»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.
00:02 nadim left 00:03 BenGoldberg joined 00:05 pmurias left, Sgeo left 00:10 Sgeo joined 00:12 raschipi joined 00:13 mcmillhj joined 00:16 Sgeo_ joined 00:17 dct left, mcmillhj left, Sgeo left 00:24 itaipu left 00:34 itaipu joined 00:41 itaipu left, BenGoldberg left 00:45 Rawriful left 00:48 raschipi left 00:52 ufobat left 00:56 Cabanossi left 00:57 Cabanossi joined 01:05 BenGoldberg joined 01:10 mcmillhj joined 01:14 mcmillhj left 01:24 profan left, APic left 01:25 profan joined 01:27 APic joined 01:39 mcmillhj joined 01:41 ash_gti left 01:44 mcmillhj left 01:46 ilbot3 left 01:51 ilbot3 joined, ChanServ sets mode: +v ilbot3 01:53 pierre_ joined 01:55 Cabanossi left 01:56 mcmillhj joined 01:57 Cabanossi joined 02:01 mcmillhj left 02:04 pierre_ left, tony-o joined 02:11 mcmillhj joined 02:15 mcmillhj left 02:17 pilne joined 02:27 zoll left 02:32 pierre_ joined 02:39 pharv_ joined, mcmillhj joined, pharv_ left 02:43 mcmillhj left 02:50 skids left 02:55 Cabanossi left 02:57 Cabanossi joined 02:58 noganex_ joined 02:59 pierre_ left 03:01 noganex left 03:02 k-man left 03:06 mcmillhj joined 03:08 jdv79 joined 03:10 Cabanossi left, mcmillhj left 03:11 Cabanoss- joined, Cabanoss- is now known as Cabanossi 03:15 cpage_ left 03:17 cdg joined, lookatme joined
lookatme good noon! 03:18
03:20 pierre_ joined 03:21 cdg left 03:25 mcmillhj joined 03:30 mcmillhj left 03:31 wigner joined 03:35 greppable6 joined, evalable6 joined, ChanServ sets mode: +v greppable6, ChanServ sets mode: +v evalable6, pilne left 03:38 wigner left 03:41 mcmillhj joined 03:46 mcmillhj left 03:54 Cabanossi left 03:56 Cabanossi joined 04:02 kevinb joined
kevinb i have a question about perl6 and "compilers", specifically, does perl6 have a "compiler" that when you're done you have an executable that can be "shipped" and no source code is available? (such as C/C++ does) 04:03
we have perl5 code and 1 of the things we don't care for is that by give our customers a "executable" we've also given away our source code 04:04
can perl6 help us?
mst that's why cpanel ship compiled perl5 code 04:05
kevinb my understanding was that the perl5 compilers were abandoned years ago, so cpanel is a current one that works? (i've never heard of it) 04:06
mst cpanel use one that apparently works for them
kevinb ah, i see, they are a company
mst I've not really looked into it, I've always preferred the radiator (radius server) philosophy of selling the source availability as a feature 04:07
kevinb for many projects, that's great and i do it to, but for a company there are preferences to keep the source code private :) 04:08
mst I've no idea what you mean by "for a company" since I was referring specifically to paid software 04:09
perhaps google what radiator is before making assumptions
BenGoldberg Generally, source code should be protected by a good license, not by hiding it. 04:12
kevinb all right, i've skimmed the article, their philosophy doesn't apply to us, if i understand it correctly, perhaps i need to give more info :)
geekosaur I think you can actually do that now, if you're willing to accept bytecode. installing a module compiles to bytecode for speed although it keeps the source around in case something needs it --- so make most of your app modules that can be precompiled (that is, don't use e.g. "augment" --- this is fairly easy, you need to specifically enable avoidance of precompilation) and leave out the source
mst there's similar tricks with perl5 as well
kevinb yes, bytecode would be acceptable, it's just far enough away from source 04:13
mst the question, really, is, do you really not trust your users, or are you just trying to get management to think tey don't need to trust your users, or somewhere in between?
04:13 AlexDaniel left
mst like, if being able to observe the behaviour of the prorgam is a problem, you probably have to move that part off the customer's machine to have any degree of certainty 04:14
kevinb when we deliver our product, it goes on to a server at the client's site, while we retain root control, no system is unhackable and we'd strongly prefer they didn't have access to our code, so they don't take it and resell it (as some people do, contracts can be ignored and damage is done)
we trust most of our customers, but there are a few markets (in other countries) where the trust is no so strong 04:15
mst the question here always becomes "did anybody figure out how often that's happened, and how many of the people who bought the hacked version would -really- have bought your code otherwise?" 04:16
then measure that against the engineering effort required to reduce that number
kevinb @geekosaur, what would i search for to read about compiling to bytecode
my searching so far hasn't yielded much that's been helpful 04:17
geekosaur er? this is done on installation, automatically (usually called precompilation or "precomp")
mst I mean metacpan.org/pod/ByteLoader exists as well
04:17 pierre_ left, BenGoldberg left
geekosaur you have to specifically disable it to avoid it ("no precompilation;" at the top of a module) 04:18
mst yeah, blowing away the non-bytecode bits from a CURLI would probably, in theory, also work fine
kevinb ok, ByteLoader gives me something to pursue, thanks! 04:19
mst terrible ideas: completely feasible in perl5 *and* perl6
kevinb "blowing away the non-bytecode bits from a CURLI", can you explain a little more? a quick search turned up nothing 04:20
mst CURLI is CompUnitRepository::Local::Installation 04:22
if you want to experiment with this, then your approach would be 'install perl6, install some modules, go look at the resulting files' 04:23
but if your codebase is currently in perl5, then perhaps look at the perl5 solutions to your problem rather than learning a new programming language first :)
kevinb: also, I'm still confused how you didn't find the perl5 bytecode stuff 04:24
like, my process was 'go to metacpan.org, type "bytecode" into search box, click search, click first link'
kevinb let's call it old info that created pre-conceived notions perl5 couldn't do it, that perlcc had been abandoned and noone had ever picked up task afterward 04:26
i know, shame on me for assuming!
but i thought i had read that perl6 would go to bytecode, hence why i'm asking here
i also thought i'd read that perl6 could run perl5 with a pragma, or that was my chain of logic, good or bad 04:27
04:28 itaipu joined
mst currently the standard approach for perl6 to run perl5 code is Inline::Perl5, which embeds a copy of the perl5 VM 04:28
04:28 mcmillhj joined
kevinb ok, as long as it works i really don't care about under the covers :) 04:29
mst the pragma you mention is theoretically possible, but I don't believe at the top of anybody's list
kevinb ok, so the pragma is on the roadmap, but unimplemented so far; that's ok, i can look at Inline::Perl5
mst yeah, but at that point your perl5 code is still executed by perl5, and you're still looking at ByteLoader and B::ByteCode to obfuscate your perl5 04:30
kevinb i'll save off this conversation and use it to research various things
mst I don't really understand what perl6 gains you here unless you're going to rewrite your codebase first
(I mean, that'd probably be fun to do ... but ...)
kevinb since i had bad info, the hope was it could help me where perl5 couldn't, but i'll abandon that line of thinking 04:31
mst maybe consider asking a perl5 channel first when you have a perl5 problem? I mean ... just a thought :P
kevinb we've got something like 75K lines of perl5 (not all code some of that is POD), so we're not going to abandon that to goto java/C++/whatever 04:32
mst (mostly because most of what I've said in here has been basically noise to this channel, and I'd've rather had it go where the perl5 people hang out)
kevinb haha, sure
04:32 itaipu left
kevinb i goofed, i'll correct that. thanks for the pointers! 04:32
mst no worries
play with perl6 sometime anyway, though. I still mostly write perl5 but it's a different sort of fun :D 04:33
04:33 mcmillhj left
kevinb i'll need to reread the docs, it's been several years since i looked; parts of it look interesting, parts make me cringe, but to be fair i'd need to go read what is current before i make a real judgement 04:34
mst I can best be described at the moment as "a perl5 programmer who occasionally noodles in perl6 because it's fun to play with" 04:35
kevinb "cringe" is a little strong and i apologize, how about make me unsure because it's so different
mst yeah, I totally had the uncanny valley effect for years
conceptualising it as "two languages, same family" and approaching it like learning a new lisp helped
kevinb yeah, i really wished Larry had called it something else, like "Oyster" :) related but different 04:36
sorry, probably a bad joke
mst more "a topic which has been beaten to death in both communities and I really don't fancy rehashing yet again" ;)
kevinb haha, sure, i've actually never participated in one of those discussions, but i really don't need to either 04:37
04:37 xtreak joined
kevinb anyway, i have to run, it's the very end of the day for me, thank you @mst again for your help! 04:37
mst aight. later. 04:38
04:39 itaipu joined, kevinb left 04:44 cgfbee left 04:46 cpage_ joined 04:48 pierre_ joined 04:53 itaipu left, cgfbee joined 04:56 zakame joined 04:58 mcmillhj joined, eythian left, eythian_ joined 05:02 mcmillhj left 05:03 pierre_ left 05:09 wigner joined 05:17 mcmillhj joined, nadim joined, cdg joined 05:21 wigner left, mcmillhj left 05:23 cdg left 05:26 wigner joined 05:27 piojo joined 05:29 mcmillhj joined 05:30 cdg joined, wamba joined 05:33 cdg_ joined, cdg left, mcmillhj left 05:35 wigner left 05:36 andrzejku joined 05:37 gigavinyl joined, cdg_ left 05:38 gigavinyl left 05:39 Cabanossi left 05:40 Cabanossi joined
andrzejku hi people 05:41
I need help with understanding some perl6 things
:P
mst don't we all 05:44
andrzejku :) 05:45
05:48 mcmillhj joined 05:52 mcmillhj left 06:04 espadrine_ joined 06:05 xtreak left 06:11 ufobat joined 06:13 mcmillhj joined
lookatme :) 06:15
06:16 pierre_ joined 06:18 mcmillhj left 06:19 gigavinyl joined 06:20 gigavinyl left 06:27 lowbro joined, lowbro left, lowbro joined 06:32 ShalokShalom_ joined 06:35 ShalokShalom left, darutoko joined 06:36 espadrine_ left, ChoHag left 06:41 lookatme left 06:42 lookatme joined, mcmillhj joined 06:44 lookatme left, lookatme joined 06:47 mcmillhj left 06:48 wigner joined 06:49 nadim left 06:52 wigner left 06:58 mcmillhj joined 07:01 nattefrost joined
moritz \o 07:01
yoleaux 17 Aug 2017 22:15Z <b2gills> moritz: I wrote a Grammar for a made up language for a code golf once, I don't know if it is helpful or not for what you are doing stackoverflow.com/a/41770672/1337
07:02 mcmillhj left
lookatme :( 07:08
07:09 mcmillhj joined, Cabanossi left
ufobat noganex_, ! 07:10
07:11 Cabanossi joined 07:12 eroux left 07:13 mcmillhj left 07:16 rindolf joined, zakame left 07:17 pierre_ left, pierre_ joined 07:18 ChoHag joined 07:22 eythian_ is now known as eythian 07:24 mcmillhj joined
lookatme andrzejku, what's up ? 07:27
07:28 mcmillhj left 07:30 pierre_ left 07:31 pierre_ joined 07:36 mcmillhj joined 07:37 leah2 left
andrzejku priv ;d 07:37
07:39 Cabanossi left 07:41 mcmillhj left, Cabanossi joined 07:45 leah2 joined 07:48 nadim joined, cdg joined 07:50 ChoHag left 07:53 cdg left 07:54 piss joined 07:56 pierre_ left, zakharyas joined, mcmillhj joined, pierre_ joined 08:00 mcmillhj left
lookatme andrzejku, seems like I can not send private message, it complaint: You must log in with services to message this user 08:09
Geth Inline-Perl5: f66e9b4734 | (Stefan Seifert)++ | lib/Inline/Perl5.pm6
No need to make $type optional in unpack_return_values

All callers supply it and optional int params are not JITed yet. So removing the default value will speed up this method a bit.
lookatme :/
andrzejku lookatme, hah lol :D
08:12 mcmillhj joined 08:13 piss left 08:14 dakkar joined, piss joined 08:17 mcmillhj left 08:18 domidumont joined 08:19 robertle joined 08:22 domidumont left 08:23 domidumont joined 08:24 Celelibi joined 08:27 mcmillhj joined 08:29 abraxxa joined 08:32 ChoHag joined, mcmillhj left 08:39 Cabanossi left 08:40 pierre_ left, pierre_ joined, lijinfeng joined, learner_of_p6 joined 08:41 Cabanossi joined, mcmillhj joined 08:46 cdg joined, mcmillhj left
andrzejku I don't understand docs.perl6.org/language/grammars#Protoregexes 08:47
how this contruction works :sym
08:48 zakharyas left, nadim left, nadim_ joined
learner_of_p6 p6: say 'word'; 08:49
camelia word
08:50 zakharyas joined, cdg left 08:51 Ven joined 08:52 mcmillhj joined, Ven is now known as Guest37379
moritz andrzejku: in a proto regex, you have to give the individual candidates some kind of identifier 08:52
andrzejku: and you do that by adding a :sym<something> to their name 08:53
and inside the candidate, you can use <sym> to match 'something', just to avoid repetition. But you don't have to
andrzejku moritz, this part I know a bit but how can I costruct something like :sym 08:54
in my code
08:54 pierre_ left
moritz you just write it 08:54
08:55 ShalokShalom_ is now known as ShalokShalom
andrzejku for example I decided to code something where I can configure more things using :fuck 08:55
08:55 pierre_ joined
moritz oh, you want to extend Perl 6 to add more stuff 08:55
I think you have to write a slang to do that, and I have no idea how that works
andrzejku moritz, okay
moritz it's easier to add your own traits 08:56
andrzejku moritz, maybe it is too early to ask for that
moritz stuff like "is rw"
08:56 learner_of_p6 left, pierre_ left 08:57 mcmillhj left, pierre_ joined 09:07 mcmillhj joined 09:11 piojo left 09:14 mcmillhj left, cdg joined 09:16 cdg_ joined 09:18 cdg__ joined 09:19 cdg left 09:21 cdg_ left 09:22 szabgab joined, cdg__ left
szabgab Hi, I was looking for the smoke test results of Rakudo itself 09:23
and how to set up Rakudo smoking
perl6.org/community/ has two entries related to smoke testing: "Perl 6 Testers" and "Smoke Testing" but I got a bit confused there. 09:24
They both seem to talk about testing the Perl 6 modules in a bit redundant way and one of them links to smoke.perl6.org/ which seems to be dead 09:25
09:28 TEttinger left
stmuk redundant is the perl way :) 09:39
testers probably hasn't worked for 3 months according to testers.perl6.org/recent.html 09:40
zoffix's toaster system ran OK last month but is being worked on currently
szabgab is toaseter testing Rakudo or the Perl 6 modules ? 09:44
*toaster
stmuk both links you refer to smoke test modules 09:46
since failures are either due to rakudo or the modules themselves I'd say toaster is testing both
I think the main motivation of toaster is to check for rakudo problems not caught by roast failures 09:48
09:49 lookatme left, Aaronepower left 09:51 cdg joined, leont joined
stmuk maybe you want travis-ci.org/rakudo/rakudo 09:53
09:53 Guest37379 left 09:56 cdg left
szabgab Travis an Appveyor are good, but I thought there are individuals testing Rakudo on their own machines with other OS-es and lots of differnt configurations. 09:56
10:03 astj_ left 10:04 astj joined 10:05 samb1 left
daxim how long does it take for a rakudobug email to show up in rt? 10:06
10:06 samb1 joined
stmuk szabgab: yes that's probably true 10:07
10:07 wamba left 10:08 astj left 10:10 pierre_ left 10:14 mr-fooba_ joined, sena_kun joined 10:16 mr-foobar left, cdg joined 10:17 Rawriful joined 10:19 szabgab left 10:21 cdg left 10:22 Aaronepower joined 10:23 mr_ron joined 10:24 noganex joined 10:25 noganex_ left
mr_ron docs.perl6.org has not updated since August 12 and there have been more than half a dozen commits to doc since. Because of some git confusion and bad clocking on my system, the first commit after August 11 appears to be mine but I cloned the latest repo and "make html" ran without problems. 10:26
10:31 piss left 10:33 wamba joined 10:38 Cabanossi left 10:40 gregf_ joined 10:41 Cabanossi joined 10:44 Exodist joined
leont daxim: it depends, rt.perl.org has a weird spam-filter mode. 10:49
yoleaux 19 Jul 2017 21:09Z <tony-o> leont: - i'm looking into it but on moar-blead i'm getting 'Iteration past end of grapheme iterator' on all tests for YAMLish
leont seems to have fixed subtest/yaml parsing in TAP::Harness o/
10:51 xtreak joined 11:06 lijinfeng left 11:07 zakharyas left 11:10 IAASIWV joined 11:15 Ulti joined 11:17 araraloren joined
Ulti any moar developments on this: stackoverflow.com/questions/392050...l-6-script 11:21
11:24 mr_ron left 11:26 sena_kun left 11:28 mr_ron joined, sena_kun joined 11:32 sena_kun left 11:33 xtreak left, xtreak joined 11:35 xtreak left, xtreak joined
ryu0 Ulti: don't think i've seen any "more" developments there. 11:42
11:48 IndexMe joined
IndexMe Hello, with lastest Rakudo Star I get zef build error when installing Inline::Python. I'm used to building perl5 and all my modules from scratch in linux but on Mac OS X and Rakudo Star, not so sure how I should go about it. Star gives very cryptic build errors like Applications/Rakudo/share/perl6/site/sources/3393EDA469A9E8925A633FF7A533AB41141495DA 11:51
Sorry the error is Building [FAIL]: Inline::Python:ver('0.3'):auth('github:niner') Aborting due to build failure: Inline::Python:ver('0.3'):auth('github:niner') (use --force-build to override) in code at /Applications/Rakudo/share/perl6/site/sources/3393EDA469A9E8925A633FF7A533AB41141495DA (Zef::Client) line 329 etc 11:52
moritz IndexMe: Inline::Python needs the python header files installed
11:53 Cabanossi left
IndexMe I see, thanks. 11:53
11:54 raschipi joined
araraloren evening 11:54
raschipi evening
IndexMe What is the recommended way to set up p6 on Mac OS X? Difficult to find the libraries and follow the build.. 11:55
11:55 Cabanossi joined
Ulti homebrew isnt a bad shout 11:58
11:58 abraxxa left
IndexMe Thanks Ulti, Moritz ! 11:58
Ulti but things like Inline::Python its hardly a normal library thats super straight forwards
12:01 mr_ron left 12:02 mr_ron joined, xtreak left 12:03 xtreak joined 12:04 xtreak left
sjn o/ 12:08
raschipi heyo
sjn quick one-liner question; is there a convenient way to get the first unused argument? (e.g. equivalent to «perl -E 'say shift' hello») 12:09
hahainternet sjn: what counts as 'unused'? 12:10
sjn in my example, both '-E' and 'say shift' are used by perl 12:11
hahainternet hmm, from my playing it looks like you'd have to change the MAIN proto or similar, so i'm honestly not sure 12:12
sjn the first one that is unused (and hence in @ARGV) is 'hello'
hahainternet sjn: perl scripts use MAIN for this sorta thing, and so i'm not sure what the semantics of -e are
perl6 scripts*
wish i could help :/ 12:13
sjn perl6 -e 'my @b=(1,1,*+*...*); say @b[@*ARGS.shift]' 12:14
so @*ARGS.shift 12:15
raschipi perl6 -e 'say shift @*ARGS' hi
hahainternet facepalms, doh
sjn would it be sensible if shift by default took from @*ARGS? 12:16
raschipi sjn: we have other mechanisms for dealing with @*ARGS by default
sjn raschipi: do tell :)
raschipi sjn: Have a look: perl6maven.com/parsing-command-line...ents-perl6 12:19
tojo hm. i can't override default value with "has" in child class see these pastebin.com/HASw8Tzm versus pastebin.com/5rUjKpE5 and the first one is what i want, but is it the correct way to do this? :) 12:24
12:30 zakharyas joined
raschipi tojo: docs.perl6.org/language/classtut#P...te_Methods 12:34
IndexMe @Ulti @moritz: I wedged it in but it was strange. I ran zef look Inline::Python, ran perl6 configure.pl and make test (it found no make install), and now zef install Inline::Python worked. Perhaps zef had trouble finding the python headers which were indeed already present. 12:35
12:38 Zoffix joined
marcusramberg 112, 12:40
Zoffix .tell szabgab RE smoking Rakudo itself. There's no real setup yet. There's Toaster ( rakudo.party/post/Perl-6-Release-Q...em-Toaster ) that's currently being run on just Debian. Eventually that'll expand to more OSes. Of course, anyone can do `make stresstest` in a Rakudo checkout to stresstest it on their own systems. 12:42
yoleaux Zoffix: I'll pass your message to szabgab.
tojo raschipi: i don't see how private methods helps. i'm trying to change default value of parent attribute from child, but i don't understand why i can't use "has" to redefine the default value 12:43
araraloren tojo, `my` is used for declaration of class static member
12:44 pierre_ joined
raschipi tojo: "To call a private method of another class the calling class has to be trusted by the called class. A trust relationship is declared with trusts and the class to be trusted must already be declared." 12:44
hahainternet here's a question, if i constrain the input of MAIN to an Int, unicode characters like roman numerals fail to work, but .WHAT on them returns Int 12:45
is there some processing going on in the repl that isn't done for MAIN or similar? 12:46
jnthn m: class A { has $.x = self.x-default; method x-default() { 1 } }; class B is A { method x-default() { 2 } }; say A.new; say B.new
camelia A.new(x => 1)
B.new(x => 2)
raschipi hahainternet: It appears to be so, not even fractions are being recognized here. 12:49
Zoffix hahainternet: you're confusing passing parameters with typing them directly in code. A roman numeral typed directly in code is an Int, because it compiles to that. sub MAIN just gets them as strings, like allomorphs
hahainternet Zoffix: the problem is that it works for latin numerals
so somewhere there's a distinction that perhaps might not be consistent
i don't want to proscribe :p
Zoffix hahainternet: because those are allomorphic 12:50
m: dd <Ⅱ >; dd <42 > 12:51
camelia "Ⅱ"
IntStr.new(42, "42")
12:51 setty1 joined
Zoffix m: "4Ⅱ".uniprops.say 12:51
camelia (Nd Nl)
tojo jnthn: thx! nice :) 12:52
hahainternet Zoffix: sorry i have to run unfortunately, but is there a correct way to constrain MAIN? IntStr? 12:53
Zoffix hahainternet: to take roman numerals? That'd probably be sub MAIN($foo where .unival.isNaN.not) 12:54
12:54 Cabanossi left, mcmillhj joined
Zoffix hahainternet: note that $foo would still have just a Str inside the sub's body; you'd need to do unival() it to get the actual numeric value 12:55
12:55 Cabanossi joined, Guest19948 joined 12:58 araraloren_ joined
tojo araraloren: aa.. never thought "my" like that inside class definition, seems that "our" works like same inside there 12:59
Geth doc: 20fdfa97e7 | (Zoffix Znet)++ (committed using GitHub Web editor) | util/trigger-rebuild.txt
debug commit
doc: 66dafe1757 | (Zoffix Znet)++ (committed using GitHub Web editor) | util/trigger-rebuild.txt
debug 2 commit
 13:00
perlpilot tojo: docs.perl6.org/language/classtut#Static_fields? 13:01
13:01 araraloren left
perlpilot tojo: granted, the docs could use a few more words on the subject. 13:02
tojo perlpilot: :D i have some how managed to skip that
13:04 lucasb joined
Zoffix from doc CONTRIBUTING.md: "highlights (optional; requires only nodejs and at least GCC-4.8 on Linux to be installed. Running make will set everything up for you.)" 13:07
"set everything up for you" what does that mean? Will install all the deps like npm?
"/bin/sh: 4: npm: not found" guess not
"nodejs is already the newest version (4.8.2~dfsg-1)" 13:08
samcv: are you around?
Geth perl6.org: tbrowder++ created pull request #82:
clarify quick install instructions		 13:13
13:13 Guest19948 left
tbrowder hi #perl6 13:15
found some glitches trying to install rakudo star, hence the PR for some more eyeballs. the issue is the continuing one of the cyclic dependencies between nqp and moar and previous installations. 13:16
Zoffix tbrowder: yeah, well, your new code will nuke user's system. 13:18
$PREFIX is unset and you're running rm -fr $PREFIX/bin as root.
moritz (moar eyeballs)++
tbrowder how so? obviously one should define $PREFIX, but it's easy to add some words. 13:19
b2gills andrzejku: I think I made a good case for proto regexes here: stackoverflow.com/a/41770672/1337
sjn: having `shift` work on @ARGV by default is a special case in Perl 5. One of the guiding principles of Perl 6 is to remove special cases
tbrowder moritz: you need to go on stage!
Zoffix tbrowder: I'd rather we not have anything of the sort in the instructions. As the saying goes, you can put all the words, but you can't make the users read them. 13:20
tbrowder well it needs something there: a warning about reusing an installation dir or something like that. the detailed instructions also don't talk about the problem, either. 13:21
Zoffix Why isn't that in the installer itself?
mspo e-using an install dir? :) 13:22
13:22 IndexMe left
tbrowder because I haven't been able to complete my work on the base installer, and it's a snake pit of dependencies. 13:22
Zoffix It'd say that's an issue in the installer, not the "Quick Start" download page. 13:23
Kinda LTA the current instructions require root too. 13:24
tbrowder I say it's both and the quick start is the place to "start" (bow to moritz)!
Zoffix I'd install into ~/.rakudo or something
13:24 rindolf left
tbrowder i'm looking at it from a system-wide view, not a local user. 13:24
13:25 Guest19948 joined
moritz then please use /opt/rakudo/ or something like that 13:25
Zoffix I'm looking at it from a point of view of a user who wants to try out Rakudo and is presented with a wall of text that requires root and knowledge of how to set env vars, just to debug some obscure issue.
moritz a location that isn't controlled by a package manager 13:26
tbrowder that's fine, but that's a sys admin's choice. the point is to use the same dir so paths don't have to be updated, etc. 13:27
raschipi Trying to put things on directories controlled by the package-manager is asking for a work of hurt. 13:28
tbrowder the fact that zef uses the same directory also complicates the issue.
Zoffix .tell samcv would you be able to unbust the doc build. After hack's upgrade it's missing some sort of npm bits and I can't figure out which: github.com/perl6/doc/issues/1457#i...-251240775 13:29
yoleaux Zoffix: I'll pass your message to samcv.
tbrowder raschipi: not arguing, opt is fine.
Zoffix mr_ron: ^ it'll get unbusted once that's fixed.
13:29 Zoffix left
raschipi either /op or /usr/local 13:29
tbrowder right. at the moment i'm using /usr/local/rakudo.d on my remote server. 13:30
Zoffix: a user has to know how to set paths, doesn't he (or she)? 13:32
raschipi If it separates it's files by type like a proper Unix program should, it goes in /usr/local/*
If it wants to have everything in a sigle directory, it goes in /opt/rakudo 13:33
tbrowder Zoffix: your quick start instructions already require root!
mspo I remember a while ago it was difficult to say "use this nqp" during install
I haven't checked recently 13:34
so the order of operations was very strict and you had to have PATH setup just-so
tbrowder raschipi: i expect a package installer to do that.
raschipi right, the package manager installs in /usr/* 13:35
b2gills I use /opt/rakudo/, and have symlinks to the executables in /home/brad/bin/
raschipi Those are just for admin installed packages.
13:36 Aaronepower left
raschipi Of course, after installing in /opt, one needs to add the proper synlinks somewhere in path, like b2gills has, but I recommend /usr/local/bin 13:37
13:37 MasterDuke left
raschipi Only debian and it's derivatives adds $HOME/bin to the path by default. 13:37
tbrowder i am the admin, and i install and set paths to where i want. again the point is rakudo star cannot configure into an existing directory without intervention.
raschipi Yeah, of course the admin is king. I'm just saying it will have consequences. 13:38
tbrowder back to the original problem: a simple warning for new users! 13:39
raschipi Debian has a system called 'diversions' to tell the package manager to let a file alone.
tbrowder and old user's...
tyil[m] My *shrc set the path from scratch so I can be sure shits working as expected wherever I am
13:40 andrzejku left 13:42 skids joined 13:47 jast left 13:50 jast joined 13:52 Kyo91_ joined 13:53 rindolf joined 14:04 cdg joined 14:05 pierre_ left 14:09 Cabanossi left 14:10 Cabanossi joined 14:11 itaylor57 joined
[Coke] can someone reset my password on www.p6c.org ? 14:14
14:15 Guest19948 left
[Coke] wonder if it would help to have rakudobrew build understand moar-nom as a build target as opposed to just "moar nom" 14:23
ugexe the -blead exists because there is no blead tag to target with `moar blead` 14:28
14:32 jonathon joined
jonathon hi! i'm getting a build failure using `zef install Linenoise`, due to development packages I don't have installed. I need to know if these are missing dependencies of (e.g. moarvm or rakudo) packages, or whether they shouldn't be necessary at all: bpaste.net/show/3b4bed826191 14:35
14:36 Aaronepower joined
jonathon e.g. is gcc incorrectly picking up the options moarvm was compiled, or should these deps be installed along with moarvm? 14:36
14:37 Guest19948 joined 14:38 leont left, leont joined
Geth infrastructure-doc: ac5d0f39d1 | (Zoffix Znet)++ (committed using GitHub Web editor) | hosts/www.p6c.org.pod
update user list on www.p6c.org
 14:43
14:48 rindolf left 14:49 jonathon left 14:53 araraloren_ left 14:54 araraloren joined, araraloren left, araraloren joined 15:00 AlexDaniel joined 15:01 araraloren_ joined 15:02 Skarsnik joined 15:04 araraloren left, rindolf joined 15:06 nadim_ left 15:08 araraloren_ left
Skarsnik hello 15:10
yoleaux 17 Aug 2017 21:47Z <Zoffix> Skarsnik: RE wtf are these VM with 100GB.... It's Google Compute Engine. 100GB is nothing. Here's my current maximum: i.imgur.com/ceRzOZm.png And I've seen news they plan to expand it to 128-core VMs (presumably with even more RAM) some time soonish
Skarsnik x)
15:16 wamba left
mspo how do you have moar without libuv? 15:18
Skarsnik I think there is a fork/branch without libuv? 15:19
jnthn You can't
Skarsnik: No, we just stopped using it for the sync I/O bits
mspo I was trying to respond to jonathon but it looks like he left 15:20
jnthn Ah :)
mspo his error is linenoise trying to compile but not finding luv, latomicops, and one other thing
which are all moar dependencies I think
geekosaur I see a -Wl,-rpath but no matching -L
jnthn Hm, yeah
mspo bpaste.net/show/3b4bed826191 15:21
geekosaur: good job spotting that
jnthn Maybe that can happen if moar is installed from a package?
And to compile libs needs -dev packages or some such 15:22
I guess the linenoise build is stealing the C flags from the Moar build
Skarsnik is atom (the code editor) lightweight?
mspo you guys should just ship with bsd-libedit and call it a day :) 15:23
Skarsnik: is that a joke or a troll?
15:23 nadim_ joined
Skarsnik It's a real question 15:23
geekosaur let's put it this way: emacs in javascript
(coffeescript, but same difference)
Skarsnik I work on VM with like 1-1.5GB memory 15:24
that why I ask
mspo might not be a great choice
Skarsnik (and p6 coloration for kate is not really updated)
15:25 lowbro left
raschipi And it offers "Find and replace" as one of it's top-6 features! 15:25
15:26 Guest19948 left
mspo Skarsnik: Atom is the flagship example for "bloated electron app" 15:26
Did you ever use ActiveState Komodo? It's about that snappy 15:27
VSCode is similarly RAM hungry but has a much better "feel"
tl;dr use vim
raschipi Skarsnik: It embedds a copy of Chromium... 15:28
Skarsnik I don't like much Vim, too much to configure x
and I found it slow on like over 200 lines of perl 6 15:29
leont What is the correct way of devnulling the stderr on a Proc::Async 15:30
jnthn .bind-stderr(open("/dev/null", :r)) should work 15:32
15:33 zakharyas left
mspo I've always wanted to build an editor that could handle many GB-size files 15:33
15:34 eliasr joined
leont Well, that's not exactly portable 15:34
Skarsnik close stderr ?
jnthn .stderr.tap; # throws it all away :) 15:35
leont Yeah that seems to work :-)
Wasn't necessary at some previous point though
Just called .stderr and program as silenced 15:36
jnthn Oh
Yeah, that was an untested behavior, and also made it impossible to do stuff like $p2.bind-stdin($p1.stderr)
raschipi jnthn: Do you know if it will do it efficiently because of the sink context or will it copy it somewhere and then just ignore it until GC comes around?
jnthn It'll end up stuck into buffers. 15:37
Ideally we'd have a .ignore-stdout and .ignore-stderr
leont s/as/was/
mspo ignore is different from close
15:38 sunnavy joined
geekosaur close stands a good chance of crashing the program if it does try to use stderr 15:38
leont devnull would be more efficient than this, but at least it's correct
jnthn++
jnthn .stderr(:bin).tap will be a bit cheaper fwiw
geekosaur there's an assumption that the standard handles are always open
jnthn Yeah, I think .ignore-stdout would just throw it away further up the pipeline and be a bit more intentful in the code 15:39
ilmari leont: according to perl5's t/test.pl «/dev/null appears to be surprisingly portable», and only windows requires "nul" instead 15:40
does rakudo have anything equivalent to perl5's File::Specl->devnull? 15:41
leont If so, in IO::Spec
Apparently so :-) 15:42
$*SPEC.devnull
ilmari src/core/IO/Spec/Win32.pm:23: method devnull { 'nul' } 15:43
src/core/IO/Spec/Unix.pm:65: method devnull { '/dev/null' }
b2gills mspo: to ignore the STDERR from a Proc::Async, just add :!err to the call to .new
leont AFAIK !err leaves it untouched, so the user will see it
jnthn Plus I think that's a flag to run/shell, not to Proc::Async :) 15:45
leont Yeah, I was wondering "this isn't documented"
ilmari docs.perl6.org/routine/devnull 15:46
15:47 Sgeo_ left
leont Using the spectest to test TAP::Harness now, I'm grateful you guys wrote a test-suite for me :-p 15:48
Test.pm is terrible :-( 15:50
15:50 alimon left
leont paste.scsys.co.uk/564908 15:50
Apparently it messes up comments in subtests 15:51
15:52 robertle left
leont Or better said, multi line descriptions 15:53
prove5 ignores any such thing, so doesn't complain about it
15:54 alimon joined, Cabanossi left 15:55 Cabanossi joined 15:56 MilkmanDan left 15:58 MilkmanDan joined
tojo i'm getting this error time to time "WARNING: unhandled Failure detected in DESTROY. If you meant to ignore it, you can mark it as handled by calling..." what it means? can i silence it somehow? 15:58
15:59 st_elmo joined, mr_ron left
raschipi tojo: What rakudo version do you have? 15:59
geekosaur a delayed exception made it through to garbage collection without being looked at (that is, you did some operation that produced a Failure/delayed exception, then never looked at the result)
tojo raschipi: this one now 2017.04.3 16:00
raschipi geekosaur: Wouldn't that mean that he sunk the failure and that should turn in in an exception? Blowing up on the GC isn't expected. 16:02
geekosaur that's why it produces that message
because it's not expected. sink normally makes Failures throw
so either this is one of the buggy versions where sink didn't always force Failures, or something else weird is wrong 16:03
16:04 lucasb left
ilmari what if it's assigned to a variable that just goes out of scope without being looked at? 16:04
raschipi tojo: You are hitting a bug and you shouldn't just quiet it. Can you try upgrading?
ilmari that won't get sunk, but will get GCed, no?
geekosaur that's more or less the case I described first, yes 16:05
raschipi m: my $a = fail;
camelia Failed
in block <unit> at <tmp> line 1
tojo pastebin.com/mW6BCqPN full error here and the code is here github.com/7ojo/perl6-powerline-pr...-daemon.p6 but.. i'm waiting currently IO::Socket::UNIX to appear and rewrite this little project
ilmari m: my $a = fail; say 42
camelia Failed
in block <unit> at <tmp> line 1
raschipi It get's sunk;
tojo raschipi: yeah i can do that
16:06 st_elmo left 16:07 Zoffix joined
ilmari m: my $a = 1/0; say 42 16:07
camelia 42
ilmari is that a different kind of Failure?
raschipi m: sub f() { fail }; my $v = f;
camelia ( no output )
geekosaur m: my $a = 1/0; dd $a 16:08
camelia Rat $a = <1/0>
ilmari m: my $a = 1/0; $a++
camelia ( no output )
ilmari m: my $a = 1/0; $a++; dd $a
camelia Rat $a = <1/0>
ilmari m: my $a = 1/0; $a++; say $a
camelia Attempt to divide 1 by zero using div
in block <unit> at <tmp> line 1
Zoffix ilmari: that's not a failure at all. You created a Rat object
16:08 astj joined
geekosaur and if you Num it, it becomes Inf 16:09
still not a failure
Zoffix tojo: that's not an error, it's a warning. Somewhere in your code you're getting a Failure, but you never do anything with it, so you have a possibly unhandled error, which is what the warning is about.
tony-o m: dd (1/0).Num;
camelia Inf
geekosaur (but Inf is not in the range of Int)
ilmari m: dd (-1/0).Num 16:10
camelia -Inf
ilmari m: dd (1/-0).Num
camelia Inf
ilmari ?!
mspo I think you are casting the warning?
16:10 mr_ron joined
Zoffix ilmari: only Num has negative zeros 16:11
geekosaur ilmari, since it's a Rat, it can't do negative zero
ilmari m: dd (1/-0.0).Num
camelia Inf
Zoffix ilmari: 0.0 is a Rat, not a Num :)
m: dd 1/-0e0
camelia Failure.new(exception => X::Numeric::DivideByZero.new(using => "/", details => Any, numerator => 1e0), backtrace => Backtrace.new)
Zoffix (in 6.d that'll follow IEEE
)
mspo is () supposed to suppress warnings?
Zoffix No
ilmari pubtimes 16:12
16:12 nadim_ left
geekosaur hm. shouldn't failed require get sunk? 16:13
raschipi Oh, I see, it throws the exception right after the warning. 16:14
Zoffix sees no warnings anywhere... :S
mspo m: (1 div 0).Num 16:15
camelia Attempt to divide 1 by zero using div
in block <unit> at <tmp> line 1
Zoffix 1 div 0 returns Failure; .Num call on it explodes it 16:17
mspo m: Failure.Num
camelia Invocant of method 'Num' must be an object instance of type 'Failure', not a type object of type 'Failure'. Did you forget a '.new'?
in block <unit> at <tmp> line 1
Zoffix leont: gonna fix that shortly. It'll do same as Perl's Test::More and insert hashes before extra lines 16:18
m: Failure.new.Num
camelia Failed

Actually thrown at:
in block <unit> at <tmp> line 1
mspo m: (Failure.new).Num
camelia Failed

Actually thrown at:
in block <unit> at <tmp> line 1
leont Zoffix: that would be very helpful :-) 16:19
Also saw a case of an empty line in the middle of a subtest, not sure how that happened, may or may not be the same issue (t/spec/S02-lexical-conventions/unspace.t "unspace can't hide space between # and opening bracket") 16:21
16:21 S007 left 16:23 nadim_ joined
tojo Zoffix: yeah i suppose, but kind of hard to debug because these warnings comes delayed after long periods of using it and fills my terminal with those :) 16:25
leont paste.scsys.co.uk/564910
It seems #SKIP is also not handled correctly
moritz yay, my book is now "#1 new release" in the "Software Programming Compilers" category 16:27
16:27 khisanth_ is now known as Khisanth
Zoffix \o/ 16:28
moritz++
tony-o moritz++
have a link handy moritz? 16:29
mspo moritz: did you pick an example thing to parse?
moritz tony-o: www.amazon.com/dp/1484228987
it says so right above the cover image
more reviews would also help (hint, hint) :-) 16:30
16:30 nadim_ left
Geth ecosystem: 0664c84616 | (Zoffix Znet)++ (committed using GitHub Web editor) | META.list
Add Toaster to the ecopsystem

  "Toaster of Perl 6 Ecosystem Modules": github.com/zoffixznet/perl6-Toaster		 16:30
Zoffix tojo: good point. I'll fix it shortly.
16:31 robertle joined
tojo Zoffix: ! awesome :) 16:33
raschipi moritz: Your book doesn't mention Perl6 can also call C++ libs. 16:34
mspo dock a star
16:35 S007 joined, mcmillhj left
tony-o 1 star 16:35
moritz raschipi: right; there are a lot of things that can't be mentioned in a mere 150 or 170 or so pages 16:36
16:37 mcmillhj joined
raschipi It's because if this isn't specifically mentioned people will get the idea that it isn't possible, because of the C++ name mangling issue. 16:37
moritz I'm starting to suspect I won't be able to mention all regex possibilites in the regex book
16:38 ChristopherBotto joined
raschipi Will it be a big book? 16:38
tony-o you'd be the homer of our time
16:38 dakkar left
moritz raschipi: probably about the same size as Perl 6 Fundamentals 16:39
16:39 Cabanossi left 16:41 Cabanossi joined, mcmillhj left, rindolf left
raschipi ShalokShalom: Do you like editing Wikipedia? Could you please add perl and Perl6 rules to en.wikipedia.org/wiki/Comparison_o...generators ? 16:44
ShalokShalom less interest in Perl currently
Skarsnik hm 16:46
16:53 raschipi left, nattefrost left, japhb joined 16:54 mcmillhj joined 16:55 raschipi joined 16:56 S007 left 16:59 mcmillhj left
TimToady m: say "foo" when "bar" given Any 17:02
camelia ( no output )
TimToady m: say "foo" when "bar" given "bar"
camelia foo
Zoffix :) 17:03
TimToady smartmatch is not generally supposed to trigger undef warnings on failure to match
Zoffix tojo: I lied. 17:04
17:04 mcmillhj joined
Zoffix tojo: it actually does print some of the backtrace. I tried a few things to make it print more, but didn't find anything helpful. 17:04
Gonna stash it into my TODO list and look at it eventuallyu (maybe)
leont Seems a spectest is failing for me (S29-os/system.t ":cwd(...) changes the directory processes are spawned in") 17:05
Zoffix leont: how are you running the tests? It's doing some awful things to figure out what the executable is: github.com/perl6/roast/blob/master...tem.t#L145 17:07
leont It fails on «./perl6 -Ilib t/spec/S29-os/system.rakudo.moar» 17:08
Zoffix ./perl6 -Ilib t/spec/S29-os/system.rakudo.moar passes for me on HEAD 17:09
17:10 abraxxa joined 17:14 abraxxa left
leont Then I have no idea what's going wrong 17:16
17:16 xinming joined
tojo Zoffix: ok! i remember that it started printing these somewhere in 2017.02-03 version of rakudo and there was ticket about this same thing, but didn't find that ticket now.. 17:17
17:22 preaction joined 17:23 S007 joined
Zoffix tojo: there was, it was closed because it wasn't a bug (just the message was improved) 17:24
preaction are the folks submitting perl 6 distribution test reports to cpan testers in the channel? i've discovered that i need to make the api more strict about the data it accepts and don't want to disrupt the reporters
17:26 andrzejku joined
Zoffix preaction: I am one of the submitters 17:26
And anyone who has Zef::CPANReporter installed is too (not sure who else uses it) 17:27
preaction: also I never came around to write the API stuff because I totally burnt myself out. currently taking a vacation (more or less) :)
leont I'm pretty sure that test is incorrect. It's executing «$*EXECUTABLE.basename», which is something that's in my path 17:28
17:28 andrzejku left
leont The test will only work if you have a perl6 installed 17:28
(and in your PATH)
preaction right now my focus is on version numbers: it was revealed to me that the Perl 5 clients prefix their version numbers with a "v", and i've now fixed that problem but need to make sure the API rejects invalid versions
Zoffix leont: oops yeah. 17:29
preaction a better question might be: are you on the cpan-testers-discuss mailing list Zoffix? or is there some medium for communication with testers (this channel, for example)?
17:30 andrzejku joined
Zoffix leont: looks like another test is also similarly aflicted: github.com/perl6/roast/blob/master...#L117-L118 17:30
17:30 user3 joined, setty1 left
Zoffix preaction: no, I'm not on the list. I might be a bit out of the loop, but I don't think we have proper testers yet 17:31
ugexe: any idea?
(for testers stuff)
preaction Zoffix: and no rush on that thing. i've been doing some thinking on how cpantesters might start treating every language as its own database all with the same schema, but that's a _long_ term solution when some short-term solutions will get y'all something to look at
Zoffix And personally, I have testers on my Toaster that smokes the entire ecosystem, but its primary use is for extended pre-release testing of the compiler.
I mean, I have that module installed on my Toaster box, so reports get submitted. 17:32
user3 why does this output an empty array instead of Nil
m: '' ~~ / [(a)+]? /; say $0;
camelia []
17:33 zakharyas joined
Zoffix user3: because you have the `+` on it, so it'll be an array of matches. Did you mean to have the `+` on the `a` instead? 17:33
user3 m: '' ~~ / [(a+)]? /; say $0;
camelia Nil
Zoffix m: 'abc' ~~ / [(.)+]? /; say $0;
camelia [｢a｣ ｢b｣ ｢c｣]
preaction Zoffix: i'm hoping to move towards those pre-release reports (testing the language against the modules, instead of testing the modules) being submitted (perl 5 does that extensively with Bleadperl Breaks CPAN) so that the website can create nice reports for that data (i just thought that a "language release report" would be pretty amazing) 17:35
Zoffix buggable: toast
buggable Zoffix, Between 2017.07-195-g8664050 and 2017.07: 35 (4.08%) modules got burnt; 36 (4.20%) got unsucced; 267 (31.16%) out of 857 modules appear unusable. See toast.perl6.party/ for details.
Zoffix ^ currently we got this
Between last release and current HEAD, 35 modules got broken by the changes 17:36
But I had to manually run it and even on a 24-core box it takes about 1-2 hours
17:37 mr_ron left 17:40 nhywyll joined, user3 left 17:42 devmikey joined 17:46 S007 left 17:51 thundergnat joined 17:53 zakharyas left
preaction Zoffix: and yeah, something like that page is like what i'm thinking, and that's really cool 17:54
thundergnat Is there somewhere where restrictions on unspace is codified?
17:55 zakharyas joined
thundergnat m: .say for (lines.lc ~~ m:g/[<[\w]-[_]>]+/)».Str.Bag \.sort(-*.value)[^5] # I would expect this to work. 17:55
camelia 5===SORRY!5=== Error while compiling <tmp>
Confused
at <tmp>:1
------> 3lines.lc ~~ m:g/[<[\w]-[_]>]+/)».Str.Bag7⏏5 \.sort(-*.value)[^5] # I would expect
expecting any of:
infix
infix stopper
moritz thundergnat: the \ can't have whitespace before it, I think
thundergnat m: .say for (lines.lc ~~ m:g/[<[\w]-[_]>]+/)».Str.Bag.sort(-*.value)[^5] # ok without the unspace 17:56
camelia und => 30
ich => 17
die => 13
der => 10
das => 8
thundergnat m: .say for (lines.lc ~~ m:g/[<[\w]-[_]>]+/)».Str.Bag\ .sort(-*.value)[^5] # ok without the unspace
camelia und => 30
ich => 17
die => 13
der => 10
das => 8
thundergnat Ah.
moritz++
17:59 rindolf joined 18:00 astj left 18:01 astj joined 18:05 astj left 18:06 cpage_ left 18:15 cdg left, astj joined
timotimo .tell masak there's still my pullrequest for Adventure::Engine to make it work again, same for crypt 18:16
yoleaux timotimo: I'll pass your message to masak.
18:19 astj left
timotimo do we know how to get things to work that "use Test::META" and i get the "Could not find META6:ver<0.0.4..*>" error? 18:20
hm, i'll just try zef nuke
Skarsnik *insert table flip fancy unicode string* I hate you libxml2, now I need to add something to resolve on what other header definition an header need :( 18:22
Zoffix timotimo: yeah, I think that's fixed with updated zef 18:23
18:27 khw joined 18:28 Svelta02 joined 18:29 mcmillhj left
mspo need XML::Grammar::PP6 :) 18:30
18:36 mcmillhj joined 18:37 spider-mario left 18:38 spider-mario joined 18:39 Svelta02 left, espadrine_ joined 18:41 mcmillhj left 18:43 aquila-chrysaeto joined 18:44 ckraniak left 18:45 zakharyas left
Zoffix m: use Test; plan 2; skip 1; ok 1, "foo\nbar" 18:46
camelia 1..2
ok 1 - # SKIP 1
ok 2 - foo
# bar
Zoffix leont: ^ both skip and newlines in descs fixed 18:47
18:47 aquila-chrysaeto left
Zoffix flies away into the sunset 18:47
18:47 Zoffix left 18:49 nadim joined 18:51 mcmillhj joined 18:53 Cabanossi left
leont Zoffix++ 18:55
18:56 Cabanossi joined 18:57 konsolebox left
leont Still seeing that newline issue in unspace.t, guess that must have been a different issue 19:01
19:03 konsolebox joined 19:06 domidumont left 19:10 wamba joined 19:11 lowbro joined, lowbro left, lowbro joined 19:20 darutoko left 19:22 domidumont joined 19:24 astj joined 19:25 domidumont left 19:28 astj left 19:31 mr-foobar joined
timotimo ZofBot: well, i did just updated zef before i tried that :) 19:32
er, wrong zof 19:33
19:33 mr-fooba_ left 19:38 S007 joined 19:39 Cabanossi left 19:40 Cabanossi joined 19:43 mr-fooba_ joined 19:44 entonian joined 19:45 ckraniak joined 19:46 mr-foobar left
gfldex jnthn: my long running Perl 6 stuff is not crashing anymore. Did you change something? :-> 19:47
19:49 nadim left, nadim joined 19:51 cdg joined 19:55 cdg left
jnthn gfldex: I fixed a couple of bugs earlier in the week where GC invariants weren't upheld, which could led the occasional crashes. Maybe one of those nailed it. 19:56
19:56 cpage_ joined 19:58 cdg joined 19:59 zorin-os joined
zorin-os can perl6 use $^X without it requiring quotes? 20:00
20:03 masak joined
geekosaur doesn't understand the question 20:04
if you're talking about perl 5's $^X, that's not a thing as such ($^ twigil is placeholder variables) 20:05
zorin-os is it possible to avoid the " colliding when using the libperl
#define SAY_HELLO "-e", "$string = "$^X"; $string =~ s/.*\///; $newstring = "$^X" ; $newstring =~ s/$string//; print $newstring;"
like does perl6 fix that issue? 20:06
geekosaur perl 5 doesn't require quotes when using $^X, so I have no idea what you are on about 20:07
20:08 cpage_ left
geekosaur also once you drag in things like cpp (not supported in perl 6) you're in a different world of hurt anyway 20:08
20:09 ckraniak left
zorin-os ok 20:09
20:10 cpage_ joined
leont zorin-os: in how many channels are you going to do this? 20:10
20:10 nadim left
zorin-os when i attempt to execute it i get this "syntax error at -e line 1, near "/;"" 20:11
wich doesnt happen when i execute it on the command line via perl -e
$ perl -e '$string = $^X; $string =~ s/.*\///; $newstristring =~ s/$string//; print $newstring;' 20:12
/usr/local/bin/
geekosaur that sounds like you are doing something involving the shell to me
nothing will fix the shell expanding $whatever in a double quoted string, that is a shell thing that happens before anything else can see it 20:13
20:13 ckraniak joined
zorin-os how do i get that to apply to this #define SAY_HELLO "-e", "$string = $^X; $string =~ s/.*\///; $newstring = $^X ; $newstring =~ s/$string//; print $newstring;" 20:13
20:13 nadim joined
geekosaur I have no idea what you are doing with that so cannot answer 20:13
that is a C preprocessor definiton
where/how are you expanding it? 20:14
zorin-os paste.pound-python.org/show/ihjjPQ...nhElIdTDk/ 20:15
geekosaur that didn't even compile correctly, so no idea what the a.out is from 20:17
your compile command got split into two lines, the first errored out because the -I it needed was on the second line, the second errored out because it started with "-fwrapv" which is not a shell command
and I would be surprised if it even looked at libperl.a because at that point the only undefined symbol is _main and it's not defined in any member of libperl.a 20:19
zorin-os must be an issue with "wgetpaste -c" then
20:20 Zoffix joined
Skarsnik geekosaur, nc can bind some basic c++ x) 20:20
Zoffix Are available code tags in docs pod listed somewhere?
How the hell do I shut up xtest about undeclared variables?
geekosaur Skarsnik, on some platforms where it knows the claling convention 20:22
20:22 ckraniak left
Zoffix Also, what's "atomicsize"? 20:23
20:23 ShalokShalom left 20:24 ckraniak joined
Zoffix *crickets* 20:25
zorin-os paste.pound-python.org/show/8bSlW6...R2SYniicb/
20:26 aquila-chrysaeto joined
Geth doc: 4b9beced0e | (Zoffix Znet)++ | 3 files
Fix some xtest failures
 20:27
Zoffix .ask jnthn what's atomicsize? The docs mention it, but never say what it is 20:28
yoleaux Zoffix: I'll pass your message to jnthn.
geekosaur zorin-os, yes, you are using a regex embedded in a string
you need extra escapes. this is because you embedded it in a double quoted string
Zoffix .tell [Coke] I fixed what xtests I could, per our fix-them-on-release agreement, but I can't find any docs on what new code pod tags were added to shut it up about undeclared variables, so those are still failing
yoleaux Zoffix: I'll pass your message to [Coke].
geekosaur again, this is not something that can be "fixed", computers cannot read your mind and switch languages mid-stream "oh this C string changed to perl I will parse as perl from here on"
zorin-os where exactly do i need to put the extra escapes
20:29 qwebirc480120 joined, ckraniak left
qwebirc480120 Hello everyone! o/ 20:29
Zoffix Hi
geekosaur basically double the existing \
I see only one, so double it
zorin-os ok
geekosaur beause C sring handling will then remove one of them and pass the other one on
qwebirc480120 I'm so embarassed. I meant to do a pull request for modifying heredoc tests and I accidentally push directly to roast (see github.com/perl6/roast/commit/e9a2...1a181999).
s/push/pushed 20:30
zorin-os yay paste.pound-python.org/show/YQs7Pr...8APAxxfeY/
moritz qwebirc480120: are the tests wrong?
Zoffix qwebirc480120: git revert e9a2bda8c4596f04b4549e416bba7b191a 20:31
qwebirc480120 moritz: No, the tests document bad behavior that needs to be fixed.
20:31 aquila-chrysaeto left
Zoffix At the very least, that skip directive is missing a comma, so after fudging, they won't compile 20:31
moritz qwebirc480120: then there is no harm
20:31 llfourn joined
Zoffix s/comma/quote/; 20:31
t/fudgeand[TAB] S02-literals/heredocs.t 20:32
moritz well, that can be fixed
Zoffix that'll fudge and run just the file and show you all failures, if there are any
geekosaur lpaste.net/3502097151834456064 20:35
20:35 ugjka left 20:39 llfourn left
Zoffix .tell [Coke] never mind. Found the docs. .oO( maybe all of these separate .md's should be in one place ) 20:44
yoleaux Zoffix: I'll pass your message to [Coke].
qwebirc480120 Zoffix: moritz: Switching to already open IRC window, so I'll now be known as ChristopherBotto instead of qwebirc480120. 20:48
Geth doc: 51b678e44c | (Jonathan Worthington)++ (committed using GitHub Web editor) | doc/Type/atomicint.pod6
Fix thinko; atomicsize was meant to be atomicint.
Zoffix qwebirc480120: OK. I'll now be known as Harry Potter
ChristopherBotto: so does the test file run? with t/fudgeandrun.t S02-literals/heredocs.t 20:49
ChristopherBotto: not a fan of those super-long lines. Even on github they don't fit on screen and require scrolling :/ 20:50
Geth doc: dfd8333fbb | (Zoffix Znet)++ | 2 files
More xtest fixage
qwebirc480120 Zoffix: Yes it runs. 20:51
Zoffix Cool.
qwebirc480120: well, the PR looks good to me. Just merge it (after getting the RT number? and maybe after making line lengths shorter? :)) 20:52
20:52 qwebirc722017 joined 20:53 qwebirc722017 left, domidumont joined, entonian left, domidumont left 20:54 nadim left, qwebirc480120 left, raschipi left 20:56 cpage_ left 20:57 ChristopherBotto left 20:58 skids left, ShalokShalom joined 21:01 cpage_ joined 21:03 nadim joined
tojo about that 'WARNING: unhandled Failure detected in DESTROY', started seeing it now with my other projects too where i'm using "require", but using "try sink require .. " suppress the warning and everything seems ok now :) 21:05
Zoffix doubts the issue is really in there as `try` fatalizes failures 21:07
Oh, your first version didn't have a `try`.
21:07 nadim left 21:09 Cabanossi left 21:10 Cabanossi joined 21:11 TEttinger joined 21:12 abraxxa joined
tojo and maybe that sink part is not needed, was seeing using that with Proc::Async on the docs 21:13
Zoffix tojo: in this case, no. In Proc::Async yeah, because the sink sinks the Proc object, which otherwise gets returned from `try` and THEN sunk, and if it were for an unsuccessfull process, then throws 21:16
so sinking it first makes the explosion happen inside the try
while `require` returns a Failure and `try` has `use fatal` in it, which makes Failures explode right away 21:17
tojo Zoffix: thanks for clarifying! 21:19
21:25 abraxxa left 21:27 mcmillhj left 21:30 cpage_ left 21:37 TEttinger left 21:39 TEttinger joined, cpage_ joined 21:40 MasterDuke joined, rindolf left 21:42 andrzejku left 21:43 rindolf joined 21:44 devmikey left
Geth doc: 73264b7d2d | (Zoffix Znet)++ | 3 files
Make xtest pass
 21:49
Zoffix k :) agreement held :)
21:51 Kyo91 joined, cdg left
timotimo overload::constant fixed \o/ 21:52
21:54 Kyo91_ left 21:55 TEttinger left, Kyo91 left, mcmillhj joined
zorin-os weird, when compiled with musl-buildchain $^X is one_perl 21:56
yet the perl executable itself remains the same result as the gcc version even though it was compiled with musl aswell 21:57
21:57 andreoss joined 22:00 APic left, mcmillhj left, TEttinger joined
zorin-os weird, when compiled with musl-buildchain $^X is one_perl yet the perl executable itself remains the same result (it displays the path to the executable) as the gcc version even though it was compiled with musl aswell, could this be a side effect of the musl toolchain? * 22:02
preaction zorin-os: this channel has nothing to do with Perl 5, at all 22:04
22:04 lucasb joined 22:17 wamba left 22:18 TEttinger left 22:21 BenGoldberg joined 22:24 TEttinger joined 22:30 Skarsnik left 22:31 Zoffix left 22:34 APic joined 22:35 skids joined 22:37 astj joined 22:41 astj left 22:46 lucasb left 22:48 ufobat left 22:53 cpage_ left 22:54 cdg joined 22:55 Cabanossi left 22:56 cpage_ joined, Cabanossi joined 23:01 nhywyll left 23:14 cdg left, espadrine_ left 23:19 mr_ron joined
mr_ron m: say chr(0x212A), " ", ord(chr(0x212A)).base(16) # unicode "kelvin" codepoint - but shouldn't ord reverse chr? 23:22
camelia K 4B
23:23 TreyHarris left 23:25 TreyHarris joined
geekosaur I think this is one of those cases where the normal form is a different character 23:30
like current Unicode spec says the Greek question mark normalizes to semicolon 23:31
mr_ron m: say chr(0x4B) # don't know much about normal forms - will try to look 23:32
camelia K
23:50 thundergnat left 23:55 nattefrost joined 23:58 mcmillhj 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!