst#$HOME/.perl6 install ...` for example 13:40
you can also edit DefaultCUR in the config to set the default value for --install-to, along with having per-user configs from attempting to load from ~./zef/config.json (if it exists - its not created by default) 13:43
El_Che thx, I'll fix it right away
ugexe also for `inst#$HOME/.perl6` you can just use `home`, i.e `zef --install-to=home install ...` 13:44
ugexe which means you can edit 'DefaultCUR' to `home` before installing zef and then all users can use the default config 13:45
El_Che ugexe: it works. Recreating the dockerfiles in order to create new packages. Thx! 13:54
dalek c: 965ab76 | gfldex++ | doc/Language/typesystem.pod6:
show mixins on class and relationship to container type
 13:56
gfldex nice illustration of mixins: i.chzbgr.com/full/3461021696/h5F7479EB/ 13:58
ugexe clearly thats inheritance: table inherits from outboard motor 14:00
[Coke] doc's bin/p6doc is slow. :| 14:29
nicq20 Hello! 14:40
o/
gfldex [Coke]: do you care to profile?
dalek doc/spellcheck: 4d09760 | (Zoffix Znet)++ | doc/Type/Mix.pod6: 15:11
doc/spellcheck: Correct that Mix.Bag/BagHash .Ints not .rounds
doc/spellcheck:
doc/spellcheck: Per today's changes[^1], the coercion now .Ints the weights instead of .round'ing
doc/spellcheck:
p3rln00b killer! :) 15:12
[Coke] gfldex: not a yak I can shave at the moment. 15:13
[Coke] updates the perl6/doc spellchecker to spell check rendered text, not raw pod6 15:14
muuuch nicer.
(but slower)
dalek c: 45544de | coke++ | doc/Language/regexes.pod6:
fix typo
 15:20
FROGGS o/
moritz \o
[Coke] anyone want to help with spellchecking the docs? 15:22
down to 714 questionable words from aspell. 15:23
[ptc] cool! There's a perl6/doc spellchecker! 15:48
[Coke] in a branch, hopefully will be able to put it into master's xtest this weekend. 15:50
er, pending a moarvm bug fix.
dalek c: df49af3 | gfldex++ | doc/Language/typesystem.pod6:
comment output
 16:11
hankache m: say "Hello #perl6" 17:13
camelia rakudo-moar 8268ff: OUTPUT«Hello #perl6␤»
p3rln00b m: 'へっぉ、 はんかちぇ'.say 17:16
camelia rakudo-moar 8268ff: OUTPUT«へっぉ、 はんかちぇ␤»
harmil_wk p3rln00b: shouldn't that "n/l" be in Katakana? ;-) 17:27
p3rln00b shrugs
TimToady p3rln00b: that looks like "hetsuo" to me, not "hello" 18:24
p3rln00b aww 18:27
Google Translate isn't good :)
TimToady and if "hankache" is meant to correspond to the Japanese for "handkerchief", they should've translated it to ハンカチ :) 18:31
hankache hein? 18:32
oh
TimToady "hankachi" in Japanese 18:33
TimToady assumes the resemblance is likely accidental... 18:34
hankache :)
El_Che ugexe: Fixed the zef installer in the rakudo packages: github.com/nxadm/rakudo-pkg/releas...serInstall 18:36
Dunearhp I've been having trouble with 'HLE/RTM' associated segfaults and best I can tell, my processor shouldn't have the the HLE bug. 18:46
Dunearhp Found github.com/andikleen/tsx-tools.git 18:47
Dunearhp there is a ignore-xend.so LD_PRELOAD library in there that stops the segfaults when I try to compile panda 18:48
dalek ecs: 5627bda | (Zoffix Znet)++ | v6d.pod:
Remove Signature/Attribute Defaults Based on Definedness

Per discussion with pmchaud, this feature can be added to current rakudo, without needing a 6.d (as it doesn't interfere with anything), and should first be implemented as a user-space module for experimentation.
Dunearhp it suggests that the segfaults occur when an already unlocked mutex is unlocked 18:49
p3rln00b wweeeeeeeeeeeeeee 18:51
harmil_wk TimToady: some of the Katakana loanwords are based on Japanese pronunciations of foreign words, not just on the original loanword. The classic example is depato (デパート) which means "department store" 18:52
p3rln00b jesus 18:56
hankache que pasa? 18:59
p3rln00b The woodruffw person join-quitting 18:59
woddruffw haw 18:59
... break the cycle.
psch except oo vs od 19:00
timotimo not if you typo it :)
woddruffw Oh, ... I need to spell it right
psch waits with baited breath for woodruffw_ 19:00
woodruffw lol 19:00
psch huh 19:01
psch ...that's just silly 19:02
woodruffw Watch me. It'll work :) 19:02
FROGGS_ looks good so far 19:03
woodruffw :D 19:03
psch :P
FROGGS_ but no excess flood
hankache hah
woodruffw Yeah, I guess I fixed it.
psch there's probably some intricacy in IRC i don't know about that makes this a solution 19:04
FROGGS_ zoffn00b++ 19:04
p3rln00b :)
FROGGS_ does not know how to l33t 19:05
psch ...i am by now getting somewhat curious what exactly the bits of the IRC protocol that made this a fix are 19:06
harmil_wk FROGGS_: It's just a jump to the left...
p3rln00b psch: my wild guess is that the client was trying to send a crapton of something on reconnect, but because 2 nicks were taken, by the time it got to the 3rd nick, whatever it was flooding with on connect got ignored by the server because the user did not yet have a nick. 19:07
psch p3rln00b: oh, so the backbuffer for commands that get send before nick negotiation is finished ran over or something like that? 19:08
p3rln00b possibly. I've no idea why it was flooding in the first place :) 19:08
psch ("ran over" as in "server said: okay that's enough we didn't even agree who you are"
)
psch ...i don't think i should try and reproduce that on freenode 19:09
p3rln00b :}
geekosaur thinks this sounds like the vnc bug 19:13
er znc not vnc
geekosaur which I think is actually a connected client and znc both doing the same channel queries, and the fact that both are doing it can lead to that kind of mismatching 19:14
dalek c/spellcheck: 45544de | coke++ | doc/Language/regexes.pod6:
fix typo
 20:04
c/spellcheck: df49af3 | gfldex++ | doc/Language/typesystem.pod6:
comment output
c/spellcheck: 3cbf2ad | coke++ | doc/Language/ (2 files):
Merge branch 'master' into spellcheck
dalek c/spellcheck: de4deb2 | coke++ | xt/.aspell.pws:
more words
 20:20
dalek c: 3974b1c | coke++ | doc/Language/subscripts.pod6:
other instances used the US version
 20:21
travis-ci Doc build errored. Will "Coke" Coleda 'more words' 20:27
travis-ci.org/perl6/doc/builds/167747432 github.com/perl6/doc/compare/3cbf2...4deb2d2a45
smls Is `$proc.in.write` supposed to block when given a large Blob to write? 20:55
geekosaur it will block if the process is not reading, yes. pipes have finite buffer space 20:57
smls ok 20:57
problem is, that my own script is doing the reading
geekosaur whether this block is happening at OS level or if libuv is supposed to be using non-blocking operations and return a "would block, try again" and some higher level is blocking based on that, I cannot say
smls and that code is never reached because the write blocks 20:58
I based my code on the example on the `Proc` p6doc page, which does `.in.say`, then closes `.in`, then `.in.close`, then `.out.slurp-rest`. 20:59
But that only works if input sent in the first step is small enough.
s/then closes `.in`,//
geekosaur correct 21:01
this is the open3 problem, sounds like
geekosaur and has the same kind of solutions (usually, multiple threads) 21:01
smls Hm. Wrapping the `.in.write` and `.in.close` in a `start { }` (and awaiting the promise after the `.out.slurp-rest`), hangs indefinitely on the `.in.close`. 21:03
while using a CPU core at 100%, no less. 21:05
demo script: gist.github.com/smls/75518c746ca65...491971eac0 21:06
To be expected, or bug? 21:08
timotimo moar sometimes gets confused, i think 21:09
or maybe it's libuv 21:10
smls ok, will RT just to be sure 21:17
But for now, what is a working way to pipe a large Buf through an external command from Perl 6? 21:18
RabidGravy shouldn't that give an eroor as other handles across threads do? 21:30
RabidGravy Proc::Async would seem to be the puppy though 21:34
gfldex smls: you get the same behaviour with rt.perl.org/Public/Bug/Display.html?id=129787 22:25

Powered by Cro and the Raku® Programming Language.

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

Thank you!