»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.
00:01 _28_ria left 00:02 _28_ria joined 00:03 kurahaupo_ joined 00:05 kurahaupo left, _28_ria left 00:06 _28_ria joined
ugexe heh, its all in rakudo or the ecosystem now 00:06
00:07 kurahaupo_ left 00:08 titsuki left, titsuki joined, _28_ria left
timotimo neat. 00:09
00:10 _28_ria joined 00:21 _28_ria left 00:23 _28_ria joined
samcv oh em gee 00:24
it's working 00:25
\o/
timotimo wow 00:26
samcv now people need to tell me if they work on other linux distros. cry.nu/files/org.perl6.rakudo-x86_64.AppImage
ooooooooohhhhhhhhh yessssssssss
timotimo i was kind of hopeful when i mentioned appimages, but you actually made it happen and that's fantastic
samcv tell me if it works on your system. i have arch btw
i can't believe i got it to work either
cause their documentation sucks
pretty bad
had to figure it out 00:27
super awesome
now what should i try
timotimo imgur.com/a/EG9dE 00:28
samcv nice
timotimo this is a fedora system
samcv no nobody will have an excuse to not try perl 6 00:30
*now
not sure how one would call zef. but it *is* installed inside the appimage 00:31
timotimo, does linenoise work properly? 00:32
also timotimo is there a delay starting it up? 00:33
timotimo linenoise works
samcv yeah
i ran the prompt before making the appimage so it's hopefully compiled in there 00:34
timotimo there's a delay, yeah
samcv dunno if it's frome linenoise or not
or the appimage
is there a way to tell if it has to compile linenoise each tme it starts?
timotimo yeah, the module debug env var
RAKUDO_MODULE_DEBUG=yes 00:35
samcv looks like it thinks its program name is ././share/perl6/runtime/perl6.moarvm
if i do --help
timotimo hehe
samcv maybe i can fix that idk
timotimo well, we can have the --exename or what it is
--execname 00:36
samcv yeah
timotimo it's already $0 in the perl6 script, but that might not go through whatever appimage does to make this work?
samcv uhm
it calls perl6
yeah it has execname=$0 00:37
can maybe clean it up in the script to not have ././
sammers samcv: works on debian sid for me
samcv nice 00:38
sammers not working with firejail, however...
samcv idk how you use that
sammers me neither
samcv :P
sammers :)
samcv time to make a new git repo for this stuff
sammers so all the modules are in the image? 00:39
samcv yes
all rakudo star modules are in there
if you want more, you have to go online and purchase an activation code
sammers I guess if we have a simple way to build an app image with our deps
samcv heh
aka install it normally
sammers oh, can't wait 00:40
samcv i'm curious if it works on like debian jessie
sammers just a sec... I will test
samcv yay
i should probably write the bash script to make the appimage today otherwise i'll probably forget all the weird things 00:41
i mean it's not, that hard. if it was properly documented
sammers yeah, worked right away. 00:42
samcv sweet
sammers base debian, just installed fuse and it worked 00:43
samcv soon we will be trendy and cool
sammers bootup is slow
samcv and everybody will dl the appimages
starting the script?
timotimo hm, gotta help a friend by providing my wheels
sammers but once it's up it seems ok
00:46 _28_ria left
timotimo sammers: maybe you can be inspired by github.com/jnthn/p6-app-installermaker-wix 00:46
samcv can you give me the output of
echo exit | RAKUDO_MODULE_DEBUG=yes ./org.perl6.rakudo-x86_64.AppImage
that
will go to stderr 00:47
on the debian jessie one or whichever older debian it was
sammers ok, please stand by... for the Wix side, I don't have windows...
timotimo that's not important
00:48 _28_ria joined
samcv hmm timotimo 00:48
won't it tell me if it recompiles things
timotimo it's only important that you give it your module and a launcher so to speak and it'll make an .msi that installs a rakudo along with your module and its dependencies and gives you a nice clickable thing to launch it
samcv or is that not possible
timotimo gotta go now!
samcv oh talking about wix
also: time echo exit |./org.perl6.rakudo-x86_64.AppImage 00:49
00:50 wamba joined
samcv need to adjust some things so you can double click on it and it will just open a prompt 00:52
sammers samcv, is there an option to have the debug output write to file? > isn't working...
gist.github.com/scmorrison/5c0c8de...896af6912d 00:54
samcv cool 00:55
ok i think i uh 00:56
messed up the checksums of the files. derp
well it still works i guess
hmm maybe will have to go through and replace checksums too. that sounds like a pain
sammers samcv, just added the time output to the gist
samcv kk thx
ok well it looks like it just precompiles them into the home directory 01:01
it's not too slow just doing -e or whatever 01:02
and people can run their files or whatever with it fine
sammers samcv, I can pipe a script to the appimage, is there any other way to pass a script? 01:07
samcv just act like it's a perl 6 binary 01:08
should work
works great actually 01:09
sammers it isn't working on my setup....
I get these messages, "Could not open deathstar.p6. Failed to stat file: no such file or directory"
samcv hm
sammers I will try from another system...
samcv yeah it works for me. maybe it's dependencies idk 01:10
wasn't working on a jessie debian install?
but REPL worked?
sammers ok, so no go for jesse either 01:12
repl works
samcv hm 01:13
timotimo i need to test appimage foo.p6, too
though it would be amazingest if we had a little gtk gui for a bunch of different things you can do
samcv with perl6? 01:14
timotimo write scripts, install modules, repl, docs
could make it in p6 with gtk or as a web app
on the road again bbiab
samcv will it have a camelia avatar, like clippy
timotimo it would be catastrophic to forget that 01:15
sammers like a perl6 command center
samcv it recommends on appimages site mounting distro iso's and using those in a chroot or whatever
to test out your images on multiple distros
sammers how about travis / docker? 01:24
timotimo oh samcv you should totally check out azawawi's farabi6 01:27
it's a web editor with a bunch of extra ide-like features
wow,there was a modernization commit 2 months ago 01:28
*neat*
i was afraid it'd have bitrotted so much it wouldn't start 01:29
oh, no actual code changes i don't think
now i'll go to bed 01:30
samcv night 01:44
01:45 awwaiid joined, ilbot3 left 01:46 BenGoldberg joined 01:48 g0d355__ left, ilbot3 joined, ChanServ sets mode: +v ilbot3 01:50 dha left 01:56 Cabanossi left 01:58 Cabanossi joined 02:25 noganex joined 02:28 noganex_ left 02:47 perigrin joined, wamba left 02:55 Cabanossi left 02:59 Cabanossi joined 03:03 Luna521 joined 03:11 Luna521 left, Luna521 joined
samcv \o/ i got relative directories working 03:20
03:27 Luna521 left 03:30 Luna521 joined
samcv
.oO(time to run roast using appimage)
03:33
03:41 jdmmmmm joined, Cabanossi left, Cabanossi joined 03:42 Luna521 left 03:45 khw left 03:47 jdmmmmm left
samcv oh no wonder. it's failing all the .rakudo.moar tests the ones that don't send it in as .t 03:51
not sure why those exist? 03:52
03:57 AlexDaniel left 04:02 jeek joined
samcv and i'm sure we can get standalone perl 6 apps working with these appimages too 04:05
04:09 cpage joined, jdmmmmm joined
samcv ok i have fixed perl6 -I t/spec/packages testfile.t # which is found several places in roast. 04:16
woo.
04:27 skids left 04:33 AndyDee left 04:39 jdmmmmm left 04:48 kurahaupo joined 04:55 Cabanossi left 04:56 curan joined, BenGoldberg left 04:58 Cabanossi joined 05:11 Actualeyes left 05:13 Actualeyes joined 05:17 domidumont joined, esh_ left 05:19 xiaomiao left, esh joined 05:21 domidumont left 05:22 domidumont joined 05:25 xiaomiao joined, xiaomiao left, xiaomiao joined 05:27 wamba joined 05:39 Cabanossi left, domidumont left 05:41 Cabanossi joined 05:42 domidumont joined 05:46 rindolf joined 05:54 RabidGravy joined 05:56 g0d355__ joined 06:00 ufobat joined 06:10 gdonald left, gdonald joined 06:15 _28_ria left 06:17 _28_ria joined 06:21 spebern joined
andrzejku hou hou :) 06:34
06:36 domidumont1 joined 06:39 domidumont left 06:48 spebern left
samcv woo i think i got travisci building appimage 06:48
so that means i guess my script is probably fully functional now
06:53 lowbro joined, lowbro left, lowbro joined 06:55 Cabanossi left 06:58 Cabanossi joined 07:01 ufobat left 07:04 abraxxa joined 07:09 darutoko joined
samcv would be cool if we could have an appimage for the next rakudo star 07:10
07:11 abraxxa left, abraxxa joined 07:12 domidumont1 left 07:16 ufobat joined 07:21 domidumont joined 07:31 kurahaupo left, kurahaupo joined 07:37 parv joined 07:38 CIAvash joined, mxco86 left 07:39 jjido joined 07:44 jjido left 07:45 wamba left 07:46 mxco86 joined, mxco86 left, mxco86 joined 07:59 dakkar joined 08:02 dakkar left 08:04 jonas1 joined, dakkar joined 08:18 robertle joined 08:24 rodarmor joined 08:25 Cabanossi left 08:27 Cabanossi joined, wamba joined 08:29 g4 joined, g4 left, g4 joined 08:31 bjz joined 08:40 rindolf left, mr-foobar left 08:41 nadim_ joined
nadim_ morning p6 08:42
timotimo: do you think you could answer my question on the patch you proposed? so I can merge it (after I understand) 08:43
08:46 rindolf joined
stmuk_ samcv: oddly I was thinking about that as well! 08:46
samcv nice
travis-ci.org/samcv/rakudo-appimage/ if you wanna see
there's a dl under releases as well 08:47
stmuk_ oh excellent I've never actually used appimage but I noticed the oddly named "otterbrowser" used it
samcv it's pretty cool once i got it to work. and it's star so has all the modules built in 08:48
stmuk_ and if you have mostly done it :) bookmarked
samcv had to fix loads of path problems and stuff but finally worked it
was difficult
stmuk_ samcv++
samcv well it's totally done
afaik. only thing is that it doesn't drop you in the pwd you launched it in. but you can do -I lib or relative files since the perl6 script resolves it 08:49
before passing it on
and some other things
it's badly documented. but i created random weird folders and then dumped %*ENV and grepped it to figure out the ENV vars that would give me the pwd 08:50
it goes into that pwd (since it starts in its own virtual system) then it resolves the directories and files on the command line args, then it goes back to where it started so the paths work
and it patches all the .moarvm files and changes the paths to be relative 08:51
it installs into --prefix /rsu then you replace /rsu with ././ (same number of bytes)
though it breaks some checksums maybe. but it precomps into the users home folder so launches after that shouldn't take too long 08:52
stmuk_ hmmm maybe upstream should support relative paths
anyway real work calls :\ I'll take a look shortly 08:53
08:54 cgfbee joined 08:59 ChoHag joined, Sound joined 09:00 Sound is now known as Guest67138, gdonald left, gdonald joined
samcv well what we really need is being able to start perl6 in a specific directory 09:01
so that $*CWD will be the expected directory
dunno if perl 5 has anything like that, or we make our own option for that 09:02
09:02 Guest67138 left 09:10 Ven joined, Ven is now known as Guest52333 09:15 bjz left 09:16 bjz joined 09:25 sjn left 09:26 zakharyas joined, sjn joined 09:28 Guest52333 left 09:30 jonas1 left 09:31 jonas1 joined 09:35 kurahaupo_ joined 09:37 kurahaupo left 09:38 Ven_ joined, mr-foobar joined, Cabanossi left, kurahaupo__ joined 09:40 kurahaupo_ left 09:42 Ven_ left 09:43 kurahaupo__ left, Cabanossi joined 09:44 wamba left
nadim_ timotimo: found the link on the other ticket. 09:45
09:46 Ven_ joined 10:09 wamba joined 10:23 ChoHag left 10:25 ChoHag joined 10:27 Ven_ left, Ven joined 10:28 Ven is now known as Guest42930 10:30 kurahaupo joined 10:31 kent\n left 10:32 kent\n joined 10:33 astj joined, astj_ left 10:35 kurahaupo left 10:36 astj left, astj joined 10:38 astj left 10:39 astj joined 10:44 Guest42930 left 10:45 Ven_ joined 10:46 kurahaupo joined 10:52 kurahaupo_ joined 10:53 kurahaupo left 10:54 pmurias joined 10:56 kurahaupo joined, kurahaupo_ left 11:03 labster left 11:04 labster joined, labster left 11:11 parv left 11:14 Ven_ left 11:15 Ven_ joined, kurahaupo left 11:20 kent\n left 11:22 kent\n joined, zakharyas left, TEttinger left 11:24 domidumont left 11:30 xtreak joined 11:44 Ven_ left 11:47 xtreak left, xtreak joined
robertle when using an actions class with a grammar, the action methods seem to get called as the grammar thingie is working it's way through the input, including the backtracking. so it executes the action and throws away the result? do I undertsand that correctly? is that desired? 11:50
gist.github.com/anonymous/0c156ced...34fc6d03e9
says "term action" three times, if you remove the "<term> '&' <expression>" line, then it calls it twice 11:51
11:51 xtreak left
timotimo yeah, it does call the action whenever something finishes parsing, but it can't know if the thing that uses the something will end up using it or not 11:51
robertle that is quite surprising, to me at least. is that desired? is that going to stay that way? 11:52
timotimo action methods are called "as things parse", the only other option would be to run all action stuff right at the very end of everything 11:53
but in that case you can also just put only a TOP action method and explore the match tree from that place 11:54
11:54 Cabanossi left
robertle it is surprising because most other parser generators allow actiosn embedded with the grammar, but only execute them after the engine has committed on a rule/token, not during lokahead 11:55
11:55 Ven joined, Ven is now known as Guest11277
timotimo i wonder if we can determine that somehow 11:56
11:56 Cabanossi joined
masak robertle: it falls on the grammar author not to backtrack in such a way that rules that succeeded and whose action methods ran subsequently get thrown away 11:57
or, perhaps more simply stated, once you're writing a grammar and not just a regex, you're meant to backtrack as little as possible 11:58
robertle masak: yeah, one of the reasons I am asking is whether perhaps this is just a consequence of an awkward grammar
masak here's how I like to think of it: for any nontrivial rule, there tends to be a "point of no return" 11:59
after which a parsefail should be a *panic*, not lead to backtracking
11:59 geekosaur left
robertle however for the language in this example, I can't see how to do that without backtracking. or for any non-trivial (= non-recursive) language 11:59
11:59 Guest11277 left
masak I mean, these are rules of thumb and there are exceptions to everything 11:59
robertle I would be very interested in how to express a grammar for that language in a non-backtracking way
masak I'd be happy to discuss your case in particular 12:00
but I'd like you to consider my point above about the "point of no return"
robertle masak: see above for paste, it's a trivial language
12:00 wamba left
masak looks 12:00
12:00 geekosaur joined
robertle you can even remove the "'!' <term>" line 12:01
masak yeah, that can totally be done without backtracking :)
you just need to refactor it a bit
something like `rule expression { ['!'? <term>]+ % ['|' | '&'] }`
the `%` op is very nice for separator-y bits 12:02
12:02 wamba joined
masak also, I'm never 100% I get the whitespace right until I back the grammar up by tests. YMMV 12:02
anyway, I guess you can see how that makes it work without backtracking 12:03
robertle right, I did see samples using both the % and the combination of multiple cases into one rule...
masak I realize you might want to factor it slightly differently yet if you plan to grab the '!' in the actions
but that should get you going
robertle thanks, I will play with that a bit 12:04
masak fun fact: we had an insidious scoping-related bug in 007 parsing because we didn't properly consider the "point of no return" thing
(i.e. side effects had time to run before backtracking happened and a completely different rule succeeded) 12:05
robertle essentially you are saying rather than using "|" in the rule, which means options and therefore potentially backtracking, use any regular sub-language that you can
masak that was "fun" to debug
robertle I think grammars are still "fun" sometimes
masak robertle: I'm saying specifically there's a way here which leads to less waste and backtracking, and should get you into less trouble generally
robertle one thing that I notice is that errors in my grammar often lead to the parser just hanging on an infinite recursion or so. that is a bit strange, should it ever build such a grammar? 12:06
masak this tends to happen when you have rules that both (a) are quantified and (b) can match zero characters 12:07
moritz robertle: that's a known error which... what masak said
masak there are ways to write a grammar *engine* such that that never happens
moritz there are also ways to structure your grammar to avoid that
masak but Rakudo's grammar engine will happily spin forever
robertle ok, understood 12:08
moritz which typically involves chaning token foo { .* } and then <foo> to token foo { .+ } and then <foo>?
12:08 leont left 12:11 leont joined 12:14 gdonald left 12:16 Ven_ joined 12:21 domidumont joined 12:23 Peter_R left 12:25 Peter_R joined 12:26 xtreak joined 12:27 regnarg joined, regnarg left 12:34 gdonald joined 12:36 CIAvash left, perlpilot joined, xtreak left 12:38 AlexDaniel joined 12:40 plicease left 12:44 Ven_ left 12:48 zakharyas joined 12:50 Ven_ joined 13:01 mcmillhj joined 13:11 mcmillhj left 13:13 gdonald left, gdonald joined 13:14 Ven_ left 13:15 mcmillhj joined 13:16 perlpilot left
pmurias masak: re write a grammar *engine* such that never happens, just like with writing an *interpreter* that never loops you would need to change the grammar language 13:16
13:16 Ven joined, kurahaupo joined
Geth doc: 0ca22958e7 | (Zoffix Znet)++ | doc/Type/IO/Path.pod6
[io grant] Reword/expand IO::Path intro prose
13:16
13:16 Ven is now known as Guest90388
masak pmurias: I don't think I'm disagreeing with that -- it would change the language 13:18
pmurias: here's an example of what I mean 13:20
perl -wE'say "abc" =~ /(a?)+bc/'
13:20 kurahaupo left
masak prints the empty string (for some reason) but doesn't look infinitely 13:20
the corresponding p6 program loops infinitely
but I mean, yeah, of course "fixing" the infinite loop also changes the semantics of the grammar language 13:21
13:21 kurahaupo joined 13:22 kurahaupo left, kurahaupo joined
robertle but surely the grammar engine can tell whether the grammar in question can loop infinitely for any finite input, and if so reject the grammar 13:23
masak ...aha. the p5 program does match, it just prints what looks like something empty as a result 13:24
ilmari because it returns the last iteration of the capture, which successfully matched the empty string after the a
masak robertle: not sure that's so. but even if it were, such analysis is not done in Rakudo at present
ilmari: ...right.
ilmari: so it matches the empty string *once*, but then figures out it's not progressing? 13:25
ilmari perl5 advances the position if a zero-width match matches in the same place twice
Geth doc: 45e84addef | (Zoffix Znet)++ | doc/Type/IO/Path.pod6
[io grant] Move IO::Path.path to attributes

Also add a warning about trying to mess with the attibute.
masak ilmari: feels kind of like a "forward-tracking" mechanism :P
13:26 plicease joined, skids joined 13:28 plicease left, plicease joined, plicease left
ilmari beats infinite looping 13:28
13:28 plicease joined
masak no argument there 13:28
masak .oO( of course, none of this would've happened in the first place if we all had just listened to Russ Cox ) :P 13:30
moritz but then we wouldn't have p6 grammars either :-)
masak we might've, but we might be talking about "procedural suffix" instead :P 13:31
13:33 curan left 13:37 nhywyll joined
Geth doc: b9de84f2b1 | (Zoffix Znet)++ | doc/Type/IO/Path.pod6
[io grant] Remove DateTime tutorial from IO::Path docs
13:43
13:46 wamba left
Geth doc: 45074858b8 | (Zoffix Znet)++ | doc/Type/IO.pod6
Fix broken URL
13:49
13:51 gdonald left 13:54 gdonald joined
Geth doc: 69b2082dc2 | (Zoffix Znet)++ | doc/Type/IO/Path.pod6
[io grant] Document IO::Path.chdir

  - Include deprecation notice for 6.d language
13:54
13:55 raschipi joined, leont left 13:56 wamba joined 13:59 leont joined 14:01 Actualeyes left 14:12 Guest90388 left 14:13 vendethiel left 14:16 jdmmmmm joined 14:25 Cabanossi left 14:28 Cabanossi joined 14:31 andrzejku left 14:32 raschipi left 14:37 cdg joined 14:39 cdg left 14:40 cdg joined 14:42 alimon joined 14:44 xinming left 14:45 xinming joined
[Coke] yawns. 14:47
14:48 ChoHag left 14:49 khw joined, ChoHag joined 14:50 zakharyas left
Geth doc: d436f3cc78 | (Zoffix Znet)++ | 5 files
[io grant] Document IO::Spec::* don't do any validation

Addresses RT#130900:
  rt.perl.org/Ticket/Display.html?id=130900
14:52
synopsebot6 Link: rt.perl.org/rt3/Public/Bug/Display...?id=130900
14:54 bjz left
pmurias [Coke]: re autofudging mostly passing tests I think it doesn't make much sense for rakudo.js as it's a (currently) more buggier rakudo rather then a independently created implementation 14:54
so if it fails a roast test it's most likely because of a bug rather then some minor unimplemented feature 14:55
14:57 cognominal joined
[Coke] pmurias: ok. 14:59
15:08 domidumont left 15:12 ChoHag left 15:13 wamba left 15:20 gdonald left, gdonald joined 15:21 ChoHag joined 15:22 nightfrog is now known as culb, xiaomiao left 15:23 cdg_ joined, sufrostico joined 15:24 cdg left 15:28 xiaomiao joined
Geth doc: 409044695b | (Zoffix Znet)++ | 2 files
[io grant] Improve chmod docs

  - Split sub and method forms, as they're very different.
  - Remove $*CWD example; bad form without `temp`ing it first.
  - Add examples for chmoding dirs too
  - Add examples with an octal in a string
15:35
doc: ca4a5d8fb6 | (Zoffix Znet)++ | doc/Type/IO/Path.pod6
Fix up example

Use method form, since the docs are for method form
15:37
15:39 Cabanossi left 15:41 sufrostico left, gregf_ joined 15:42 Cabanossi joined 15:43 sufrostico joined 15:45 rindolf left 15:56 abraxxa left, lowbro left 16:04 nadim_ left 16:05 robertle left, mcmillhj left 16:07 Dunearhp left 16:19 wamba joined 16:20 mcmillhj joined 16:21 dakkar left 16:22 g4 left 16:25 mcmillhj left
ugexe github.com/CurtTilmes/Perl6-GraphQL 16:25
16:27 sufrostico left, zakharyas joined, sufrostico joined
Geth doc: 1527d321ba | (Zoffix Znet)++ | doc/Type/IO/Path.pod6
[io grant] Document :completely arg to IO::Path.resolve

Rakudo impl: github.com/rakudo/rakudo/commit/6a8d63d1b6 Tests: github.com/perl6/roast/commit/a809f0f1f5
16:31
16:31 mcmillhj joined 16:36 mcmillhj left 16:42 st_elmo joined 16:45 mcmillhj joined 16:47 robertle joined 16:48 raschipi joined 16:49 cdg_ left, cdg joined 16:50 raschipi left, mcmillhj left
Geth ecosystem: 354d3b3376 | (Nick Logan)++ | META.list
Add GraphQL

  github.com/CurtTilmes/Perl6-GraphQL
16:55
16:58 pmurias left
jsimonet Hello, I would like to know how objects destructors works in Perl6. I found the submethod DESTROY but is only called by the GC. Is there a way to properly close a resource before the end of the program/a block? 17:01
17:01 jonas1 left 17:04 mcmillhj joined 17:07 titsuki left
timotimo well, for blocks you can use LEAVE blocks 17:07
which will get run for any kind of leaving
though not when the program exits
17:08 titsuki joined 17:10 TimToady left 17:12 TimToady joined
jsimonet Thank you. In objects, this doesn't work ( class Foo { LEAVE { note "leaving" } }; for ^100 { my $f = Foo.new } . It prints "leaving" only once, since LEAVE is for blocks. 17:12
timotimo yup 17:13
the body of a class gets run, but not when you instantiate the class
jsimonet I'm trying to develop a project where I have a channel, and some operations are done on the channel data. I would like to manage these data when the program ends. 17:15
(asynchronously) 17:16
timotimo have you heard of "react" blocks? they manage subscriptions for you. not 100% sure how they interact with channels, i've mostly used them with supplies and promises 17:18
jsimonet yes, but i'm currently using Channel.Supply.tap. I suppose they work the same way? 17:20
I wrote a previous version with react, but I needed to write something like 'start { react { whenever Channel { do-stuff; } } }' 17:21
Geth perl6-most-wanted: CurtTilmes++ created pull request #38:
Remove GraphQL
17:23
perl6-most-wanted: 31795a73ac | (Curt Tilmes)++ | most-wanted/modules.md
Remove GraphQL (#38)

Now available: github.com/CurtTilmes/Perl6-GraphQL
17:24
timotimo not the same way, i.e. you'll have to do subscription management yourself, which is what the react block would give you "for free" 17:26
subscription management meaning, among other things, what happens when channels close or throw errors or some such 17:27
17:29 domidumont joined
jsimonet ok, with QUIT phaser as mentionned in the concurency page :) 17:30
timotimo for example, yeah 17:33
timotimo BBL 17:34
jsimonet So, I think I'll have to wrote a "flush" method for the object instance to flush and close the resource, and in case of a crash, the data will be lost. 17:36
Thank you for your responses BTW :) 17:37
17:38 labster joined 17:54 Cabanossi left 17:57 Cabanossi joined 17:59 sufrostico left 18:01 sufrostico joined 18:04 cdg left 18:07 zakharyas left 18:10 gdonald left, gdonald joined
timotimo sure 18:14
DrForr o/ 18:18
timotimo yo 18:22
RabidGravy BOOM! 18:23
18:24 Cabanossi left
RabidGravy I scared myself today, I almost suggested a distributed transaction manager in a meeting, but I bit my tongue 18:25
DrForr Don't they call those databases? :)
Also good timig.
*timing 18:26
RabidGravy twas timig and the slithy toves
18:26 Cabanossi joined
DrForr gimbls. (keeping with the theme) 18:28
RabidGravy nothing like a spotting of gimballing on a Monday evening 18:31
timotimo i have no idea what you're talking about :) 18:33
DrForr: i may have asked this before, but does anything keep perl6/doc from using Perl6::Parser for pod? 18:34
RabidGravy Jabberwockey
lizmat starts on the Perl 6 Weekly
timotimo i meant the transaction manager thing :D
perigrin RabidGravy: is that a borogrove in your pocket or are you just mimsy to see me?
RabidGravy :) 18:35
perigrin flees before a jubjub bird eats his children.
samcv good morning o/ 18:36
RabidGravy I actually pulled myself back from the brink of making a binding for Enduro/X the other week
DrForr timotimo: You're welcome, but at the moment it just treats it as block of text, so you won't get much in the way of detail. The POD needs to be reentrant, and it's not yet. 18:37
18:37 jjido joined
DrForr *blocks - The keyboard on this laptop must have been affected a little more than I thought by the screen crack... 18:37
18:39 jjido left
timotimo reentrant? 18:39
DrForr I.E. use Perl6::Pod's parser. 18:40
timotimo you mean rakudo's, or Perl6::Parser's POD Parser? 18:42
DrForr The former.
Perl6::Parser is something of a misnomer as it doesn't actually parse anything. 18:43
timotimo oh, huh?
i thought the whole project was there so we can parse perl6 code without executing anything :) 18:44
18:44 perlpilot joined, domidumont left
DrForr If I were to do that I'd never get around to actually getting anything done. 18:45
Also you practically have to do that in the presence of custom operators. 18:47
timotimo damn~ 18:49
DrForr Script A can use custom operators that are created conditionally in a file that gets use'd, which means I'd have to load all of _that_ code and parse it. 18:51
18:53 xiaomiao left
unclechu how do i disable stdout buffering in perl6? like i do in perl5: `$| = 1;` 18:53
timotimo unclechu: i think you use the termios module for that?
geekosaur that doesn't disable whatever buffers libuv has 18:54
18:54 nadim_ joined
DrForr I'd love to be able to statically search for '{sub|multi} infix:<+>(){}' but if I search for the static text I'll likely miss variants, and why should I rewrite Perl6::Grammar when it's already been written? Not to mention it's pretty much impossible to untangle the parsing from the compilation portions, altough I really didn't get much beyond pulling out files and utterly failing to find entry points. 18:55
The only thing that actually *executes* are the 'use' statements and BEGIN blocks, and I'm happy with that tradeoff for not having to rewrite the entire parser and .. well, never get anything actually done. 18:56
18:59 xiaomiao joined
DrForr I realize the sound I just heard was a crest falling, but it was a choice between the seemingly Sisyphean task of rewriting a parser that's already *there* and keeping it up-to-date with changes in the main code *and* getting it to output useful data vs. getting something done. 19:01
19:01 g0d355__ left
RabidGravy I don't suppose anyone knows how big __syscall_slong_t is ? or do I have to ask a compiler 19:04
long int apparently 19:07
alphah hello perl6, I'm trying to use look-around assertions in my grammar to to arange tokens order, which comes after which and eventually to meet the test cases in the grammar, for some reason when I use look-around assertions some tokens don't match, more details in the comments on code gist.github.com/anonymous/f563aee0...90ff0afc5b 19:11
19:11 raschipi joined 19:16 raschipi left, sufrostico left
DrForr alphah: It's probably incorrectly failing because version can look like tag. 19:19
moritz uhm, does using <( and )> in subregexes work? 19:20
alphah I can make tag token to not start with digit and will test that 19:21
moritz (and suggesting to use .split('-') likely won't make you happy, right?)
DrForr And I'd really encourage you to use the Test module.
"ok Pkg.parse('rakudo-star'), 'name';" # Simpler way to write your tests. 19:22
alphah mortiz: I'm not sure what you mean, but I use <( and )> to limit the regex to what is in between
mortiz: if .split(-) will work sure will make me happy 19:23
19:23 nadim_ left
alphah DrForr: I will look into the test module, perl pilot helped me for this test cases (becuase I was writing them in comments and wa sdoing the tests manually) 19:24
DrForr Give me a sec and I'll finish rewriting this. 19:25
19:25 darutoko left
DrForr gist.github.com/drforr/f424f8187f4...a02592ddb7 19:26
alphah Thanks DrForr I will use that 19:27
DrForr That replaces your block and test loop, and also works with Travis integration.
timotimo DrForr: so what does Perl6::Parser actually do? i'm confused. 19:28
DrForr Like I said it's a bit of a misnomer, in that instead of parsing everything on its own, it uses Perl6's internal parser. 19:29
timotimo right
and then? :)
samcv not sure if it's my system or not but getting linenoise errors with appimage here: github.com/samcv/rakudo-appimage/releases
not sure if it's my system or not though. cause i am pretty sure the exct same one worked for me before 19:30
DrForr Spits back a tree of Perl 6 objects that correspond to your perl code.
samcv can somebody test it for me? dl the org.perl6.rakudo AppImage, chmod +x then run it. should pop up perl6 repl prompt with no errors
19:37 ChoHag left
DrForr alphah: gist.github.com/drforr/f424f8187f4...a02592ddb7 updated and passes your basic test suites. 19:39
19:40 Cabanossi left
DrForr Crude, and expanding package-name to ['-' \w+] doesn't do what you want. But it's a different approach. 19:40
19:41 avalenn joined
DrForr Though it's apparently an issue with \w also containing \d. 19:42
alphah DrForr: tested that and worked. 19:43
19:43 Cabanossi joined 19:44 sufrostico joined
alphah yeah I'd probably keep package-name as regex, as some cases will be "package" , "long-package" , "very-long-package-name" 19:44
DrForr Let me paste this...
gist.github.com/drforr/f424f8187f4...a02592ddb7 - fixed and updated so it'll also work with rakudo-perl6- # as well. 19:46
alphah Awesome Thank you! 19:47
19:47 mcmillhj left
DrForr The 'arch' token will need expanding as well, doubtless. 19:47
19:48 mcmillhj joined 19:49 pecastro joined
DrForr All I ask is that it end up on the ecoystem eventually :) 19:51
19:53 rindolf joined
alphah sure if it becomes decent enough :) 19:56
19:57 jjido joined
DrForr Done, and done. 19:57
20:00 cdg joined
RabidGravy DrForr. while you're on a roll - I've got these modules ..... ;-) 20:02
20:03 jjido left
samcv guys how do you feel about having appimages created every day for moar-blead 20:09
that would be cool
looking at setting up github pages and having travis upload them to github pages
gfldex is there anything travis cant do? o.O 20:10
samcv yes
have dependencies sanely installed and use an up to date distribution 20:11
it can't do that
El_Che samcv: funny you ask
AlexDaniel
.oO( travis is not doing allah is doing )
gfldex also, its api can't sync the repo list from github
El_Che samcv: I spent a few days reading stuff on Flatpack, Snap and appimage 20:12
I wasn't impressed :(
samcv well
gfldex what happend to good ditro packages?
20:12 sufrostico left
El_Che gfldex: building ubuntu 17.04 i386 as we speak 20:12
gfldex El_Che++ 20:13
El_Che I extremely dislike the Apple Appstore model of snap (and it only works as advertised on Ubuntu)
samcv idk krita has appimages. and they're trendy and cool. maybe we can be trendy and cool too :S
El_Che (I am an Ubuntu user)
samcv hahaha
oh yeah 20:14
snap sounds lame
flatpak seems nicest, but you still have to install something
El_Che thing is, it 'works' elsewhere
but whitouth the security sandboxing
samcv but appimages don't need any install
El_Che so lame
samcv oh flatpaks? i thought they needed something to be installed?
El_Che flatpack only works on fedora :)
they have framework deps 20:15
so it's not completely self contained
which is not bad as such
I was reading up on appimage, but it never really went mainstream 20:16
and all the press is now for snaps and flatpack
AlexDaniel samcv: by the way, any chance of putting zef into it?
samcv idk appimages are most mainstream
El_Che (it does not have sandboxing either)
samcv it has zef in it
atm you can only launch perl 6 but you can use any * modules
El_Che I think appimage is the most interesting of the 3
samcv yeah. tho it still sucks 20:17
well not that bad but
20:17 cyphase left
samcv flatpak is better 20:17
but appimages are more practical in $CURRENT_YEAR
El_Che I wonder though if app image is the right format for a programming lang
samcv why not
El_Che it is for a perl6 app
samcv you dl it and you run it
El_Che bundling the runtime and modules
samcv and can launch scripts with it or use the repl
i mean. El_Che i can make a .tar.gz that you extract and can run 20:18
El_Che because of modules
AlexDaniel … and then if you need to install a module, then…
samcv you have to launch it from a certain path though
El_Che you need to integrate it with external stuff
samcv but i guess i could make a helper file or something like i did with appimages
El_Che while an application is self contained
samcv i mean.
would just a rakudo that you unarchive and run
would anybody want that
cause i can do that too. 20:19
but appimages seem a good way to get people to try perl6 out
10mb dl, run it. then if you like it you can install it fully
El_Che as a linux user, I would *always* prefer native packages, and if not available I will be happy with the distro independent tar.gz (like Go does)
this said, my pkgs are really lacking a repo to handle upgrades automatically 20:20
I really need to look into that
AlexDaniel samcv: why is it 10MB, by the way?
I thought it would be around 7 20:21
samcv it has all modules
of rakudo star
idk
AlexDaniel what does it use for compression?
samcv uh
the same thing linux iso's use
El_Che a rakudo star app image
nice
samcv El_Che, try out the one i have under releases
github.com/samcv/rakudo-appimage/releases 20:22
i think it works. it worked for me yesterday. but i think i broke my fuse
cause it works on travis
maybe it's just my build and i should build somewhere else. we will see
El_Che samcv: would it work in a Docker container? (I don't want to break stuff, just migrating to Gnome from unity :) ) 20:23
20:24 Cabanossi left
samcv dunno 20:24
sounds scary 20:25
idk appimages seem the least hassle for the casual person checking out perl 6
idk my hope is to get people to try it since they don't have to compile it etc etc
El_Che samcv: do you need to install something to support kik? 20:26
samcv appimages?
El_Che the original name
samcv you need nothing to support appimages other than having fuse installed. which like everything do
?
El_Che I like to call stuff by their old name like old people do
20:26 Cabanossi joined
El_Che how is life in the colonies? 20:26
:)
ah fuse 20:27
AlexDaniel samcv: OK I have question! If I have rakudo installed somewhere, can I easily turn it into appimage?
El_Che so they changed it
AlexDaniel samcv: I'm thinking of sharing all of whateverable builds 20:28
samcv ALL?
like all builds of rakudo ever
AlexDaniel well, not all of them together, by one by one :)
El_Che they used to do a regular iso mounting so they e^Hwere limited to 9 images in totoal
AlexDaniel so that you can ask a bot to give you an image of some particular build, yes
El_Che crappy connection here 20:29
AlexDaniel but*
20:29 Sound joined, Sound left, Sound joined
El_Che I see a good use case: app bundling 20:29
the bad thing is that server probably don't have fuse 20:30
20:30 _28_ria left
El_Che deploying perl6 like java or go would be sweet 20:30
AlexDaniel samcv: 「echo "Replacing path in binaries"」 oh wow…
samcv lolllllllllllllllllllllll
yep
that's why i install into /rsu so i can replace it into ././ (same number of bytes) 20:31
ninja
AlexDaniel yea I noticed that
can't it be any other path?
samcv if there's too many ././ then it breaks everything
20:31 _28_ria joined
samcv uh 20:31
no
how else to make it relative?
we don't support relative builds
AlexDaniel hmm dunno just trying to understand
samcv like if it was /home/samantha/blah/blah 20:32
you'd have to replace it with
AlexDaniel whateverable installs everything in /tmp/whateverable/something
samcv ./././././././././././././././././
to be the same number of bytes
El_Che :)
samcv and that breaks things
AlexDaniel and thinking about it, maybe no appimage is actually required in that case… just extract it “properly” and it should work… hmmm 20:33
El_Che github.com/nxadm/rakudo-pkg/releas...2017.03_02 <-- added Ubuntu 17.04 i386 pkgs (OS out next week)
samcv what case?
oh so if you put it in that folder you mean
well. can always uh 20:34
El_Che samcv: what's the problem with modules? Zef can find a valid install path?
samcv is there a problem with modules?
El_Che of the run path changes every time it runs?
samcv uh
El_Che samcv: oh, I thought you wrote that. I misunderstood
samcv there is no zef
also it resolves relative directories
given to it
and some hackery inside the perl6 script inside usr/bin/ inside the appimage 20:35
see the perl6 script in the repo
also have you tried dling and running it?
El_Che I have the page open 20:36
was seeing if I could run in a container
but my connection is crappy here
(at the coast ith the family)
20:36 st_elmo left
El_Che (connected to a provider public wifi that kind of works if I sit in the right place :) ) 20:37
20:40 setty1 left 20:41 nhywyll left 20:45 TEttinger joined
timotimo yay my template6 PR was merged \o/ 20:47
it should now test fine again
samcv argh looks like it can't find Test module on Travis 20:51
maybe i did break something
well on star at least. i think nom probably works since it has no modules
timotimo Test is inside rakudo 20:54
if it can't find Test, then rakudo is gone :P
20:55 nhywyll joined
samcv nah 20:55
Could not open /tmp/.mount_vGeETt/usr/(Test)
hm
timotimo what, in parenthesis? 20:56
samcv yeah that's what it shows
timotimo weird
samcv use NativeCall;
Could not open /tmp/.mount_vGeETt/usr/(NativeCall). Failed to stat file: no such file or directory
timotimo yeah, that doesn't look like it's possible to work
can you get the RAKUDO_MODULE_DEBUG env var up in there, or is that already set?
samcv uh linenoise and other modules WERE working 20:57
but idk what i did
lizmat and another Perl 6 Weekly hits the Net: p6weekly.wordpress.com/2017/04/10/...aboom-⁽¹⁾/ 20:58
samcv yay
hmm 20:59
$*CWD is /usr
well /tmp/.mount_ckZEtL/usr 21:00
21:00 skids left
gfldex I spy with my little eye and I see a p6weekly.wordpress.com/2017/04/10/...%e2%81%be/ 21:00
lizmat: you may have to republish with a better URL 21:01
samcv use NativeCall;
Must specify something as a path: did you mean '.' for the current directory?
in any statement_control at ././share/nqp/lib/Perl6/Grammar.moarvm line 1
arghhh
timotimo yeah, we can change the url itself without changing the title
lizmat timotimo: how ? 21:03
ah, the slug?
timotimo think so, yes
lizmat p6weekly.wordpress.com/2017/04/10/...15-kaboom/ 21:04
gfldex: seems both url's work, at least for me 21:05
gfldex they do for me too
i'm gonna tell reddit
timotimo cool 21:06
El_Che samcv: just run the app 21:07
I see, it's meant as a repl, not as a install? 21:08
samcv yeah
i mean does it give errors on launching the repl?
i mean you can run scripts with it
El_Che i complains about Linenoise 21:09
I ran into a problem while trying to set up Linenoise: Could not open /tmp/.mount_OpJshC/usr/(Linenoise). Failed to stat file: no such file or directory
timotimo yeah, wtf is that about
El_Che image -v works 21:10
but this doesn't:
samcv yeah
El_Che ./org.perl6.rakudo-x86_64.AppImage -e 'say "foo"'
samcv argh
El_Che Could not open /home/foo/"foo". Failed to stat file: no such file or directory
samcv oh that makes sense
lol.
cause it thinks that is a directory and it resolves it in my helper script 21:11
will need to fix that
wish i didn't overwrite my working binary lol
oh well
El_Che github.com/samcv/rakudo-appimage/b...ster/perl6 <- that one?
samcv if i did it once hopefully can do it again
timotimo working binary, as in: one of the appimages?
samcv the first one i gave people 21:12
timotimo i still have that
samcv i overwrote it on my server
ok
timotimo hold on
samcv then please gimme :)
heh
cause that one actually worked
and didn't have the path issues
timotimo hack.p6c.org/~timo/org.perl6.rakudo...4.AppImage 21:14
samcv thanks :)
sweet. that works 21:15
-e works too, but relative paths are busted though since i fixed that recently 21:16
21:16 alimon left
samcv but if you give it full paths then it should work pretty great 21:16
21:17 RabidGravy left 21:18 bjz joined, mcmillhj left 21:22 jdmmmmm left
samcv ok yay. it looks like it's just the bash script that is the issue. and taking the bash script from the old one timotimo gave me made new one work :) 21:23
i was scared there was an issue patching the binaries and started to freak out
AlexDaniel s: 'unique', \(1, 2, 3) 21:31
samcv oh btw guys 21:36
do we want a 'help' command on our repl
python has one
looks like the maker of AppImages tried my package github.com/samcv/rakudo-appimage/i...-293085980 21:38
was interesting to see somebody type help
but it makes sense. it's what i do if on some new prompt i have no clue wtf it is. like grub prompt or efi firmware cmd which i have no clue what the commands are
21:38 wamba left
gfldex .oO( We wont `help` you but we will let you `exit`. ) 21:42
python wont let you go that easily! 21:44
>>> exit
Use exit() or Ctrl-D (i.e. EOF) to exit
samcv ...
hah
i think we should have a help though 21:45
gfldex i agree
samcv would make it more friendly
what should it say though
El_Che samcv: I'll test it when you put the new img online
samcv like it doesn't matter what, but as long as something not nothing
it is online
gfldex and it should have Camelia ASCII art
El_Che I give up for now, the connection is too crappy
samcv El_Che, RC2 github.com/samcv/rakudo-appimage/releases 21:46
El_Che (was trying to add alpine pkgs, but just surfing the pkg archive of alpine is too painfull)
gfldex but before you can have that p6doc would need to work properly
samcv yeah but should direct you to like docs.perl6.org or say something encouraging or something
heh www.nntp.perl.org/group/perl.perl6....34171.html 21:47
gfldex we could provide a helpful link when the repl is invoked
samcv that is too big though
hm
El_Che I think that I can minimize JJ's alpine rakudodocker img to around 15meg if I can get the pkg built 21:48
(he uses rakudobrew with all the dev deps)
however, the appimage has the advantage of not needing root!
samcv yep 21:49
El_Che that's huge
samcv yugeeee
tremendous!
it is pretty great though.
the only real issue though. is it lands you in $*CWD that is in the appimage temp mount
and not in where you start it 21:50
21:50 TimToady left
samcv so relative dirs don't work. but i have a hack for that, just need to make sure it doesn't hose the modules (still not sure how it did that) 21:51
21:52 TimToady joined 21:55 Cabanossi left 21:58 Cabanossi joined 22:10 rindolf left
gfldex lizmat: the link to Slippy::Semilist is a 404 22:11
lizmat oops, lemme check
22:11 Sound left
lizmat should be fixed now, gfldex++ 22:12
gfldex thanks
22:14 bjz left
gfldex while de-nqp-lifying Rakudo::Slippy::Semilist to Slippy::Semilist I found that performancy drops by only 15% these days if I call .EXISTS-KEY, .BIND-KEY and friends directly 22:14
what also means that postcircumfix:<{ ]> is terribly slow (factor 25 IIRC) 22:15
samcv ok. i made better debugging of my stuff 22:17
and i figured out why it's dying on using modules
gist.github.com/samcv/401ddc20a5d9...3adcc44233 22:18
oh my
hehe
lizmat gfldex: yeah, postcircumfix:<{ }> will be hard to optimize until we have a way to indicate a signature *without any* named parameters 22:19
something like sub a(\a, !%) { }
samcv semiconfused looking at that...
lizmat well, bikeshed away 22:20
perhaps !%_ ?
samcv oh i see what it's doing... it thinks (Test) is a directory. and finds the path of it 22:21
that's funny actulaly
gfldex sub s($a! where { !$a.foo}, !%_) 22:22
i not sure if I would like ! as a prefix and a suffix in sigs
I don't have a better idea tho 22:23
sub s($a! where { !$a.foo} --> Int; !%_)
that would make it stand out more
it's a thing that is not part of the Signature because it must be missing 22:24
.uni 擯 22:25
sub s($a! where { !$a.foo}, ¬%_) 22:27
timotimo samcv: why is it even trying "(Test)", i.e. where do the parens come from?
samcv uh the argv's
they're in there 22:28
i guess telling it it's for the Test module or something
gonna fix it the same way i'm gonna fix -e, gonna have a list of command line args that if it sees those, the next one is not gonna be interpreted as a path
shell scripting is so weird/bad. there's like 8 ways to do things 22:30
all with caveats and weird things
also: gist.github.com/samcv/be88a800d54a...122c062bcc 22:31
bashhhhh why do you do this to me
gotta duble check it's not the appimage mangling args tho 22:32
mst that looks very odd
samcv nope it's bash 22:33
22:33 kurahaupo left
samcv i hate shell scripting 22:33
mst how so?
samcv it's messing up my arguments
when it's quoted it should not be seperate
geekosaur ?
mst samcv: uhh, bash won't split those 22:34
samcv like you quote things, and then it passes the quoted section as one command arg
mst yes, that's how bash works
samcv apparently ${BASH_ARGV[*]} does
(split them)
mst oh, ah, not exactly
22:34 sufrostico joined
mst what's happening is word splitting is happening afterwards 22:34
samcv oh. 22:35
mst how are you using ${BASH_ARGV[*]} ?
samcv not quoting it. argh
mst exactly
22:35 bjz joined
mst hence why generally shell scripts that are wrappers to stuff like 22:35
exec $program "$@"
I suspect your weird bash specific thing is a fancy $@
hard to tell since there's lots of blaming bash and no bash code in the backlog so far :)
samcv heh 22:36
ok here gist.github.com/bc009f92a764fffe8f...f0d34e3bb2
22:36 pytuger left
samcv i quoted it. now need to track somewhere else it's spliting 22:36
well wait 22:37
now i get my args all messed up
now that i quoted it
Got Arg: 1 is ‘say hi -e’
it uh. quoted it. and they're ordered in reverse order
mst what is this code actually supposed to do? 22:38
you're using bash-specific weirdness everywhere 22:39
samcv i agree
mst I always write either shell scripts that work on multiple OSes
samcv please help me
mst or I use perl5 to mangle things
AlexDaniel 「for a in ${NEWARGS[*]}」 that would normally be 「for a in "${NEWARGS[@]}"」
samcv :(
i tried to get it to work on posix shell
AlexDaniel but depends on what exactly you're trying to achieve, I haven't looked closely
samcv but i did not kno whow
mst well it certainly doesn't work in bash atm :P
samcv i wanted to iterate $@
but i couldn't figure out how
after searching online for a long time. since most was related to bash 22:40
geekosaur for f in "$@"; do
mst bash -c 'for i in "$@"; do echo -- $i; done' -- -e "say 'hi'"
samcv ok using @ instead of the *
mst works fine
samcv for some reason that didn't work. but. i will do that
22:40 kyan joined
mst please can we try and make it a real shell script? 22:41
geekosaur has to be quoted, and has to be $@ not $*
samcv i must have messed something up idk
yes
AlexDaniel mst: you forgot to quote "$i"
samcv that was my end goal
mst AlexDaniel: was just showing that the 'for' worked
AlexDaniel still forgot to quote it :P
mst AlexDaniel: alright, you write it then
samcv ok thanks guys
will get it less bashy
AlexDaniel for i in "$@"; do printf '%s\n' "$i"; done 22:42
:P
timotimo bash also needs the GLR 22:43
AlexDaniel samcv: well, in practice, you'd probably want 「cd -- "$OWD"」 and stuff like that
22:44 bjz left
samcv in practice what 22:44
i have that in there? not sure
what you mean by that
AlexDaniel samcv: does not mean that there's necessarily a problem everywhere, but generally if you have no -- then it is quite possibly a bug
samcv also timotimo++ please tell them to get on that
AlexDaniel there are other shells that do a better job
samcv oh
why do i need --
AlexDaniel samcv: because if your variable starts with - then most commands will treat it as a parameter 22:45
for example
samcv yes
very true
AlexDaniel a='-n'; echo "$a"
samcv thank you
i will also quote the "--" as well :P /s
AlexDaniel (for that reason you cannot actually use echo to print variables, it does not support -- and you have to use printf instead)
samcv: also, if you're using bash anyway, you probably want [[ ]] instead of [ ] everywhere 22:46
samcv no i want not bash
but i'm going to keep it as bash and get it working first
once it's working i'll posix it
i think i'm almost there
22:47 sufrostico left
samcv hmm i need to iterate the bash argv's the opposite direction 22:47
uh mst how can i access one element of $@ 22:48
that's the original thing i tried to do so i could iterate how i wanted
AlexDaniel "$1" ?
22:49 sufrostico joined
samcv no 22:49
i'm not using those i need to iterate them
not access individual ones
because idk how many args i will get
geekosaur POSIX shell doesn;t provide a decent way to do it programmatically
AlexDaniel hm I wonder if "${@[1]}" is going to work
geekosaur there are bash-isms that can do it
like that
POSIX shell requires horrors like $(eval echo -- "\$$i") 22:50
hm, not quite, quoting :/ 22:51
$(eval echo -- "\"\$$i\"")
AlexDaniel /o\
geekosaur now you understand why so many folks punt to bash... or perl etc. 22:52
samcv save me
AlexDaniel samcv: you sure you cannot use bash?
samcv no i'm using bash now
i want to make it work first 22:53
AlexDaniel samcv: then just iterate it backwards? $# is the number of arguments, access them with "${@[$i]}"
samcv yeah that's what i'm gonna do
22:56 cdg left
samcv ${@[$i]} does not work 22:56
printf "this command is executed at least once $i %s\n" "${@[$i]}" # bad substitution 22:57
22:58 sufrostico left 22:59 sufrostico joined
samcv ok 23:00
the problem is. i can't access $@ inside the function
AlexDaniel samcv: right, because that's how you get function args
samcv ooook 23:01
now things are working
all my problems are solved 23:03
well. these ones are 23:04
23:10 skids joined
samcv yay it's working now :) 23:16
-e now works
sjn lizmat: typo in p6weekly: "&nbnsp;" 23:19
timotimo how did Ninja Sex Party™ (only the best band in the world) get in there?!? 23:21
Geth doc: 372545c059 | (Zoffix Znet)++ | doc/Type/IO/Path.pod6
[io grant] Straighten up file test docs
23:26
23:27 nhywyll left 23:39 ufobat left, Cabanossi left 23:42 Cabanossi joined 23:51 gurwall_ joined, gurwall_ left 23:54 BenGoldberg joined