»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.
redfd done? 00:00
timotimo cool, then you have zef
redfd why nor work command zef?
timotimo you have probably not added the path to you rpath yet
redfd how add? please
timotimo you're on linux? 00:01
redfd yes
timotimo OK, typically, you'd add paths in your ~/.profile
redfd this? ~/rakudo-star-2017.01/modules/zef# ls bin lib LICENSE logotype META6.json README.pod resourcest xt 00:02
00:02 mcmillhj left
timotimo no, more like /home/timo/perl6/install/share/perl6/site/bin/zef 00:02
depending on where you installed rakudo-star to
i have this perl6 folder in my home where i put everything perl6-related
i don't know where rakudo-star installs to by default
what do you get from "which perl6"? 00:03
redfd stop, please_ I can use p6 without root?
mst redfd: of course you can
redfd: how did you install rakduo-star ? 00:04
redfd which perl6 /usr/local/bin/perl6
timotimo oh, you installed it into your system
mst ok, so you installed it as root
why did you install it as root if you want to use it without root?
redfd i dont know
))
brb
i want install perl6 00:05
with zef
mst what happens if you run 'which zef' ?
redfd nothing 00:06
no zef
00:08 jdmmmmm left, mcmillhj joined
ugexe m: say $*REPO.repo-chain.grep(* ~~ CompUnit::Repository::Installable).map(*.prefix.child("bin").absolute) 00:09
camelia (/home/camelia/.perl6/bin /home/camelia/rakudo-m-inst-2/share/perl6/site/bin /home/camelia/rakudo-m-inst-2/share/perl6/vendor/bin /home/camelia/rakudo-m-inst-2/share/perl6/bin)
ugexe zef is installed to one of those paths already, but its not in your PATH
redfd I'm reinstallling again 00:10
how make p6 without root? with zef? please help 00:11
step to reproduce)
00:11 Cabanossi left
ugexe github.com/ugexe/zef/blob/master/....ml#L20-L27 00:12
that is how to install rakudo and set the PATH properly. then you can install zef and it will work because PATH will already be set
redfd ok I do rm rakudo
ugexe note its line 20 to 27 specifically
redfd so 00:13
ugexe you dont have to do the two export ZEF_* lines tho
redfd i dont know
timotimo when you use that "export", you'll only have the right path in your PATH until you close that shell, and also only in that shell, not in others
00:13 mcmillhj left
redfd step 1 for install perl6? 00:13
00:14 pmurias left
redfd i want rakudo-star 00:14
not *brew 00:15
because in brew not work zef
00:15 Cabanossi joined
redfd zef cant download LWP::Simple 00:15
and so on
test with errors 00:16
ugexe zef works with rakudobrew. your setup sounds broken
redfd wtf? its normal for p6?
or I stupid?)
ugexe it sounds like you have multiple perl6 installs in your PATH, both rakudobrew and regular 00:17
00:17 labster joined
redfd so, I download this rakudo.org/downloads/star/rakudo-st...est.tar.gz rigth?? 00:18
ugexe I don't use rakudo-star other than on windows, so I can't help you there 00:19
redfd hmm
maybe rakudobrew?? 00:20
timotimo rakudobrew isn't meant for end-users
redfd what meant for end users? 00:22
timotimo just the steps ugexe linked to from the .travis.yml file 00:23
00:24 kurahaupo left 00:25 kyan joined
redfd ok 00:25
00:25 bjz left
redfd perl Configure.pl $VM_OPTS --gen-nqp --backends=$BACKEND 00:27
what is it?
timotimo ah i didn't look closely 00:28
one thing you may be interested in is using a --prefix=somepath
like, my prefix is /home/timo/perl6/install
redfd perl Configure.pl --gen-moar --gen-nqp --backends=moar -- its ok?
timotimo well, you may want a --prefix 00:29
redfd so.. perl Configure.pl /home/user/perl6/install --gen-moar --gen-nqp --backends=moar ?
timotimo you have to put --prefix= in front of the path 00:30
redfd so.. perl Configure.pl --prefix=/home/user/perl6/install --gen-moar --gen-nqp --backends=moar ?
timotimo that looks good
redfd ok
one momet 00:31
timotimo i'll go to bed real soon
but there's good instructions in the README
redfd i ubderstand English
very bad
timotimo wait, maybe the instructions you need are in INSTALL, not in README 00:32
okay it's INSTALL.txt
redfd now 00:33
I just want perl6 and zef
i complet perl Configure.pl --prefix=/home/user/perl6/install --gen-moar --gen-nqp --backends=moar |||| what next? 00:34
timotimo "make" and then "make install"
redfd ok
timotimo after that do the thing with zef again: go into zef's directory and "perl6 -Ilib bin/zef install ."
i wish you the best of luck :) 00:35
good night!
redfd thank \ u2)
00:41 AlexDaniel left 00:42 cognominal left
llfourn Just got my TPF 50min talk approved \o/. Going to be talking about: github.com/spitsh/spitsh 00:42
01:09 bjz joined 01:13 gdonald left, gdonald joined 01:28 Cabanossi left 01:29 Cabanossi joined 01:30 espadrine left 01:31 bjz left 01:33 lowbro left 01:42 Sgeo joined 01:46 ilbot3 left 01:48 ilbot3 joined, ChanServ sets mode: +v ilbot3
tbrowder ref docs, routine symlink: it seems that the p6 symlink args are reversed from the *nix link cmd, or am i misreading it? 01:48
Voldenet > ln [OPTION]... [-T] TARGET LINK_NAME (1st form) 01:51
> sub symlink(Str $target, Str $name --> Bool) 01:52
doesn't seem reversed
but I do dislike that it's not consistent with cp, which basically is 'cp name target' 01:54
tbrowder but look at the docs, 2nd arg is name of file, not link, oppo from ln 01:55
grondilu thinks the target should be a named parameter 01:56
tbrowder ln -s filename link
redfd help please
perl6 -Ilib bin/zef install . All candidates are currently installed No reason to proceed. Use --force to continue anyway
Voldenet oh
yeah, naming choices there
redfd which zef - no result
Voldenet *cough*
mst redfd: please uninstall all perl6s you already installed and start from a clean system 01:57
redfd: you are guessing and we don't know what's on your system 01:58
redfd no
mst then no help
01:59 mr-foobar joined
redfd ok 02:03
tbrowder well, docs are correct, symlink is reversed from *nix ln, bummer!
although, to be fair, i always thought ln should be that way, too 02:04
when you do ls -l on a link it show link then tgt 02:05
redfd i dont kbow use symlink)
BenGoldberg m: dd symlink("/nope", "/nope") 02:07
camelia symlink is disallowed in restricted setting
in sub restricted at src/RESTRICTED.setting line 1
in sub symlink at src/RESTRICTED.setting line 18
in block <unit> at <tmp> line 1
BenGoldberg m: dd symlink("/nope", "/nope") if 0; 02:08
camelia ( no output )
redfd please 02:14
02:15 redfd left 02:16 redfd joined
redfd have derectory /home/xxxxx/perl6/install/share/perl6/site/bin/ 02:17
./zef running
how facking add symlink?
dmaestro tbrowder: I suppose an idiom like this is intended: perl6 -e 'my $fn = "file.pl6".IO; $fn ~~ :f && say "new-link".IO.symlink($fn)'
redfd please give me command 02:18
dmaestro redfd: what file do you want to link to, and where do you want to put the link? 02:21
redfd ./zef
in /home/xxxxx/perl6/install/share/perl6/site/bin/
mst why is /home/xxxxx/perl6/install/share/perl6/site/bin not in your $PATH 02:22
if you want to be able to run that?
redfd how add ? 02:25
02:27 noganex_ joined 02:30 noganex left 02:43 redfd left 02:49 bjz joined 02:50 labster left 02:54 BenGoldberg left
u-ou hi 02:56
02:59 skids joined 03:04 g0d355__ left 03:41 Cabanossi left 03:44 mr-foobar left, Cabanossi joined 03:48 labster joined 03:56 khw left 04:07 mcmillhj joined 04:12 mcmillhj left 04:20 kaare__ joined 04:21 kaare_ left 04:22 gdonald left 04:23 gdonald joined 04:24 mcmillhj joined 04:25 Cabanossi left 04:27 Trioxin joined 04:28 Cabanossi joined, mcmillhj left
samcv what in the actual.... how the hell 04:43
how does this happen
look at the numbers i.imgur.com/ePs6Ygc.png 04:44
that is not in my code
geekosaur terminal's auto font switching to get the glyphs and not switching back properly 04:45
samcv maybe
04:46 gdonald left
samcv it happens every time i run it 04:46
geekosaur because every line it happens on has a glyph, and the lines that don't do it render nothing
samcv hmm
04:47 gdonald joined
samcv yeah 04:47
uh my other terminal stops showing at 171 04:48
really weird
geekosaur lots of possibilities, sadly. probably not an ft2 issue but could be fontconfig, likely pango for VTE-based terminals, or the terminal itself 04:50
many versions of pango are buggy for non-ISO, non-CJK characters :/
geekosaur keeps running into that rendering hebrew and arabic 04:52
samcv yeah 04:53
oh i think it actually is just rando characters 04:54
and it just happened to be the same number.
my code's not properly resetting something
and it iterate the chacaters forever maybe
yay i think i fixed it 04:56
really weird that it somehow lined up with actual numbers
05:00 rurban joined, rurban left 05:02 wamba joined 05:09 Trioxin left, Trioxin2 joined 05:10 Cabanossi left 05:14 Cabanossi joined
u-ou im a baby perl6er 05:14
skids o/ 05:25
u-ou oh, man you can make special variables your arguments 05:27
I love that
like for $/ or $_
05:32 Sgeo left 05:37 titsuki left 05:38 titsuki joined 05:58 ufobat joined 06:01 titsuki left, titsuki joined 06:02 labster left 06:04 titsuki left, titsuki joined 06:10 RabidGravy joined 06:11 mcmillhj joined 06:17 rurban joined 06:18 mcmillhj left 06:22 ChoHag joined 06:28 mcmillhj joined 06:33 mcmillhj left 06:36 ufobat left
samcv magic 06:39
perl is just special
06:41 Cabanossi left 06:43 Cabanossi joined 06:50 Actualeyes left 06:51 ufobat joined 06:54 rindolf joined, g0d355__ joined 06:56 skids left 07:03 darutoko joined 07:15 rindolf left 07:21 cyphase left, labster joined 07:23 rindolf joined 07:30 Sgeo joined 07:48 cyphase joined 07:55 Cabanossi left 07:59 Cabanossi joined
moritz \o 08:05
08:08 spebern joined 08:11 domidumont joined 08:16 domidumont left 08:17 domidumont joined 08:19 szabgab joined, e1 joined
simcop2387 m: 1+1 08:27
camelia WARNINGS for <tmp>:
Useless use of "+" in expression "1+1" in sink context (line 1)
TEttinger m: say 1+1 08:31
camelia 2
samcv ok. it's coming 08:33
optimized string search :) regex gonna get much faster 08:35
about to bunch moar for git rakudo now 08:36
gonna get some ice cream to celebrate 08:38
08:39 CIAvash joined 08:46 rumble joined
u-ou :D 08:48
szabgab I was looking at smoke.perl6.org/report but it seems to be out of date. Is it still functioning? 08:55
yoleaux 1 Apr 2016 13:39Z <[Coke]> szabgab: - Can you please double check rt.perl.org/Ticket/Display.html?id=123497 against a 2016 R*?
08:55 espadrine joined
samcv ok done. now i will enjoy this ice cream 08:55
08:56 grumble left
szabgab yoleaux: p6doc works on the 2017.01 I have. AFAIK that ticket can be closed 08:56
samcv yoleaux, is a bot. it cannot hear you ;) 08:57
08:58 wamba left 08:59 wamba joined
szabgab samcv: oh just nagging the rare visitor :) 08:59
samcv damn p6doc doesn't work on os x. that's lame 09:00
or maybe it does
it's been 1 year :)
ticket is still open tho
09:02 tneeds joined
szabgab Oh I see now I reported it for osx. I am now using Perl 6 inside a Docker image so I don't know if p6doc works on OSX or not. 09:03
samcv can you not install modules or whatever? 09:05
szabgab checking again, apparently I have also installed Rakudo on OSX p6doc works ther
there
09:05 kyan left
szabgab that ticket can be closed 09:06
samcv cool 09:07
want me to close it or you
szabgab pls do
samcv kk 09:08
09:10 spebern left
szabgab so back to smoke testing: smoke.perl6.org/report seems to be stuck in 2015. testers.perl6.org/ seems to report for released verisions of Rakudo 09:18
Is there a page for smoke testing Rakudo with the spectests? Is there a page for smoke testing modules with the development version of Rakudo? 09:19
09:22 ChoHag left
gfldex szabgab: testers.perl6.org/ 09:25
szabgab gfldex: does that test on the latest development version or only on released versions? 09:26
gfldex don't think so, seams to be Star 09:27
if module authors setup travis cron jobs, the ecosystem tests against bleeding 09:28
(very few do)
samcv .o(should check if i have travis cron jobs)
gotta take advantage of all that free cpu time
oh i do actually 09:29
it ran 2 days ago. cool
travis-ci.org/samcv/URL-Find/builds
gfldex samcv: see 2nd halve of gfldex.wordpress.com/2017/01/20/once-a-week/
samcv haven't updated it since a few months ago but still running tests
gfldex, uhm is testers.perl6whatever does that go off of travis at all? 09:30
like if we got everybody to setup cron jobs
szabgab what do you mean travis cron jobs?
samcv it runs every set amount of time
and can run against development and release version 09:31
gfldex testers wont but modules.perl6.org does
samcv yeah
szabgab, have you seen our travis matrixes
we have them for multiple os's too
09:31 rindolf left
samcv very neat 09:31
szabgab samcv: I don't think so
I found this now: docs.travis-ci.com/user/cron-jobs/
gfldex you can tell travis where to get rakudo from. So you could have a meta package that tests against release 09:32
samcv this is for moarvm travis-ci.org/MoarVM/MoarVM
here's moar's matrix
has loads of configurations of moarvm and two os's mac and linu
szabgab I wonder if travis has a API that could be used to tell to run all the test when a new commit arrived to Rakudo
samcv and hopefully we'll have windows automated soon
it does
every time
gfldex the travis api is rather shitty
samcv oh, do you mean modules?
modules probably not easily unless you set it up yourself 09:33
gfldex it does not run on commits in a dep
samcv nope
that would be nice. but until then. we have cron
09:34 e1 left, bjz left
samcv which part did you want me to read 09:36
i read it all tho
szabgab samcv: I guess "latest" in the .travis.yml refers to the latest released Rakudo. Is there a way to add "nightly" or "blead" or some other keyword that would use a Rakudo build that was created from the most recent push to the Rakudo repo? 09:38
samcv yeah. uh
i think there way
*was but i forget it
i thought i remember doing it. but dunno have many repos 09:39
well not that many
i will try now on my module though 09:40
yeah i think it worked 09:41
travis-ci.org/samcv/URL-Find
it has two different jobs now. release and development
oh ok. i needed to change it to `nom` not moar-blead. it tries to check out that branch i guess 09:43
09:45 szabgab left 09:51 rurban1 joined 09:52 Trioxin2 left 09:53 rurban left 09:54 andrzejk_ joined 09:57 Cabanossi left, Cabanossi joined 10:02 rumble is now known as grumble 10:14 szabgab joined
szabgab samcv: so if one adds "nom" to the travis.yml file that will run the tests on the most recently pushed code? 10:18
hmm the doc here: docs.travis-ci.com/user/languages/perl6/ seems to indicate otherwise 10:20
10:31 szabgab left 10:33 titsuki left 10:34 titsuki joined 10:35 AndyDee left 10:36 AndyDee joined 10:38 TEttinger left 10:45 ufobat left 10:50 tomboy64 left 10:58 zakharyas joined 11:00 holyghost joined 11:01 holyghost left 11:10 tomboy64 joined 11:14 labster left 11:22 rindolf joined 11:34 leah2 left 11:43 awwaiid left 11:51 leah2 joined 11:55 Cabanossi left 11:57 mr-foobar joined 11:59 Cabanossi joined 12:00 Trioxin2 joined 12:05 zakharyas left 12:15 rurban1 left
tbrowder dmaestro: thanks, but 12:18
uh, no buts... 12:19
12:26 Cabanossi left 12:28 Cabanossi joined 12:30 rurban joined 12:34 leah2 left
Geth doc: 6bd0f98376 | (Zoffix Znet)++ | 5 files
[io grant] Dissuade readers from using IO::Spec*
 12:37
12:47 robertle left 12:49 leah2 joined 12:52 AlexDaniel joined, robertle joined 12:54 Guest15811 joined 12:56 rurban1 joined 12:58 rurban left 13:00 Guest15811 left
Geth doc: 7afd9c4279 | (Zoffix Znet)++ | doc/Type/IO/Path.pod6
[io grant] Remove unrelated related classes
 13:12
13:16 domidumont left
Geth doc: a43ecb9d3e | (Zoffix Znet)++ | doc/Type/IO/Path.pod6
[io grant] Document IO::Path's $.SPEC and $.CWD
 13:28
doc: fadf4e930b | (Zoffix Znet)++ | doc/Type/IO/Path.pod6
Fix typo
 13:31
13:33 CIAvash left
andrzejk_ hello perl6 :) 13:37
13:45 espadrine left
masak hi andrzejk_ 13:47
andrzejk_ hey 13:48
masak how is your sunday?
Geth doc: e9b6809260 | (Zoffix Znet)++ | 4 files
[io grant] Document IO::Path::* subclasses
 13:56
14:03 ajr_ joined 14:04 Actualeyes joined
Geth doc: 9102b51380 | (Zoffix Znet)++ | doc/Type/IO/Path.pod6
[io grant] Fix up IO::Path.basename

  - it doesn't just return filename, since IO::Path can reference
   more than just files
  - Point out semantics with Windows shares		 14:04
ajr_ Can anyone tell me where and how the Rakudo* installation sets the .../install/share/perl6/site/bin and install/share/bin values in PATH? 14:06
moritz ajr_: which platform are you on?
ajr_ There's a message at the end of the install process that says "please set them", but it seems to have been done automatically 14:07
Linux
moritz I'm pretty sure it can't 14:11
maybe you have done it previously?
andrzejk_ moritz how the perl6 is parsing every line till terminate sign ;
Geth doc: 5c1d3b6558 | (Zoffix Znet)++ | doc/Type/IO/Spec/Unix.pod6
[io grant] Document IO::Spec::Unix.basename
 14:12
ajr_ Moritz - that's entirely possible, (as the fact that I've forgotten doing it), but I didn't want my update script to be fighting the official process pointlessly. 14:16
Geth doc: aa293f9a9e | (Zoffix Znet)++ | doc/Type/IO/Spec/Unix.pod6
Fix typo
 14:22
doc: a1cb80bd02 | (Zoffix Znet)++ | doc/Type/IO/Spec/Win32.pod6
[io grant] Document IO::Spec::Win32.basename
 14:23
moritz andrzejk_: token by token 14:34
andrzejk_ but what terminate parser
I mean it is JIT compiler
so he parse first line and push to jit
it looks like every perl6 guarantee that ends with ; or EOF 14:35
or it is the line before closing parenthess
timotimo no, the whole program is parsed before it goes to code-generation
andrzejk_ timotimo lulsh why?
why so?
why not a line 14:36
14:36 rurban1 left, rurban joined
timotimo well, one reason is we allow you to call subs before you declare them - but if you don't declare them we refuse the parse 14:36
m: this-sub-doesnt-exist(1)
camelia 5===SORRY!5=== Error while compiling <tmp>
Undeclared routine:
this-sub-doesnt-exist used at line 1
timotimo m: this-sub-doesnt-exist(1); sub this-sub-doesnt-exist($_) { say "haha, fooled you" }
camelia haha, fooled you
timotimo you can't run that code if you didn't finish parsing 14:37
andrzejk_ timotimo ok now I get it
timotimo most things aren't sensibly runnable before you have parsed a whole lot more of the program, either
we do compile-time selection of multi candidates, too
so even if we're using sub only-works-with-string(1) near the start, how is the compiler supposed to know if there's a multi candidate for Int later in the source? 14:38
andrzejk_ do you mean overloading?
timotimo multiple dispatch 14:39
andrzejk_ multi is perl6 overloading method equivalent
timotimo it's similar
andrzejk_ yeah
okay I got it a bit
you parse whole file prepare syntax tree and then generate assembler
timotimo no
andrzejk_ ?
timotimo we don't generate assembler 14:40
andrzejk_ oh okay
:D
timotimo we go from code text to QAST, to MAST, to moar bytecode, and then the JIT might turn parts of it into x86_64 machine code
andrzejk_ timotimo okay 14:41
that's okay I think
:)
what about syntax can I ask some questions you? 14:42
timotimo you can certainly try
andrzejk_ ok
as I know all things are certainly objects in perl6 14:43
moritz all things that can go into variables are objects
andrzejk_ which correspond to 6model
6model is meta things which consist in each object 14:44
14:44 Trioxin2 left
timotimo "consist in"? 14:44
andrzejk_ come in
timotimo sorry my english is fucked except when I hurry to write smth 14:45
:P
especially
moritz I'd say 6model is an implementation of the low-level parts of the p6 object system
andrzejk_ sure
so lets say first line "hello world".say 14:46
so the parser create anonymous variable "hello world"
and he knows that characters in quotes correspond to Str 14:47
and then he call method say on Str
timotimo no variable involved
andrzejk_ that how I understand it
timotimo have you ever tried perl6 --target=ast -e 'your code here'?
andrzejk_ wait let me try x)
Array? 14:50
timotimo sorry, what?
moritz m: say Array
camelia (Array)
timotimo m: say so Array 14:51
camelia False
andrzejk_ I don't undestand it well
it is String
- QAST::Stmt <sunk final> \"hello\".say 14:52
14:55 kurahaupo joined 14:59 khw joined
andrzejk_ fuck 15:00
it looks very hard to implement this
:D
15:02 ajr__ joined 15:05 ajr_ left
timotimo ... implement what? 15:07
and what do you mean by "it is String"?
15:08 ajr__ left, ajr joined, ajr is now known as Guest76122 15:10 Guest76122 left 15:11 ajr_ joined
andrzejk_ I am trying to implement perl6 15:15
and just trying to figure out from what to start lexer implementation
ajr_ andrzejk - are you trying to re-write the compiler? 15:16
andrzejk_ ye exactly
ajr_ Why?
andrzejk_ I think I can write a better one 15:17
ajr_ Have you read the original source code?
andrzejk_ no but there no some functionality which I need
grondilu andrzejk_: could you write one that targets webassembly?
andrzejk_ I want to have Perl6 on embedded devices
and faster though 15:18
in any way I am searching of start point
15:20 ajr__ joined 15:21 ajr_ left 15:22 ajr__ left, itaipu joined, ajr_ joined
ajr_ andrzejk - if you can write a better Perl6 compiler singlehandedly, there's clearly a space on your shelf for a Turing Award 15:24
andrzejk_ ajr_ :D
I want to start and maybe someone help me in the future
ajr_ Studying the existing one would be a good start
andrzejk_ I don't like read C code :s 15:25
15:26 Cabanossi left
timotimo have you looked at arane? 15:26
andrzejk_ what is arane?
timotimo a p6 compiler written in cpp
incomplete
andrzejk_ that sounds good 15:27
x)
I will start reading it
timotimo thnks for help timotino
15:29 Cabanossi joined
andrzejk_ timotimo what happend with bizzarecake? 15:30
is he alive? 15:31
15:31 rurban1 joined
timotimo i certainly hope he didn't die 15:31
hasn't been on irc for a long time
15:31 rurban2 joined, rurban2 left
andrzejk_ timotimo ohh 15:32
timotimo .seen BizarreCake 15:33
yoleaux I saw BizarreCake 30 Aug 2014 14:30Z in #perl6: <BizarreCake> Is an int array implicitly castable to an Int array?
andrzejk_ timotimo my friend die like a 3 month ago 15:34
he was programmer
15:34 rurban left 15:35 rurban1 left 15:37 cgfbee left, Rawriful joined
timotimo oh shit, that sucks :( 15:38
andrzejk_ yeah 15:40
that happens to many programmers here
i suppose
he got contract b2b and his client needs a code urgently but he won't lose money so he program even when he was sick
and one day he just die 15:41
timotimo it looks like he just implement this 15:47
without any idea how to add Mu object to it etc.
15:48 zakharyas joined
timotimo i don't know what b2b means 15:51
andrzejk_ they pay you good money but you have no vacances etc. 15:52
ne to go
for a bit
15:52 andrzejk_ left, skids joined
gfldex timotimo: as requested: github.com/gfldex/perl6-meta6-bin/...087fe3a06b 15:52
timotimo: it does not know about authors yet and doesn't do anything fancy either 15:53
15:56 Cabanossi left
timotimo fantastic! i'll install that straight-away 15:57
15:58 Cabanossi joined
timotimo it'd be cool if the repository description or the readme mentioned that the module name is actually META6::Bin 15:59
15:59 domidumont joined
gfldex timotimo: you may have to --force, i forgot to dump version 15:59
timotimo i didn't have META6::Bin yet 16:00
16:01 leont joined
gfldex timotimo: hold on, I have a typo 16:02
timotimo META6::bin, though
first it's all caps, then it's no caps at all
16:04 Ven joined, Ven is now known as Guest99305 16:08 jdmmmmm joined, Guest99305 left 16:15 tokomer joined 16:24 TEttinger joined
Geth doc: 28b6283b96 | (Zoffix Znet)++ | 4 files
[io grant] Document IO::Spec::*.canonpath
 16:28
16:30 ajr_ left 16:32 jdmmmmm left 16:48 cdg joined 16:50 ufobat joined 16:53 andrzejk_ joined
Geth doc: 50e5565699 | (Zoffix Znet)++ | 3 files
[io grant] Document IO::Spec::*.catdir and .catfile
 16:53
17:11 zakharyas left 17:14 itaipu left 17:17 cdg left 17:19 ajr_ joined 17:20 ajr_ left
Geth doc: 9dd7399f15 | (Zoffix Znet)++ | doc/Type/IO/Spec/Cygwin.pod6
Remove IO::Spec::Cygwin.catfile

The class doesn't provide it, but merely inherits		 17:21
doc: dbdc9959ab | (Zoffix Znet)++ | 3 files
[io grant] Document IO::Spec::*.catpath
 17:24
17:30 itaipu joined 17:41 Cabanossi left, jjido joined 17:44 Cabanossi joined, stmuk_ joined 17:46 stmuk left 17:59 spebern joined
samcv .tell szabgab yes if you put nom in .travis.yml it does devel build of rakudo. Worked for me here travis-ci.org/samcv/URL-Find 18:02
yoleaux samcv: I'll pass your message to szabgab.
18:13 spebern left 18:24 jjido left 18:35 Sound joined, Sound is now known as Guest9049 18:37 Guest9049 left 18:38 Sound_ joined 18:40 Woodi_ is now known as Woodi 18:47 spebern joined 18:48 kaare__ left 19:05 jjido joined 19:06 titsuki left 19:07 titsuki joined 19:19 domidumont left 19:26 itaipu left 19:28 mxco86 left 19:29 labster joined 19:31 darutoko left 19:33 itaipu joined, titsuki left 19:39 jjido left, titsuki joined 19:41 kaare__ joined 19:42 Cabanossi left 19:43 Cabanossi joined 19:44 as_ joined
as_ Hi guys. I wonder if the way modules are stored in file system is invention of Rakudo or zef or Perl 6 19:45
19:47 imcsk8 left, jjido joined
lizmat Perl 6 19:48
19:55 imcsk8 joined 20:06 titsuki left
moritz rakudo, no? 20:06
timotimo what, the compunitrepo stuff isn't perl6? 20:08
moritz the API is, but the file layout isn't, afaict 20:09
20:09 titsuki joined
moritz I mean, are there are tests for the actual file locations and contents in roast? 20:09
timotimo oh, ok
that makes sense 20:10
moritz a different compiler could store a complete distribution or compunitrepo or so in an sqlite db
20:13 mxco86 joined
timotimo right 20:13
sjn that sounds kinda weird to me
timotimo in fact, if it weren.t needed to load NativeCall to bind sqlite, i'd totally dare someone to write that over the course of an evening
sjn doesn't that mean that one can't switch compiler/VM and continue using the modules that are already installed? 20:14
lizmat moritz: yeah, rakudo, you're right :-) somehow I missed the first of 3 options 20:17
timotimo sjn: um, how is the jvm supposed to load .moarvm files? 20:18
i mean, sure, custom classloader, but do we really want to implement a moarvm in java to make that work?
well, different vms can share precomp folders for different formats 20:20
and nothing prevents a different compiler from implementing the same CUR as rakudo has, maybe even using the code almost 1:1
20:27 spebern left 20:30 _28_ria left, BenGoldberg joined 20:31 _28_ria joined
lizmat ok, so I have a patch that will make Set.WHICH 8x faster (for a 50 element Set) 20:33
20:34 as_ left
lizmat it uses nqp::sha1 to create a unique fingerprint of the .WHICH representations of the objects in the Set 20:34
it lets 1 TODO test pass
and it breaks another that specifically checks for a specific .WHICH value 20:35
(which in my opinion is a faulty test)
I figured since nqp::sha1 is good enough for CUR, it should be good enough for set identification
20:37 leont left
timotimo do you do some kind of sorting or something to make sure equal sets get the same which? 20:38
lizmat yes, sorting is involved :-)
El_Che kkk
oops
laggy connection 20:39
20:40 jjido left
timotimo sounds like a good idea to me, to use sha1 20:41
20:42 leont joined
AlexDaniel lizmat: “sha1” and “unique fingerprint” sound a bit weird when put into the same sentence :) 20:44
lizmat AlexDaniel: why ?
:-)
it you're worried about collisions, maybe we should have nqp::shaX that's better :-) 20:45
AlexDaniel that would make me feel better, I guess… 20:46
timotimo .o( love shaX, la la love shaX~ )
lizmat we currently use nqp::sha1 for CUR
.oO( B52's :-)
timotimo we can somewhat easily add another sha* algo in there
then we just generate sha1 and shaWhatever in CUR and if one file exists, mv it to the other filename 20:47
lizmat timotimo: could we have an nqp::sha that takes a list_s ?
so we could bypass the nqp::join step ?
timotimo i think we can do that, yeah
lizmat or would that still concat internally anyway ?
timotimo but i'm really wondering why we even allow Str to be pushed into sha1 at all
no, almost all hash algos support pushing arbitrary amounts of data in piecewise and getting the result at the end (or anywhere in-between) 20:48
20:48 itaipu left
lizmat cool! 20:48
timotimo suggest a name for it?
listsha1? 20:49
though maybe that could also be how we allow passing Buf and friends in?
so ... objsha1 perhaps?
it'd internally dispatch based on the REPR of the thing passed in, then
(can't make the regular sha1 op do that because its signature says it has to be a string) 20:50
20:51 BenGoldberg left
lizmat nqp::sha ? 20:51
timotimo haha, like "sha1 means sha one, sha means sha many"? :)
lizmat hmmm 20:52
:-)
not how I meant it, but yeah :-)
timotimo do we want to give that a parameter to choose hash type, too?
so we don't have to add a new opcode every time we decide to add another hash function?
samcv ok gonna get back to work on trying to make perl 6 appimages 20:57
timotimo i have a sha256 here, but no sha512 to go with it 21:02
21:02 itaipu joined
timotimo github.com/mycelium-com/entropy - this project has a public domain sha256 and sha512 in C 21:02
maybe i should grab those
samcv nice 21:05
timotimo / Total length of the entire message must be provided, as it will be appended 21:07
/ at the end according to the specification.
???
21:08 RabidGravy left, wamba left
timotimo oh, is the length of the message appended to the message before it gets hashed? 21:08
why did i claim you could feed in arbitrary amounts of data? it has to be in terms of the hashes block size for many (including 256 and 512 versions of sha2) 21:09
eeehhh this code also uses an endiannes swapping that they implemented like this: github.com/mycelium-com/entropy/bl...b/endian.h 21:10
and i have no idea if that works with all our compilers?
we surely don't want to have atmel support in there, how is that supposed to compile moarvm anyway :) 21:11
i don't really feel like chasing this around right now :| 21:13
21:18 andrzejk_ left, Sound_ left
lizmat timotimo: sorry, I was distracted 21:26
all ideas sound nice, also the not chasing around right now part 21:27
timotimo hah :) 21:29
samcv timotimo, do you know if moar bytecode will get messed up if the alignment of it changes 21:30
if i replace some directory names inside it
oh looks like it might be... 21:32
ok so with moar i can just set LD_LIBRARY_PATH=./install/lib and not worry too much
but nqp's /share/nqp files are an issue
timotimo oh btw INSTALL.txt still refers to panda 21:33
samcv nice 21:34
timotimo and maybe we need to change more stuff in there?
samcv probably
rakudostar mentions zef
in readme at least
timotimo i'm not sure, but i'd assume changing the length of some strings in the string heap will make things terribly confused 21:35
we do store the offsets of sections near the beginning
21:35 titsuki left, titsuki joined
samcv yeah 21:36
it got angry
gonna go for round 2
gonna --prefix="/r/us" and then gonna use sed to replace that to ./usr
and hope that works
might not for moar. but hopefully for nqp stuff it will 21:37
timotimo well, you see ... strings are stored with their length in front in the string heap 21:38
samcv yeah that is smart 21:39
timotimo if you adjust the sizes of sections, you can rewrite strings in the heap
sizes/offsets/samethingreally
samcv but replacing /r/us with ./usr is much less bad than ./././././././././././. +10 more times
hopefully will reduce our shame.
i am looking at some travis build files for making appimages 21:40
for different projects and things
also timotimo are there uh. ENV vars i can change to make nqp work in another directory like i can for moar
with LD_LIBRARY_PATH
the /share/nqp/blahhhhh.moarvm files 21:41
timotimo uuuuhh
samcv also idk if there's a way we could get *real* moveable support 21:42
by actually not putting the full path in places
for nqp building and such
timotimo well, the paths have to go somewhere. maybe as flags in our launcher shell scripts :\
samcv well yeh they go somewhere
"./usr" or whatever
timotimo did you see the ModuleLoader.nqp inside src/vm/moar? 21:43
samcv yea
21:43 Rawriful left
timotimo certainly if we have a dynamic var that we set from the very main function after the commandline parser gave it to us ... that could perhaps help? 21:44
anyway, how do other langs make their binaries movable? like python?
samcv probably like that
relative directories
linux why you do this :( 21:45
make it so hard
timotimo wait, does it, though? 21:48
samcv uh
not that much i guess you have preload
timotimo if i mv my `which python` somewhere else it'll b0rk?
samcv not really linux's fault
but libraries at least
idk
21:48 robertle left 21:54 ufobat left 21:55 Cabanossi left
samcv wow 21:56
i did it
wooo
it's... working
21:59 Cabanossi joined
timotimo cool! 21:59
you're using sed and stuff? to replace paths and everything?
samcv yeah. 22:00
it's not loading linenoise though
can't find the .so file
timotimo ah
samcv but otherwise works
gonna get a bit more agressive tho
i've been doing it in stages 22:01
timotimo well, that's in perl6-land where we have much more freedom
you know what really, really annoys me?
samcv .moarvm files all can be replaced with ./usr
maybe
timotimo you can't load the contents of a so file. it has to be a file on disk
samcv oppression!
hmm. uh
timotimo i don't want to write my own dynamic linker?!?!
samcv make a fake disk
timotimo what, like with fuse? :)
samcv ok 22:02
now linenoise works
wooooo
ok now to try making an appimage with this
timotimo, is there a command line argument for perl 6 to change paths upon launching? 22:13
m: say $*CWD 22:16
camelia "/home/camelia".IO
MasterDuke samcv, timotimo: can't stick around to chat about it, but is what you've been talking about (relocating rakudo) the same thing mst was working on?
samcv i get "/r".IO 22:17
weird
oh
i am in /r
durp
22:17 jeek left
samcv thought it was \r for a sec 22:17
no clue
timotimo MasterDuke: i think mst was working on things mostly in the makefile 22:19
we're talking about taking an already-compiled rakudo and making it movable to anywhere in the filesystem
samcv: it'd be fantastic if we could get something that also works on windows, so that the next rakudo star msi could get a path selector for installation 22:20
samcv well you have to compile it in a specific directory
well.
./say /u/sr -> ./usr
otherwise theings get weird with multiple levels of ././././
as i experienced
some things don't support it
or something. dunno. it didn't work. and was ugly 22:21
nice. ok i have a script automated to make the install relocatable
now to make appimage tho
gfldex timotimo: it would be nice if you could test meta6 --pull-request. My tests are rather synthetic. (see github.com/gfldex/perl6-meta6-bin#...t-options) 22:24
samcv modules work too timotimo 22:26
timotimo cool 22:28
gfldex: not before bed, i think. but cool thing to have !
it'd probably be cool to have a link to github.com/settings/tokens from the readme 22:29
El_Che lizmat: the green leaf stickers you haveat the booth are mongodb 22:34
22:35 pecastro_ left
gfldex timotimo: the github api can create tokens. I may add meta6 --create-github-token tomorrow. 22:36
22:48 bjz joined 22:49 kurahaupo left 22:51 kurahaupo joined 22:55 Cabanossi left 22:58 Cabanossi joined 23:02 titsuki left, titsuki joined
u-ou is it ok to just say for ITERABLE -> $_ { ... }? 23:02
AlexDaniel u-ou: well, if you're writing an arrow, why not give it a useful name then? 23:03
u-ou oh, right. you don't need to -_-
actually wait. I was thinking of it with whenever 23:04
23:04 rindolf left
AlexDaniel u-ou: but you can write it like that, no problem 23:04
whenever?
u-ou yeah
AlexDaniel like, with supplies?
u-ou like whenever $whatever.Supply -> $_
AlexDaniel oh
23:05 titsuki left
AlexDaniel u-ou: a block is a block, should work just fine 23:05
u-ou ok
23:05 titsuki joined 23:06 kurahaupo left
samcv ok well i made a working perl 6 that doesn't use appimage. but i think i need to change the paths for appimage to work 23:10
i think it wants to "be" in /usr but the binaries and other files to think that they are in ./
so gonna install in /m and then sed it to ./
23:10 bjz left 23:24 kurahaupo joined 23:25 dha joined 23:26 Cabanossi left 23:29 Cabanossi joined
samcv krita has appimages and i hear they're popular 23:36
timotimo krita is great in general
23:37 itaipu left
samcv yeah 23:37
people are lazy. and if they can run it by just downoalding one file. that would be neat 23:38
dunno if installing new modules will work. but. should be great to play around with it for sure 23:39
since you have the star modules
timotimo i'd assume there's a mechanism to store stuff on the user's system
samcv uhm
it's just a program
i dunno 23:40
it colud write to ~/.config if it were like krita or something
timotimo i mean with appimages
samcv inside the appimage?
maybe not
timotimo nah
23:40 cpage__ left
samcv or could just have them downoald a .tar.gz 23:40
and put it wherever they want. and as long as they don't install more modules 23:41
everyhing should work if moved. but if they make a module it'll probably lock it into that directory
maybe. idk.
i guess i'll find out 23:42
timotimo i mean we'd do module installation like krita does image saving 23:44
you wouldn't want to use krita if it can't save away images you've made, right?
samcv yeah i mean
you can write files where you want
it's just a program
timotimo well, it's run in its own little FS, no? 23:45
samcv but can't write to its own folders maybe except in memory
yeah
squashfs
timotimo it's not chrooted?
samcv uh
timotimo like, how does it give you a /usr?
oh wait, squashfs, is that like an overlay over the normal FS?
samcv i do not know
timotimo i mean there's #appimage right here on freenode, we could probably ask ... or maybe i'll find out in the github wiki or something 23:46
ugexe: i still have your Installation2.pm6 gist open :D 23:49
hm. if i make an appimage out of, like, bash or fish ... 23:50
then i could just explore what the inside of an appimage looks like
23:52 _28_ria left 23:53 _28_ria joined 23:54 _28_ria left 23:56 _28_ria joined 23:57 kurahaupo left, kurahaupo joined 23:58 kurahaupo left, kurahaupo 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!