»ö« Welcome to Perl 6! | perl6.org/ | evalbot usage: 'p6: say 3;' or rakudo:, std:, or /msg camelia p6: ... | irclog: irc.perl6.org or colabti.org/irclogger/irclogger_logs/perl6 | UTF-8 is our friend!
Set by masak on 28 November 2015.
00:00 clkao left
Herby_ That was it. I'm a dummy 00:00
00:01 clkao joined 00:03 Juerd left 00:04 RabidGravy left 00:05 znpy` left 00:08 PotatoGim left, PotatoGim_ joined
Herby_ thanks 00:09
00:09 Herby_ left 00:10 thundergnat joined 00:11 Juerd joined 00:15 znpy` joined
dalek href="https://modules.perl6.org:">modules.perl6.org: 82d9214 | (Zoffix Znet)++ | / (6 files):
Do not purge dists from DB when fetching META file fails

  RabidGravy++		 00:15
00:16 csd_ left
Zoffix There! Fixed forever :) 00:16
00:17 PotatoGim_ left
MadcapJake i feel like i'm doing something wrong, I'm trying to iterate over a `JSON::Tiny.from-json` array, and I am getting all sorts of strange results. For some reason the array is wrapped in an array, and if i try and print a newline before each array entry, it just prints one newline and then dumps the whole array out. 00:17
00:17 pierre-vigier joined
MadcapJake `for @p6[0] { say ""; .say }` shouldn't that print a newline between each array item? 00:18
00:19 znpy` left
lucs panda build problem: fpaste.scsys.co.uk/502164 00:19
Skarsnik yes
00:20 Actualeyes joined, Actualeyes left
hartenfels m: use NativeCall; sub printf(Str, Str is encoded("utf16")) is native(Str) { * }; printf("%p\n", "") 00:20
camelia rakudo-moar 0ede33: OUTPUT«(signal SEGV)»
Zoffix m: my @p6 = ([4, 5, 6], 45); for @p6[0] { say ""; .say }
camelia rakudo-moar 0ede33: OUTPUT«␤[4 5 6]␤»
hartenfels Uh
Skarsnik hm I think vararg does not work like this?
00:21 PotatoGim_ joined
lucs Both panda errors appear to be related to a commit 7cd00ffc3 00:21
Zoffix lucs, yeah, there've been a merge that's expected to keep stuff in turmoil for 1-2 days. I think your error looks same as this one: github.com/retupmoca/P6-Find-Bundled/pull/2 00:22
00:22 brabo joined
lucs Sure does :) 00:22
Skarsnik m: my @p6 = ([4, 5, 6], 45); for @p6[0] { say ""; say $_ } 00:23
camelia rakudo-moar 0ede33: OUTPUT«␤[4 5 6]␤»
Skarsnik m: my @p6 = ([4, 5, 6], 45); for @p6[0] { say "l"; say $_ }
camelia rakudo-moar 0ede33: OUTPUT«l␤[4 5 6]␤»
Skarsnik m: my @p6 = ([4, 5, 6], 45); for @(@p6[0]) { say "l"; say $_ }
camelia rakudo-moar 0ede33: OUTPUT«l␤4␤l␤5␤l␤6␤»
Zoffix m: my @p6 = ([4, 5, 6], 45); for @p6[0].List { say "l"; say $_ }
camelia rakudo-moar 0ede33: OUTPUT«l␤4␤l␤5␤l␤6␤»
MadcapJake sweet, so even though .WHAT returns array, I still have to turn it into a list to iterate ove rit? 00:24
Skarsnik m: my @p6 = ([4, 5, 6], 45); for @p6[0] -> $a { say "l"; say $a }
camelia rakudo-moar 0ede33: OUTPUT«l␤[4 5 6]␤»
Zoffix MadcapJake, no idea :) I don't know much of P6 :)
But from what I see, it ends up thinking that you wish to iterate over just one item, which turns out to be [4, 5, 6] 00:25
00:25 Juerd left
MadcapJake ok cool, I feel like I'm at a serious disadvantage learning Perl 6 with zero experience in Perl 5 :S 00:26
00:26 Juerd joined
Skarsnik Not really 00:26
Zoffix MadcapJake, I've been coding Perl 5 for a decade and Perl 6 looks alien to me :P
Skarsnik perl5 is messy for this kind of stuff x)
Zoffix MadcapJake, actually, I feel at a disadvantage because I constantly have to force myself to forget how P5 does stuff :P
lucs Zoffix: I checked out the commit preceding the one shown, and it builds okay -- I suppose I can live with that for now :)
Zoffix sweet 00:27
MadcapJake lol i suppose there's pros and cons on both sides of the fence
Zoffix :)
lucs Zoffix: Looks alien to you? Hmm... To me it looks like an incredible step up, feels completely perlish, but way way more powerful. 00:28
00:28 telex left
Zoffix lucs, ¯\_(ツ)_/¯ same for me, but the power is from the "alien" stuff :) Like all sorts new of operators and core method calls. And of course all the typed stuff, on which I'm not yet entirely sold. 00:31
00:31 znpy` joined
lucs Well, you know, try before you buy :) 00:32
Zoffix I do have much more fun coding P6 than P5 though.
Not sure if that's just 'cause I code P5 for a living, so it just feels like "work" :) 00:33
lucs I remember when I went from perl4 to perl5 and all its cool new features.
Now I have the same feeling, but at least an order of magnitude or two stronger. 00:34
Zoffix perl4 0.o
lucs Yessir
Zoffix I don't think I was even born back then :P
lucs On windows no less.
Zoffix :o
m: 378/462 00:35
camelia ( no output )
Zoffix curses
We really need to fix the bot to output last statement
00:35 clkao left, clkao joined
Skarsnik m: 40/10.say 00:35
camelia rakudo-moar 0ede33: OUTPUT«10␤»
ShimmerFairy MadcapJake: I still don't know any Perl 5, to give you an example of someone who's learned Perl 6 without P5 knowledge :P
Zoffix ha-ha
Mouq m: my @p6 = ([4, 5, 6], 45); for |@p6[0] { say ""; .say } # MadcapJake 00:36
camelia rakudo-moar 0ede33: OUTPUT«␤4␤␤5␤␤6␤»
Zoffix Ah Mouq++
Skarsnik | is .flat?
Zoffix I believe so 00:37
dalek kudo-star-daily: 09c9bb8 | coke++ | log/ (2 files):
today (automated commit)
 00:38
jdv79 i thought | was more like a slip
Mouq Skarsnik: In this case, it's the same
m: my @p6 = ([4, 5, 6], 45); say @p6[0].flat; say |@p6[0] # one difference
camelia rakudo-moar 0ede33: OUTPUT«(4 5 6)␤456␤»
skids Hrm, shouldn't @a[0] be a single-element-rule triggerer? Also isn't "for" supposed to go further than single-element-rule? 00:39
00:39 thundergnat left, hartenfels left
skids m: my @a = [42, 3], "a"; for @(@a[0]) { "l".say; $_.say; } 00:40
camelia rakudo-moar 0ede33: OUTPUT«l␤42␤l␤3␤»
ShimmerFairy Mouq, I didn't share this yesterday, but here's my Pod parser :) shimmeryfairy.wordpress.com/2015/1...supernova/
Mouq skids: tbh I'm still catching up on the exactitudes of GLR. I was away for awhile..
ShimmerFairy++ 00:42
00:42 Juerd left, BinGOs left, BinGOs joined
Zoffix k. modules.perl6.org db has been rebuilt. The dist-purging bug should be fixed now. 00:42
00:42 Juerd joined
Mouq 100% agree with having a separate Pod-lang 00:42
ShimmerFairy skids: my guess is that since each Array element is a Scalar container, you're getting scalar behavior from a single-elem access 00:43
Mouq in the braid, that is
skids Possibly some of that "should decont" work may have changed the behavior since immediately post-glr. I wonder what some older builds do.
00:44 Zoffix left
ShimmerFairy Mouq: another thought in that direction I've had is that we should have separate files per language, e.g. MAIN-Grammar.nqp, Quote-Actions.nqp, and so on, just to keep things more logically separate (we have technology to concat files, at least) :) 00:44
skids: I know that being able to have for $[1,2,3] { } mean loop over the list as one element (that is, force non-single-arg-rule-ing) was added because trying to stop single-arg when a leading $ didn't was too much of a headache 00:45
Mouq ShimmerFairy: What, and make the grammar file less insane? why would we do that? ;P
00:47 TEttinger left, sQuEE` left 00:48 sQuEE joined, dayangkun joined 00:49 TEttinger joined, skaji_ left
Mouq just realized how awesome it is that due to NFG, table parsing shoudn't have to do special unicode contortions to DWIM for a table with funny symbols in it 00:49
00:50 dalek left 00:51 dalek joined, ChanServ sets mode: +v dalek, PotatoGim_ left
ShimmerFairy Mouq: how do you mean? I don't immediately see where NFG would make that particularly easier per se. 00:52
00:52 PotatoGim joined
Mouq Tables a character-based, aren't they? Someone writes what looks like X characters, that table column should be considered to be X characters 00:55
are*
00:55 [Tux] left
ShimmerFairy Mouq: ooh, yes! NFG is so nice I forgot that I'd be forced to handle combining chars and the like appropriately otherwise :P 00:56
Mouq haha 00:57
ShimmerFairy Mouq: the only place where that doesn't count is with tab characters, but I don't yet count tabs right for margins, much less in a general fashion for, say, inside a table cell. (How would you even tweak regexes to do that? O_o) 00:58
00:59 dalek left, clkao_ joined
ShimmerFairy For virtual margins, it's just a matter of counting "visual" spaces when getting the margin and checking for it on the next lines (which just means encountering and handling a tab specially). Generalizing support for \t as a variable-length character sounds really tricky, though. 00:59
00:59 clkao left, clkao_ is now known as clkao, dalek joined, ChanServ sets mode: +v dalek
Mouq ShimmerFairy: Better to veto mixed spaces and then, if wanted, add support later 01:00
ShimmerFairy Mouq: yeah, for now margins are lazily just space characters IIRC (or else pretends tab is one space) :)
Mouq: I just thought of a $*TAB_DRIFT variable (reset per line) that holds the number of spaces tab characters have pretended to be (so if a tab pretends to be 4 spaces, then $*TAB_DRIFT == 3), to aid in calculating cell positions 01:01
01:02 telex joined
ShimmerFairy (keeping the TAB_DRIFT separate means we could still get the "real" column of a line if needed, and generally feels nicer than lumping it all together) 01:02
01:02 rgrau left 01:03 kipd left, pierre-vigier left
Mouq ShimmerFairy: Except when the next person opens the file in an editor with different tab display and can't figure out why their docs don't look right :P 01:03
ShimmerFairy Mouq: that's what $?TABSTOP is for :P 01:05
moral of the story: don't use tabs inside table cells, or in fact anywhere beyond the left margin (i.e. follow the "tabs for indent, space for alignment" convention I've heard before) 01:06
01:06 kipd joined
Mouq s/look/compile 01:06
but yeah
skids
.oO(replace all normal characters from previous cols with spaces, then run indent)
 01:08
01:09 samb1 left, cognominal left
ShimmerFairy mixed space margins should of course be supported however. Like I said, it's just a matter of calculating width according to $?TABSTOP when taking a margin, and then carefully counting spaces as you consume them on subsequent lines :) 01:10
jdv79 how can an import break custom precedence on an op?
[Coke] ok, I was collecting names manually. I've gotten enough requests to add more repos that I'm now redoing the whole list of historic contributors as a program. I thought I was being the right kind of lazy, but I guess not. :) 01:11
ShimmerFairy Mouq: interesting question: should configuring para and code blocks affect their implicit variants too? For example, what would happen with =config code :margin<|> to implicit code blocks? Would they still need more indentation, would a starting | *be* the more indent, or what?
01:12 BinGOs left, BinGOs joined
ShimmerFairy And then what happens if you see =config code :margin<|> \n =begin MyThing :margin<+> \n + Stuff \n\n +|code? \n =end MyThing ? 01:13
jdv79 [Coke]: where will this list go?
01:13 kipd left
ShimmerFairy (That reminds me, I forgot to add the :nested option to blocks that have it by default) 01:13
01:15 kipd joined
Mouq ShimmerFairy: Probably be better if implicit code/paras had seperate =configs from their explicit siblings 01:15
01:15 [Tux] joined
Mouq is busy fighting the repo manager :P 01:16
ShimmerFairy Mouq: true, I've so far been able to say "literally a code/para, but, well, implicit". But I think perhaps a little separation is in order :) (for sure, the renderer *shouldn't* have to care about explicit/implicit)
Mouq: I already came up with =config item* and =config head* to generally configure all heads and items (though they're at least now just fallbacks, not any of that cascading business :P), so more special options shouldn't be impossible 01:17
[Coke] jdv79: into the 2015.12 release announcement.
a work in progress is already there.
01:20 pierre-vigier joined
Mouq ShimmerFairy: I'm just glad you got =config working at all :P 01:20
ShimmerFairy Mouq: oh, it doesn't in rakudo? Huh. 01:21
Mouq: if you're wondering why I invented some new config options for implicit para/code and so on (Actions.pm6), it's because I had to figure out a way for custom blocks to get that kind of stuff if desired, and this felt like the natural solution.
I used to have the custom blocks/formatting codes specify those properties via methods, but that was back when I built the AST as I went. I did away with that when it felt too clunky and potentially terrible to rely on a tree of P6 objects for parse-time knowledge :) 01:23
01:24 yeahnoob joined 01:26 boegel left 01:27 Juerd left, Juerd joined 01:29 clkao left, skaji joined 01:32 kipd left 01:33 PotatoGim left, roguelazer left, boegel joined 01:34 KCinJP joined
Mouq ShimmerFairy: line 66 of Grammar::Parsefail, .first-index needs to be .first(* >= $at,:k) 01:34
ShimmerFairy Mouq: lemme know if you spot anything else of interest in SUPERNOVA, and if you'd like a commit bit :) . I'd like to hear your thoughts on my idea to handle L<scheme:> and P<scheme:> in the same way M is handled, via subclasses. (I know the Http and etc. subclasses are a bit eh right now, but they can always be improved)
Mouq: yes, I _just_ finally fixed that :)
01:34 clkao joined
Mouq It's also extremely hard to get new copies of a module installed. Even incrementing the version number seems to use the old version? 01:35
01:35 Peter_R left
Mouq ShimmerFairy: Absolutely :) 01:35
01:35 roguelazer joined
Mouq ShimmerFairy: (Once I wrasle panda/CURLI enough to get it working) 01:35
ShimmerFairy Mouq: yeah, for whatever reason panda won't let me install my own Parsefail module (non-CURLI), it's weird. 01:36
Mouq ShimmerFairy: It's your source line, I'm pretty sure
Try git: instead of https:
*In META6.json 01:37
01:37 dalek left, dalek joined, ChanServ sets mode: +v dalek
ShimmerFairy Mouq: I'm fairly sure it worked before though, and I think Term::ColorText also uses https and yet works 01:37
Mouq Hm 01:38
ShimmerFairy oh, nevermind, it uses git:
but still, it worked before, of that I'm certain. Since I didn't change the source url, it can't be my fault, can it?
01:39 kipd joined, PotatoGim joined
ShimmerFairy Just tried someone else's https: using module, failed too. So weird, it obviously worked before if I'm not the only one to do it. 01:40
skids Hrm maybe once CURLI gets all sorted it will be high time for a "module maintainers howto" 01:41
01:42 clkao left, clkao joined
Mouq I got it to run \o/ 01:43
01:44 integral_ left
jdv79 --ast seems a bit cranky 01:46
oh, --target=ast i mean
ShimmerFairy Mouq: \o/ I'll hand a commit bit to 'Mouq' unless you want me to wait :)
01:51 csd_ joined 01:52 integral joined 01:53 samb11 joined
timotimo how does it crank, jdv79? 01:55
01:56 PotatoGim left
timotimo oh, there's a mail in my inbox about this 01:56
01:57 PotatoGim joined
timotimo wow, that looks very broken 01:57
oh, it's because of | head
02:01 yeahnoob left 02:03 Skarsnik left 02:08 kipd left 02:09 kipd joined
ShimmerFairy Failed to create directory '/usr/local/share/perl6/vendor/dist' with mode '0o777': Failed to mkdir: 13 02:12
oh man, lovin' this CURLI business :V
02:12 eiro left
timotimo why does it go there? o_O 02:13
02:13 hobbs left
ShimmerFairy 1) I used to be able to run panda and have it install to ~/.perl6 , no sudo business required, and 2) like hell I'm putting a 0o777 in /usr/local without a good explanation 02:13
02:14 hobbs joined
ShimmerFairy timotimo: I install my perl6 to /usr/local, universal default install location except for MoarVM. I'm guessing that's why. 02:14
Or more clearly: I'm guessing nobody who tried CURLI ever tried installing to the Universal Default Install Location™ before :)
02:16 pnu left 02:17 cdg joined 02:19 BinGOs left, pnu joined, solarbunny left 02:20 roguelazer left 02:21 hobbs left, [Tux] left 02:22 hobbs joined
ShimmerFairy My question is why it didn't create that directory (and all the other ones) when I typed 'sudo make install' in rakudo. Also, does it really need to be 777? My computer tried really hard to find a single regular file in /usr/ with a 777 permission (stopped it due to lack of patience :P) 02:22
timotimo yeah, clearly wrong behaviour there 02:25
jdv79 timotimo: no. that's just to get any large output out of hte way ^H 02:26
02:26 BinGOs joined 02:27 FROGGS_ left 02:28 roguelazer joined 02:29 [Tux] joined
ShimmerFairy timotimo: seems that it happens when using a module that's not installed. -MTest works, but something like perl6 -Ilib -Mfoo will fail with that error. 02:29
02:31 pnu left
flussence what's weird is mine's been writing precomp files under both $install_path/ and ~/.perl6/$rakudo_version/. it should just use the latter in all cases, imo 02:32
diakopter hopes it can just install to ./install under rakudo, in addition 02:33
ShimmerFairy flussence: well, having a system-wide thing is I think pretty much necessary, but something is definitely screwy (I do have a new directory in ~/.perl6 filled with new CURLI stuff, at least)
diakopter: well, my hypothesis is that nobody has yet to do sufficiently extensive testing beyond the ./install default MoarVM and friends have :) 02:34
flussence iirc, mono used a location under ~/.cache/ for this stuff...
02:34 kipd left
dalek href="https://modules.perl6.org/raiph-patch-1:">modules.perl6.org/raiph-patch-1: 5c38f9f | raiph++ | templates/root/index.html.ep:
Suggested shift in emphasis to present status

  - All of them have been working on <a href="rakudo.org/">Rakudo</a> some point.
  + The Travis column shows results of testing each module using <a href="rakudo.org/">Rakudo</a>.		 02:34
02:35 kid51 left 02:36 kipd joined, PotatoGim left 02:37 PotatoGim joined 02:41 yeahnoob joined
dalek href="https://modules.perl6.org/raiph-patch-1:">modules.perl6.org/raiph-patch-1: 3e59050 | raiph++ | templates/root/index.html.ep:
Link to github.com/ugexe/P6TCI
 02:41
02:42 hobbs left
ShimmerFairy shall file the rakudobug :) 02:43
02:46 ilbot3 left 02:50 Herby_ joined 02:51 vendethiel joined 02:52 BinGOs left
ShimmerFairy bug here, for those interested: rt.perl.org/Ticket/Display.html?id=126765 02:54
02:57 bpmedley left 02:58 boegel left
[Coke] . 03:00
It would be nice if the desired mode was given in the ticket.
ShimmerFairy [Coke]: I don't know for sure what the desired mode should be, but "let EVERYBODY modify this directory!" sure isn't it :) 03:03
03:04 BinGOs joined, hacst left 03:06 Exodist left, boegel joined, solarbunny joined, hobbs joined 03:07 pnu joined, Ben_Goldberg joined 03:08 hobbs left, hobbs joined, hacst joined
grondilu finally managed to use escape characters to display a chess board properly: i.imgur.com/tbiUqwK.png 03:09
geekosaur I was wondering if 1777 was intended, but that sounds wrong too
03:10 pierre-vigier left, cdg left
grondilu is still wondering how he could make sure the squares really are "squared". 03:11
[Coke] grondilu: without a custom font, that's unlikely, I'd wager.
geekosaur can't, without knowing the terminal and the font metrics
ShimmerFairy [Coke]: more specifically, it seems that for me at least 755 is normal, and 644 for files.
755 for directories, I mean. 03:12
Herby_ skarn, you around?
geekosaur or cheating: look for $WINDOWID in the envionment and scribble directly in the window with graphics primitives (X11 only of course) 03:13
03:13 Juerd left, Exodist joined, vendethiel left 03:14 noganex_ left 03:15 PotatoGim left, kipd left 03:16 PotatoGim joined 03:17 kipd joined, Herby_ left, bpmedley joined 03:19 Juerd joined 03:21 eiro joined, integral left, samb11 is now known as samb1 03:23 integral joined
skids ｆｕｌｌｗｉｄｔｈ characters might be squarer, but I'd be surprised if chess symbols are available. 03:34
lucs In perl5, to pause my program until I pressed return (to look around the filesystem while debugging), I just placed a line with "<STDIN>;" 03:39
03:39 emdashcomma left, lestrrat left
lucs What's an equivalent way to do that in Perl6? 03:39
03:40 dalek left
skids lucs: $*IN.get 03:41
lucs Thanks skids
03:42 dalek joined, ChanServ sets mode: +v dalek, zengargo1le left 03:45 Gardner left
AlexDaniel Interesting! Text::Markdown does not use grammars. I wonder why would it be so? 03:47
03:48 znpy` left 03:49 noganex joined 03:50 FROGGS joined, Gardner joined 03:51 emdashcomma joined
ShimmerFairy There's really no reason why it shouldn't, the only thing I can see is perhaps a connection masak's markdown thing, which has a similarly bewildering aversion to grammars. 03:51
03:52 TEttinger left
Mouq Idk, I imagine it's probably faster 03:54
03:55 [Tux] left 03:58 KCinJP left, moznion left 04:04 Ben_Goldberg left, skids left 04:05 geekosaur left 04:06 geekosaur joined 04:08 hobbs left, eiro left 04:09 hobbs joined 04:11 BenGoldberg joined, mrsolo joined 04:12 pnu left 04:14 kipd left
[Coke] if github reports an author email like "Author: zeriod <zeriod@c213334d-75ef-0310-aa23-eaa082d1ae64>" - is there anything I can do to convert that to a "real" email? 04:15
04:15 BinGOs left
[Coke] ah. probably a holdover from a git-svn conversion. 04:15
04:16 kipd joined 04:19 BinGOs joined 04:20 [Tux] joined 04:21 kaare_ joined, solarbunny left, solarbunny joined 04:23 pnu joined
MadcapJake why do I have to use `*-2` to remove the last item of a list? 04:23
04:24 Timbus_ left
MadcapJake m: my $selector = "comment.block.delimited.perl6fe"; given $selector { .split(".")[2..*-1].join(".").say } 04:24
camelia rakudo-moar 0ede33: OUTPUT«delimited.perl6fe␤»
MadcapJake m: my $selector = "comment.block.delimited.perl6fe"; given $selector { .split(".")[2..*-2].join(".").say }
camelia rakudo-moar 0ede33: OUTPUT«delimited␤»
ShimmerFairy MadcapJake: *-1 refers to the last element of the list
MadcapJake ohh, so it's including it, i see
ShimmerFairy it's short for something like @foo[{$^size - 1}] 04:25
in other words, * - 1 becomes a block that takes a single parameter, and when array indexing gets a block it passes the array size to it :) 04:26
MadcapJake neat! thanks ShimmerFairy!
04:26 Gardner left 04:27 kaare_ left
MadcapJake wow that's super slick! Love all these little gems in Perl 6! 04:27
04:28 kipd left
ShimmerFairy m: say (1,2,3,4,5)[(*/2).Int] # another cool way to use WhateverCode in array indexing, getting the middle elem of a list with odd number of elements :) 04:29
camelia rakudo-moar 0ede33: OUTPUT«3␤»
Mouq m: say (1,2,3,4,5)[* div 2] #? 04:30
camelia rakudo-moar 0ede33: OUTPUT«3␤»
ShimmerFairy ah yes, forget about integer division :) 04:32
dalek href="https://modules.perl6.org:">modules.perl6.org: 5c38f9f | raiph++ | templates/root/index.html.ep:
Suggested shift in emphasis to present status

  - All of them have been working on <a href="rakudo.org/">Rakudo</a> some point.
  + The Travis column shows results of testing each module using <a href="rakudo.org/">Rakudo</a>.		 04:33
href="https://modules.perl6.org:">modules.perl6.org: 3e59050 | raiph++ | templates/root/index.html.ep:
Link to github.com/ugexe/P6TCI
href="https://modules.perl6.org:">modules.perl6.org: 70f02bb | (Zoffix Znet)++ | templates/root/index.html.ep:
Merge pull request #46 from perl6/raiph-patch-1

Suggested shift in emphasis to present status
04:34 Gardner joined, clkao left 04:35 skarn left 04:44 pnu left, solarbunny left, ilbot3 joined 04:48 khw left 04:51 csd_ left 04:52 flussence left, clkao joined
lucs Some panda code reads CompUnitRepo.new("file#$where", :next-repo($*REPO)) , yet new() takes a single Str argument, according both to S22 and to .../rakudo/src/core/CompUnitRepo.pm. So how does that work? 04:52
04:53 flussence joined 04:57 boegel left 04:58 hacst left 04:59 boegel joined
[Coke] lucs - are you sure you're looking at a recent copy of that file? 05:01
I see: method new(Str $spec, CompUnit::Repository :$next-repo) {
05:02 skarn joined
lucs Not the most recent, no, you're right. I checked out about 4 commits earlier, as the current one "didn't work" (forgot what exactly). 05:04
I'm speaking of panda actually.
05:05 Gardner left, Juerd left
lucs As for rakudo, I cloned it today. 05:05
05:05 hacst joined 05:06 PotatoGim left
lucs [Coke]: So, sorry, I was talking about panda. I'll check to see which CompUnitRepo.pm I have exactly. 05:06
05:08 skarn left 05:12 bpmedley left 05:13 bolangi left
lucs [Coke]: Sheesh, just did a git pull, 41 files have changed since this morning ?! 05:15
(including, yes, CompUnitRepo.pm)
lucs realizes he has to keep up! 05:18
05:19 roguelazer left 05:20 kipd joined, roguelazer joined, Gardner joined 05:24 clkao left 05:30 lestrrat joined, BenGoldberg left 05:37 hobbs left 05:41 Timbus joined 05:45 skarn joined 05:47 Juerd joined 05:50 bpmedley joined, sQuEE left 05:51 AlexDaniel left 05:54 Juerd left, Juerd joined 05:55 sQuEE joined 05:56 kipd left 05:58 roguelazer left 05:59 Zoffix joined 06:00 hobbs joined 06:01 roguelazer joined 06:06 hacst left, [Sno] left 06:07 [Tux] left 06:17 hacst joined 06:19 hobbs left 06:22 roguelazer left 06:23 Gardner left 06:24 roguelazer joined 06:29 BinGOs left 06:30 BinGOs joined, mrsolo left 06:31 mrsolo joined 06:33 hobbs joined, solarbunny joined, hobbs left, hobbs joined 06:37 [Tux] joined, Gardner joined 06:40 hobbs left 06:46 kipd joined 06:47 xfix joined 06:49 hobbs joined 06:50 geraud left 06:51 TEttinger joined 06:52 solarbunny left 06:53 solarbunny joined, balazs left 06:54 pnu joined 06:55 bpmedley left 06:59 hobbs left 07:01 pnu left, solarbunny left, pnu joined 07:05 TEttinger left 07:09 vendethiel joined 07:10 hobbs joined
[Tux] gist.github.com/Tux/1c3a7f4ff4718020b945 :( 07:12
yoleaux 29 Nov 2015 23:09Z <lizmat> [Tux]: say (10101..99999).pick(44).sort should now work
07:12 moznion joined
[Tux] :) 07:12
[Tux] starts from scratch
07:13 FROGGS left
lizmat yeah, that seems to be the solution :-) 07:13
dalek kudo/nom: 2aad18c | (Jimmy Zhuo)++ | .gitignore:
update .gitignore
 07:14
kudo/nom: 85e359c | FROGGS++ | .gitignore:
Merge pull request #606 from zhuomingliang/patch-8

update .gitignore
07:15 xinming left
dalek kudo/nom: e62545b | (Brad Gilbert)++ | src/core/Cool.pm:
Cool.substr-eq should return a Bool
 07:15
kudo/nom: 8c001e0 | FROGGS++ | src/core/Cool.pm:
Merge pull request #605 from b2gills/patch-3

Cool.substr-eq should return a Bool
07:19 solarbunny joined
[Tux] $ panda install Inline::Perl5 Slang::Tuxic File::Temp CSV::Parser 07:19
make: execvp: panda: Permission denied
60038056 -rwxrwxr-x 1 merijn 5297 2015-11-24 09:12:49 /pro/3gl/CPAN/rakudobrew/jvm-nom/panda/bin/panda 07:21
59904848 -rwxrwxr-x 1 merijn 5297 2015-11-30 08:18:38 /pro/3gl/CPAN/rakudobrew/moar-nom/panda/bin/panda
59245098 -rwxrwxr-x 1 merijn 5297 2015-11-24 09:12:59 /pro/3gl/CPAN/rakudobrew/panda/bin/panda
Ah, all is clear now. Starting afresh used the wrong base folder :( 07:22
07:24 raiph left 07:26 wtw left, wtw joined
nine [Tux]: the panda executable is now in a different directory 07:26
07:26 pnu left
yoleaux 29 Nov 2015 23:08Z <lizmat> nine: it looks like trying to build with uncommitted changes, gives the "already installed" error if you built on the latest commit before 07:27
nine .tell lizmat I know. I hope that soon someone will become annoyed enough to have an idea how to avoid that. 07:28
yoleaux nine: I'll pass your message to lizmat.
07:29 pnu joined
lizmat .botsnack 07:29
yoleaux 07:28Z <nine> lizmat: I know. I hope that soon someone will become annoyed enough to have an idea how to avoid that.
:D
07:29 vendethiel left
lizmat nine: the error message doesn't really state which file(s) are offending 07:31
it would be nice if it would be clear which file(s) would need to be zapped, instead of having to clear all of install
nine It's the distribution, not individual files.
lizmat so killing the dist dir should do it ? 07:32
nine Well right now an rm -rf install/share/perl6 is the easiest way around. Note that panda installs into ~/.perl6 right now, so you don't lose anything
But I didn't want to volunteer that information, because the simpleness of the workaround has kept me from finding a proper fix :) 07:33
lizmat well, maybe it is as simple as just overwriting the setting dist always ? and not warn about it ? 07:34
I mean, if you're not a core dev, that situation would never occur, would it ?
nine Probably not, yes. 07:35
So, in other words a :force flag that just doesn't do the installed check
lizmat well, this would only apply to the settings dist, no ?
nine: also, is Missing test file: t/spec/S28-named-variables/inc.t something that rings a bell ? 07:37
07:39 xfix left 07:42 eiro joined, kipd left 07:43 kipd joined
dalek kudo/nom: b9bb064 | lizmat++ | t/01-sanity/52-parse-include-spec.t:
Remove test for CREATE-INCLUDE-SPECS
 07:45
07:45 pnu left
[Tux] I don't think the new panda install path is an improvement 07:46
gist.github.com/Tux/c48a7bf99612f3e9116f
dalek kudo/nom: a31235c | lizmat++ | src/core/CompUnitRepo.pm:
Remove CREATE-INCLUDE-SPECS fossil
07:46 moritz_ is now known as moritz 07:47 Gardner left, pnu joined
llfourn so what cool stuff can I do now that nine++'s work has been merged? 07:49
07:49 sammers joined
dalek kudo/nom: 0cff007 | (Stefan Seifert)++ | t/spectest.data:
S28-named-variables/inc.t is gone. Tested the obsolete @*INC

Thanks to lizmat++ for pointing this out!		 07:50
nine lizmat: indeed ^^^
lizmat llfourn: panda should have precomp again
llfourn lizmat: oh nice.
lizmat bare startup time has dropped from 115 msecs to 104 msecs on my machine 07:51
llfourn is precomp easier for my non-panda modules as well?
lizmat llfourn: not sure atm, eventually yes
dalek kudo/nom: 81b4a5d | (Stefan Seifert)++ | / (2 files):
Force installation of the CORE dist

This relieves core developers of the annoyment of "CORE already installed" messages on every but the first make install.
llfourn kk
lizmat nine++ :-) 07:52
07:52 Gardner joined
llfourn nine++ I just saw that error for the first time 5 secs before seeing that 07:52
nine [Tux]: installing into ~/.perl6 is temporary. Plan is to install into share/perl6/site if possible and only fall back to ~/.perl6. 07:53
07:53 pnu left
[Tux] I install into /pro/3gl/CPAN/rakudobrew, never in $HOME 07:54
my $prefix = rakudobrew
07:54 pnu joined
nine Probably shows that I've never used rakudobrew myself 07:54
[Tux] should try to create a ~/.perl6 with mode 0000 to ensure those unwanted fallbacks are signalled early 07:56
07:56 lestrrat left
nine llfourn: yes, FileSystem repositories (use lib "lib" et al.) automatically precompile on first use. Should for example reduce time to run tests. 07:56
[Tux] indeed, lots of stuff in ~/.perl6 again :( :(
llfourn nine: nice++ 07:57
nine [Tux]: would be happy if you could help fix that! All we need is a way to identify the "site" repo and have panda pick that one preferrably and only fall back to others. 07:58
07:58 pnu left
[Tux] I just cleared out ~/.perl6 and removed write-rights to it. Lets have a looks where the process now breaks 07:58
07:58 pnu joined 07:59 eiro left
moritz nine: we used to have %*CUSTOM_LIB<site> for that 07:59
nine moritz: was that its purpose? Then I'll just have to replace the stringified paths by the Repository instances to make it usable for panda 08:00
moritz nine: the purpose was to make it easier for panda and the likes to find where to install site packages 08:02
nine moritz: excellent! Then I do have a plan :) 08:03
Will have a look at it tonight
[Tux] gist.github.com/Tux/64423a1cfa9210bac873
dalek kudo/nom: 7b78887 | lizmat++ | CURLI:
Remove original CURLI design fossil
 08:04
08:04 eiro joined, telex left 08:06 bpmedley joined 08:07 domidumont joined 08:10 PotatoGim joined 08:11 kipd left, kipd_ joined 08:13 cognominal joined, hobbs left 08:14 kensanata joined 08:15 bpmedley left
cognominal I want to explain what is bytecode and try to demonstrate : perl6 --target=mbc -ohi.moarvm -e 'say "hi"' and I get "cannot dump that object; no dump method". What am I doing wrong? 08:16
08:18 FROGGS joined, dalek left, darutoko joined, dalek joined, ChanServ sets mode: +v dalek
masak morning, #perl6 08:19
08:20 hobbs joined, csd_ joined, lestrrat joined, TEttinger joined
moritz cognominal: write it to a file, --output=hi.moarvm 08:21
cognominal: and then open the file in a hex editor, or something
08:21 geekosaur left, geekosaur joined 08:22 BinGOs left, BinGOs joined 08:23 sitaram left, telex joined, sitaram joined, sitaram left, sitaram joined 08:24 xinming joined
[Tux] so, with ~/.perl6 disabled, I cannot install 08:25
nine [Tux]: how about making .perl6 a symlink to your normal install directory? Or wait till tonight ;) Or being faster than me at fixing panda 08:26
[Tux] that last option is not an option ($work)
08:26 ely-se_ joined 08:27 Humbedooh left, Humbedooh joined 08:28 Humbedooh left, Humbedooh joined 08:29 mrsolo left, zakharyas joined 08:30 mls left, mls joined, pdcawley joined, abraxxa joined, pnu left, pnu joined, PotatoGim left, PotatoGim joined, kipd_ left, kipd_ joined, Woodi joined 08:31 hacst left
cognominal moritz, the -o form is advertised in perl6 -h but not supported? 08:32
08:33 [Sno] joined 08:34 hacst joined 08:38 BinGOs left, Gardner left, roguelazer left
lizmat catches up on some sleep 08:38
08:38 BinGOs joined 08:39 sQuEE left
moritz cognominal: then we should fix the docs 08:39
08:41 roguelazer joined
[Tux] nine, now it installs in rakudobrew/2015.11-292-g7b78887 08:41
08:41 mrsolo joined 08:42 sQuEE joined 08:43 Gardner joined
[Tux] which - of course - is not in my $PATH 08:44
moritz which explains why panda doesn't work anymore on travis
08:47 CurtisOvidPoe left
nine I would actually love if panda installed scripts into /usr/local/bin and fall back to ~/bin. The generated scripts are independent of installed versions of the modules anyway. 08:48
08:48 Juerd left 08:49 Juerd joined 08:50 kipd_ left
[Tux] oh my, this migh be the most tedious install ever 08:52
$ panda install Inline::Perl5 Slang::Tuxic File::Temp CSV::Parser
don't know if it hangs or if it takes ages in ==> Testing Pod::Coverage
JSON::Tiny installed fine
FROGGS nine: we also need some way to tell rakudo (or rakudo via panda) to install scripts into install/bin for rakudo star 08:53
08:54 cognominal left, domidumont left, eiro left
Mouq [Tux]: I also had Pod::Coverage hangs 08:54
[Tux]: Looks to rely on old CUR stuff, needs to be updated 08:55
08:55 Gardner left, zakharyas left
[Tux] it blocks all other stuff, so I cannot give timeings 08:55
lemme try to force skip that
08:56 Gardner joined 08:57 sQuEE left
[Tux] *WTF*????? 08:58
Missing required term after infix
at /pro/3gl/CPAN/Text-CSV6/t/10_base.t:9
------> my $csv = Text::CSV.⏏new;
expecting any of:
dotty method or postfix
nine During GLR I used PERL6LIB instead of panda for dependencies
08:59 sQuEE joined, xpen joined 09:00 kipd_ joined, M-eternaleye is now known as eternaleye, eternaleye left, eternaleye joined, ShimmerFairy left 09:02 zakharyas joined
[Tux] test 50000 23.206 23.097 09:03
test-t ===SORRY!===
Cannot find method 'run_alt'
*I* do not have a run_alt
nor do I knowingly invoke it
09:07 RabidGravy joined
RabidGravy yikes 09:07
09:08 johan___ is now known as johan 09:11 Gardner left
FROGGS [Tux]: run_alt sounds like something regex interval fwiw 09:14
09:14 zakharyas left 09:15 molaf joined
[Tux] lizmat, pick/sort verified. top 09:16
09:16 kipd__ joined 09:17 kipd_ left, kipd__ is now known as kipd_, kipd_ is now known as kipd, dayangkun left 09:20 Gardner joined 09:21 dayangkun joined 09:23 zakharyas joined
lab Hi, I'd like to do: for @a Z 0 .. Inf -> $element, $index but Z does not flatten, if I get it right, what is needed to get that behavior? 09:24
moritz lab: for flat @a Z 0..* -> $element 09:26
arnsholt_ Or you can destructure
m: my @a = <a b c d e>; for @a Z 0 .. Inf -> ($x, $y) { say $x, $y }
camelia rakudo-moar 7b7888: OUTPUT«a0␤b1␤c2␤d3␤e4␤»
09:26 arnsholt_ is now known as arnsholt 09:28 dalek left
lab arnsholt: thanks, I was looking for a solution without flat, perfect 09:28
RabidGravy is travis-ci being particularly flaky today or is it something weird about the current rakudo build or just my code being shit?
09:29 clkao joined
moritz RabidGravy: panda on travis doens't really work 09:29
RabidGravy this is what I am seeing
moritz RabidGravy: due to the curli ( precomp ) branch merge yesterday
09:32 dalek joined, ChanServ sets mode: +v dalek
[Tux] $me.commute ($work) 09:34
09:36 Ven joined
Ven fwiw, is there any request topics for the advent calendars? I don't have many ideas. Maybe one on introspections in Perl 6, like I did in some of my small modules? 09:37
I'm not sure that'd interest enough people to see how stuff like sixcheck or hydrate6 works.. 09:38
RabidGravy Ven, github.com/perl6/mu/blob/master/mi...instorming
Ven amazing, thanks RabidGravy 09:39
I'm not sure we actually want an article on perlito? afaik the 6 part isn't active 09:40
moritz Ven: I'd find it interesting
09:41 nebuchadnezzar joined, iH2O joined
masak links to github.com/perl6/mu/blob/master/mi...5/schedule to signal-boost it a bit 09:41
09:43 emdashcomma left
dalek : ec4e564 | (Carl Masak)++ | misc/perl6advent-2015/schedule:
claim four slots

I just took some days. If someone feels like they wanted that day, I'm easily convinced to switch.		 09:43
09:44 iH2O left
masak people: help fill the advent calendar! surely we can keep this hallowed tradition alive *on the year of release*? :P 09:44
this will be the seventh year we do an advent calendar. wow.
09:45 yeahnoob left
masak people seem to feel that the topics are running dry, but in my experience that's an illusion. 09:45
there are always things to talk about. doesn't have to be features as such.
I notice no-one is talking about the new supply syntax yet.
that's one of the things to be most proud of this year, and maybe in Perl 6 overall.
09:45 emdashcomma joined
stmuk I don't think the new suppy syntax is even in the doc repo? 09:46
^ supply
masak I also see no slot claimed to talk about modules and the new curli stuff
Ven well then, I'm gonna go ahead and claim a slot
dalek : 43d57ec | ven++ | misc/perl6advent-2015/schedule:
claim a spot
masak \o/
let's have a claim-a-spot party today :D
slot* 09:47
moritz masak++
Ven++
masak Ven: you forgot to write your nick.
Ven masak: plan to claim a spot about 007?
masak: I'm bad at stuff
dalek : 29f3f1f | ven++ | misc/perl6advent-2015/schedule:
add my nick
 09:48
masak Ven: I'll have to think about whether to mention 007. if I do, it'll probably be in the wider context of macros.
moritz masak: please do
masak ok, cool.
09:48 kjs_ joined
moritz I think the scope of the advent calendar can be quite wide 09:49
09:49 bitmap left
RabidGravy stmuk, I may look at that today as the examples in the concurrency doc don't even work now 09:49
masak ++RabidGravy
moritz language features, "how do I retrieve a web page with Perl 6?", cool projects, editor support, philosophy, design of the JIT/VM/Compiler/whatever
dalek : 3120b06 | (Carl Masak)++ | misc/perl6advent-2015/schedule:
by popular demand, status-update on macros/007
 09:50
09:50 dakkar joined 09:51 Juerd left
dalek : e2dd10b | (Steve Mynott)++ | misc/perl6advent-2015/schedule:
Benchmarking - stmuk (itz)
 09:51
09:51 Gardner left, bitmap joined
RabidGravy get in there 09:51
stmuk the examples in the docs probably should be unit tested somehow .. maybe based on expected output in a comment 09:53
09:53 Ven left
stmuk eg. say 2+2 # => 4 09:53
09:53 Gardner joined 09:54 n0xff joined
stmuk not sure if such a framework or similar exists yet in ecosystem 09:54
09:56 znpy` joined 09:57 Juerd joined, g4 joined, g4 left, g4 joined
grondilu pulls and rebootsrap pand, hopes it's ok after the curli merge 09:58
09:58 Ven joined
stmuk it isn't :) 10:00
10:00 joli joined, MilkmanDan joined, joli is now known as aigan
RabidGravy grondilu, I've just nuked the whole rakudo and started again 10:01
nope still broken
is zef better?
grondilu so far it seems to work fine for me
I haven't pulled rakudo since yesterday though 10:02
10:02 rurban joined
RabidGravy nope the whole thing is borked with zef as well 10:04
stmuk and probably many ecosystem tests using @*INC as well 10:05
RabidGravy I'm sure it'll get fixed at some point
moritz hopes this is the last breaking change before Christmas
Ven well well, that dalek commit link crashed my irc client 10:07
moritz Ven: sounds like a security vulnerability waiting to be found 10:08
grondilu oh yeah it crashed indeed
10:08 xpen_ joined
stmuk whats the current recommended way of setting include path for tests BTW? 10:08
moritz stmuk: use lib
10:09 skaji left 10:11 xpen left 10:12 skaji joined
grondilu I do $ PERL6LIB=$path perl6 program.p6 10:13
(usually PERL6LIB=./lib) 10:14
moritz perl6 -Ilib program.p6 # also works
grondilu that's shorter, I'll do that next time 10:15
Woodi hallo #perl6 :) 10:16
10:16 znpy` left, espadrine joined
Woodi I remember some fuzzy testing of (probably) rakudo code... any results or maybe already merged ? 10:17
10:17 clkao_ joined 10:18 clkao left, clkao_ is now known as clkao
Ven moritz: well.. it's closed source so idk 10:18
10:20 donaldh joined 10:25 rindolf joined
donaldh r: my @msgs = <one two three>; my %args = 'oh' => 'my', 'what' => 'fun'; @msgs.push(%args.map: { .key ~ '=' ~ .value }); say @msgs.perl 10:28
camelia rakudo-moar 7b7888, rakudo-jvm e3c591: OUTPUT«["one", "two", "three", ("oh=my", "what=fun").Seq]␤»
donaldh hmm, not what I'm getting from rakudo-star 10:29
moritz donaldh: try .append instead of .push 10:30
donaldh moritz: thanks. I thought push was the wrong thing to use but didn't spot append. 10:31
10:38 dayangkun left
jnthn stmuk: I've got several flights/train journeys coming up, so if I have the energy may try writing up the supply syntax for the docs during one of them. :) 10:41
10:41 domidumont joined 10:42 flaviusb joined
moritz donaldh: note that doc.perl6.org/routine/push even mentions method append :-) 10:43
10:43 eiro joined
moritz (the link is broken through :/ ) 10:44
10:44 YP-QMUL-W joined
donaldh must get into the habit of searching doc.perl6.org instead of desgin.perl6.org 10:45
arnsholt Yeah, I'm starting to learn that too 10:46
10:47 ely-se_ left
donaldh what surprised me was that with star 2015.09 push did flatten the array parameter. I thought it was new enough to have the latest semantics. 10:48
10:48 xpen_ left 10:49 YP-QMUL-W left
jnthn The GLR changes took a while to become equally distributed. :) 10:50
stmuk jnthn: that would be cool but I'd also recommend some relaxation :) 10:51
moritz waterbed theory of GLR changes :-)
stmuk or beanbag 10:53
10:53 eiro left 10:56 eiro joined
RabidGravy jnthn, while we're on the subject of supplies, would: 11:02
m: my $supply = supply { for 1 .. 10 { emit($_); } }; $supply.tap( -> $v { say $v });
camelia rakudo-moar 7b7888: OUTPUT«1␤2␤3␤4␤5␤6␤7␤8␤9␤10␤»
RabidGravy be the new canonical version of the first example in docs.perl6.org/language/concurrency#Supplies
jnthn Well, they're different. 11:03
A supply block is an on-demand supply 11:04
11:04 pdcawley_ joined
jnthn It is an easy first example, though :) 11:04
RabidGravy ah so the "live" one always requires the Supplier and the Supply
jnthn Aye
RabidGravy right so the real equivalent is 11:05
m: my $supplier = Supplier.new; my $supply = $supplier.Supply; $supply.tap( -> $v { say $v }); for 1 .. 10 { $supplier.emit($_); }
camelia rakudo-moar 7b7888: OUTPUT«1␤2␤3␤4␤5␤6␤7␤8␤9␤10␤»
11:05 pdcawley left
jnthn Yeah. 11:06
Good to contrast the two 11:07
(Live and on-demand) 11:08
RabidGravy yeah, that document doesn't really dwell on that much
11:08 rgrau joined
RabidGravy all good, I'll fix the examples forthwith then :) 11:10
11:11 Axord joined
RabidGravy and add a doc for Supplier while I'm playing 11:11
11:13 rgrau left 11:14 kjs_ left 11:15 zakharyas left 11:17 eiro left, eiro joined 11:18 zakharyas joined 11:22 luis` is now known as luis 11:24 Skarsnik joined
masak 14 unclaimed slots in github.com/perl6/mu/blob/master/mi...5/schedule ! 11:25
11:25 znpy` joined
masak don't be shy, people ;) 11:25
11:26 TEttinger left
grondilu doesn't quite understand what this advent thing is about 11:27
masak grondilu: one post per day in December. something cool about Perl 6.
grondilu ok 11:28
RabidGravy masak, I'll do another one if there aren't any takers ( I don't think I'll be up for 14 though) 11:29
tadzik hmm, I wonder what can I write about
maybe Grammar::BNF and how it generates grammars 11:30
11:31 rurban left
tadzik I think it's the only new thing I created this year :o 11:31
masak tadzik: sounds great
dalek : 91fcc78 | RabidGravy++ | misc/perl6advent-2015/schedule:
Update schedule
 11:32
11:32 Ven left
RabidGravy Only other conceivable date for me 11:32
dalek : 6575ccd | tadzik++ | misc/perl6advent-2015/schedule:
Claim 8th for Grammar::BNF
masak I wouldn't mind if someone wrote about hypens and apostrophes in identifiers. I don't think anyone's every taken that as a theme. 11:33
if there are no takers, I can write about it ;)
eiro not skilled enough for writting but i think we can setup a french translation team (i have few people in mind)
stmuk mmm panda uses @*INC
masak eiro: cool.
did we ever do an advent post on Date and DateTime? 11:34
moritz masak: perl6advent.wordpress.com/2010/12/...-in-perl6/
masak ok.
maybe advent blog about the :k thing replacing the -index methods?
nine stmuk: upgrade panda
masak (and the discussion around adverbs connected to that)
jnthn airport & 11:35
masak ooh! someone needs to advent blog about the one-arg rule! :)
and maybe a separate post about GLR and its consequences in general 11:36
stmuk oh I'm on the wrong branch
11:37 xpen joined
stmuk I need a zsh git prompt :/ 11:37
11:37 Spot__ left, xpen left
masak an advent post about the new 'with' keyword 11:37
grondilu btw what's the status on perl6pod? Will the synopsis be written in 6pod someday? 11:38
11:38 xpen joined
stmuk I could probably do a GLR post if someone code reviews it first :) 11:38
masak an advent post about the more liberal and more useful when/default semantics
(inspired by real-world usage!) :D 11:39
11:39 Spot__ joined
Zoffix ugexe, would it make more sense to list your travis file in github.com/ugexe/P6TCI on docs.travis-ci.com/user/languages/perl6/ instead? This way it'll be more visible and not in an unrelated repo... 11:39
11:39 mprelude left
masak an advent post about Levenshtein-based error messages. 11:39
a post about various IO improvements? 11:41
Zoffix lols... 11:42
stmuk newio :>
Zoffix Yak shaving? And for a second I thought we were doing P6 tuts :P
stmuk trolls
Zoffix Hm. I wanna sign up, to this advent thing, but I fear what I talk about will be so dumb, compared to other posts. 11:44
'cause I don't know much P6
stmuk it's a good chance to learn
RabidGravy not as dumb as mine will be ;-)
Zoffix I suppose.
dalek : 3b0f0ff | (Steve Mynott)++ | misc/perl6advent-2015/schedule:
claim GLR
stmuk I'll regret that :/
masak Zoffix: we are pretty good at reviewing each others' posts
stmuk++ 11:45
Zoffix Very well, I'll take the 13th
stmuk maybe the Dec 25th post should be a UUENCODED release tar 11:46
dalek : 4bc3831 | (Zoffix Znet)++ | misc/perl6advent-2015/schedule:
13 IRC and Robots - Zoffix
 11:47
Zoffix Since I'll be writing my New Years bot anyway... 11:48
(and there will be more other robots)
masak secretly hopes that someone with a big cowboy hat claims 24th
Zoffix++
I want to recommend medium.com/@betable/tifu-by-using-...c308c4fd9d from last week. not Perl 6-specific, but very interesting about PRNGs. 11:49
also, made me wonder how good MoarVM's PRNG is... :)
stmuk there is some PRNG audit library to check that which plots random points 11:50
11:51 burnersk joined 11:52 TEttinger joined
dalek : 1120076 | (Zoffix Znet)++ | misc/perl6advent-2015/schedule:
rename post
 11:55
Zoffix There, now we have mystery, innuendo, and alliteration :P
dalek : aadf5e9 | moritz++ | misc/perl6advent-2015/schedule:
Add missing author information (infered by "git blame")
 11:57
11:58 zakharyas left
masak ooh! ooh! nine, please make an advent post where you blow all our socks off by combining Perl 5 and Perl 6 in yet another amazing way :) 11:59
dalek : 197b1d0 | (Nigel Hamilton)++ | misc/perl6advent-2015/schedule:
Nigel Hamilton (aka Trexy) volunteering for the 20th.
 12:01
: 01dcf7b | (Nigel Hamilton)++ | misc/perl6advent-2015/schedule:
Merge branch 'master' of github.com/perl6/mu
: 50173d9 | (Nigel Hamilton)++ | misc/perl6advent-2015/schedule:
Merge branch 'master' of github.com/perl6/mu
Skarsnik I am tempted to do a small thing about NC but I am bad, and I already have too much stuff to do x) 12:02
dalek : d10d7b8 | (Zoffix Znet)++ | misc/perl6advent-2015/schedule:
Change title again: 13th "Going Raw with Rogue Robots" - Zoffix
moritz Skarsnik: NC?
Zoffix NativeCall
Skarsnik NativeCall
masak NativeCall
Skarsnik damn Zoffix too fast
masak that's a good thing if you're doing NativeCall :P 12:03
moritz Skarsnik: you can try to write a post, and only claim a spot when you're confident that it works out
anybody who can't log in on perl6advent.wordpress.com/wp-admin/ but wants to write an advent post, please /msg me your email address
masak moritz++ 12:04
moritz I'll invite you with the necessary privileges
12:05 moznion left 12:06 moznion joined 12:08 ggoebel7 left 12:09 ggoebel7 joined 12:11 eiro left 12:14 cognominal_ joined 12:15 Exodist left 12:17 eiro joined 12:19 eiro left, burnersk left, domidumont left
cognominal_ m: say :(:a).WHAT 12:19
camelia rakudo-moar 7b7888: OUTPUT«(Pair)␤»
cognominal_ m: sub foo(:a) {} 12:20
camelia rakudo-moar 7b7888: OUTPUT«5===SORRY!5=== Error while compiling /tmp/93jQNq0fT4␤Missing block␤at /tmp/93jQNq0fT4:1␤------> 3sub foo(:7⏏5a) {}␤»
cognominal_ m: sub foo(666) { say "evil" }; foo 666 12:21
camelia rakudo-moar 7b7888: OUTPUT«evil␤»
12:22 Ven joined
cognominal_ why the pair :a is not accept as a parameter? 12:22
Skarsnik :$a
12:23 lucasb joined
cognominal_ I don't want to bind a parameter variable, just use it like I did with 666. 12:23
Skarsnik foor(:a) means set the a named parameter to True
12:23 kid51 joined
masak cognominal_: good question. 12:24
12:24 eiro joined, burnersk joined, domidumont joined
masak cognominal_: I can't think clearly about it, but maybe it collides with some other syntax? 12:24
FROGGS masak: hmm, can't imagine that
cognominal_ writing an article, I am testing all sort of edge cases
masak cognominal_: if it doesn't, then what you say makes sense to me
cognominal++
12:24 donaldh left
cognominal_ m: :( :$a ) 12:27
camelia ( no output )
12:28 ShimmerFairy joined 12:29 burnersk left
ShimmerFairy [backlog] I can't imagine why sub foo(:a) { } should declare a variable, since when has :a done something like that? Is it sigilless, does it get an implied $a, would it affect :a($a-long-opt), etc. 12:29
cognominal_ you mean, it except a subsignature or something like that? 12:31
12:31 hartenfels joined, Peter_R joined
ShimmerFairy :a($a-long-opt) is how you make -a and --a-long-opt for sub MAIN, for example 12:31
masak ShimmerFairy: I didn't get the sense that cognominal_ wanted to declare a variable.
ShimmerFairy masak: then why was it used is a sub declaration? 12:32
12:32 burnersk joined
lucasb he wanted to pattern match it with a value 12:32
masak ShimmerFairy: because he wanted to sigbind on the *presence* of an :a parameter, but didn't care about its value
cognominal_ like my evil sub, to pattern match
12:33 sufrostico joined
masak either that, or he wanted to match on :a(True), I dunno 12:33
cognominal_: which one did you intend? :)
cognominal_ masak: well it tests the value otherwise
ShimmerFairy I hate our actually-ternary "boolean" named params :)
cognominal_ m: sub foo(666) {}; foo 42
camelia rakudo-moar 7b7888: OUTPUT«Constraint type check failed for parameter '<anon>'␤ in sub foo at /tmp/UAaykgUEjH:1␤ in block <unit> at /tmp/UAaykgUEjH:1␤␤»
12:34 TEttinger left
masak cognominal_: in `sub foo(:a) {}; foo(:a(False))`, would you expect successful sigbinding? 12:34
cognominal_ m: sub foo(666 $a) { say $a}; foo 666 # not useful, just trying
camelia rakudo-moar 7b7888: OUTPUT«666␤»
Skarsnik sub foo(:$a where $a = True) {say "true"}; sub foo(:$a where $a = False) { say "false"}; foo(:a);
m: sub foo(:$a where $a = True) {say "true"}; sub foo(:$a where $a = False) { say "false"}; foo(:a); 12:35
camelia rakudo-moar 7b7888: OUTPUT«5===SORRY!5=== Error while compiling /tmp/e5PeG20ZeN␤Redeclaration of routine foo␤at /tmp/e5PeG20ZeN:1␤------> 3foo(:$a where $a = False) { say "false"}7⏏5; foo(:a);␤ expecting any of:␤ horizontal whitespace␤ postfix…»
ShimmerFairy masak: I think the issue here is that matching against 666 is matching against a value, while :a is matching against a variable name, so I don't think they're actually comparable cases.
lucasb m: say my (1,2,3) # just crazy :)
camelia rakudo-moar 7b7888: OUTPUT«(Any)(Any)(Any)␤»
Skarsnik hm, I don't get my error x) 12:36
aside it should be $a == True probably
cognominal_ :a is not a variable, it is the pair a => True
it is a value and I want to pattern match against the value. 12:37
ShimmerFairy Skarsnik: you need 'multi sub'
RabidGravy the best I can do is
m: sub foo(Pair $ where *.key eq "a" ) { }; foo("a" => True)
camelia ( no output )
Skarsnik Oh I did not see the first part
ShimmerFairy cognominal_: it's not a Pair object, it's just a named parameter, which is a subtle, annoying, and important
Skarsnik m: multi sub foo(:$a where $a = True) {say "true"}; multi sub foo(:$a where $a = False) { say "false"}; foo(:a); 12:38
camelia rakudo-moar 7b7888: OUTPUT«true␤»
Skarsnik m: multi sub foo(:$a where $a = True) {say "true"}; multi sub foo(:$a where $a = False) { say "false"}; foo(:a(False));
camelia rakudo-moar 7b7888: OUTPUT«Cannot call foo(:!a); none of these signatures match:␤ (:$a where { ... } = { ... })␤ (:$a where { ... } = { ... })␤ in block <unit> at /tmp/9_m0gMvbso:1␤␤»
ShimmerFairy also, I believe the equality test is ==, not = :)
cognominal_ ShimmerFairy, it may want to parse as a named parameter but it is not one.
Skarsnik m: multi sub foo(:$a where $a == True) {say "true"}; multi sub foo(:$a where $a == False) { say "false"}; foo(:a(False));
camelia rakudo-moar 7b7888: OUTPUT«false␤»
ShimmerFairy (or rather === for us)
Skarsnik m: multi sub foo(:$a where $a == True) {say "true"}; multi sub foo(:$a where $a == False) { say "false"}; foo(:a(True));
camelia rakudo-moar 7b7888: OUTPUT«true␤»
ShimmerFairy cognominal_: doesn't really matter what you want, unfortunately. In this case it is a named parameter. 12:39
Skarsnik I did not think the contraint was part of the signature
masak cognominal_: I just want to note that I disagree with ShimmerFairy's reasoning on this one. :)
cognominal_ what surprises me is that :( :a ) parses correctly but not foo(:a) {}
masak cognominal_: and I think your use case is sound (if a bit unusual)
cognominal_ definitvely unusual :) 12:40
lucasb so, in other words, P6 can pattern match against all other values, *except* pair values... right?
masak cognominal_: I could see it working quite well in multi MAIN parameter lists
cognominal_ as I said I probe edge cases.
lucasb: strange indeed :)
ShimmerFairy m: sub classify-pairs(*@a, *%b) { say "PAIR: @a[]\nNAMEDS: %b<>" }; classify-pairs(1 => 2, :a, "B" => True, b => True);
camelia rakudo-moar 7b7888: OUTPUT«PAIR: 1 2 B True␤NAMEDS: a True␤b True␤»
ShimmerFairy masak: AFAICT, this is the named/pair issue that I and lizmat have complained about before, so note that I'm not agreeing. I still think the conflation of named params and Pairs in syntax is a mistake, and this shows why :) 12:41
cognominal_ m: say :(:a)[0] # more edgy 12:42
camelia rakudo-moar 7b7888: OUTPUT«a => True␤»
ShimmerFairy m: sub classify-pairs(*@a, *%b) { say "PAIR: @a[]\nNAMEDS: %b<>" }; classify-pairs(a => 1, 1 => 2); # and just to show that autoquoting isn't the only rule for determining a named...
camelia rakudo-moar 7b7888: OUTPUT«PAIR: 1 2␤NAMEDS: a 1␤»
cognominal_ so here it is a value in the positional part of the signature. 12:43
masak ShimmerFairy: a big difference though is that in the above, you show named/pair differences in an *argument list*, but cognominal_'s requested use case is in a *parameter list*. 12:45
ShimmerFairy cognominal_: really? That just makes it worse
cognominal_ I am not it. But I am not a bad guy, just a worse guy :) 12:46
ShimmerFairy m: say :(:a).list; say :(:a).pairs; 12:47
camelia rakudo-moar 7b7888: OUTPUT«(a => True)␤(a => True)␤»
12:47 kid51 left
ShimmerFairy m: say :(:$a).list; say :(:$a).pairs; 12:47
camelia rakudo-moar 7b7888: OUTPUT«((:$a))␤(0 => (:$a))␤»
ShimmerFairy m: say :($z, :$a).list; say :($z, :$a).pairs;
camelia rakudo-moar 7b7888: OUTPUT«(($z, :$a))␤(0 => ($z, :$a))␤»
RabidGravy well I have all the existing Supply examples in the concurrency doc working 12:52
now to get the words to make sense in the new order 12:53
12:56 znpy` left
cognominal_ in both foo(:a) and :(:a), the ":a" should be a parameter of the signature rule. So what context make them behave differently? 12:59
13:00 eiro left, domidumont left 13:01 domidumont joined, eiro joined
ShimmerFairy Y'know, I can't yet figure out why rakudo insists on creating a system directory now, where before I could just have modules go in ~/.perl6 and be on my merry way. 13:02
stmuk RabidGravy++ 13:03
cognominal_ [13:34] <masak> cognominal_: in `sub foo(:a) {}; foo(:a(False))`, would you expect successful sigbinding?
I missed your question. My answer is no. That's a different value. 13:04
masak cognominal_: what about `foo(:a(42))` ? 13:05
cognominal_: what about `foo()` ?
lucasb no too! because a=>True is different than a=>42 :)
masak the reason I ask the last one is that nameds are optional by default 13:06
m: sub foo(:$a) { say "see? works!" }; foo() 13:07
camelia rakudo-moar 7b7888: OUTPUT«see? works!␤»
13:07 burnersk left
El_Che Hi, I am trying to change set the record serparator in perl6 like in perl5: 13:07
this will separate records by empty lines: perl -e 'local $/ = "\n\n"; open(my $fh,"<","auth2_200"); while (<$fh>) { print "___\n" . $_ . "__\n" }'
cognominal_ it is not a named, it is a value used as a type.
El_Che I tried this in p6, but to no avail: my $fh = open('auth2_200', nl => "\n\n"); for $fh.lines -> $line { say "___\n" ~ $line ~ '___' }
cognominal_ m: sub foo(666 $a) {} # the 666 here is a value used as a type in a signature. 13:08
camelia ( no output )
cognominal_ m: sub foo(666) {} # same thing but not parameter var is bound when called. 13:09
camelia ( no output )
cognominal_ * no parameter var
lucasb ^^ these 2 subroutines *are* different!
the last one pattern matches against 666, the first one don't 13:10
m: my ('oh' $x, 'hai' $y)
camelia ( no output )
lucasb that is just strange syntax that happens to be valid, but has no meaning, I think
cognominal_ m: sub foo(666 $a) { say $a }; foo 42
camelia rakudo-moar 7b7888: OUTPUT«Constraint type check failed for parameter '$a'␤ in sub foo at /tmp/Ju3yphn8Wq:1␤ in block <unit> at /tmp/Ju3yphn8Wq:1␤␤»
cognominal_ m: sub foo(666 $a) { say $a }; foo 666 13:11
camelia rakudo-moar 7b7888: OUTPUT«666␤»
13:11 domidumont left
lucasb cognominal_: oops, you are right, sorry :) 13:11
Skarsnik El_Che, hm no error?
lucasb I was going by the 'my' declaration...
cognominal_ lucasb, what is interesting is the various expectation here :) 13:12
13:12 abaugher left
El_Che Skarsnik: no, it separated the records on the same way as "\n" 13:12
Skarsnik: so it keeps separting by line instead of by record
13:13 abaugher joined
cognominal_ m: my ('ho' $x) = 'so bad' 13:13
camelia ( no output )
cognominal_ m: my 'ho' $x = 'so bad' 13:14
camelia rakudo-moar 7b7888: OUTPUT«5===SORRY!5=== Error while compiling /tmp/YGc3f1alfc␤Malformed my␤at /tmp/YGc3f1alfc:1␤------> 3my7⏏5 'ho' $x = 'so bad'␤»
Skarsnik Defaults to "EOL", which implies accepting any combination of "\n", "\r\n" or "\r" or any other Unicode character that has the Zl (Separator, Line) property.
hm I wonder if seeing \n\n it fallback to the new line graphem
cognominal_ m: my ('ho' $x) := 'so bad'
camelia rakudo-moar 7b7888: OUTPUT«Too few positionals passed; expected 1 argument but got 0␤ in block <unit> at /tmp/fAzYAjD8y0:1␤␤»
El_Che it looks like it
Skarsnik try with like an arbitrary letter? 13:15
to see if nl work
El_Che it does
cognominal_ m: my 'ho' $x := 'so bad'
camelia rakudo-moar 7b7888: OUTPUT«5===SORRY!5=== Error while compiling /tmp/VhWoafZuOZ␤Malformed my␤at /tmp/VhWoafZuOZ:1␤------> 3my7⏏5 'ho' $x := 'so bad'␤»
13:15 eiro left 13:16 znpy` joined
El_Che Skarsnik: single char and even strings work 13:16
13:16 eiro joined
cognominal_ So there is also an inconstency, values as types are not syntactically accepted everywhere a type is expected. 13:17
13:17 oka_ left
ilmari m: my ('now' $x) = ('now', 'what') 13:17
camelia ( no output )
ilmari m: my ('now' $x) := ('now', 'what')
camelia rakudo-moar 7b7888: OUTPUT«Too many positionals passed; expected 1 argument but got 2␤ in block <unit> at /tmp/vCiWB3F3OP:1␤␤»
ilmari m: my ('now' $x) := ('now')
camelia rakudo-moar 7b7888: OUTPUT«Too few positionals passed; expected 1 argument but got 0␤ in block <unit> at /tmp/P6I8r7mxRl:1␤␤»
ilmari counting is hard?
cognominal_ limari: :)
moritz ilmari: did you mean to put a comma between the 'now' and the $x? 13:18
Skarsnik El_Che, hm fill a bug on rt.perl.org or wait for someone more knowledgable of this stuff to answer x)
ilmari moritz: no
El_Che it looks like whenever you put a \n in the nl it ignores other chars
cognominal_ moritz, we are testing values used as types.
moritz ah 13:19
ilmari m: my (Str $x) := ('wat')
camelia rakudo-moar 7b7888: OUTPUT«Too few positionals passed; expected 1 argument but got 0␤ in block <unit> at /tmp/ESq0ilqfXo:1␤␤»
ilmari m: my (Str $x) := ('wat', 'now')
camelia rakudo-moar 7b7888: OUTPUT«Too many positionals passed; expected 1 argument but got 2␤ in block <unit> at /tmp/2GHl1UGADQ:1␤␤»
cognominal_ btw, I don't neceserally complain about the inconsistency. Just noting its existence.
ShimmerFairy please tell me I'm not seeing the full picture, but is src/core/Process.pm seriously setting up a %CUSTOM_LIB that nobody uses? 13:21
13:21 eiro left, eiro joined
cognominal_ Also in my ('ho $x ) := 'so bad', it should complain that 'so bad' is not 'ho'. 13:22
moritz ShimmerFairy: nine++ has already announced that he'll fix that later
13:22 Celelibi joined
cognominal_ getting deeper in the maze :( 13:22
ShimmerFairy moritz: that won't fix my problem though, and it's not helping my understanding of said problem
13:23 llfourn left, hanekomu joined
cognominal_ notes he should uses tests with more positive connotations 13:24
ilmari ShimmerFairy: panda uses it
ShimmerFairy ilmari: it uses %*CUSTOM_LIB, which rakudo doesn't make, not %CUSTOM_LIB, which (AFAICT) nobody uses :P 13:25
ilmari m: say %*CUSTOM_LIB
camelia rakudo-moar 7b7888: OUTPUT«Dynamic variable %*CUSTOM_LIB not found␤ in block <unit> at /tmp/R58EOM6ANe:1␤␤Actually thrown at:␤ in block <unit> at /tmp/R58EOM6ANe:1␤␤»
moritz ShimmerFairy: %CUSTOM_LIB in GLOBAL becomes %*CUSTOM_LIB 13:26
ilmari PROCESS::<%CUSTOM_LIB> := %CUSTOM_LIB;
13:26 burnersk joined
ShimmerFairy moritz: but it's being bound to PROCESS 13:26
ilmari $ p6 -e 'say %*CUSTOM_LIB'
home => /home/ilmari/.perl6/2015.10-268-g31784a7, perl => /home/ilmari/.rakudobrew/moar-nom/install/share/perl6, site => /home/ilmari/.rakudobrew/moar-nom/install/share/perl6/site, vendor => /home/ilmari/.rakudobrew/moar-nom/install/share/perl6/vendor
lucasb m: PROCESS::<$oh-hai> = 42; say $*oh-hai # ShimmerFairy: PROCESS also prepends a "*" 13:27
camelia rakudo-moar 7b7888: OUTPUT«42␤»
ShimmerFairy moritz: the bug I'm encountering btw is the fact that trying to use non-installed libraries (e.g. those found via -I) makes perl6 want to make a world-writable system directory.
lucasb: ah. I'm not familiar with cases where a wild * appears, so I wouldn't have known :P 13:28
13:30 burnersk left
moritz ShimmerFairy: it should try to store the precompiled files in ~/.perl6 or so 13:30
ShimmerFairy oh, but it don't. Not even --target=parse will succeed :D 13:31
timotimo o/ 13:32
RabidGravy: i'm very glad you're taking on the supply docs 13:33
13:34 NativeCallUser joined
NativeCallUser hi all 13:34
timotimo oh hai
masak hi, NativeCallUser
RabidGravy :) I was just procrastinating until I was certain that it was all settled
Skarsnik Interesting name x) 13:35
masak .oO( I wonder what FFI he uses? )
RabidGravy timotimo, the difficulty with the new Supply world order is that there are too many ways to do the same examples ;-) 13:37
13:37 xpen left
NativeCallUser A question about NativeCall and Win32 API. I read the NativeCall manual and tried the example with MessageBox and it works. So I want to use the Win32 API function GetCursorPos. The function signature is '__Out__ LPPOINT lppoint'. and the return value 'BOOL WINAPI'. Both is not mentioned in the NativeCall help. Someone an idea what to write here? 13:37
13:37 rindolf left 13:39 kjs_ joined
RabidGravy well they won't be mentioned as they are application specific, for the LPPOINT you will be looking at a CStruct I guess and BOOL maybe an int8 maybe something else 13:39
stmuk ufo++ # can be hacked up to work as a fallback
RabidGravy you will need to find out from the windows api documentation 13:40
NativeCallUser I tried Pointer[CArray].
13:40 eiro left, eiro joined, ShimmerFairy left
FROGGS NativeCallUser: sub GetCursorPos(CArray) returns uint8 { * }; my $lppoint = CArray.new; $lppoint[0] = 0; say GetCursorPos($lppoint); say $lppoint[0] 13:40
NativeCallUser: something like that 13:41
NativeCallUser Win API said a pointer to a POINT struct.
Skarsnik NativeCallUser, bool is probably unsigned char so try int8. for llpoint no idea what it's supposed to be x)
dalek osystem: 9ffffb6 | Demayl++ | META.list:
Add Email::Valid

See github.com/Demayl/perl6-datetime-format
osystem: dd57afa | (Zoffix Znet)++ | META.list:
Merge pull request #97 from Demayl/master

Add Email::Valid: github.com/Demayl/perl6-Email-Valid/
Skarsnik just do OpaquePointer then
*use
RabidGravy class LPPOINT is repr('CStruct') { has int32 $.x; has int32 $.y; } 13:42
FROGGS Skarsnik: we can merge the bool stuff today when I got time, though I want to modify your PR before merging it
RabidGravy google is your friend
FROGGS ohh
RabidGravy (those longs may be int64 depending on the actual OS long int size I guess) 13:43
Skarsnik nc as the long type I think 13:44
13:44 abaugher left
timotimo that'? right 13:44
that's right, i mean.
NativeCallUser @all thanks for the help 13:45
Skarsnik FROGGS, I try to think a bit more about the bool, but only rakudo has a configure stuff?
13:46 burnersk joined
FROGGS Skarsnik: we've got build/probe, which is what I prefer to #ifdefs 13:46
13:46 abaugher joined
Skarsnik *self promotion* You can try NativeCall::TypeDiag when trying to figure how to map cstruct 13:46
13:46 ggoebel7 left
stmuk oh that sounds useful 13:46
13:47 YP-QMUL-W joined
Skarsnik FROGGS, configure.pl refuse to work on my forked repository (it try to git upstream). that why I can't never test this code directly in my fork ~~ 13:48
garu hi everyone! I'm having some trouble updating my perl6 installation with rakudobrew + panda => paste.scsys.co.uk/502180
timotimo how do typical json parsers handle string literals starting in a combining character?
13:48 YP-QMUL-W left, Celelibi left, hartenfels left
timotimo garu: you'll have to nuke your install/ folder under .rakudobrew/moar-nom/ i'm afraid 13:49
garu: we just merged the magical curli branch and were under too much time pressure to put in proper deprecation stuff and such
garu timotimo: I was thinking about it but thought I'd rather show this to you first 13:50
maybe I can help?
timotimo yup, it's okay
i don't know unfortunately
13:50 aigan left
NativeCallUser RabidGravy++ your hint with the class was the right one. It works :-) 13:50
RabidGravy Yay!
garu that's fine, I'll just reinstall - you keep doing all the great stuff you've been doing! :D 13:51
Skarsnik Ooh curli is merged, nice
stmuk garu: you could try specifying an older version of rakudo and panda and waiting maybe 2 days
garu: you can pass sha1 hashs into rakudobrew 13:52
13:52 ShimmerFairy joined
garu stmuk: naah, it's fine, really. Just thought if it was simple enough that maybe I could help in some ways. You know, clear the path so you can keep working on making things awesome for christmas :) 13:53
timotimo damn, i must have thrown away that big json file i used for performance analysis of JSON::Fast
13:53 Ven left
garu but if it's a known issue with not enough value to be fixed, I'll just reinstall and be happy 13:53
stmuk its a major change which isn't fully changed yet 13:54
Skarsnik timotimo, profile a big perl6 script and use the ouput json as test? x) 13:55
timotimo omg.
well, i want to improve the nom-ws function ... the profiler doesn't output any additional whitespace whatsoever
13:55 n0xff left
garu timotimo: maybe this helps? raw.githubusercontent.com/zemirco/...ylots.json 13:55
timotimo mtgjson.com/ - using some of those now 13:56
13:56 zacts left 13:57 NativeCallUser left
timotimo hm, i'll have to pretty print that first 13:57
it's also devoid of whitespace
ugh, these precomp file paths are really ugly for our profiler UI 13:58
hum. nom-ws only takes 8.3% of the time when reading AllCards.json (with added whitespace) 14:00
tadzik ooh, mtgjson
garu timotimo: citylots.json is 181M large and pretty-printed (with whitespaces and such) 14:01
anyway, let me get back to reinstalling moar-nom 14:02
thanks everyone!
14:02 emdashcomma left
timotimo :) 14:03
14:03 ggoebel7 joined
timotimo garu: sadly, AllCardsWhitespace.json (which i generated with json_reformat) already takes 15s to parse 14:04
grr. i don't want to have a block inside nom-ws, but i can't have CATCH without curlies
moritz use try without block 14:06
timotimo i fear that'll be slower, though
i will measure.
yeah, makes the whole thing 5s slower :( 14:07
garu try? You mean as a keyword? I thought it didn't exist (can't see it in docs.perl6.org/language/exceptions)
timotimo it does
garu sorry, I don't want to sidetrack you
nevermind me! 14:08
timotimo no worries
moritz: changing the code to use try accidentally prevented jit compilation of nom-ws
14:08 ShimmerFairy left, burnersk left
moritz :( 14:08
timotimo but nom-ws won't get much better than this. it's already comparatively cheap 14:09
do you want to know how long JSON::Tiny takes? :P 14:10
moritz yes 14:11
timotimo i'm measuring it now
moritz you can tell me tomorrow :-)
timotimo get yourself a cup of coffee and a few books or solitaire games! :P
1:19.74 14:12
moritz wow
timotimo m: say "JSON::Fast was able to get it in { (60 + 19.74) R/ 15.08 } of the time" 14:13
camelia rakudo-moar 7b7888: OUTPUT«JSON::Fast was able to get it in 0.189115 of the time␤»
timotimo i expect any "real" json parser would do this in a milisecond instead
the memory usage of JSON::Fast was also quite a bit better 14:15
m: say "::Fast took { 231452 / 931076 }x as much memory "
camelia rakudo-moar 7b7888: OUTPUT«::Fast took 0.2485855x as much memory ␤»
timotimo not really proud of the fact moar balloons up to 230 MB to parse a 8.2 MB json file ... 14:16
tadzik looks like you at least have some decent profiling material now :)
masak I'm running a home-brewed coverage tool on the 007 source code. been running it for about 3 consecutive days. 14:17
timotimo thing is, the function that parses a string takes far too many closures :( 14:18
14:18 lucasb left 14:19 oka_ joined
dalek rl6-roast-data: b9876ee | coke++ | / (9 files):
today (automated commit)
 14:19
El_Che at the moment I am making my VM swap, slurping a 140MB ldif dump and putting it into a hash :)
14:19 hartenfels joined
timotimo masak: how did you build it? 14:20
tadzik: 4924342 BOOTCode objects allocated from the parse_string function's inner loop alone
Woodi El_Che: with Perl6 ? :)
El_Che Woodi: yep :)
timotimo but also: 4924342 Int objects
tadzik hm 14:21
timotimo i'm not entirely sure why those aren't natives, tbh
tadzik sounds like it could use an Int table of sorts
or that
timotimo we do have an Int table built-in, but only for small ints
14:22 ggoebel7 left
masak timotimo: it's just a Perl 6 script that goes through the code base line by line, and checks if things build and pass the tests without that line. github.com/masak/007/blob/fuzzer/tools/fuzzer 14:22
14:23 kivutar joined
masak timotimo: it's only a little bit clever, in that it knows that a line begins a block, it removes the whole block 14:23
RabidGravy would it be fair to say that "supply { }" is pretty much sugar for "Supply.on-demand({})" ?
masak timotimo: it's already found a number of pieces of the code that simply didn't do anything. :) 14:24
14:26 domidumont joined
timotimo oh, cute! 14:26
tadzik: turns out spesh isn't clever enough yet to remove the p6bool before eq's return, even after inlining 14:27
tadzik: so all the "eq" operators that check for weird stuff will actually box an int (which is what Bool now is)
14:28 burnersk joined
ilmari m: Bool.^mro.say 14:28
camelia rakudo-moar 7b7888: OUTPUT«((Bool) (Int) (Cool) (Any) (Mu))␤»
tadzik masak: hah, that's not very far from mutation testing :)
ilmari oh, I thought it was an Enum?
tadzik timotimo: oh gawd
14:29 balazs joined
masak tadzik: I think I was inspired by that, yes. 14:29
tadzik: I may decide to try some mutation testing, too. but "removed lines" was the lowest-hanging fruit.
tadzik right
and doesn't require AST mangling :)
masak no, but I'd like for it to know the AST better :)
masak .oO( someone should specify a format for that ) 14:30
timotimo ilmari: enums are Int by default. Bool used to be "not a true Enum" for a long time, because of bootstrap difficulties 14:32
ilmari timotimo: ah 14:33
masak oh, that's fixed now? 14:34
whoever++
007 solved that issue by omitting Bool; it only has Int ;) 14:35
though if/when 007 gets enums, it'll probably also get a real Bool
timotimo that satisfying feeling when you fix the problem and the problem is still there ... >_>
masak why is the problem still there? 14:36
[Coke] panda borkage with r-latest, even after removing ~/.perl6
timotimo masak: good question! 14:37
i'm investigating
[Coke] Could not find Panda:ver<True>:auth<True>:api<True> in: /Users/williamcoleda/.perl6/2015.11-292-g7b78887
hartenfels I just meant to post that too.
Same error in panda, despite full reinstall and rm -rf ~/.perl and adding install/share/perl6/site/bin to PATH. 14:38
masak timotimo: I should have asked, rather, "what is it that you see that constitutes the problem still being there?" :)
14:39 moznion left
nine [Coke]: can you gist panda's boostrap output? 14:39
lizmat .botsnack
yoleaux :D
14:39 ZoffixW joined 14:40 skaji left
ZoffixW Do we have a large version of Camelia with transparent background? Like this one, but larger: perl6.org/camelia-logo.png 14:40
14:40 abaugher_ joined 14:41 abaugher left
timotimo hm, i somehow get to 12.9s now 14:41
14:41 kivutar left, Skarsnik left, dakkar left
timotimo masak: the profile showed lots and lots of Int allocations still 14:41
those are gone now \o/
14:41 Axord left
timotimo 262 GC runs now instead of the ~310 as before 14:42
14:42 znpy` left
havenwood ZoffixW: upload.wikimedia.org/wikipedia/com...amelia.svg 14:43
ZoffixW Thanks.
masak timotimo: cool! 14:44
14:45 Some-body_ joined, DarthGandalf left, Some-body_ is now known as DarthGandalf
ZoffixW Does anyone hate the modern look on perl6advent.wordpress.com/ ? Sadly WP doesn't let you edit almost everything... unless you pay them. I'm tempted to revert back to the old, cuter look—too bad it's so narrow and not mobile-friendly. 14:45
14:46 zakharyas joined
leedo personally, i like it more than the previous layout 14:46
timotimo masak: can hardly feel the improvement so far :\ 14:47
it feels like i'm fighting an impossible battle
no matter how good i micro-optimize this, it'll still be 100x slower than what people expect
ZoffixW :( 14:48
timotimo or rather, even if i get it to be 100x faster, people will be all like "you call this JSON::Fast ?!?!"
14:48 khw joined
masak timotimo: I promise not to be all like that :P 14:51
timotimo: I will be all like "whoa! you made it 100x faster! here, have a beer!" 14:52
14:52 hobbs left, pnu left, solarbunny left, ponpon is now known as ponbiki
timotimo haha 14:53
hm, today's improvements were a bit better than i felt them to be
[Coke] ZoffixW: no hate, no.
14:53 retupmoc_ is now known as retupmoca
MadcapJake ZoffixW: any way to get some more color when viewing an article? 14:53
timotimo m: say "got down to { 12.65 / 16 }x as much time as before"
camelia rakudo-moar 7b7888: OUTPUT«got down to 0.790625x as much time as before␤»
timotimo 20%? that's way more than i thought 14:54
[Coke] arglebargle. failure mode has now changed to reflect the "doesn't work great with http proxy" issue. 14:55
jdv79 well, if my clone ever completes i'll volunteer for some advent slots... how big is that repo?
ZoffixW oh gawds... I forgot the name of the original theme.. Dyrson, Dyson, Dysron.. something :S
Woodi timotimo: 21% even. what is new % if thing need to be 100x faster ? ;)
[Coke] Could not download module metadata: Failed to resolve host name
ZoffixW MadcapJake, well, yes, if the article author adds a header pic
[Coke] jdv79: you can edit a file via the github web ui.
MadcapJake ah ok that's cool then!
jdv79 oh...
Woodi btw. is there some algoritm to parse JSON ?
timotimo m: say "needs to be { 16 / 100 }s, is 12.6s. still { (16 / 100) / 12.6 }x to optimize down to." 14:56
camelia rakudo-moar 7b7888: OUTPUT«needs to be 0.16s, is 12.6s. still 0.012698x to optimize down to.␤»
abraxxa rakudo.org/2015/11/28/announce-raku...w-in-beta/ contains a typo is anyone likes to fix it: 'since the ast Rakudo Star release' should be 'last'
timotimo tadzik: maybe you want to include the last JSON::Fast commit in panda? 14:57
m: say "JSON::Tiny used to be { (60 + 19.74) / 16 }x slower than JSON::Fast. now that ratio is { (60 + 19.74) / 12.6 }x" 14:58
camelia rakudo-moar 7b7888: OUTPUT«JSON::Tiny used to be 4.98375x slower than JSON::Fast. now that ratio is 6.328571x␤»
ZoffixW man. I could use a git history or something right now -_
[Coke] 'make test' on JVM very broken.
14:59 Axord joined
ZoffixW Wait. 14:59
14:59 znpy` joined
[Coke] nine: ah. the error is hugely different between bootstrap and rebootstrap 14:59
ZoffixW Oh, I thought [Coke] said they hate it.
dalek : b6ce0b5 | jdv79++ | misc/perl6advent-2015/schedule:
add cpan for advent on 22
 15:00
[Coke] bootstrap gives me the host name error (because no http proxy support out of the box)
abraxxa please don't start the ::Tiny and ::Fast war already!
timotimo jdv79: cool :)
[Coke] ZoffixW: I am coke, and I did not say that.
moritz it's not a war
jdv79 hopefully installing will work by then
timotimo abraxxa: there is no war here; you either want performance or you want readable code :)
moritz ist's simply optimizing for differnt things
ZoffixW [Coke], yeah, misparse on my part :)
.oO( I am Commander Sheppard and I support this store... )
[Coke] jeez, and now the rebootstrap error has changed:
jdv79 [Coke]: thanks. much faster. 15:01
[Coke] gist.github.com/coke/b2ce9de03b57c475777d
hartenfels [Coke]: I had that error before, rm -rf ~/.perl6 fixed it I think. 15:02
[Coke] hartenfels: as I said, that didn't help.
hartenfels Oh, you already re-tried it after this error?
[Coke] but there's probably no point in trying to fix the other things if I'm still stuck behind an http proxy. 15:03
abraxxa timotimo: performance
readable code can be compiled to performing code
timotimo yeah ... at some point :)
you're free to work on the compiler, optimizer and spesh! :)
masak decommute & 15:04
nine [Coke]: I suspect, that rebootstrap is kind of obsolete, since you won't have to reinstall the modules after upgrading rakudo anymore. 15:05
15:05 skids joined 15:06 hobbs joined, pnu joined, solarbunny joined
hartenfels Panda be working for me again at least. 15:06
dalek kudo/nom: 0862289 | (Stefan Seifert)++ | src/core/Process.pm:
Store CompUnit::Repositories in %*CUSTOM_LIB

This makes the CompUnit::Repository objects easily available in panda for more intelligent picking of the target repository.		 15:08
nine Now all that's missing is a trivial patch to panda to check %*CUSTOM_LIB<site>.can-install and use that preferrably 15:09
dalek kudo/nom: 266f5d2 | lizmat++ | src/core/Str.pm:
Make Str.samespace about 50x faster

Which should have noticeable effects on ss/.../.../		 15:11
timotimo oooh, shiny! 50x!
ZoffixW lizmat++ 15:13
RabidGravy more! whip the hamsters harder!
15:14 hobbs left, pnu left, solarbunny left
lizmat just noticed the new perl6advent theme mangles paragraph headers into uppercase 15:18
which makes paragraph headers like ":exists" turn up as the incorrect and confusing :EXISTS
ZoffixW lizmat, I'm changing it. I kinda hate it.
lizmat e.g.perl6advent.wordpress.com/2013/12/...-only-way/ 15:19
ZoffixW I just wish Wordpress theme browser weren't so annoying.
timotimo i notice we don't have anything to give us low-level || and && on int vars
not sure how to tease something good out of our current compiler for that
oh, actually, that's not really about the candidates we have in the setting 15:20
dalek kudo/nom: 9848c20 | grondilu++ | src/core/native_array.pm:
use a Seq for permutations

use a proper iterator, with the lexicographic method. See: L<en.wikipedia.org/wiki/Permutation#..._order>		 15:21
kudo/nom: 43c3ab9 | grondilu++ | src/core/native_array.pm:
remove unnecessary die line

as suggested by psch on IRC
kudo/nom: db5397e | grondilu++ | src/core/native_array.pm:
use :end in permutations
kudo/nom: b71084b | grondilu++ | src/core/native_array.pm:
using a named argument to new in permutations

Not sure why travis did not catch the error here.
kudo/nom: 8fb631a | lizmat++ | src/core/native_array.pm:
Merge pull request #596 from grondilu/patch-1

use a Seq for permutations
timotimo fortunately the code i'm looking at right here is only called in an error state and so is not hot at all 15:22
but ... hot damn, this bytecode we generate! %)
jdv79 where are the places that import/export stuff happens? is that nqp? 15:23
ilmari m: class Foo { has $!n; submethod BUILD(:$!n) {}; methdo enn { $!n } }; Foo.new( n => 42).enn.say 15:26
camelia rakudo-moar 7b7888: OUTPUT«5===SORRY!5=== Error while compiling /tmp/m0hmxL7LQI␤Variable $!n used where no 'self' is available␤at /tmp/m0hmxL7LQI:1␤------> 3bmethod BUILD(:$!n) {}; methdo enn { $!n7⏏5 } }; Foo.new( n => 42).enn.say␤»
jdv79 masak: what about when and default are you referring to above?
ilmari m: class Foo { has $!n; submethod BUILD(:$!n) {}; method enn { $!n } }; Foo.new( n => 42).enn.say 15:28
camelia rakudo-moar 7b7888: OUTPUT«42␤»
lab after half a day trying to put something together in P6, I am not worried about P6 speed anymore; I am more worried about my own speed! 15:30
timotimo jdv79: i think it's all in CompUnitRepo and friends?
let me look
grondilu also regarding permutations, please consider looking at github.com/perl6/roast/issues/79
timotimo look for REQUIRE_IMPORT for example, jdv79 15:31
dalek c: 2aa797d | RabidGravy++ | doc/Language/concurrency.pod:
Make the existing examples for Supply work again.
c: d0d1113 | RabidGravy++ | doc/Language/concurrency.pod:
Explain the difference between live and on-demand
c: 0171c24 | RabidGravy++ | doc/Type/Supply.pod:
Supply is no longer a role
timotimo that should give you a place to start
dalek c: fb3a9d0 | RabidGravy++ | doc/Type/Supplier.pod:
Add some docs for Supplier
15:31 cygx joined
jdv79 timotimo: i'm trying to triage rt.perl.org/Public/Bug/Display.html?id=126761 a bit 15:31
timotimo for some reason that's in operators.pm 15:32
cygx o/
jdv79 but pretty sure a day oe my time is like 1min of jnthn or someone else that knows that code
timotimo hm. probably not that strange to put it there
jdv79 *of
lab It will get just good :). in the meantime, having an array ref in a variable, doing $v.WHAT gives me an error which I think I kinda understand (Cannot lookup attribute in a type object) how do I get that it is an array ref, like P5 ref did?
cygx .tell jnthn irclog.perlgeek.de/perl6/2015-11-20#i_11574374 was about a non-existing dynamical variable, cf github.com/rakudo-p5/v5/commit/a41...d0ddf215c4
yoleaux cygx: I'll pass your message to jnthn.
15:32 g4 left, burnersk left 15:33 rurban joined
RabidGravy m: my $v = [1,2,3]; say $v.WHAT; 15:33
camelia rakudo-moar 8fb631: OUTPUT«(Array)␤»
15:33 domidumont left
RabidGravy lab, ^ 15:33
15:34 Exodist joined
lab RabidGravy: I get the message I posted above. Cannot lookup attributes in a type object 15:34
RabidGravy m: my $v = [1,2,3]; say so $v ~~ Array; # though that is more idiomatic
camelia rakudo-moar 8fb631: OUTPUT«True␤»
15:34 domidumont joined, skaji joined
RabidGravy lab, show us your code then 15:34
15:34 moznion joined
RabidGravy or rather the smallest code that exhibits that behaviour 15:35
15:35 burnersk joined
ilmari m: my $v = Array; say $v.WHAT 15:36
camelia rakudo-moar 8fb631: OUTPUT«(Array)␤»
15:36 bpmedley joined 15:37 emdashcomma joined, PotatoGim left, PotatoGim joined 15:38 kivutar_ joined
lab first thing I write in P6 (ever) so it is short nopaste.linux-dev.org/?869894 15:38
15:38 Celelibi joined
ZoffixW k, *now*, I'm happy. perl6advent.wordpress.com/ 15:39
And it only took an hour, yey CMSs :P 15:40
</sarcasm>
15:40 hobbs joined, pnu joined, solarbunny joined
PerlJam #perl6++ (for filling out the advent calendar) 15:41
ZoffixW \o/ 15:42
15:43 vytas joined
PerlJam ZoffixW++ for giving it an updated look :) 15:43
ZoffixW \o/
MadcapJake ZoffixW++ # love the hat! 15:44
The blog title is left aligned on the article page, looks a bit wonky 15:45
15:45 zengargoyle joined
ZoffixW Yeah. And there's no way to change that without paying that I see :/ 15:47
(or hosting WP yourself)
PerlJam &nbsp; x 10 ;-) 15:48
ZoffixW That's not mobile friendly :P
And that would mis-align it on the home page lol :)
15:48 ZoffixW left
flussence
.oO( new in Unicode 10: text alignment and formatting characters )
 15:49
RabidGravy lab, change the $v.WHAT there to $v.^name and all will be good 15:50
ilmari m: my $v = 42; say so $v ~~ any(Int, Str) 15:51
camelia rakudo-moar 8fb631: OUTPUT«True␤»
ilmari m: my $v = []; say so $v ~~ any(Int, Str)
camelia rakudo-moar 8fb631: OUTPUT«False␤»
ilmari m: my $v = []; say so $v.^name ~~ any(Int, Str)
camelia rakudo-moar 8fb631: OUTPUT«True␤»
ilmari the type name is a string 15:52
timotimo yeah
Zoffix: the new advent design looks good! 15:53
15:53 rindolf joined
dalek kudo/nom: 437ef18 | lizmat++ | t/04-nativecall/11-cpp.t:
Fix spelling, spotted by Rares Aioanei++

Fixes RT #126762		 15:54
15:55 xpen joined 15:56 lab is now known as nadim 15:57 telex left, ggoebel7 joined
nine Please revert my last commit to rakudo 15:57
it's wrong and breaks panda
nadim RabidGravy: any quick explanation why .WHAT failed? 15:58
15:58 telex joined
lizmat nine: will do 15:59
15:59 lucasb joined, kaare_ joined
dalek kudo/nom: 0a61b9a | lizmat++ | src/core/Process.pm:
Revert "Store CompUnit::Repositories in %*CUSTOM_LIB"

This reverts commit 08622892426fb4a758ee8eae2f13f3d2c43ba2dc.
As requested at irclog.perlgeek.de/perl6/2015-11-30#i_11628270		 16:00
RabidGravy m: say "-" ~ Array.WHAT
camelia rakudo-moar 8fb631: OUTPUT«Cannot look up attributes in a type object␤ in block <unit> at /tmp/EYUqT030Nv:1␤␤»
RabidGravy because that
16:01 xpen left
timotimo oh 16:01
lizmat RabidGravy: looks like an infix:<~> signature issue to me
RabidGravy yeah something like that
lucasb m: Array.Str 16:03
camelia rakudo-moar 8fb631: OUTPUT«Cannot look up attributes in a type object␤ in block <unit> at /tmp/lxekskIiJ3:1␤␤»
16:03 cognominal_ left
lizmat m: Array.Stringy # similar 16:03
camelia rakudo-moar 8fb631: OUTPUT«Cannot look up attributes in a type object␤ in block <unit> at /tmp/UTqPUgbhLx:1␤␤»
RabidGravy but "multi method Str(Array:U:) { self.^name }" kinda sucks 16:05
16:06 burnersk left
nine lizmat: thanks! 16:08
lizmat yw, nine++ :-)
dalek osystem: e7deabf | (Carsten Hartenfels)++ | META.list:
Add JavaScripdt::SpiderMonkey to ecosystem

See github.com/hartenfels/Javascript-SpiderMonkey		 16:09
osystem: 2674ab1 | RabidGravy++ | META.list:
Merge pull request #98 from hartenfels/master

Add JavaScripdt::SpiderMonkey to ecosystem
16:09 brrt joined 16:10 YP-QMUL-W joined 16:20 domidumont left, eiro left 16:21 lestrrat left 16:25 lestrrat joined 16:26 sjohnson` is now known as sjohnson
RabidGravy is docs.perl6.org built by a cron job? 16:29
it doesn't seem to have been built since Saturday
16:32 YP-QMUL-W left 16:33 YP-QMUL-W joined, n0xff joined, uruwi left
lucasb seems there was some error generating the Supplier pod 16:34
zengargoyle probably still parsing... :P
16:35 YP-QMUL-W left 16:37 YP-QMUL-W joined
lizmat FWIW, I'm working on the P6W, if anybody has something interesting to mention that you think I may miss, let me know! 16:38
[Coke] ah. jvm make test is failing due to RT#126765 16:39
nine: ^^
16:40 eiro joined, kjs_ left, burnersk joined 16:41 pmurias joined
hoelzro o/ #perl6 16:41
lizmat hoelzro o/ 16:42
hoelzro o/ lizmat
16:43 CIAvash joined
[Coke] we should consider not installing -gdb- and -valgrind- executables unless they're asked for, maybe. 16:43
lucasb the "=end" pod directive has changed recently? was it "=finish" before?
tadzik timotimo: probably :) 16:44
[Coke] rakudo-m test dying on t/04-nativecall/11-cpp.t 16:45
(on OS X)
hoelzro timotimo: OT question, but do you have your i3 config posted anywhere? I'm giving it a shot this week, and I remember oohing and aahing over your setup
16:47 n0xff left 16:48 n0xff joined
pmurias hoelzro: \o 16:49
dalek c: a60d7ec | RabidGravy++ | doc/Type/Supplier.pod:
Forget the pod on =end
 16:52
RabidGravy me-- # didn't test the file 16:53
hoelzro o/ pmurias
lizmat RabidGravy++ # compensation :-)
RabidGravy tony-o_, I've slapped you a PR on Event::Emitter cause the Supply changes done broke it 16:54
16:57 abraxxa left 16:59 balazs left
PerlJam blah 17:00
someone give me feedback on perl6advent.wordpress.com/?p=2867&...eview=true
I've lost what I wanted to say and I'm wondering if I should just scrap the thing.
stmuk lizmat: maybe mention the perl 6 talks at LPW? (which you probably will anyway)
lizmat stmuk: yes, I will :-) 17:01
lucs PerlJam: "Oops! That page can’t be found."
pmurias same a lucs here
s/a/as/
PerlJam I suppose you have to be logged in to perl6advent.wordpress.com to see it.
17:02 raiph joined
stmuk PerlJam: that looks good .. short and sweet .. maybe mention curli as well? 17:02
lucs Logged in?
lizmat PerlJam: looks fine as an introduction: maybe also highlight some of previous years advent posts that stood the test of time ? 17:03
lucs PerlJam: How does one log in?
PerlJam lucs: if you don't have an account I can send you an invite if you give me an email address
lucs Oh, okay, thanks.
PerlJam lucs: but you have to submit an article on Perl 6 sometime between now and the heat death of the universe ;) 17:04
lizmat stmuk: I will mention curli in the P6W
lucs PerlJam: Whoa! Slow down there!
lizmat PerlJam: especially mentioning perl6advent.wordpress.com/2014/12/...ong-right/ seems appropriate at this time :-) 17:06
arnsholt lizmat: It's from last week, technically, but could you include a mention of the IPerl6 kernel too? If someone gets hyped by the idea, there are some LHF tickets in the issues list that they can start with
lizmat arnsholt: you mean github.com/timo/iperl6kernel/ ? 17:07
arnsholt Yup!
timotimo hoelzro: i wouldn't really call it "setup", but here you go: github.com/timo/perl6-vortrag
lizmat arnsholt: I'll try to wriggle it in :-)
zengargoyle lucasb: «=finish» will be «=eof comment» | «=eof pod» | «=eof data» 17:08
«=end» is still closer for «=begin» AFAIK
RabidGravy yeah it was the "=begin pod" "=end" that shagged it up 17:09
stmuk was ist ein "krankem framework"? sick framework? :)
arnsholt lizmat: Cheers! =)
lizmat stmuk: I think "diseased framework" gives better imagery :-)
arnsholt I'm trying to adopt the jnthn approach of showing the way for something cool and waiting for someone to come along and pick it up ;) 17:10
moritz stmuk: yes, "krank" can mean "sick" but also "crazy"
zengargoyle but i doubt that bit of change has made it into code anywhere yet.
lizmat zengargoyle: did I miss the memo on the end of =finish ?
zengargoyle ah, forget vs forgot. :) 17:11
RabidGravy ==> Successfully installed panda 17:12
==> Please make sure that inst#/home/jonathan/.perl6/2015.11-301-g0a61b9a/bin is in your PATH
zengargoyle lizmat: maybe... there was a discussion with TT a few days back about some pod stuff.
RabidGravy yay! and whoah! 17:13
lizmat but no implementation for it yet, no ?
17:14 balazs joined
zengargoyle about what =finish means vs what people think it means etc. decided that =eof could be used to signify the ...goes to EOF without needing =end thing. 17:14
lizmat: yeah, pretty sure no actual implementation yet.
=eof would default to mean «=eof comment» sorta free-form ignore all this stuff 17:15
«=eof pod» would be like ... rest of this file is POD document (no code)
«=eof data» would be like __DATA__ 17:16
RabidGravy that /home/jonathan/.perl6/2015.11-301-g0a61b9a/bin path is it possible to obtain it somehow from the installed perl6?
[Coke] RabidGravy: that seems like an difficult path to require. ew.
17:16 n0tjack joined
[Coke] every time you install a new version, you need to update your path? 17:17
RabidGravy yeah which would suck somewhat
timotimo i put a symlink in ~/.perl6 called "current"
RabidGravy it could be mitigated if the installed perl could be made to 'fess it up
[Coke] what is repo/ ?
timotimo don't worry, nine said it'd only take a small patch to fix that stuff 17:18
17:20 hartenfels left, kjs_ joined
timotimo how do we reach takadonet? the Algorithm::Diff module has been b0rked at least since the GLR 17:20
17:21 cygx left, hartenfels joined
PerlJam needs to eat lunch 17:21
17:22 Oatmeal joined, Oatmeal left 17:23 raiph left, Oatmeal joined, Oatmeal left, zakharyas left, csd_ left
lizmat m: my \Δ = my $ = 1; say Δ++; say Δ++ # sigilless variables :-) 17:24
camelia rakudo-moar 0a61b9: OUTPUT«1␤2␤»
lucasb zengargoyle: hey, thanks for the note. (I was away)
17:25 vytas left 17:26 hankache joined
hankache hola 17:27
lucasb aloha
17:29 csd_ joined
lucs allo là 17:30
17:34 AndyDee joined, csd_ left
zengargoyle commutes & 17:34
autarch thanks to whoever posted my blog post to HN and made my server cry ;) 17:35
hankache allo oui? ;) 17:36
autarch++
flussence it looks like panda's bootstrap.pl works fine, but rebootstrap.pl doesn't. It gets even more confused if you run it after bootstrap.pl... 17:39
17:40 rurban left
RabidGravy has something changed in the way that Attribute traits are applied in the last couple of days? 17:45
17:46 n0tjack left
RabidGravy the "is marshalled-by" in JSON::Marshal has just stopped working 17:47
TimToady RabidGravy: probably
jnthn++ was working on making attributes work more like normal variables
RabidGravy Hmm 17:48
it looks like it just isn't seeing the trait anymore 17:49
dalek kudo/nom: 632f75b | usev6++ | src/core/Mu.pm:
Avoid NullPointerException on rakudo.jvm

  'say my @a' lead to NPE; timotimo++ for suggesting nqp::existskey		 17:51
kudo/nom: c0edfcb | peschwa++ | src/core/Mu.pm:
Merge pull request #598 from usev6/jvm_npe_gistseen

Avoid NullPointerException on rakudo.jvm
psch hm, wonder if i can bend interop back into shape *and* add enough new bits for an advent post... 17:52
would that already make a tradition..? :)
timotimo christmas: that magical time of the year where you can use JVM interop on rakudo
psch oh be fair 17:53
it worked until the GLR :P
timotimo you're right 17:54
and you get all my ++ es for working on it
psch++ # jvm interop maintaining
psch not much maintaining recently, with all the glr/jvm fallout :/
timotimo well, you're working on making the jvm not be exploded
that's worth a whole lot, too
psch and all those pitfalls that open themselves /o\ 17:55
timotimo ;(
psch like that we don't have a REPR.initialize on jvm, like at all
r: say nqp::create(Int) # which is why this
camelia rakudo-jvm e3c591: OUTPUT«===SORRY!===␤java.lang.NullPointerException␤»
..rakudo-moar 0a61b9: OUTPUT«0␤»
timotimo right
psch hm, maybe that's an ok topic 17:56
TimToady RabidGravy: probably 6a45feccb459d5643b4475f98efb59949155fd2b
psch although it's again lots of object.asm heh
TimToady our bot should look up things that are valid git ids 17:57
17:57 Skarsnik joined
hartenfels I was thinking on doing an advent post on something NativeCally, since I been working on gluing Mozilla's JavaScript to Perl 6. Would that be something fitting? 17:58
17:58 rurban joined
hartenfels Maybe try and inspire some more languages or libraries being hooked together. 17:58
JimmyZ nqp-j: my $a := nqp::list(); nqp::say($a); 17:59
camelia nqp-jvm: OUTPUT«#␤# There is insufficient memory for the Java Runtime Environment to continue.␤# pthread_getattr_np␤# An error report file with more information is saved as:␤# /tmp/jvm-1278/hs_error.log␤»
psch hartenfels: a walk-through-by-example of NativeCall bindings sounds like a neat topic to me
JimmyZ nqp-j: my $a := nqp::list(); nqp::say($a);
camelia nqp-jvm: OUTPUT«#␤# There is insufficient memory for the Java Runtime Environment to continue.␤# pthread_getattr_np␤# An error report file with more information is saved as:␤# /tmp/jvm-1284/hs_error.log␤»
JimmyZ nqp-j: my $a := nqp::list(0); nqp::say($a);
camelia nqp-jvm: OUTPUT«(signal ABRT)#␤# There is insufficient memory for the Java Runtime Environment to continue.␤# pthread_getattr_np␤# An error report file with more information is saved as:␤# /tmp/jvm-1290/hs_error.log␤»
psch hartenfels: although that does depend on the scope of the library, of course
JimmyZ: nqp-j always ABRTs
nqp-j: 1 18:00
camelia nqp-jvm: OUTPUT«#␤# There is insufficient memory for the Java Runtime Environment to continue.␤# pthread_getattr_np␤# An error report file with more information is saved as:␤# /tmp/jvm-1310/hs_error.log␤»
hartenfels psch: Well, it'd probably pick some things out that haven't been described in previous advent posts, like calling functions back and forth.
JimmyZ oh my ..
psch JimmyZ: it's only via camelia... :)
18:00 z8 joined
psch hartenfels: oh, yeah, i'd like callbacks and such, yeah 18:00
hartenfels: i have a thing or two i could need them for, haven't bothered to try and figure them out yet though 18:01
JimmyZ anyway, I think NPE should be fixed in nqp/jvm, not rakudo.
18:01 kensanata left
psch m: my %h; nqp::existskey(%h, 'a') 18:01
camelia rakudo-moar 0a61b9: OUTPUT«(signal SEGV)»
hartenfels psch: Guess I'ma make a pull request then.
psch JimmyZ: both backends do bad things in some cases, and it's generally agreed that nqp-level should check for a key before trying to access it
[Coke] JimmyZ: the ideal is that rakudo has no jvm/moar conditionals and that nqp hides all implementation details. 18:02
JimmyZ hmm, why moarvm doesn't do SEG without that patch..
psch m: my %h = a => 1; say nqp::existskey(%h, 'a') # something about initialize or so 18:03
camelia rakudo-moar 0a61b9: OUTPUT«1␤»
psch m: my %h = a => 1; say nqp::existskey(%h, 'b') # actual example :P
camelia rakudo-moar 0a61b9: OUTPUT«0␤»
psch JimmyZ: i agree that we shouldn't NPE on r-j at all, but this case specifically is close enough to "undefined behavior" territory for me to accept the patch instead of trying to fix it properly - for now 18:04
JimmyZ m: use nqp; my %h; nqp::existskey(%h, 'a') 18:05
camelia rakudo-moar 0a61b9: OUTPUT«(signal SEGV)»
18:05 AlexDaniel joined 18:06 z8 left
JimmyZ looks moarvm does something different about hash init 18:06
18:07 brrt left, znpy` left 18:10 hankache left 18:12 pyrimidi_ left 18:13 pyrimidine joined
[Coke] Work in progress on script to harvest everyone's names (which I had done manually for about 5 repos before I realized it would be so painful and that I missed some commits in those repos) - gist.github.com/coke/63c7bbb82f075c7bb1f5 18:15
Skarsnik hartenfels, I wanted to do that. My idea was to use a lib made for the article and cover some usual case 18:16
but feel free to do it x)
18:16 pdcawley joined, sQuEE` joined 18:17 Some-body_ joined
psch nine: r-j HEAD complains about repo/precomp/.lock not existing..? 18:17
18:17 DarthGandalf left, roguelazer left
hartenfels Skarsnik: Which library did you want to implement on it? 18:17
18:18 Some-body_ is now known as DarthGandalf, sQuEE left
[Coke] biggest thing you could do to help right now: add any git repos I've missed to the list at the top of gist.github.com/coke/63c7bbb82f075...storic-pl6 18:18
18:18 mspo left, espadrine left
JimmyZ Can I haz Chinese name ? :P 18:18
18:19 pdcawley_ left
[Coke] if it's in CREDITS, it'll happen automagically. 18:19
Skarsnik hartenfels, change this github.com/Toyunda/toyunda/blob/ma...ndaparse.c to make it a lib. it parse a weird subtitle format, so there is a 1-2 struct to bind and I wanted to add callback for the global container
nine psch: haven't had the time to look at jvm yet :/ 18:20
JimmyZ [Coke]: what about nqp-old
18:20 sQuEE` left
Skarsnik hartenfels, like in the end you will have to bind a ctruct, handle callback, a function call. can probably add some arrays for fun 18:20
[Coke] JimmyZ: url?
18:20 lichtkind joined
psch nine: no worries, i'm just wondering if the file not existing is a problem in lockfh itself or earlier 18:21
nine: as in, does r-m lockfh create a file, i suppose
18:21 sQuEE joined
hartenfels Skarsnik: Sounds like that'd be more fitting for an advent post, since you can put the entire library into it. 18:21
As opposed to bits and pieces of one.
JimmyZ [Coke]: github.com/perl6/nqp-rx and github.com/perl6/std ? 18:22
18:22 skarn left, Hotkeys left
JimmyZ [Coke]: what about github.com/perl6/doc ? 18:22
dalek href="https://perl6.org:">perl6.org: c17ce8a | (Zoffix Znet)++ | source/index.html:
Add a link to view "more" recent blog posts
 18:23
18:23 shmibs joined
Skarsnik hartenfels, that why I said, that was my idea for an advent post about NC ^^ 18:23
[Coke] whoops, I thought I had doc in there.
hartenfels Skarsknik: Lemme unpull my request then :P 18:24
JimmyZ and github.com/perl6/perl6.org ? # not sure about this one though :P
18:24 diakopter left
[Coke] JimmyZ: yikes, glad I asked. :) 18:25
Skarsnik hartenfels, no, do it. I probably don't have time to write something serious in time x)
dalek : 53ba209 | (Carsten Hartenfels)++ | misc/perl6advent-2015/schedule:
Add NativeCall advent calender entry

Some NativeCall bits that haven't been treated in previous advent posts, such as callbacks, Bufs and maybe LibraryMake if Panda unbreaks in time.
: f958c0e | RabidGravy++ | misc/perl6advent-2015/schedule:
Merge pull request #12 from hartenfels/master

Add NativeCall advent calender entry
timotimo Zoffix: is it also easy to subscribe to the RSS feed from the front page of perl6.org?
Skarsnik hartenfels, it was nust if you wanted an angle for the article
18:25 diakopter joined
hartenfels Skarsnik: Well, I have my JavaScript binding that uses them :P 18:26
And I guess too late to unpull it now.
JimmyZ [Coke]: Does it 6.c contain panda and modules?
[Coke] JimmyZ++ that's about 40 new names. 18:27
JimmyZ: this list will probably not include the ecosystem.
Skarsnik someone should write a module to compile C code x)
[Coke] but now that I have a script, it's doable.
JimmyZ ok, then no panda or so yet ;)
nine psch: no, thebfile is opened with :create, then locked
timotimo Skarsnik: we do have that already
Skarsnik: LibraryMake would do that for you 18:28
Skarsnik: also, Inline::C does that, too
18:28 Hotkeys joined
[Coke] JimmyZ: added all those repos, re-ran, re-pushed. 18:28
hartenfels LibraryMake is failing to install for me right now, with all the Panda craziness going on.
timotimo :(
jdv79 who let the crazy panda out?! 18:29
Skarsnik I don't need a makefile for generated C file x)
18:30 sQuEE left
jdv79 is it panda, rakudo, or both that are crazy? 18:30
hartenfels Probably both.
18:31 n0tjack joined, sQuEE joined, profan joined
psch nine: right... i'll have a look 18:31
18:32 mspo joined
Skarsnik Should the VM.config be normalised or something? 18:33
18:33 diakopter left
Skarsnik when patching NC to add api version, there is like 5 ways to get the lib extension 18:33
18:34 xfix joined
JimmyZ oh, if it is about rakudo star 6.c , then github.com/rakudo/star too :P 18:35
[Coke]: I found new face in this repo :P 18:36
psch nine: apparently the problem is simpler - the path itself doesn't exist at all, i.e. repo in the build directory isn't there
psch takes another look at the xmas tickets instead vOv
[Coke] new face? 18:37
JimmyZ yeah,, in the star repo
18:37 kivutar_ left
[Coke] one new person there, plus a few aliases, good catch. :) 18:38
JimmyZ++
18:40 pmurias left
jdv79 [Coke]: why not just put a call out to contact you in the next p6weekly if ones name is missing from your list. 18:40
18:40 ZoffixW joined, dha joined
ZoffixW timotimo, I don't use RSS, so I've no idea how easy it is to "subscribe" to anything :P 18:40
18:40 emdashcomma left
[Coke] Because this script isn't going to be done by then, probably. 18:40
18:40 Exodist left
hahainternet rss is basically great 18:41
i mean let's face it, it's the best of a bunch of really dodgy options
dha If you use RSS, it's pretty easy to subscribe to things. If not... not. :-)
yoleaux 28 Nov 2015 17:05Z <jdv79> dha: maybe you could talk to mr. foy about what timo is talkin bout
hahainternet there's probably a decent gap in the market
for something that combines the facilities of google reader
google news
[Coke] I mean, we have David H. Adler on the list, we're good. :)
hahainternet and twitter
together with RSS
ie 'news but without the social waffle'
jdv79 put a fork in it 18:42
ZoffixW likes their news with the waffle
hahainternet also hi all, good evening, another HN post about p6 :D
18:42 Exodist joined
hahainternet only a smattering of pointless negativity too 18:42
and i'm downvoting what i can see
18:42 emdashcomma joined
dha jdv79 - Sadly, I don't know what timo was talkin bout on that day. :-) 18:42
Hotkeys Morning everyone
Er
Afternoon I guess
timotimo Happy UGT, Hotkeys
dha [Coke] - What list am I on now?
jdv79 stale content on p6 on a site of his i think
18:43 roguelazer joined
dalek href="https://modules.perl6.org:">modules.perl6.org: ec0342f | (Zoffix Znet)++ | t (3 files):
Fix stargazers and Issues links

Closes github.com/perl6/ecosystem/issues/99		 18:43
dha Oh, ok. You could probably talk to him yourself, but I can certainly do that if you like. I'll need some details, though.
ZoffixW dha, this website has extremely outdated info that shows up at the top on Google for certain searches. We wish for that info to be removed or at least slapped with a giant warning that it is outdated: www.learningperl6.com/Chapters/ 18:44
JimmyZ [Coke]: do we got all names from RFC too ?
*get
[Coke] Yes, RFC is on my list of things to process... but that's not a git repo, so more manual coding there. 18:45
JimmyZ I think that is all then :P
[Coke] (I mean, it might be a repo, but I doubt that each RFC is committed by the person who worked on it. will have to process author info there.)
jdv79 dha: thanks
dha It does say, right at the top, "These chapters are a work-in-progress. They are incomplete, out-of-order, and very far from their final form." 18:46
Hotkeys timotimo: ugt?
dha You want to add "and possibly outdated" to that?
timotimo "universal greeting time"
[Coke] JimmyZ: added to the readme. thank you so much for your feedback!
Hotkeys Ah
Nice
Is that a thing?
Or did you make it up just for me
timotimo dha: what was i talking about when?
JimmyZ [Coke]: iirc, there may be some names in specs authors but not in git
timotimo oh 18:47
[Coke] JimmyZ: I think the author names mostly got removed
dha timotimo - on Nov 28, apparently.
JimmyZ oh, then we miss it
[Coke] If you can find one in a doc at HEAD, I can parse it out.
timotimo yes, i see now
dha: the problem is not Chapters/ itself
18:47 domidumont joined
timotimo because i don't think anything links to that 18:47
[Coke] Or maybe with a git grep over the history.
JimmyZ [Coke]: maybe parse a old version specs?
timotimo the individual html files you can get from there don't say "DO NOT READ THIS PLEASE" at the top
ZoffixW dha, yeah, but that notice is not very obvious nor is it present here: www.learningperl6.com/Chapters/11.f...ndles.html 18:48
timotimo all they say is in small text "last modified ... 2007" at the far bottom
18:48 diakopter joined
ZoffixW dha, and if I just googled for something I'll go skimming straight for the code examples, not try to eye some tiny "this maybe outdated" text at the top :P 18:48
JimmyZ [Coke]: before the removal time :P
dha ZoffixW - I'm not sure I see a solution for that particular problem. Putting "don't do that" on every paragraph doesn't seem like a good solution, for instance. 18:49
timotimo easy: the text gets put into a div that's positioned relative to the screen 18:50
ZoffixW meh. No need
timotimo other way: delete Chapters/ completely
ZoffixW CSS. I have the code but my browser is too freakin' slow -_- 1 sec 18:51
JimmyZ anyway, 02:51 am here ,good night
dha In any case, I imagine I'll talk to brian soonish. I'll mention it.
timotimo thank you!
the reason why i complain about this was someone on twitter complained about outdated documentation
awwaiid make it a BIG "don't read this" instead of a small one
[Coke] can you add elements to a set at runtime?
psch [Coke]: probably the same way as adding characters to a string at runtime 18:52
timotimo twitter.com/je_turcotte/status/670...0008237057
psch where "the same way" means "with same semantics", not "same operators" :)
timotimo [Coke]: sets are immutable, but you can use SetHash if you want a mutable Set 18:53
ZoffixW dha, html:before{background:#f00;color:#fff;font-size:14pt;position:fixed;padding:20px;z-index:9999;content:"WARNING: content of this page is outdated";}
timotimo m: SetHash.new();
camelia ( no output )
timotimo m: my $mutsh SetHash.new(); $mutsh<foo> = 1; say $mutsh
camelia rakudo-moar c0edfc: OUTPUT«5===SORRY!5=== Error while compiling /tmp/8ONd_xDbnR␤Missing infix inside S␤at /tmp/8ONd_xDbnR:1␤------> 3my $mutsh S7⏏5etHash.new(); $mutsh<foo> = 1; say $muts␤ expecting any of:␤ infix␤ infix stopper␤»
timotimo m: my $mutsh = SetHash.new(); $mutsh<foo> = 1; say $mutsh
camelia rakudo-moar c0edfc: OUTPUT«SetHash.new(foo)␤»
timotimo m: my $mutsh = SetHash.new(); $mutsh<foo> = 1; $mutsh<bar>++; say $mutsh
camelia rakudo-moar c0edfc: OUTPUT«SetHash.new(foo, bar)␤»
[Coke] timotimo: danke. 18:54
timotimo bitteschön :)
18:54 roguelazer left
ZoffixW dha, or better: html:before{background:#f00;color:#fff;font-size:14pt;position:fixed;padding:20px;z-index:9999;content:"WARNING: content of this page is outdated";} 18:55
timotimo what if there's something with a z-index of 99999?
18:55 diakopter left
timotimo i think the industry standard is "five nines" 18:55
dha Well, I'll leave the implementation to brian, but I'll outline the problem. 18:56
ZoffixW This is what it looks like: i.imgur.com/CNsFePx.png
k
18:56 rvchangue left
hahainternet timotimo: someone should have played you a rimshot for that 18:56
timotimo thank you :)
dha ZoffixW - But that looks workable. I'll probably toss it in.
18:57 espadrine joined
ZoffixW dha, add "text-align: center; width: 100%" at the end tho :) I copy-paste failed that second snippet :P 18:57
dha So, <ZoffixW> dha, or better: html:before{background:#f00;color:#fff;font-size:14pt;position:fixed;padding:20px;z-index:9999;content:"WARNING: content of this page is outdated";}text-align:center; width:100% 18:58
dha isn't good at this stuff.
18:59 skarn joined 19:00 roguelazer joined
ZoffixW dha, html:before{background:#f00;color:#fff;font-size:14pt;position:fixed;padding:20px;z-index:9999;content:"WARNING: content of this page is outdated";text-align:center; width:100%;} 19:01
19:01 espadrine left
dha Ta. 19:01
timotimo can i have someone with C pointer arithmetic chops help me a little bit?
ZoffixW dha, it'd can be placed in this file and that should be enough from what I see: www.learningperl6.com/Chapters/chapters.css
hahainternet timotimo: don't ask to ask :p 19:02
19:02 rvchangue joined
timotimo fair enough :) 19:02
RabidGravy okay, what in the last few days what make it need github.com/jonathanstowe/JSON-Mars...8e4edc60ae to see the traits defined in the module?
totally confused by that 19:03
19:03 diakopter joined, ZoffixW left, hankache joined
timotimo gist.github.com/timo/dab3afea0b5362d5e6fa - i added the four ops with "_deref_" in their names. i need them to do basically the same thing as the ones without _deref_ in the name, but the offset (what's passed in GET_UI16(...)) points at a pointer, not at the value i want to get/set 19:04
i'm not quite sure how to tease the right behaviour out of C 19:05
nine retupmoca: you have a pull request fixing installation of LibraryMake
19:08 water_melon joined
timotimo i think i have to have ((MVMint64 **)((char *)o ... 19:08
hahainternet timotimo: i'm going to ask a friend to take a look
retupmoca nine++ # PR is merged 19:09
timotimo hahainternet: it seems liek i might have solved it
nope, segfault :D 19:10
19:10 oka_ is now known as oka^, zengargoylew joined
timotimo tadzik: please don't include the latest JSON::Fast yet. TimToady is going to break it :D 19:13
grondilu timotimo: shouldn't you cast the output of GET_UI16 as well before adding it? 19:14
timotimo grondilu: hm, why?
i can't imagine what i'd cast it to that would be more correct
grondilu I don't know, does it return char *
?
Skarsnik Not sure you need the * in GET_REG(cur_op, 0).i64 = *target 19:15
timotimo no, it's not supposed to be char *
dalek kudo/nom: ce69f30 | TimToady++ | / (6 files):
rename SEQ() to STATEMENT_LIST()

The term "sequence" is already badly overloaded without SEQ().
grondilu then I'm surprised you can add it to a char*
timotimo i'll be AFK for a tiny bit
grondilu: that's just regular pointer arithmetic
like a[5]
grondilu yeah, my bad 19:16
[Coke] ok, collapsed a lot of duplicate committer names on that list. 19:18
RabidGravy tadzik, JSON::Unmarshal borked in today's rakudo github.com/tadzik/JSON-Unmarshal/pull/14 fixes 19:19
hoelzro new docker image for Rakudo Star 2015.11 is up: hub.docker.com/_/rakudo-star/ 19:21
hahainternet timotimo: my friend says not enough context for him to help sorry 19:22
timotimo: shout me when you're back and lemme see if we can context it up 19:23
19:25 lestrrat left
timotimo hahainternet: i'm back :) 19:26
19:26 qcoh_ joined
timotimo i'll paste some more crap 19:26
hahainternet timotimo: i've asked my friend to drop in if he has time, if not i'll paste more to him
he's much MUCH more of a C/C++/low level expert than I am
grondilu timotimo: didn't you just add intermediate variables? 19:27
(o and target, that is)
water_melon Hello all, I see on doc.perl6.org/language/nativecall under Typed Pointers it says TBD. Does that mean it's not possible to use IntPtr type in NativeCall? I'm having trouble using SystemParametersInfo from user32.dll. Gives error "Cannot locate symbol" 19:30
19:30 lestrrat joined
Skarsnik it's mostly not implemented? (It probably exist) 19:30
grondilu m: use NativeCall; my IntPtr $p;
Skarsnik but you can work around with nativecast 19:31
camelia rakudo-moar c0edfc: OUTPUT«5===SORRY!5===␤Type 'IntPtr' is not declared. Did you mean 'IntStr'?␤at /tmp/UHtZl0Vry2:1␤------> 3use NativeCall; my IntPtr7⏏5 $p;␤Malformed my␤at /tmp/UHtZl0Vry2:1␤------> 3use NativeCall; my7⏏5 IntPtr $p;␤␤»
Skarsnik m: use NativeCall; my Pointer[int32] $p;
camelia ( no output )
TimToady m: SEQ(123; 456)
camelia rakudo-moar ce69f3: OUTPUT«5===SORRY!5=== Error while compiling /tmp/yKFK6p6vgH␤Use of unsupported SEQ macro; please change to STATEMENT_LIST␤at /tmp/yKFK6p6vgH:1␤------> 3SEQ(7⏏05123; 456)␤»
grondilu (if anything the name should be more intPtr than IntPtr, btw) 19:32
TimToady wonders how much other undocumented internal stuff has been used in the ecosystem... 19:33
Skarsnik water_melon, what you can do, work with OpaquePointer and and when you need the cstruct behind, do this my c_struct_type_foo $p = nativecast(c_struct_type_foo, $pointer);
timotimo just when i wrote that, the internet connection died
grondilu: in the latest local version i also put ** instead of *
Skarsnik Did you try removing the * when you affect to GET_ ? 19:34
TimToady of course, now STATEMENT_LIST in undocumented internal stuff, but I can live with that :)
timotimo hahainternet: i added a second file that has an example invocation of one of the ops in question in pseudocode. hopefully that makes things a little clearer? 19:36
TimToady *is
hahainternet timotimo: will pass it across
timotimo thank you kindly 19:37
at some point i really have to learn this by heart
flussence does anyone have a pre-written bashrc incantation for putting the post-curli perl6 bin path in $PATH? I really don't like the idea of having to parse `perl6 --version` output in there. 19:38
timotimo grondilu: if you f5 that gist, you'll see the version with ** 19:39
FROGGS water_melon: TBD just means "To Be Documented" there
timotimo grondilu: but it doesn't compile because i probably need to add ** on the LHS of the last assignment in bind 19:40
FROGGS water_melon: a typed pointer is something like Pointer[int32], as shown
m: use NativeCall; constant IntPtr = Pointer[int32]; my IntPtr $a = 42; # water_melon 19:41
camelia rakudo-moar ce69f3: OUTPUT«Type check failed in assignment to $a; expected Pointer but got Int␤ in block <unit> at /tmp/SHzTTch14_:1␤␤»
FROGGS ups
okay, that assignment was bogus :o)
m: use NativeCall; constant IntPtr = Pointer[int32]; my IntPtr $a .= new; 19:42
camelia ( no output )
timotimo oh damn it. did i put the wrong operator order in there?
grondilu if what you want is returning the pointer, what about just GET_REG(cur_op, 0).i64 = ((MVMint64 *)((char *)GET_REG(cur_op, 2).o + GET_UI16(cur_op, 4)));
timotimo hahainternet: problem solved! 19:43
hahainternet timotimo: oh damn he just figured it out i think
timotimo hahainternet: the version i updated the gist with last is the one that works
all i needed was ** everywhere
the reason i got the segfault was because i confused the second and third operator in code-gen
but the journey isn't over yet! 19:44
i still have to implement the same thing in the JIT :)
and i suck at ASM :D
water_melon Thanks for the help so far. SystemParametersInfo takes (int32, int32, Pointer[int32], int32) and I was thinking that the "Cannot locate symbol" error was because of wrong argument types, but it still cannot be located. I'm porting some code from perl 5 and it works fine there with Win32::API
Skarsnik FROGGS, that really work? how you use it to get the struct behind? like my int32 $foo = $ 19:45
put OpaquePointer?
hahainternet timotimo: i suck at C, so this is already way over my head ;)
water_melon lol I learned perl because I didn't want to learn C 19:46
dha I try to do all my programming in hq9+ :-)
hahainternet i know a little C but trying to deal with macros and complex pointer ops, no thanks
19:47 vincent__ joined
timotimo :D 19:47
19:47 vincent__ is now known as Guest1854
Skarsnik m: use NativeCall; my Pointer[int32] $p; my int32 $b = $p; 19:47
camelia rakudo-moar ce69f3: OUTPUT«Cannot unbox a type object␤ in block <unit> at /tmp/SOyN92QaXr:1␤␤»
timotimo hahainternet: i still appreciate you connected me to your friend; tell him i'm sorry to cause him work and then invalidate it right at the end :( 19:48
Skarsnik Not even sure how to put a value in this lol
RabidGravy water_melon, you did do " SystemParametersInfo( ..... ) is native('user32')" ?
timotimo m: use NativeCaall; my Pointer[int32] $p; say $p.^methods; 19:49
camelia rakudo-moar ce69f3: OUTPUT«===SORRY!===␤Could not find NativeCaall:ver<True>:auth<True>:api<True> in:␤ /home/camelia/.perl6/2015.11-304-gce69f30␤ /home/camelia/rakudo-m-inst-2/share/perl6/site␤ /home/camelia/rakudo-m-inst-2/share/perl6/vendor␤ /home/camelia/rak…»
[Coke] m: say "a" ~~ /a/;
camelia rakudo-moar ce69f3: OUTPUT«｢a｣␤»
RabidGravy specifying the library is important
timotimo m: use NativeCall; my Pointer[int32] $p; say $p.^methods;
camelia rakudo-moar ce69f3: OUTPUT«(of deref perl of gist Int Numeric new deref)␤»
water_melon RabidGravy: my code is "sub SystemParametersInfo(int32, int32, Pointer[int32], int32) 19:50
returns int32
is native("user32")
{ * }"
FROGGS m: use NativeCall; constant IntPtr = Pointer[int32]; my IntPtr $a .= new; # if the pointer got malloced, you could do: say $a.deref
19:50 kyclark joined
camelia ( no output ) 19:50
FROGGS Skarsnik: ^^
flussence
.oO( shouldn't those error messages be printing “:ver<*>:auth<*>:api<*>”? )
FROGGS though, there is no way yet to nicely malloc/free or to set the value a pointer is pointing to
water_melon crap i didn't know my irc client would post each line as its own message
RabidGravy weird
kyclark lpaste.net/146228 19:51
Cannot find method 'CALL-ME'
Skarsnik m: use NativeCall; my Pointer[int32] $p .= new; $p.deref = 42;my int32 $b = $p; say $b;
camelia rakudo-moar ce69f3: OUTPUT«(signal SEGV)»
timotimo kyclark: you want to use $john."$att"()
kyclark I'm just naively trying some Perl6 object stuff, I was hoping to called "$obj.$method"
FROGGS water_melon: "Cannot locate symbol" means that "SystemParametersInfo" is not found
RabidGravy might need to try "is native('User32.dll')"
timotimo $obj.$method is when you have an actual method or routine object in $method
kyclark Aha! Thanks.
FROGGS water_melon: or are you trying to bind a C++ lib? 19:52
n0tjack going through the Rakudo source is a good way to learn more obscure-ish p6 features
Skarsnik RabidGravy, nc put the .dll
timotimo m: sub do_it($foo) { say "doing it to $foo!" }; my $thingie = &do_it; my $subject = 1000; $subject.$thingie
camelia rakudo-moar ce69f3: OUTPUT«doing it to 1000!␤»
timotimo kyclark: ^ little example
FROGGS kyclark: needs to be $john."$att"() in your case 19:53
Skarsnik FROGGS, commit my apiversion patch xD
FROGGS Skarsnik: what patch?
timotimo FROGGS: already helped 'em :) 19:54
kyclark timotimo, what would it look like to have a method? my $method = &$john.age; ??
water_melon RabidGravy: I'm porting the following line from perl5 to perl6 - "my $syspinf = Win32::API->new('user32','SystemParametersInfo', [I,I,P,I], I)" NativeCall replaces Win32::API for uses like these, correct?
Skarsnik FROGGS, rt 126645, the last comment x)
water_melon, what is the C code 19:55
don't look at perl5
timotimo kyclark: it's a bit more involved. here's an example:
[Coke] current listing of names has 699 people.
dalek : ec9866f | TimToady++ | misc/perl6advent-2015/schedule:
claim Christmas Eve
 19:56
timotimo m: my $meth = Str.^find_method("uc"); say $meth.perl; say "hello".$meth
camelia rakudo-moar ce69f3: OUTPUT«method uc (Str $: | is raw) { #`(Method|46153792) ... }␤HELLO␤»
timotimo kyclark: the ".^" means "call a method on the meta-object", which is what allows us to ask the Str class for what methods it has
kyclark Huh, it works. my $method = Human.^find_method("name"); 19:57
TimToady perhaps the 25th should just be the release notes :)
kyclark Thanks.
water_melon Skarsnik, SystemParametersInfo is documented at msdn.microsoft.com/en-us/library/aa932539.aspx 19:58
timotimo kyclark: you're quite welcome :) 19:59
[Coke] TimToady: or a "ha! made you look, just kidding."
kyclark timotimo, can I use a similar method to ask a class/object all of its methods?
FROGGS Skarsnik: ahh, that api version 20:00
[Coke] (just kidding!)
kyclark And where are docs on this so I can haz the readz?
TimToady m: say DateTime.^methods
camelia rakudo-moar ce69f3: OUTPUT«(check-time now clone clone-without-validating Instant posix offset offset-in-minutes offset-in-hours later earlier truncated-to whole-second in-timezone utc local Date Str day-of-month check-value IO daycount-from-ymd get-daycount day-of-week check-date d…»
water_melon Skarsnik, msdn.microsoft.com/en-us/library/w...85%29.aspx is more helpful actually
kyclark Whoa, nice.
TimToady m: say DateTime.^methods[1]() 20:01
camelia rakudo-moar ce69f3: OUTPUT«Too few positionals passed; expected 1 argument but got 0␤ in block <unit> at /tmp/fwpEOaWLj_:1␤␤»
[Coke] any thoughts on trying to get Šuppa to sort near Sośnierz instead of dead last?
TimToady m: say DateTime.^methods[1](DateTime)
camelia rakudo-moar ce69f3: OUTPUT«2015-11-30T21:01:24+01:00␤»
kyclark What do I get back from .^methods? It will stringify, but I can't do say "all methods = ", join ', ', Human.^methods; 20:02
TimToady [Coke]: international collation is a world of hurt :)
Skarsnik water_melon, I think your issue is it does not find the symbol at all. does the dll is in standard path?
psch m: Any.^methods[0].WHAT.say 20:03
camelia rakudo-moar ce69f3: OUTPUT«(Method+{<anon|74369136>})␤»
RabidGravy kyclark, a list of Method objects
TimToady but as a first approximation, sort on NFD?
psch m: my Str $all-methods = map *.Str, Any.^methods; say $all-methods
camelia rakudo-moar ce69f3: OUTPUT«Type check failed in assignment to $all-methods; expected Str but got Seq␤ in block <unit> at /tmp/myASEHqFBz:1␤␤»
psch m: my Str $all-methods = join ', ', map *.Str, Any.^methods; say $all-methods
camelia rakudo-moar ce69f3: OUTPUT«Method object coerced to string (please use .gist or .perl to do that) in whatevercode at /tmp/Nn2L4YJSEx:1␤Method object coerced to string (please use .gist or .perl to do that) in whatevercode at /tmp/Nn2L4YJSEx:1␤Method object coerced to string …»
psch oh, right, that warning 20:04
masak TimToady: happily, "statement list" is a concept used both by STD.pm6 and 007, with pretty much that meaning
psch m: my Str $all-methods = join ', ', map *.perl, Any.^methods; say $all-methods
camelia rakudo-moar ce69f3: OUTPUT«Method+{<anon|74369136>}.new, Method+{<anon|74369136>}.new, Method+{<anon|74369136>}.new, method cache ($: *%_) { #`(Method|71151664) ... }, Method+{<anon|74369136>}.new, Method+{<anon|74369136>}.new, Method+{<anon|74369136>}.new, Method+{<anon|74369136>}.…»
[Coke] TimToady: at this point, I just have these 4 gist.github.com/coke/63c7bbb82f075...s-out-L696 which I can manually shuffle if needed. If anyone wants to try to sort them better, please feel free to try. :)
psch m: my Str $all-methods = join ', ', map *.name, Any.^methods; say $all-methods # probably usefuller
camelia rakudo-moar ce69f3: OUTPUT«EXISTS-KEY, DELETE-KEY, DELETE-POS, cache, list, flat, eager, List, Slip, Array, hash, Hash, elems, end, keys, kv, values, pairs, antipairs, invert, pick, roll, iterator, classify, categorize, rotor, reverse, combinations, permutations, join, nodemap, duck…»
psch kyclark: ^^^ i suppose that's kind of what you were looking for?
lucasb isn't it better to sort by the first name, not the last name? 20:05
water_melon Skarsnik, the error when the library can't be found is "Cannot locate native 'library foo.dll': error 0x7e, while the error I'm getting is "Cannot locate symbol 'SystemParametersInfo' in native library user32.dll" So it does find user32.dll just fine
lucasb those emails addresses will be removed, won't they? 20:06
kyclark psch, yeah, I guess so. I need to figure out how to play with these Method objects. Can the perl6 REPL source a file?
[Coke] lucasb: yes, although they're all public already.
dalek : 4f10588 | (Matthew Wilson)++ | misc/perl6advent-2015/schedule:
claim a slot
[Coke] it's just going to be the name that appears outside of the ｢｣ quotes.
kyclark And is there builtin help in the REPL?
20:07 spider-mario joined
flussence m: put sprintf(‘PATH="%s:%s:$PATH’, $*VM.config<bindir>, ???); # golfed my bashrc question. What do I replace «???» with that'll let me run panda again? :) 20:07
camelia rakudo-moar ce69f3: OUTPUT«Stub code executed in block <unit> at /tmp/qld0a9qmru:1␤PATH="/home/camelia/rakudo-m-inst-1/bin:0:$PATH␤»
psch kyclark: the REPL can EVALFILE (iirc), but the only help you get is introspection (and minimal tab-complete if you have Linenoise installed)
m: say so &EVALFILE 20:08
camelia rakudo-moar ce69f3: OUTPUT«True␤»
20:09 |Sno| joined
Skarsnik water_melon, hm, the signature should look like foo(uint32, uint32, OpaquePointer, uint32) 20:09
so it should work
kyclark psch, the 2nd one was right one. I'd tried a "map" but was using $_ instead of *.
FROGGS Skarsnik: forget about OpaquePointer, use Pointer instead
Skarsnik it's not the same? x) 20:10
psch kyclark: you can use $_ in map, but only in an explicit block. lots of the opaque magic is gone
[Coke] lucasb: I'm just including them for diagnostics.
psch m: map { $_.say }, (1,2,3,4) # like here
camelia rakudo-moar ce69f3: OUTPUT«1␤2␤3␤4␤»
psch kyclark: as for help, there's doc.perl6.org which also has a CLI tool afair, at github.com/perl6/p6doc 20:11
lucasb [Coke]: ok, thanks [Coke]++
what about sorting by the first name, not the last?
psch oh, wrong :l
RabidGravy I don't suppose anyone's written a thing to parse a crontab 20:12
20:12 [Sno] left
psch kyclark: it's just github.com/perl6/doc apparently... :) 20:12
[Coke] lucasb: that's easy to do, but IME, it's typically (in US terminology) last name, first name, middle name. 20:13
lucasb hm, ok
20:14 darutoko left
water_melon Skarsnik, changing my int32s to uint32s and changing Pointer[int32] to OpaquePointer or Pointer do not change the error. Is it just a guessing game at this point as to what types to provide the function or is something actually wrong with my code? Am I using NativeCall correctly? My full code is at github.com/Djent-/Windows-Desktop/...Desktop.pm 20:14
[Coke] .u VINCENT 20:15
yoleaux No characters found
20:16 CIAvash left
RabidGravy water_melon, if it can't find the symbol the changing the signature won't help 20:18
n0tjack has anyone come up with a Grammar to match only valid Roman Numerals?
n0tjack has no idea why his fingers decided to capitalize roman numerals
retupmoca water_melon: "is nativeconv('stdcall')" maybe? I'm not sure if that changes the symbol lookup 20:19
20:19 balazs left
retupmoca (or however nativecall spells WINAPI / __stdcall calling convention) 20:19
20:20 geraud joined
[Coke] n0tjack: see p5's Regexp::Common::Number /roman 20:20
dalek : dc07a2a | (Zoffix Znet)++ | misc/perl6advent-2015/schedule:
Include info on topics covered, for other authors to know if needed (bit.ly/1NEaeSn)
 20:21
water_melon retupmoca: A new error at least, "Type check failed in binding (anon): expected 'Pointer" but got 'Str" 20:22
n0tjack [Coke]: thanks
20:22 cognominal_ joined
[Coke] it was surprisingly small. 20:23
20:25 TEttinger joined 20:26 domidumont left
masak wow, schedule is filling up nicely 20:27
I like this community <3
n0tjack doing little katas in perl6 is a lot of fun 20:28
it really is like the old perl, it just gets out of my way
though the fact that it has a typing system seriously tests my self-restraint on over-engineering before I know anything.... 20:29
masak n0tjack: I think that last bit is a matter of habit.
n0tjack: I generally start out largely untyped these days, and then add in types as the thing grows 20:30
FROGGS nine: there?
lizmat hmmm... looks like the perl 6 advent calendar site is already showing 2 new advents posts? one on 30 nov, and one on 2 dec ?
masak to the point of upgrading from untyped hashes to typed objects, sometimes
nine FROGGS: yep
n0tjack masak: yeah, I think falls under "discipline". The language isn't making me do these things, but it has enough treats to tempt me into doing unhealthy stuff. 20:31
masak lizmat: only if you happen to be logged in
FROGGS nine: do you have an idea how we want to make the alternative bin install paths work?
flussence m: put sprintf(‘PATH="%s:%s:$PATH"’, $*VM.config<bindir>, ‘$HOME/.perl6/’ ~ comb(/‘rakudo version ’<(\S+/, run($*EXECUTABLE, <--version>, :out).out.get).Str ~ ‘/bin’); # answering my own question... I guess
camelia rakudo-moar ce69f3: OUTPUT«run is disallowed in restricted setting␤ in sub restricted at src/RESTRICTED.setting:1␤ in sub run at src/RESTRICTED.setting:14␤ in block <unit> at /tmp/Z9OujjwcdE:1␤␤»
masak lizmat: I was surprised too
lizmat aahhhh... ok :-)
masak n0tjack: you have to train yourself to be more eager to see the next test pass, so that you don't spend time on typing unnecessary tokens :P 20:32
nine FROGGS: I'm currently working on making panda install into %CUSTOM_LIB<site> again if possible. That will at least give us highly predictable bin/ paths again.
20:32 balazs joined
n0tjack masak: hmm, I should probably get into the habit of writing tests then ;) 20:32
FROGGS nine: to it feels like I want to tell the repo to have one path for packages, and one for scripts
[Coke] m: my @a = ('masak', 'Carl Masak', 'Carl Mäsak'); @a.sort({-$_.chars * 100_000_000 - ([+] $_.ords)}).perl.say;
camelia rakudo-moar ce69f3: OUTPUT«("Carl Mäsak", "Carl Masak", "masak")␤»
masak n0tjack: well, duh ;) 20:33
FROGGS to me*
dha Oh good. I'm actually getting my travel plans together for LPW.
20:33 lucasb left
n0tjack I'm using to REPLing stuff until my confidence masks my ignorance and I persist whatever I wrote in a script. IOW, I'm using to doing testing, but I don't keep the tests. 20:34
masak the new perl6advent layout is very nice, but I wonder if the headings for the posts could be made less light gray. it looks like the posts are switched off or somethign when they're that light gray.
something*
TimToady +1
mst somebody Ctrl-S'ed the blog 20:35
masak if I want bad contrast, I can arrange that locally, on my reading device
FROGGS hehe 20:36
[Coke] wow, we have 2 ilmaris that have contributed. 20:37
dha They're multiplying?
20:38 Timbus left 20:39 Timbus joined
RabidGravy so it's PATH=`perl6 -e 'say $*REPO.Str'`/bin:$PATH for the time-being 20:42
20:43 water_melon left 20:45 dbohdan_ is now known as dbohdan 20:46 kaare_ left 20:50 bzipitidoo joined
TimToady hmm, doc.perl6.org/language/faq#What_is_...gnature%3F is kind of incorrect 20:51
lunch & 20:52
20:54 water_melon joined
bzipitidoo I'm back, with more questions. How does ~ work in regexs? Seems to be an operator with 3 operands. Far as I can tell, the 1st operand is a set of characters to use for an opening bracket, the 2nd is the closing bracket, and the 3rd is everything in between the brackets. 20:55
I didn't find any documentation about this use of ~ in doc.perl6.org/language/regexes 20:56
masak bzipitidoo: that's correct, as far as I know
bzipitidoo: note that if you have more than one atom for either of the operands, you need to bracketize
20:57 cbk joined
masak I think I just did a first: a method containing a lexical class with attributes whose *defaults* are some of the method parameters :D 20:58
(and it actually simplifies the code quite a bit)
jnthn Additionally, if it fails to find the terminator, it calls a method FAILGOAL, which you can define in your grammar if you want to do good error reporting
yoleaux 15:32Z <cygx> jnthn: irclog.perlgeek.de/perl6/2015-11-20#i_11574374 was about a non-existing dynamical variable, cf github.com/rakudo-p5/v5/commit/a41...d0ddf215c4
bzipitidoo ok, so far so good. In the example examples.perl6.org/categories/inter.../lisp.html there is this: token string { \c[QUOTATION MARK] ~ \c[QUOTATION MARK] [ <str> | \\ <str=.str_escape> ]* } 20:59
PerlJam bzipitidoo: see S05:872 for some info about it.
20:59 water_melon left 21:00 synbot6 joined
PerlJam S05:872 21:00
synbot6 Link: design.perl6.org/S05.html#line_872
PerlJam hello synbot
bzipitidoo I suppose QUOTATION MARK is a set of characters that includes single and double quote. Is ~ supposed to be smart, and only match like characters, or would it match a double quote with the next quote, whether single or double? 21:01
jnthn \c[...] is actually just a way to refer to a char by its name
So it is a literal "
nine When we install into %*CUSTOM_LIB<site> precompilation on installation will not actually be all that usefull because $*REPO will be %*CUSTOM_LIB<home> and that's where we get a precomp store from.
psch nine: do you perhaps have a hint or two for me to figure out the case of the missing repo/precomp on r-j? 21:04
nine: i recall you complaining about a mkdir difference between the backends, could that be it?
[Coke] Cannot find method 'has_compile_time_value' 21:05
TEttinger m: say "\c[APOSTROPHE]"
camelia rakudo-moar ce69f3: OUTPUT«'␤»
jnthn TIL that if you are on a plane and it flies through a blizzard, it looks like you're in a space movie going full speed... :) 21:06
RabidGravy++ # updating the Supply docs
RabidGravy I wasn't sure that the language is entirely correct 21:07
moritz jnthn: I hope the plane withstood the blizzard without damage :-)
dalek kudo/nom: ab6f591 | (Stefan Seifert)++ | src/core/Process.pm:
Store CompUnit::Repositories in %*CUSTOM_LIB

This is the second and much more correct attempt. The previous commit with this name re-introduced the bug where using -I install/share/perl6/site would make us loose most of the repo chain. This implementation retains the uniqueness filter when adding repos to the chain.
RabidGravy Supplier is like a factory and something else
jnthn moritz: Easily, I think. Wasn't even turbulent.
nine psch: oh, indeed, that may be it! I thought I fixed it but I couldn't actually try out the fix due to other JVM brokeness that you since fixed. 21:08
jnthn RabidGravy: I'll have a look over it, but a bit tired from the trip.
moritz wow, only 5 free slots in the advent calendar
jnthn Wow
psch nine: right, i'll have a closer look there :)
colomon \o/
jnthn Did anyone sign up to do one about the supply syntax?
moritz ... before we even published the first one
jnthn I could do one of those, but after LPW :P
RabidGravy it's fine, no-one noticed the examples weren't working for a month ;-)
moritz jnthn: the 14th is still free (that's the lastest that's still available) 21:09
jnthn: quick, claim it now :-)
jnthn Where's the list? :)
bzipitidoo thx everyone. Trying to understand perl6 grammars. Be nice to have a list of reserved (or special) names. So far I have seen TOP and actions, and now jnthn mentioned FAILGOAL. Back to playing around with them.
gfldex jnthn: github.com/perl6/mu/blob/master/mi...5/schedule
moritz jnthn: perl6/mu on github, file misc/perl6advent-2015/schedule
jnthn Ah, guess I can edit it on GitHub rather than haivng to clone that repo :)
flussence m: say Grammar.^methods.map(*.name) 21:10
camelia rakudo-moar ce69f3: OUTPUT«(parse subparse parsefile FAILGOAL upper CAPHASH CALL_SUBRULE !reduce_with_match !protoregex DYNQUANT_LIMITS !fresh_highexpect !BACKREF-LATEST-CAPTURE !DYNQUANT_LIMITS orig !cursor_start !protoregex_nfa !protoregex_table MATCH_SAVE RECURSE !cursor_more !cu…»
21:12 rindolf left
RabidGravy ooh I didn't realise it output the "private" methods as well 21:12
[Coke] m: $_="Bruce Gray"; my $x="Andrew Egeler"; m:i:m/$x/; 21:13
camelia rakudo-moar ab6f59: OUTPUT«Invalid string index: max 4, got 5␤ in block <unit> at /tmp/rbJOtklMkQ:1␤␤»
[Coke] ^^ that's LTA. 21:14
ISTR tripping over this earlier, not sure if it's RTd
jnthn More than LTA, a bug
[Coke] near the comment "# ignorecase+ignoremark" 21:15
grondilu didn't something like 'my %h = { foo => "bar" };' use to work?
(instead of 'my %h = foo => "bar"' that is)
[Coke] kind of wants $a eq:i:m $b 21:16
lizmat m: my $a = "FOO"; my $b = "foo"; say $a eq:i $b 21:17
camelia rakudo-moar ab6f59: OUTPUT«5===SORRY!5=== Error while compiling /tmp/jntRgsIcbf␤Two terms in a row␤at /tmp/jntRgsIcbf:1␤------> 3y $a = "FOO"; my $b = "foo"; say $a eq:i7⏏5 $b␤ expecting any of:␤ infix␤ infix stopper␤ postfix␤ …»
dalek : aa128b5 | jnthn++ | misc/perl6advent-2015/schedule:
Sign up to do a post
lizmat [Coke]: if that would parse, I don't see why we wouldn't be able to make that work :-)
psch huh 21:18
lizmat [Coke]: but what would the :m mean ?
psch trying to fix mkdir on r-j i noticed that r-m doesn't care if we try to create an existing directory
lizmat ah, ignoremark :-)
[Coke] lizmat: ignoremark, same as regexp
psch ...which i like, because it means i can do the less-code fix for r-j :P
[Coke] jnthn: the ignorecase/ignoremark code is also using nqp::lc to compare things. 21:19
psch well, except if we *should* complain when trying to create an existing directory
[Coke] psch: shouldn't we complain only if we tried to create a directory and then a directory didn't exist?
jnthn [Coke]: Well, fc it...
RabidGravy but that's more assert-dir rather than mkdir but hey 21:20
[Coke] is thinking in terms of "an install process" not literally just on "mkdir"
jnthn [Coke]: I think I wired up fc on the JVM to use lc as a cheat, so it may be as easy as "replace the op"
21:20 cdg joined
psch [Coke]: oh. i'm talking about literal &mkdir 21:20
flussence m: put “PATH="{($*VM.config<bindir>, ($*REPO.repo-chain X~ ‘/bin’), ‘$PATH’).flat.join(‘:’)}"”; # I guess this is as correct as it's gonna get... 21:21
camelia rakudo-moar ab6f59: OUTPUT«PATH="/home/camelia/rakudo-m-inst-2/bin:/home/camelia/.perl6/2015.11-305-gab6f591/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:$PATH"␤»
psch [Coke]: as in, r-m doesn't complain about < my $path = "foo".IO; mkdir $path; mkdir $path >
(this is motivated by r-j being broken currently, which seems to be at least related to mkdir, because r-j mkdir doesn't create directories recursively) 21:22
the easy fix for r-j (as in, nqp-j::mkdir) is using Files.createDirectories instead of Files.createDirectory, but the former doesn't throw on trying to create an existing directory 21:23
which moar also doesn't
which is why i was about moar not complaining :)
21:24 |Sno| left
psch +glad 21:24
nine Who needs a specification when you can emulate an implementation ;)
[Coke] jnthn: added RT #126771 21:25
synbot6 Link: rt.perl.org/rt3/Public/Bug/Display...?id=126771
psch well, moar's the favored backend for a reason, i try to think :P
hankache anyone ever used multidimensional arrays?
ugexe m: say %*CUSTOM_LIB; # is this not supposed to be used outside core now? 21:26
camelia rakudo-moar ab6f59: OUTPUT«Dynamic variable %*CUSTOM_LIB not found␤ in block <unit> at /tmp/zUtypmG5DM:1␤␤Actually thrown at:␤ in block <unit> at /tmp/zUtypmG5DM:1␤␤»
jnthn hankache: If "writing hundreds of tests for them" counts as using, I guess I have... :) 21:27
[Coke] hankache: Doubt they've been out long enough for anyone to do anything really big, but sure - what's your question?
... aside from jnthn who wrote everything and I was trying to avoid throwing him under the bus. :)
hankache hi jnthn [Coke]
[Coke] m: use lib 'lib'; use lib 'lib'; use lib 'lib'; say @*INC.grep(* eq 'file#lib').elems 21:28
camelia rakudo-moar ab6f59: OUTPUT«Dynamic variable @*INC not found␤ in block <unit> at /tmp/kT3E65THEn:1␤␤Actually thrown at:␤ in block <unit> at /tmp/kT3E65THEn:1␤␤»
[Coke] ^^ nine: that's from RT #126718
synbot6 Link: rt.perl.org/rt3/Public/Bug/Display...?id=126718
hankache is it normal that both arrays are the same? gist.github.com/hankache/481ddf2834b0eb0947f0 21:29
[Coke] hankache: you're getting bitten by "say" 21:30
jnthn Yes, they'll gist the same
[Coke] say calls ".gist" and gives you an overview of the object. try "say @thing.perl" to see a closer-to-the-metal version.
jnthn And what [Coke] said 21:31
[Coke] with .perl you get:
Array.new(:shape(2, 2), [1, "x"], [2, "y"])
[[1, "x"], [2, "y"]]
TimToady or use "put" to mean "print with a newline"
[Coke] TimToady: that doesn't help show the difference. 21:32
they put the same, also.
TimToady didn't actually read the backlog :)
hankache ah i see, i should have used .gist
jnthn We noticed :P
[Coke] no worries. :)
21:32 parisba_ is now known as parisba
[Coke] hankache: no, you DID use .gist 21:33
hankache hi TimToady
[Coke] say @foo is the same as say @foo.gist
TimToady
.oO(which is the same as put @foo.gist)
 21:34
[Coke] my @a=1,2,3;say @a.gist.gist.gist.gist.gist.gist.gist.gist.gist.gist.gist.gist.gist.gist.gist; 21:35
m: my @a=1,2,3;say @a.gist.gist.gist.gist.gist.gist.gist.gist.gist.gist.gist.gist.gist.gist.gist;
camelia rakudo-moar ab6f59: OUTPUT«[1 2 3]␤»
[Coke] (was wondering why that was so slow the first time. :)
21:35 dha left
hankache [Coke] meant .perl ;( 21:36
21:36 Guest1854 left
hankache is there any methods/subs that i can use only on multidimensional arrays and not on normal ones? 21:37
jnthn hankache: No, the difference is more in their behavior 21:38
m: my @a = [1,2], [3,4]; say @a.values;
camelia rakudo-moar ab6f59: OUTPUT«([1 2] [3 4])␤»
jnthn m: my @a[2;2] = [1,2], [3,4]; say @a.values;
camelia rakudo-moar ab6f59: OUTPUT«(1 2 3 4)␤»
hankache if i consider a multidim array as a table, can i push a whole row instead of cell by cell? 21:39
thanks jnthn 21:40
I am trying to see how i can use multidimensional arrays to store data fetched from a DB
gfldex m: my @a[3;3]; @a[0..2;1] = (1,2,3); dd @a; 21:42
camelia rakudo-moar ab6f59: OUTPUT«Partially dimensioned views of arrays not yet implemented. Sorry. ␤ in block <unit> at /tmp/e8EfNZ6kOY:1␤␤»
hankache on that note, what do you use if you want to store data fetched from a DB? in atabular format 21:45
moritz well, an arrays isn't a bad choice :-)
21:45 bjz left, emdashcomma left 21:48 emdashcomma joined
jnthn For now, indexing into multi-dim arrays will be a little tedious, as gfldex just noticed. Eventually that should work. 21:48
Not sure we've time to make it happen this side of 6.c, but that's fine; the syntax is reserved and throws an NYI, we can come back to it in the future. 21:49
hankache thank you jnthn moritz 21:50
21:50 diakopter_______ joined
hankache thanks gfldex 21:51
n0tjack just coming back to this, but shaped arrays and indexing is relevant for a larger project I'm considering - are they slated for 2016?
dalek c: 4eed22e | TimToady++ | doc/Language/faq.pod:
--> is not checked at compile time
 21:54
21:54 bjz joined
psch n0tjack: native shaped arrays are already there i gather 21:55
m: my int @a[2;3];
camelia ( no output )
21:56 xfix left
jnthn n0tjack: Shaped arrays are there, just not as convenient to work with or as performant as you might hope. Yes, improvements on both are probably something for 6.d 21:56
Well, performnce isn't part of the language spec. :) 21:57
But my expectation is that we'll be doing a lot of perfoormance engineering in 2016, and not a lot of feature.
psch secretly plots a brainfuck impl of perl6
jnthn (And robustness, and so on)
n0tjack thanks psch, jnthn
psch fwiw, i'll probably do a lot of features, but those'll likely be r-j confined 21:58
jnthn As in, get r-j up to parity? :)
psch well, yes, that too
21:58 skids left
jnthn suspects NFG on JVM will be one of the trickier ones. 21:59
psch yeah
not sure i can tackle that one
diakopter_______ just extend JVM with some Unsafe-heavily-depending modules; everyone else does it
psch interop is kinda my hobbyhorse
jnthn I suspect I can figure out how to get NFG in 22:00
22:00 sno joined
jnthn One of those "the second impl is eaiser than the first" things :) 22:00
Zoffix diakopter_______, every day I see you grow another _ :)
jnthn I'm totally fine with interop features going in. :)
22:00 kjs_ left
psch jnthn: yeah, i suspect that'd hold. i don't have anything close to a handle on unicode spec, so i'll readily defer to you there :) 22:00
nine ugexe: %*CUSTOM_LIB is initialized together with $*REPO. So currently it's only available after the first access to $*REPO (like a use or require statement) 22:01
psch nine: updating nqp::mkdir to include subdirs and not throw on existing gives me "Missing serialize function for REPR ContextRef" for install-core-dist.pl
(as in, it dies in the make step that runs install-core-dist.pl with that error) 22:02
just as an fyi 22:03
jnthn fwiw, Moar doesn't implement serializing those either 22:04
22:04 cognominal_ left
ugexe nine: thanks. i take it thats likely to change eventually? 22:05
22:05 kmel joined 22:06 hankache left
kmel how do you turn an array into a list? 22:06
psch jnthn: so that's nqp-j over-eagerly throwing on missing serialize for a given REPR..?
kmel: assuming "a list" means "a list of parameters" you can use |@myarray to use an array as a list of positional parameter for routine invocation 22:07
nine psch: can you run that script with --ll-exception?
jnthn psch: No, I think Moar would throw too
masak 'night, #perl6
jnthn psch: So I think there's a bit more going on
psch jnthn: ah, so more along the lines of trying to serialize where we shouldnt, i gather 22:08
diakopter_______ masak: gnikht
psch 'night masak o/
kmel night masak
22:08 diakopter_______ left
psch nine: --ll-exception doesn't really add anything 22:08
nine odd
kmel psch i think @array.list does it 22:09
psch nine: subnames are "throw, throw, , , sink-all, sink, protect, install, <unit>..."
m: my @a = 1,2,3; say @a.list.WHAT
camelia rakudo-moar ab6f59: OUTPUT«(Array)␤»
psch kmel: well, it depends on what you want
kmel: if you want an honest-to-the-interpreter List, .List is probably your best bet :) 22:10
m: my @a = 1,2,3; say @a.List.WHAT
camelia rakudo-moar ab6f59: OUTPUT«(List)␤»
psch kmel: if you're using list in the perl5 way of a-list-of-parameters, using |@a to lower your array into positional parameters is what you want 22:11
kmel psch yes |@a does the job
22:12 YP-QMUL-W left
nine Good night all! 22:12
22:12 balazs left
kmel night nine 22:13
m: my @tbl2; @tbl2 = [1,"x"],[2,"y"]; say @tbl2.perl;
camelia rakudo-moar ab6f59: OUTPUT«[[1, "x"], [2, "y"]]␤»
kmel m: my @tbl2; @tbl2 = [1,"x"],[2,"y"]; say |@tbl2;
camelia rakudo-moar ab6f59: OUTPUT«[1 x][2 y]␤»
kmel m: my @tbl2; @tbl2 = [1,"x"],[2,"y"]; say @tbl2.list;
camelia rakudo-moar ab6f59: OUTPUT«[[1 x] [2 y]]␤»
kmel m: my @tbl2; @tbl2 = [1,"x"],[2,"y"]; say @tbl2.List;
camelia rakudo-moar ab6f59: OUTPUT«([1 x] [2 y])␤»
22:13 kyclark left 22:14 diakopter_______ joined
dalek c: 1ce53ae | TimToady++ | doc/Language/faq.pod:
more FAQ clarifications
 22:14
[Coke] TimToady: "instead throws an exception instead of returning" could probably lose the first instead. 22:16
kmel psch the flattening operator | is what I was looking for. thanks
TimToady [Coke]: thanks 22:17
kmel time to hit the road 22:18
see you later
22:19 kmel left, kyclark joined
[Coke] time to hit the gym. blargh. 22:20
gtodd If someon is reading example Perl6 code and can't wait for Christmas and you see: class Whatever { has $.thing; has $!privatething; .... } should they be able to look up what $. and $! are by searching for "$!" in the docs or should they know they need to search for classes attributes etc? 22:21
how "granular" (?) should searching on doc.perl6.org be ? 22:22
oh hmm you can searc hfor $! ... it gives only one result though (variables) 22:23
22:23 donaldh joined
gtodd I guess you would find it be searching for "has" 22:24
22:26 raiph joined 22:27 balazs joined 22:29 Praise joined, Praise left, Praise joined 22:30 zengargoylew left, cgfbee left, water_melon joined, kjs_ joined 22:31 qcoh_ left
Hotkeys gtodd: I think being able to search for that would be useful 22:31
dalek c: 9e90d1d | TimToady++ | doc/Language/faq.pod:
more FAQ tweaks
donaldh psch: re "Missing serialize function for REPR ContextRef" try setting NQP_VERBOSE_EXCEPTIONS
Hotkeys TimToady: can we implement time travel in perl 6? 22:32
donaldh m: sleep 1 # time travel in one direction
camelia ( no output )
Hotkeys I mean apart from real-time forward time travel 22:33
22:33 cgfbee joined
diakopter_______ that's a self-fulfilling prophecy request prophecy 22:34
22:36 khw left
RabidGravy anyone got a thought of how to make a subset of JUnction which the range of values is constrained? 22:37
like
m: subset Minute of Junction where * ~~ all(0 .. 59); my Minute $m = any(0 .. 59); my Minute $m2 = any(10,11) 22:38
camelia rakudo-moar ab6f59: OUTPUT«Type check failed in assignment to $m2; expected Minute but got Junction␤ in block <unit> at /tmp/IMC_wHfCSJ:1␤␤»
RabidGravy except the second one works
22:38 espadrine joined
RabidGravy but "my Minute $m = any(10, 70)" doesn't work 22:39
22:40 llfourn joined 22:44 njmurphy_ joined 22:49 mephinet left 22:50 bjz left 22:51 khw joined, uruwi joined
AlexDaniel m: say ‘hello’; sleep -10; say ‘time travel’ 22:52
camelia rakudo-moar ab6f59: OUTPUT«hello␤time travel␤»
AlexDaniel m: say ‘hello’; sleep -Inf; say ‘time travel’ 22:53
camelia rakudo-moar ab6f59: OUTPUT«hello␤time travel␤»
AlexDaniel m: say ‘hello’; sleep NaN; say ‘time travel’
camelia rakudo-moar ab6f59: OUTPUT«hello␤time travel␤»
AlexDaniel m: say ‘hello’; sleep ‘five hours’; say ‘time travel’
camelia rakudo-moar ab6f59: OUTPUT«hello␤Cannot convert string to number: base-10 number must begin with valid digits or '.' in '7⏏5five hours' (indicated by 7⏏)␤ in block <unit> at /tmp/LUoAZmllEm:1␤␤Actually thrown at:␤ in block <unit> at /tmp/LUoAZmllEm:1␤␤»
Hotkeys m: subset Minute of Junction where * ~~ all(0 .. 59); my Minute $m = any(0 .. 59); my Minute $m2 = any(10,11); my Minute $m3 = any(0, 59); my Minute $m4 = any(1 .. 58);
camelia rakudo-moar ab6f59: OUTPUT«Type check failed in assignment to $m2; expected Minute but got Junction␤ in block <unit> at /tmp/V09CmB6TfW:1␤␤»
Hotkeys m: subset Minute of Junction where * ~~ all(0 .. 59); my Minute $m = any(0 .. 59); my Minute $m3 = any(0, 59); my Minute $m4 = any(1 .. 58); 22:54
camelia rakudo-moar ab6f59: OUTPUT«Type check failed in assignment to $m3; expected Minute but got Junction␤ in block <unit> at /tmp/wW3VBJMC7Y:1␤␤»
Hotkeys m: subset Minute of Junction where * ~~ all(0 .. 59); my Minute $m = any(0 .. 59); my Minute $m4 = any(1 .. 58);
camelia rakudo-moar ab6f59: OUTPUT«Type check failed in assignment to $m4; expected Minute but got Junction␤ in block <unit> at /tmp/lGeeQyDQl1:1␤␤»
lizmat And another Perl 6 Weekly hits the Net: p6weekly.wordpress.com/2015/11/30/...ly-splits/
AlexDaniel Hotkeys: arguably sleep with negative values should throw NYI warning
Hotkeys m: say any(10,11) ~~ any(0..59);
camelia rakudo-moar ab6f59: OUTPUT«True␤»
Hotkeys oh 22:55
uruwi m: say any(10, 11) ~~ 0 .. 59
camelia rakudo-moar ab6f59: OUTPUT«any(True, True)␤»
Hotkeys m: say any(10,11) ~~ all(0..59);
camelia rakudo-moar ab6f59: OUTPUT«False␤»
uruwi m: say True ~ any(10, 11) ~~ 0 .. 59
camelia rakudo-moar ab6f59: OUTPUT«False␤»
uruwi m: say True ~~ any(10, 11) ~~ 0 .. 59
camelia rakudo-moar ab6f59: OUTPUT«True␤»
uruwi m: say True ~~ (any(10, 11) ~~ 0 .. 59)
camelia rakudo-moar ab6f59: OUTPUT«True␤»
uruwi m: say (True ~~ any(10, 11)) ~~ 0 .. 59
camelia rakudo-moar ab6f59: OUTPUT«True␤»
uruwi what 22:56
Hotkeys Why is minute being defined with all(0..59)
uruwi:
m: say so 0..59
camelia rakudo-moar ab6f59: OUTPUT«True␤»
Hotkeys probably why
RabidGravy: Why is minute being defined with all(0..59) instead of any(0..59) 22:59
lizmat hmmm... looks like day 2 of the advent calendar is visible to the world ? 23:00
sorry, crying wolf
23:00 kyclark left
RabidGravy for no particular reason, I just can't work out a way of saying "any subset of this junction" 23:01
AlexDaniel m: say $*TOLERANCE
camelia rakudo-moar ab6f59: OUTPUT«1e-15␤»
23:02 lichtkind left
AlexDaniel m: ^100 .rand.say 23:02
camelia rakudo-moar ab6f59: OUTPUT«Use of uninitialized value of type Any in numeric context in block <unit> at /tmp/Uh8pq0Nbyo:1␤Use of uninitialized value of type Any in numeric context in block <unit> at /tmp/Uh8pq0Nbyo:1␤0␤»
23:03 telex left
AlexDaniel m: ^100.WHAT 23:03
camelia rakudo-moar ab6f59: OUTPUT«Potential difficulties:␤ Precedence of ^ is looser than method call; please parenthesize␤ at /tmp/4Hsdbrbf4n:1␤ ------> 3^1007⏏5.WHAT␤WARNINGS:␤Useless use of "^" in expression "^100.WHAT" in sink context (line 1)␤Use of uninitia…»
AlexDaniel m: ^100 .WHAT
camelia rakudo-moar ab6f59: OUTPUT«===SORRY!===␤Method call must either supply a name or have a child node that evaluates to the name␤»
23:03 uruwi left, n0tjack left
AlexDaniel m: (^100).rand.say 23:03
camelia rakudo-moar ab6f59: OUTPUT«Use of uninitialized value of type Any in numeric context in block <unit> at /tmp/qHG1yYclJ_:1␤Use of uninitialized value of type Any in numeric context in block <unit> at /tmp/qHG1yYclJ_:1␤0␤»
RabidGravy that is I want a subset of Junction that is a Junction that will match some minutes.
AlexDaniel m: say (^100).rand 23:04
lizmat AlexDaniel: hmmmm... looking into that
camelia rakudo-moar ab6f59: OUTPUT«Use of uninitialized value of type Any in numeric context in block <unit> at /tmp/iL8LwR7QQi:1␤Use of uninitialized value of type Any in numeric context in block <unit> at /tmp/iL8LwR7QQi:1␤0␤»
23:04 telex joined
lizmat hmmm... it should fail 23:04
m: my $a = ^100 .rand; say $a 23:05
camelia rakudo-moar ab6f59: OUTPUT«Use of uninitialized value of type Any in numeric context in block <unit> at /tmp/FVd4UEMbyw:1␤Use of uninitialized value of type Any in numeric context in block <unit> at /tmp/FVd4UEMbyw:1␤0␤»
AlexDaniel it works on g308b9b7 though
lizmat hmmm.... wietrd
m: say (0.1 .. 0.3).rand
camelia rakudo-moar ab6f59: OUTPUT«0.255097074306016␤»
RabidGravy I think I'll just leave that as Junction for the time being 23:06
AlexDaniel uh what is the difference?
lizmat m: my $a = ("a".."z").rand; say $a 23:07
camelia rakudo-moar ab6f59: OUTPUT«Can only get a random value on Real values␤ in block <unit> at /tmp/gtI5_dU7Ux:1␤␤Actually thrown at:␤ in block <unit> at /tmp/gtI5_dU7Ux:1␤␤»
lizmat hmmm....
Hotkeys is excited for the perl 6 calendar 23:08
Er
Advent calendar
I assume it is happening again this year
I hope it is
hartenfels Hotkeys: github.com/perl6/mu/blob/master/mi...5/schedule 23:09
Hotkeys Woo
23:11 lucasb joined 23:13 khw left 23:14 kid51 joined
lizmat m: say (1..10).rand 23:16
camelia rakudo-moar ab6f59: OUTPUT«5.1859874970708␤»
23:16 hankache joined
Hotkeys lizmat: 23:17
lizmat AlexDaniel: seems the bug is in if any of the endpoints are excluded
fixing now
Hotkeys er
nvm
lizmat Hotkeys++ :-)
AlexDaniel lizmat++ :-)
Hotkeys I saw
<lizmat> m: my $a = ("a".."z").rand; say $a 23:18
and thought you meant .pick
but I guess now
not
lizmat no, .pick works, .rand should fail
23:18 diakopter_______ left
Hotkeys right 23:19
also
'2 bind or !2 bind - gfldex' I like this title
23:20 geraud left
lucasb Are the LPW talks getting recorded and put on youtube? I hope yes 23:20
23:20 uruwi joined
lizmat lucasb: I have no idea 23:21
can you find any from last year ?
uruwi /msg NickServ VERIFY REGISTER uruwi ryibyirffpen
lizmat then probably yes :-)
23:21 Peter_R left
lucasb just read the weekly, lizmat++ :) 23:21
uruwi Whoops
23:21 espadrine left
stmuk lucasb: usually the main ones are 23:21
23:21 diakopter_______ joined
lucasb stmuk: thanks, I'll wait then 23:22
I'm curious to know the content of the 'the emperor's new programming language' talk :D
lizmat well, I'm sorta looking forward sitting on row #1 on that one :-) 23:23
23:23 geraud joined
lizmat not sure if I'll be feeling naked or not :-) 23:23
stmuk . o O ( look at the tags for that talk ) 23:24
lizmat
.oO( that whine is definitely not perlsecco )
23:24 water_melon left 23:25 mr-foobar left, adhoc left, RabidGravy left
dalek kudo/nom: 20718a9 | lizmat++ | src/core/Range.pm:
Fix Range.rand, spotted by AlexDaniel++
 23:27
23:27 mr-foobar joined 23:28 sitaram_ joined, sitaram_ left, sitaram_ joined
zengargoyle 's laptop starting to give up the ghost :( 7½ year run is probably not so bad... 23:28
23:29 khw joined, sitaram left
zengargoyle magic smoke production not keeping up with demand 23:30
lizmat good night, #perl6!
AlexDaniel m: say “{7+½} year run is not so bad”
camelia rakudo-moar ab6f59: OUTPUT«7.5 year run is not so bad␤»
lizmat m: say 7½ # would be brill if that would work :-) 23:32
camelia rakudo-moar 20718a: OUTPUT«5===SORRY!5=== Error while compiling /tmp/pSF8ejLT7t␤Bogus postfix␤at /tmp/pSF8ejLT7t:1␤------> 3say 77⏏5½ # would be brill if that would work ␤ expecting any of:␤ infix␤ infix stopper␤ postfix␤ s…»
zengargoyle y'know, How to input Unicode on all the things... would make a good advent post.
hankache lizmat++ #perl6weekly 23:33
lizmat sleep&
diakopter_______ timotimo: can you correct a typo in that p6weekly
23:34 Skarsnik left
diakopter_______ DrForr talk should be act.yapc.eu/lpw2015/talk/6465 23:34
lizmat diakopter_______: fixed
really sleep&
hankache bonne nuit lizmat
AlexDaniel “Tim Toady added an approximately-equal operator, infix:<≅>” 23:35
AlexDaniel honestly wonders why it is not ≈
.u ≅≈ 23:36
yoleaux U+2245 APPROXIMATELY EQUAL TO [Sm] (≅)
U+2248 ALMOST EQUAL TO [Sm] (≈)
AlexDaniel hmm
diakopter_______ someone asked that on irclog and they said it looks too much like smartmatch
that's the royal they
timotimo okay, caught up with the backlog again
well, skipped a few lines here and there ... :S
lizmat: how do you feel about pointing people to hub.docker.com/_/rakudo-star/ from the weekly? 23:38
you know, since we also pointed out the rakudo star release?
AlexDaniel .u ≇
yoleaux U+2247 NEITHER APPROXIMATELY NOR ACTUALLY EQUAL TO [Sm] (≇)
timotimo that's a beautiful name 23:39
23:41 kjs_ left
ilmari .u ≹ 23:41
yoleaux U+2279 NEITHER GREATER-THAN NOR LESS-THAN [Sm] (≹)
geekosaur "almost, but not entirely, completely unlike..."
uruwi m: my int32 $a = 9; say $a;
camelia rakudo-moar 20718a: OUTPUT«9␤»
zengargoyle wishes for a fixed bitmap font with slightly better coverage of weird characters. 23:42
geekosaur s/completely//
AlexDaniel zengargoyle: unifont?
ilmari zengargoyle: not bitmap, but symbola is good for coverage
uruwi Hmm, after declaring an int32 variable on the REPL, any subsequent input produces "Unhandled lexical type in lexprimspec for '$a'" 23:43
zengargoyle yeah, want fixed. good ole plain x font
can't deal with truetype stuff. never looks good.
jnthn 'night, #perl6 23:44
ilmari likes anonymous pro for normal fixed-width stuff
AlexDaniel zengargoyle: right, soooo… unifont?
lucasb uruwi: interesting, I can reproduce this error here
23:45 rurban left
uruwi Do I have to explicitly cast int32 into int64 if I want to get a 64-bit product of two 32-bit integers? 23:45
zengargoyle lucasb: unifont looks horrible at the size i normally use. 23:46
23:46 yqt joined
lucasb zengargoyle: well, unicode chars for me all show up as question marks... I guess I live in the 90s :) 23:47
uruwi ツーバッド
AlexDaniel truetype looks horrible, unifont looks horrible…
timotimo uruwi: int32 and friends only really work properly when you do NativeCall
in other instances, it's usually just a 64bit integer instead 23:48
AlexDaniel actually, I kinda agree. The only font that I can stand is terminus, but the coverage is duh…
zengargoyle good old misc-fixed works mostly, get ≹ ≅ ... most things. but not say 💡 or 🔌
uruwi So what should I use instead?
timotimo the % operator perhaps? :(
23:51 skids joined 23:52 mr-foobar left, hankache left, sitaram_ left
uruwi So int is 32 bits on 32-bit platforms, right? 23:53
timotimo nah, we have 64bit integers everywhere
23:53 mr-foobar joined, csd_ joined 23:54 sitaram joined
uruwi So I can always trust int to hold 64. 23:54
TEttinger AlexDaniel: I'm pretty used to unifont. most pan-unicode fonts are not great 23:55
gentium is very nice in print
it's pretty good on screens
lucasb m: my int32 $x = 2**31-1; say $x++; say $x
camelia rakudo-moar 20718a: OUTPUT«2147483647␤-2147483648␤»
dalek kudo-star-daily: d7cbc90 | coke++ | log/ (2 files):
today (automated commit)
TEttinger (not pan-unicode, not fixed-width, but not well-known and really it's a great times new roman replacement for ordinary office stuff) 23:56
timotimo oh, huh? that surprises me
lucasb idk if they are int64 internally, but they behave like int32 with regard to wrap around, don't they?
timotimo i didn't know we implement this
23:57 spider-mario left
zengargoyle hrm, my rakudo from last week gives the lexprimspec error when using int32 23:58
diakopter_______ m: my int32 $x = 2**31; say $x
camelia rakudo-moar 20718a: OUTPUT«-2147483648␤»
zengargoyle maybe it's really new....
23:58 hartenfels left
uruwi Difference between Num and num? 23:58
timotimo uruwi: boxed vs not boxed
huh. does our regex engine per chance ever emit a call to postfix:<++>?
uruwi When should I use Num instead of num?
23:59 n0tjack joined
zengargoyle i've been using Source Code Pro in browser for scalable stuff. it's pretty nice. 23:59
sorta can't go wrong with Adobe.

Powered by Cro and the Raku® Programming Language.

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

Thank you!