»ö« Welcome to Perl 6! | perl6.org/ | evalbot usage: 'p6: say 3;' or rakudo:, std:, or /msg camelia p6: ... | irclog: irc.perl6.org | UTF-8 is our friend! | feather will shut down permanently on 2015-03-31
Set by jnthn on 28 February 2015.
00:04 Alina-malina left, Alina-malina joined 00:05 tinyblak left 00:10 skids joined 00:16 lucasb left 00:33 aborazmeh joined, aborazmeh left, aborazmeh joined 00:36 gfldex left 00:46 raiph joined 01:01 Rounin left 01:05 yqt left 01:07 aborazmeh left, laouji joined 01:15 yeahnoob joined
tony-o thanks for the PR zoffix 01:17
Zoffix np
01:19 aborazmeh joined, aborazmeh left, aborazmeh joined 01:22 dtscode joined
dtscode if i wanted to build the standard perl 6 interpreter from source, is this all i need? github.com/perl6/std/ 01:22
colomon dtscode: no. 01:23
dtscode: that’s just the parser
and not the parser that rakudo actually uses, it’s more of a standard reference implementation of a p6 parser
TimToady see rakudo.org/how-to-get-rakudo/ 01:24
dtscode oh so i should use rakudo? 01:25
colomon if you want to actually use perl 6, yes. :) 01:26
dtscode thanks :D 01:27
#perl was right. you guys are cool
colomon is just glad TimToady++ thought of the how-to-get-rakudo link 01:28
TimToady I went to perl6.org and clicked on the big DOWNLOAD button to see what it'd say :) 01:30
Zoffix woohoo. *now* I'm done. All non-compliant S11 dists received a RP :) ... "April 1, 2015: 124 contributions" on github lol :D 01:31
Actually did not see that big download button there the two times I wanted to install perl6.. It sorta blends into the background among those shadowy blocks 01:32
TimToady we should maybe put it in the top row rather than the bottom 01:35
afk
01:42 aborazmeh left, Vlavv_ left 01:46 dj_goku_ joined
dalek href="https://perl6.org:">perl6.org: 3aa4c7f | (Zoffix Znet)++ | source/index.html:
Move download button to top row, to make it more obvious
01:46
01:48 dj_goku left 01:55 Vlavv_ joined 02:06 chenryn_ joined 02:07 aborazmeh joined, aborazmeh left, aborazmeh joined 02:17 noganex_ joined, Juerd joined 02:18 synopsebot joined, hugme joined, ChanServ sets mode: +v hugme 02:20 noganex left 02:27 telex left 02:28 telex joined 02:43 jack_rabbit joined 02:47 tinyblak joined, tinyblak left 02:48 chenryn_ left 02:51 chenryn_ joined 03:01 tinyblak joined
Zoffix What do you think about fixing this github.com/perl6/doc/issues/45 (add missing variables to 'Search') by adding the missing variables directly to github.com/perl6/doc/blob/master/t...emplate.js ? 03:17
They're compile and dynamic variables (doc.perl6.org/language/variables#Co...variables) and they're in a table. I don't mind doing the work.
Unless there a way to make htmlfy to index them somehow? 03:18
Ah... design.perl6.org/S26.html#Indexing_terms 03:29
Zoffix makes a note to look at that tomorrow
03:33 BenGoldberg left
skids gist.github.com/skids/dbe151cb35a4de8c4b9a # could someone with a full test environment fold this in? 03:43
03:48 chenryn_ left
skids (I don't think we should be expecting new users to figure out that RAKUDO_MODULE_DEBUG exists.) 03:49
03:49 chenryn_ joined 03:53 chenryn_ left 04:02 aborazmeh left 04:03 dtscode left 04:12 raiph left 04:15 prime left 04:17 kaare_ joined
raydiak after building a fresh clone of raydiak/MoarVM, which is identical to MoarVM/MoarVM, moar --version gives "2015.01-190-gc3bf959"...anyone have a clue what is up with the ".01"? the commit part is correct :P 04:18
04:28 aborazmeh joined, aborazmeh left, aborazmeh joined
raydiak ah fixed it, needed a git pull --tags upstream master (and then git push --tags too) 04:35
04:37 skids left 04:46 aborazmeh left 04:50 chenryn_ joined 04:54 chenryn_ left 04:58 sirdancealot left 05:01 uvb joined 05:05 uvb left, uvb joined 05:31 uvb left 05:37 chenryn_ joined 05:38 yeahnoob left 05:47 xfix joined 05:55 Sqirrel joined
jnthn morning, #perl6 06:12
dalek kudo/nom: 10cd401 | jnthn++ | src/Perl6/Grammar.nqp:
Give decent error for | used outside of arglist.

Before we just called it out as a call to a non-existing prefix:<|> sub, which was LTA.
06:13
kudo/nom: 1457816 | jnthn++ | src/ (2 files):
Typed exception for mis-placed |.
bartolin_ \o/ jnthn 06:14
dalek ast: 72ca1c1 | jnthn++ | S32-exceptions/misc.t:
Test for X::Syntax::ArgFlattener exception.
jnthn Folks can feel free to word-smith the message; there may be a better one :) 06:15
moritz "Illegal use of prefix | outside of an argument list"? 06:17
not sure that's better
anyway, jnthn++ for throwing the error in the first place
ah, I'm glad to see we're not the only one having trouble with high system load (on hack, I mean) 06:18
06:18 brrt joined
moritz the debian mirror we use at $work has a load avg. > 220 right now :-) 06:18
06:18 brrt left
jnthn loads of trouble... 06:18
06:22 jack_rabbit left 06:35 FROGGS joined 06:44 Sqirrel left 07:02 darutoko joined
moritz uhm, did the 'sillymoose' github account rename to 'dnmfarrell'? 07:13
Woodi hallo #perl6 07:15
FROGGS moritz: yes
Woodi I tried to use Bailador for simple app, but fall into misc troubles... eg. pastebin.com/fcbCSM1G 07:16
dalek osystem: 29353d7 | moritz++ | META.list:
Track a github username change
07:17
Woodi I think Buf thingies should be drowned much deeper
moritz FROGGS: github.com/dnmfarrell/URI-Encode/c...7a403be094 introduced a trailing comma in build-depends 07:18
Woodi getting simple %env is much harder then it need to be...
FROGGS eww
jnthn request.env<$_>
Surely that's not what was meant? :)
{$_} 07:19
Woodi jnthn: works for request.env<PATH_INFO>;
jnthn Yes, that's a literal.
foo<...> is foo{'...'}
So you're doing request.env{'$_'} there
FROGGS moritz++ # I PR'd it
Woodi hmm :) 07:20
moritz PerlJam: raw.githubusercontent.com/perlpilo.../META.info is invalid JSON
Woodi but why that Bufs popping up ? 07:21
also: reload on ^key in running Bailador app would be nice :) 07:22
jnthn Woodi: I didn't design it so I don't know the actual answer, but "we can't rely on what encoding we have at that level" would be a fair enough answer.
07:23 zakharyas joined
jnthn You'd have to ask somebody who knows Bailador better than I do :) 07:23
Woodi jnthn: same thing on commend line works... or it is other level ?
jnthn Woodi: I don't know what request.env actually corresponds to 07:24
Woodi jnthn: as usual, method calls :) 07:25
jnthn :P
.oO( Not all that is accurate is useful... )
07:26
07:26 foRei joined, foRei left 07:27 eli-se joined 07:35 eli-se left 07:38 pdcawley joined
moritz moritz@pat:~/p6/all>ls */*/META.info|wc -l 07:40
297
moritz@pat:~/p6/all>grep -L provides */*/META.info|wc -l
TimToady m: my %a; say %a.of; my $a = {}; say $a.of
moritz 89
camelia rakudo-moar 145781: OUTPUT«(Mu)␤(Any)␤»
moritz TimToady: that looks worng
TimToady m: my @a; say @a.of; my $a = []; say $a.of
camelia rakudo-moar 145781: OUTPUT«(Mu)␤(Any)␤»
TimToady so does that
07:41 Sqirrel joined 07:42 eli-se joined, eli-se left
raydiak FROGGS++ thanks for the merge :) 07:43
07:45 fhelmberger joined
FROGGS pleasure :o) 07:46
raydiak allows Compress::Zlib to not require Inline anymore, which is great for installing on Windows 07:48
which seemed important to me all of a sudden when I merged the PNG support into Pray yesterday and then realized you can't install it without a c compiler 07:49
07:52 tinyblak_ joined
FROGGS ahh yes 07:53
I also need Compress::Zlib for panda/CPAN
that's awesome
raydiak funny thing I noticed about panda on windows, seemes to require "prove", which means I had to install strawberry perl too, unless we have our own replacement 07:55
07:56 tinyblak left
raydiak oh but if anyone feels up to porting github.com/MoarVM/MoarVM/commit/7a...a2e70ad5de to JVM...precisely how to do that isn't immediately obvious to me, I'd be guessing 08:00
08:03 eli-se joined, laouji left
eli-se I can make Perl 6 look like C++! 08:03
08:04 laouji joined
moritz is that a good thing? :-) 08:04
eli-se m: sub static_cast { { int => sub ($x) { $x.Int } } }; static_cast<int>(1.23)
camelia ( no output )
eli-se m: sub static_cast { { int => sub ($x) { $x.Int } } }; static_cast<int>(1.23).say 08:05
camelia rakudo-moar 145781: OUTPUT«1␤»
eli-se :)
FROGGS raydiak: you need to handle it here: github.com/perl6/nqp/blob/master/s....java#L178
raydiak: though, there is a caveat
eli-se Does a slang have to be a Grammar? 08:06
or based on
FROGGS raydiak: and it results in the question: "how the heck do I get a pointerish thing from the underlined storage of a VMArray?!"
08:07 laouji left
FROGGS eli-se: I'd say so 08:07
raydiak FROGGS: yeah that's where I stopped when I was looking in to it :)
08:07 laouji joined
eli-se meh 08:07
FROGGS raydiak: that's also where I stopped when I attempted that months ago .O9 08:08
err, :o)
eli-se: why meh?
eli-se that means you can't use any existing parser
such as libclang :P 08:09
jnthn Well, the real issue is having the other parser know when to stop and hand back control...
eli-se Can't just brace-match because of C preprocessor.
FROGGS exactly
eli-se jnthn: right
FROGGS if you could have a main entry token that knows where to stop you could hand off that piece to the other parser 08:10
08:10 RabidGravy joined
jnthn afaik, the thing calling a slang mostly wants to get back a Cursor saying "I matched and got to this point in the text", 08:10
08:13 kjs_ joined
raydiak FROGGS: I'll try to remember to look in to it again soon if nobody else does it sooner; I was trying to read about it and got confused but it's after 1 AM here so that's kinda expected 08:14
FROGGS yeah
sadly it still wont be a LHF after two cups of coffee in the morning :/
timotimo o/ 08:15
eli-se I find grammars difficult to use though 08:16
raydiak g'mornin timo
08:16 eli-se left
timotimo hey raydiak 08:17
Compress::Zlib working without Inline::C is fantastic news
also, leont was working on a TAP::Harness implementation in perl6 some time ago
that could be used to provide "prove" if there's no system-wide prove installed (i.e. on windows) 08:18
raydiak nice
timotimo github says leont/tap-harness6 was updated 7 hours ago, but the latest commit is somewhat older 08:20
might just have been a push of already-committed stuff
oh 08:21
no, that's from the "provides" pull request
moritz timotimo: also, could be another branch, or a tag
timotimo Zoffix: i don't know where these "6" come from in github.com/Leont/tap-harness6/pull/2/files 08:22
is that your code or just pandas gen-meta?
moritz uk.businessinsider.com/reddit-april...015-4?r=US muhaha 08:24
tadzik www.reddit.com/r/thebutton/comments...n_so_many/ :D 08:28
nine_ So what happened after April 1st was over? 08:30
yoleaux 1 Apr 2015 22:29Z <japhb> nine_: I seem to recall someone (you, I think?) mentioning that there was something unsafe about running multiple Python interpreters in a single process as a way of working around the unsafe/poorly scaling threading. Can you give specifics?
moritz nine_: www.reddit.com/r/thebutton seems like it stopped working :( 08:32
08:43 tinyblak_ left 08:44 eli-se joined 08:48 kjs_ left 08:49 rurban_ joined
raydiak good night #perl6 08:49
eli-se ybe
timotimo gnite raydiak 08:50
tadzik goonite
08:50 sirdancealot joined 08:52 albert joined 08:53 kjs_ joined, tinyblak joined, Rounin joined
albert Hello, I wonder is there a Perl 6 IDE with code completion and debugger? I tried Padre with Perl6 plugin,, but it doesn't work. 08:54
08:55 kjs_ left
moritz albert: what do you mean by "it doesn't work"? 08:56
08:57 tinyblak left
moritz albert: if it looks like a bug in Padre, please submit a bug report; your chances of getting it fixed are higher than finding another IDE 08:57
osfameron someone's tried embedding perl6 into eclipse too 08:58
hmmm, that's probably not the same thing 08:59
albert moritz: I already submitted a bug report: github.com/PadreIDE/Padre-Plugin-P.../issues/1. Maybe someone here can advise how to fix this bug?
nine_ albert: I'd guess that you're missing some Module::Install Padre plugin 09:00
nwc10 b 09:01
nine_ albert: try it again after installing Module::Install::PadrePlugin 09:02
09:02 gfldex joined
albert Ohh, Thanks! 09:02
eli-se I'm gonna rewrite my compiler in Perl 6 today. 09:03
it has some ugly hacks in it that I want to get rid of anyway 09:04
jnthn eli-se: Have fun! :) Hope it comes out nicely :) 09:05
eli-se the hardest part will probably be learning Perl 6 regexen
timotimo they are pretty much worth it, IMO
eli-se so I can translate this regex to Perl 6: [ \n]|#(\((?:(?-1)|.)*?\))|#.*?\n
timotimo using perl6-debug-m (plus installing Debugger::UI::CommandLine) is very nice for playing around with regexen 09:06
eli-se also, is 09:07
oops
09:07 eli-se left 09:08 eli-se joined
vendethiel- is someone on reddit really complaining about "FatRat.nude"? 09:08
timotimo may be 09:09
link?
eli-se lol reddit
pack seems to be similar in Perl 6
vendethiel- timotimo: it's directly linking to doc, r/programming/new "I know that it’s ... that doesn’t change the fact that Perl 6 has a method called FatRat.nude() (doc.perl6.org)"
tadzik vendethiel-: after Pillars of Transphobia there is very little that can surprise me
timotimo that's not what the method is called? 09:10
jnthn Rat and FatRat do have a .nude method
Because who wants to write .numerator-and-denominator? :)
eli-se has never seen a fat rat that isn't nude
tadzik hmm, well, the TMNT master was not fat iirc 09:11
timotimo i mean, the method is not called "FatRad.nude", at least not to me
eli-se it's not called until you call it!
timotimo tadzik: but he wasn't fat, was he? i seem to recall he was rather thin?
nine_ jnthn: Yesterday I did some digging on rt.perl.org/SelfService/Display.html?id=123989
tadzik yes, that's what my memory tells me too
eli-se tadzik makes me want tzatziki
tadzik hehe
nine_ jnthn: I guess a solution may be to copy the previous callframe's $*DISPATCHER to the current callframe. Does that sound feasible? 09:13
eli-se does Perl 6 have the equivalent of Perl 5 my $x = ''; open my $fh, '>', \$x; ?
timotimo is that what Tux built with IO::String?
jnthn nine_: Well, that should already be being done
nine_: That's what the setdispatcher/takedispatcher nqp ops are about
nine_ jnthn: where?
|Tux| IO::String was for reading, not for writing 09:14
timotimo oh, OK
nine_ jnthn: ok, thanks. Will look further into that direction
eli-se I want to write binary data, not text.
|Tux| though writing will be on my todo
jnthn nine_: Search for nqp::setdispatcher in src/Perl6/Metamodel/Dispatchers.nqp 09:15
09:17 eli-se left
|Tux| eli-se: gist.github.com/Tux/8ebe8af1949c2aced18b 09:17
vendethiel- I think you're supposed to use buffers then
|Tux| Buf is binary, Str is string 09:18
vendethiel- |Tux|: "str!" <- that ! shouldn't be needed for a positional parameter 09:19
timotimo right, but: Buf is rw, Str is immutable
(though there is substr-rw)
vendethiel- isocpp.org/blog/2015/04/the-c-languages-merge ahhaha
09:20 mr-foobar left
|Tux| April fool? 09:20
jnthn substr-rw doesn't change Str being immutable :) 09:21
timotimo right
jnthn It makes a new Str and assigns it to the container
timotimo it works on the scalar instead
yeah
jnthn Blob is an immutable buf 09:22
|Tux| adds Blob to notes
timotimo .o( and every last word he said was adorable! he said blob, blob, blob blo-blo-blo-blo-blo-blo-blob )
09:23 espadrine_ joined
timotimo aaw, stack egg 09:26
that's cute
sort of like a democratic tamagochi
masak antenoon, #perl6 09:29
moritz \o masak 09:30
Zoffix timotimo, fixed, sorry: github.com/Leont/tap-harness6/pull/2/files 09:31
moritz Zoffix++ # all those pull requests 09:32
Zoffix timotimo, they came from my code that figures out the names that stripped 'pm' before 'pm6' so the 6 got left over :} github.com/zoffixznet/perl6-util-M...run.pl#L68
\o/
timotimo hah :D
thanks a whole lot for doing this! :)
Zoffix np
09:34 eli-se joined 09:36 wryk joined 09:39 kjs_ joined 09:56 haroldwu left 09:57 tinyblak joined, haroldwu joined 09:59 prl joined
prl Hi guys, 09:59
masak \o
prl if I use a start block,
should i expect that the promised is kept? 10:00
my $p = start {1}; say $p.status
i got Planned in the latest Rakudo star
moritz prl: eventually, yes
prl: well, it's only kept after the code of the promise has actually run
prl: a thread pool scheduler runs it when it's convenient 10:01
and until then, it's Planned
prl aha, sleep 1 before reading $p.status helps
but
moritz don't sleep
if you want to wait for the result, that is
access $p.result; it'll block until the promise is fulfilled (or broken) 10:02
masak .oO( blog post: bad habits of a pre-promises era )
10:02 tinyblak left
prl yes true) 10:02
moritz doc.perl6.org/type/Promise
prl what if i divide by zero inside the start block?
it is still kept
moritz prl: why don't you try it and see?
masak: now listening to: No Doubt - Don't Sleep :-) 10:03
masak :)
prl it is still kept
moritz m: my $p = start { die "OH NOEZ" }; await $p; say $p.result
camelia rakudo-moar 145781: OUTPUT«===SORRY!===␤OH NOEZ␤»
prl it is still kept
10:03 albert left
prl urm:) 10:03
moritz prl: because 1/0 produces a Rat instead of dieing 10:04
prl my $p = {1 / 0}; .... say $p.status prints Kept
оу
ок
ok
cool
thanks
masak .u к 10:05
yoleaux U+043A CYRILLIC SMALL LETTER KA [Ll] (к)
masak hugme: hug the letter к 10:06
hugme hugs the letter к
masak I love our current collection of bots :>
jnthn m: my $a = 1/0; say $a.WHAT 10:07
camelia rakudo-moar 145781: OUTPUT«(Rat)␤»
jnthn Heh, such lazy :P
An actual exception in a Promise will make it Broken, though 10:08
masak .oO( but what happens if you wrap the Promise inside an Unbreakable Vow...? ) :P
masak .oO( 2037: Perl 7 is announced. it is the merge of Perl 6 and the HPMoR community ) 10:09
vendethiel- HPMoR?
moritz Hilarious Perl and the Methods of Rats?
tadzik :D 10:10
masak vendethiel-: massive time sink advisory.
you have been duly warned.
moritz vendethiel-: hpmor.com/
a Harry Potter fanfic (and a good one), with the same amount of text as ~4 original Harry Potter books
masak .oO( my Lock $lock .= new; $lock.alohamora ) 10:11
moritz masak: no, you must first .colloportus before you can .alohamora it :-) 10:12
DrForr sub AVADAKADAVRA() { self.free }
masak o.O
DrForr: you... you... you just *used self in a sub*! :( 10:13
moritz that's fine, if it's declared as an outer lexical
DrForr s/self// # ;...
masak also, be glad you got a vowel wrong there. otherwise you might've gotten a stern talk with Moody.
10:13 raydiak left
DrForr Never read the books. 10:14
tadzik you can't tell me what to do!
english so unambigous
DrForr I just assumed she ripped off 'abracadabra'.
10:16 chenryn_ left, prl left
DrForr "Well, see, they just took Latin, and ... ruined it." 10:16
masak what, you're saying "wingardium leviosa" is not actually Latin? 10:17
10:17 chenryn_ joined
jnthn Let's ask a native Latin speaker...oh, wait... :P 10:17
arnsholt Well, the Harry Potter spells are a reasonable facsimile of Latin, IMO =) 10:19
masak of course, they've been meticulously catalogued by the Wikipedia community: en.wikipedia.org/wiki/List_of_spel...rry_Potter 10:20
up to and including actual intersections with actual Latin. 10:21
tadzik nine_: I get an error when building Inline::Perl5: /usr/bin/ld: /home/tsosnierz/.plenv/versions/5.18.1/lib/perl5/5.18.1/x86_64-linux/CORE/libperl.a(op.o): relocation R_X86_64_32S against `PL_sv_yes' can not be used when making a shared object; recompile with -fPIC
10:21 chenryn_ left
tadzik is my perl5 miscompiled? 10:21
10:22 tinyblak joined, Akagi201_ left
moritz tadzik: look like 10:23
tadzik plenv---
moritz tadzik: see the README.md in Inline::Perl5 10:24
perl 5 needs the -Duseshrplib configure option
tadzik aha! 10:25
thank you
10:27 laouji left 10:30 laouji joined 10:43 eli-se left 10:44 eli-se joined, lichtkind joined 10:45 eli-se left 10:47 grondilu joined
grondilu what's this "provides" section in META.info about? github.com/grondilu/clifford/pull/1 I can't see any "provides" word in S11. 10:50
nine_ tadzik: hit me, too a couple of times already ;) 10:52
timotimo this cat likes stepping onto my touchpad 10:54
FROGGS grondilu: design.perl6.org/S22.html#provides
grondilu: and we all do that for this: perl6advent.wordpress.com/2013/12/...g-modules/ 10:55
and we talk about "S11 support" because it is about supporting design.perl6.org/S11.html#Versioning 10:56
10:56 kurahaupo1 joined 10:59 chenryn_ joined
bartolin_ while trying to make Test.pm report correct line numbers for failing tests I found the following: 11:05
11:05 eli-se joined
bartolin_ m: for 1 .. 200 { }; { say callframe(1) } 11:05
camelia rakudo-moar 145781: OUTPUT«(signal SEGV)»
bartolin_ m: for 1 .. 199 { }; { say callframe(1) }
camelia rakudo-moar 145781: OUTPUT«CallFrame.new(level => 3, annotations => ("line" => "1", "file" => "/tmp/azHtS40kZu").hash, my => EnumMap.new("\$!" => Mu, "!UNIT_MARKER" => Mu, "GLOBALish" => Mu, "\$=pod" => Mu, "\$_" => Mu, "\$?PACKAGE" => Mu, "\$/" => Mu, "EXPORT" => Mu, "::?PACKAGE" =…»
eli-se I wonder how VMs catch stack overflows.
like JVM and CLR do to throw exceptions
bartolin_ MVM_SPESH_DISABLE=1 makes the segfault go away.
eli-se in Windows it's trivial due to SEH
bartolin_ is the above a known problem or should I report it (and where: github issue for moarvm?) 11:06
timotimo bartolin_: that seems quite similar to a bug i saw recently
bartolin_ timotimo: masak digged for a segfault with "for": RT #124191 11:07
synopsebot Link: rt.perl.org/rt3//Public/Bug/Displa...?id=124191
bartolin_ timotimo: but this one does not go away with MVM_SPESH_INLINE_DISABLE=1. furthermore it looks different in so far as my for loop does nothing. 11:08
timotimo good catch
11:10 wryk left, SamuraiJack joined
bartolin_ with latest rakudo star it segfaults with 100 iterations (99 run fine): 11:11
star-m: for 1 .. 100 { }; { say callframe(1) }
camelia star-m 2015.02: OUTPUT«(signal SEGV)»
bartolin_ star-m: for 1 .. 99 { }; { say callframe(1) }
camelia star-m 2015.02: OUTPUT«CallFrame.new(level => 3, annotations => ("line" => "1", "file" => "/tmp/ERb3GGEtqd").hash, my => EnumMap.new("EXPORT" => Mu, "\$!" => Mu, "\$=pod" => Mu, "\$?PACKAGE" => Mu, "\$/" => Mu, "GLOBALish" => Mu, "::?PACKAGE" => Mu, "!UNIT_MARKER" => Mu, "\$_" => Mu…»
11:12 prl joined
prl guys, another question about promises 11:12
11:12 synopsebot left
prl my $p = start{sleep 5}; my $t = Promise.new($p, Promise.in(3)}; say $t.result; say $p.statusж 11:12
my $p = start{sleep 5}; my $t = Promise.new($p, Promise.in(3)}; say $t.result; say $p.status;
behaves differently from what I expect. I expect it to stop after 3 sec and print Planned for $p 11:13
instead it waits 5 sec and prints Kept
timotimo bartolin_: yeah, we very recently tuned the thresholds für spesh and jit to kick in a bit later
11:13 hugme left
timotimo the first argument to Promise.new is supposed to be code to run, you're passing in a Promise into Promise.new 11:14
FROGGS für :P
timotimo FROGGS: damn it %)
FROGGS timotimo: never mind, he does understand :o)
bartolin_
.oO( ist mir gar nicht aufgefallen ... )
11:15
prl no no, it was Promise.anyof
timotimo ah, ok
prl my $p = start{sleep 5}; my $t = Promise.anyof($p, Promise.in(3)}; say $t.result; say $p.status;
moritz m: my $p = start{sleep 1; say 'done sleeping'}; my $t = Promise.anyof($p, Promise.in(0.5)); say $t.result; say $t.status 11:16
camelia rakudo-moar 145781: OUTPUT«5===SORRY!5=== Error while compiling /tmp/xteei0iYJk␤Calling start() will never work with declared signature (&code, :catch(&catch))␤at /tmp/xteei0iYJk:1␤------> 3my $p = 7⏏5start{sleep 1; say 'done sleeping'}; my ␤»
moritz m: my $p = start {sleep 1; say 'done sleeping'}; my $t = Promise.anyof($p, Promise.in(0.5)); say $t.result; say $t.status
camelia rakudo-moar 145781: OUTPUT«done sleeping␤True␤Kept␤»
moritz m: my $p = start {sleep 3; say 'done sleeping'}; my $t = Promise.anyof($p, Promise.in(0.5)); say $t.result; say $t.status 11:17
camelia rakudo-moar 145781: OUTPUT«done sleeping␤True␤Kept␤»
prl yes, exactly what i have
timotimo m: my $p = start { say "starting to sleep"; sleep 3; say 'done sleeping'}; say "$p created" my $t = Promise.anyof($p, Promise.in(0.5)); say $t.result; say $p.status
prl Kept is for both $p and $t
camelia rakudo-moar 145781: OUTPUT«5===SORRY!5=== Error while compiling /tmp/p8MHCfYoxt␤Two terms in a row␤at /tmp/p8MHCfYoxt:1␤------> 3; say 'done sleeping'}; say "$p created"7⏏5 my $t = Promise.anyof($p, Promise.in(0.␤ expecting any of:␤ infix␤ infi…»
timotimo m: my $p = start { say "starting to sleep"; sleep 3; say 'done sleeping'}; say "$p created"; my $t = Promise.anyof($p, Promise.in(0.5)); say $t.result; say $p.status 11:18
camelia rakudo-moar 145781: OUTPUT«Promise<139684156545648> created␤starting to sleep␤done sleeping␤True␤Kept␤»
timotimo m: my $p = start { say "starting to sleep"; sleep 3; say 'done sleeping'}; say "$p created"; my $t = Promise.anyof($p, Promise.in(0.1)); say $t.result; say $p.status
camelia rakudo-moar 145781: OUTPUT«Promise<140493355297432> created␤starting to sleep␤done sleeping␤True␤Kept␤»
timotimo huh.
11:18 chenryn_ left
timotimo m: my $p = start { say "starting to sleep"; sleep 3; say 'done sleeping'}; say "$p created"; my $t = Promise.anyof($p, start { 1 }); say $t.result; say $p.status 11:18
camelia rakudo-moar 145781: OUTPUT«Promise<140333017106400> created␤starting to sleep␤True␤Planned␤» 11:19
prl nice:)
but why
timotimo good question 11:20
seems like a bug
11:20 eli-se left
timotimo tony-o: my web app seems very stable with latest HSA, but i still seem to have no way to detect when the client closes the connection for my EventSource connection 11:20
tony-o: and for some reason it seems like chrome keeps reconnecting over and over 11:21
11:24 KCL joined
timotimo tony-o: it's quite suspiciously close to exactly 30 seconds before each reconnect 11:25
or "close" for that matter
11:27 chenryn_ joined 11:28 KCL_ left
tadzik FROGGS: it seems like panda doesn't install test-depends 11:32
I had to manually install File::Temp for Inline::Perl5 11:33
FROGGS yes, test-depends has nothing to do with depends
panda could be made smart enough to only temporarily fetch test-depends, but until now it just ignores it 11:34
tadzik well, but now panda can't install Inline::Perl5 :)
FROGGS this commit is bogus I'd say: github.com/niner/Inline-Perl5/comm...fb1cd0f10c 11:35
for now at least
tadzik I think it's panda's fault to not understand it
I say we should make it behave like it used to, so install both build-depends and test-depends
FROGGS 'used to'? there never was a test-depends 11:36
tadzik yes, I know 11:37
what I mean is that modules that are now correctly written according to S11 are not installable with panda
and that's panda's fault
11:37 prl left
tadzik in this case panda is like cpanm, which also just installs whatever test-depends it finds for the sake of things working 11:38
(except that panda does not)
FROGGS m: my $depends = ['Foo']; say $depends (+) Any
camelia rakudo-moar 145781: OUTPUT«bag(Foo)␤»
FROGGS m: my $depends = ['Foo']; say ($depends (+) Any).list
camelia rakudo-moar 145781: OUTPUT«Foo => 1␤»
tadzik . o O ( it depends )
FROGGS m: my $depends = ['Foo']; say ($depends, Any).unique.list
camelia rakudo-moar 145781: OUTPUT«Foo (Any)␤»
11:41 coffee` joined 11:43 kjs_ left 11:45 Sqirrel left
FROGGS hmmm, my patch is not working yet :/ 11:48
11:50 chenryn_ left 11:53 SamuraiJack left 11:54 wryk joined 12:00 chenryn_ joined
FROGGS tadzik: I pushed a fix 12:03
tadzik FROGGS++ \o/
in return, have this floppy drive song: www.youtube.com/watch?v=4bnakHWlRAI
FROGGS that's totally worth it :D 12:04
tadzik oh it is :D
up from a certain point, at least
FROGGS *g* 12:05
lol: www.youtube.com/watch?v=cM_sAxrAu7Q 12:06
tadzik have you seen the one on dot printer?
www.youtube.com/watch?v=u8I6qt_Z0Cg 12:07
FROGGS wow 12:12
12:13 chenryn_ left, chenryn_ joined, raydiak joined 12:18 chenryn_ left 12:21 chenryn_ joined 12:27 laouji left
colomon www.scaleway.com/pricing # hmmm, might be worth it just to have someplace decent to try Rakudo on ARM? 12:35
12:36 chenryn_ left 12:40 rindolf joined
RabidGravy m: class Foo { }; say Foo.^ver; 12:44
camelia rakudo-moar 145781: OUTPUT«No such method 'gist' for invocant of type 'NQPMu'␤ in sub say at src/gen/m-CORE.setting:18729␤ in block <unit> at /tmp/uebvqBad6T:1␤␤»
[Coke] Zoffix++
RabidGravy is that expected, and if so how does one determine the version of a class?
12:45 rmgk left, raiph joined
RabidGravy colomon, got to be quicker than a RaspberryPI ;=) 12:46
colomon yeah, I figure 2GB ram would be a huge help there. :)
12:47 rmgk joined, telex left 12:48 telex joined 12:53 chenryn_ joined 12:54 chenryn_ left 13:05 aborazmeh joined, aborazmeh left, aborazmeh joined 13:06 kurahaupo1 left 13:15 sirdancealot left 13:21 skids joined 13:22 kjs_ joined 13:38 Hor|zon left 13:43 Hor|zon joined 13:50 captain-adequate left 14:01 rindolf left 14:03 tinyblak left 14:09 yqt joined 14:10 Perl6_newbee joined 14:15 captain-adequate joined, lucasb joined 14:17 aborazmeh left 14:21 captain-adequate left
[Coke] merges zoffix++'s PR for coke's dumb module. 14:21
14:23 Hor|zon left 14:24 pdcawley left 14:27 captain-adequate joined, pdcawley joined
lucasb perl6 -e '@*ARGS.shift.IO.lines' <(seq 10) #=> Unhandled exception: Failed to seek in filehandle 14:32
I know I can't seek a pipe, but why does the lines method need to seek?
moritz it shouldn't 14:33
14:35 Hor|zon joined 14:45 azawawi joined
azawawi hi 14:45
FROGGS: ping
.tell timotimo i think your problem with farabi6 is related to the new panda installation scheme (i.e. using CompUnitRepo). Please take a look at your share/site folder 14:46
yoleaux azawawi: I'll pass your message to timotimo.
azawawi .tell FROGGS I managed to fix farabi6 to use CompUnitRepo. Please see github.com/azawawi/farabi6/blob/ma....pm6#L182. FROGGS++ 14:47
yoleaux azawawi: I'll pass your message to FROGGS.
dalek kudo/nom: 3a8ecbc | usev6++ | lib/Test.pm:
Report correct line numbers for failing tests

The hard coded level for callframe we had before did only match in some cases. Now we're looking outwards until we leave Test.pm
Additionally we can grab the call frame from within conditional, since the bug which prevented that seems to be gone (cmp. commit 4111fa43ca).
14:51
kudo/nom: b4576d7 | PerlJam++ | lib/Test.pm:
Merge pull request #401 from usev6/fix_test.pm

Report correct line numbers for failing tests
timotimo hmm 14:52
yoleaux 14:46Z <azawawi> timotimo: i think your problem with farabi6 is related to the new panda installation scheme (i.e. using CompUnitRepo). Please take a look at your share/site folder
colomon woah, getting the line number of failing tests would be pretty huge… 14:53
lucasb perl6 -e '@*ARGS.shift.IO.slurp.perl.say' <(echo hello) # returns empty string 14:54
moritz colomon: haven't we had that for a while? 14:55
lucasb So, my guess is that some IO code doesn't like pipes
colomon moritz: may mean it’s been a while since I was debugging something in roast. ;)
14:56 kjs_ left
PerlJam moritz: modulo RT#124227, yes 14:56
moritz PerlJam: well yes, but I mean the feature in general 14:57
14:57 rindolf joined
moritz which hoelzro++ added in 2014-07-24 (commit 4a1fb9fa67790cd040ec98fdf5806e517dcbc28a) 14:58
colomon remembers long, frustrating periods pouring over roast source trying to figure out where the failing test was....
hoelzro++
moritz can't help but wonder how long it would have taken to find that commit with svn 14:59
15:00 azawawi left
bartolin_ hmm, yes the commit message was a bit misleading, sorry. hoelzro++ for adding that feature ("line numbers for failing tests"). 15:02
15:02 yqt left
[Coke] we need a test for that. :P 15:02
15:04 kjs_ joined 15:06 rindolf left 15:08 vytas joined
moritz bartolin_: I don't think it was a bad commit message 15:12
15:20 fhelmberger left
ugexe testers isn't generating HTML pages still 15:21
moritz FROGGS: you can re-enable your cron jobs, I think
15:26 Possum left
FROGGS moritz: shall I enable all of them at once? 15:27
yoleaux 14:47Z <azawawi> FROGGS: I managed to fix farabi6 to use CompUnitRepo. Please see github.com/azawawi/farabi6/blob/ma....pm6#L182. FROGGS++
15:28 zakharyas left
moritz FROGGS: yes, though with flock proctection against multiple executions at the same time 15:28
FROGGS sure
15:34 LonelyGM joined, Perl6_newbee left
japhb nine_: about? 15:36
15:38 coffee` left 15:39 coffee` joined, coffee` is now known as coffee, coffee left 15:40 coffee` joined 15:45 robinsmidsrod left, robinsmidsrod joined
lichtkind cheers i just decided to convert my personal private toolchain to perl 6 15:46
15:46 sirdancealot joined
[Coke] lichtkind++ 15:49
15:49 RabidGravy left
skids perl6 is such a tool :-) lichtkind++ 15:49
lichtkind i just started ++ when i have done 15:50
its also jsut a training becasue i will give next p6 talks on german perl workshop and yapc
PerlJam lichtkind: no, the ++ is because you've started :)
lichtkind :) 15:51
PerlJam lichtkind: Is there a reason this toolchain is personal and private? Could you make some modules out of it and share?
lichtkind no i also thought about to release it they are general useful for fast movements between different projects and such things 15:52
it just wrote it for fun but should be usable by any friend of command line 15:53
is github mandatory or would bitbucket accptable? 15:54
masak lichtkind++ 15:55
moritz lichtkind: I think part of our toolchain depends on the github URL schemes 15:56
lichtkind: so if you want it the easy way, use github; if not, we can try to flush out those github specifics
lichtkind but should be no problem to change sme strings?
moritz once you find what needs changing, it shouldn't be a big problem 15:57
lichtkind i just ask becasue i have currently same diskussian wiht mishin to convert github because metacpan
abraxxa FROGGS: did NativeCall already gain is rw support? 15:59
16:03 RabidGravy joined 16:09 eli-se joined 16:10 spider-mario joined 16:11 LonelyGM left
FROGGS abraxxa: no, it is trickier than I thought :/ 16:11
abraxxa FROGGS: i see. do you start a branch I could look at?
16:12 Psyche^ joined
eli-se hi 16:12
FROGGS abraxxa: I just have diffs on my disk yet... 16:13
abraxxa: I might put it in branches tonight
hi eli-se
16:13 LonelyGM joined
abraxxa FROGGS: ok 16:13
jnthn FROGGS, abraxxa: If this is a "should be done soon", I can be pondering it :) 16:14
nine_ japhb: hi
abraxxa jnthn: would be great so I can continue with DBDish::Oracle
16:14 Patterner left
jnthn And work on it in the next days. 16:14
Yes, unblocking Oracle support sounds worthwhile :)
abraxxa once the basics work and it's merged to the main repo I plan on working on DBIish itself too
jnthn Nice! 16:15
abraxxa things like AutoCommit should be a Bool, the github discussion about not making DBI's faults regarding PrintError/RaiseError
nine_ japhb: about your Inline::Python question: Python does not support multiplicity the way perl does. So using multiple interpreters concurrently probably is out of the question 16:16
abraxxa nine_: fork + IPC :P 16:17
16:17 LonelyGM left
dalek kudo/param-is-rw: 00674de | FROGGS++ | lib/NativeCall.pm:
implement "is rw" params in NativeCall
16:17
FROGGS jnthn: I just pushed my stuff to MoarVM/param-is-rw and rakudo/param-is-rw 16:18
though that does not mean that I hand it over directly... if I manage to get further today I'll push :o) 16:19
TimToady agh, S11 now says: Compilation units posted to an authority (such as CPAN) 16:21
CPAN is not an authority!
CPAN authors can be authorities though 16:22
jnthn FROGGS: I'm still recovering from sleeping 2 hours the night before last, so it'll be at least tomorrow before I'm up to it :) 16:23
FROGGS jnthn: just tell me
jnthn FROGGS: Will do :)
abraxxa FROGGS: looks good as far as I understand the code 16:24
FROGGS abraxxa: problem is "just" that we decontainerize the arguments in nqp when a native sub call is made, which strip off the NativeRef thingies 16:25
and that means we cannot assign to it anymore
that's the current showstopper 16:26
abraxxa FROGGS: what arguments? the 'is rw'? 16:27
FROGGS abraxxa: you apply 'is rw' to params... but the arguments you pass later when you call the native subs, these get deconted 16:28
abraxxa FROGGS: you mean that you don't know any more which argument correlated to which param? 16:29
jnthn No, we lost the fact it's a container. 16:30
abraxxa doesn't know what a container is
jnthn my $a = 42; # here, $a is a Scalar container, and 42 is the value stored inside of it
abraxxa why isn't it called a variable? 16:31
jnthn heh
Better to maybe say that a Scalar contianer is bound to the variable $a :)
*container 16:32
abraxxa a struct which hold type + value?
jnthn The distinction matters in Perl 6 because containers are first class
That's pretty much what a Scalar is, yes.
Hand-waving a bit, though, since actually it points to a descriptor containing the type, the variable name, and some ro-ness info 16:34
Uh, rw-ness actually
16:35 wryk left
abraxxa that MoarVM internal details, right? 16:37
another implementation might do it differently
dalek ecs: 093fdd7 | TimToady++ | S11-modules.pod:
Clarify that CPAN is not itself an authority

Also explain that "standard Perl 6 library" does not just mean publicly shared libaries.
jnthn abraxxa: Scalar is spec, but how it looks on the inside is not 16:39
timotimo i've been slightly annoyed by having decont_all generated for every nativeinvoke anyway
but what really annoys me about it is probably not having to do it ... it's not knowing what the callsite looks like at that point 16:40
FROGGS timotimo: and we have to tweak exactly that I think
bbl
16:40 FROGGS left 16:41 noganex_ left
jnthn timotimo: Yes, decont_all was certainly a "get stuff working" thing :) 16:41
OK, I'm informed it's dinner time.
&
timotimo i'm slightly jealous; i'll have to go grocery shopping before dinner can be made here :S
16:46 lizmat joined 16:47 kjs_ left
lizmat is glad that TimToady finally takes the time to read what I wrote over a year ago 16:47
yoleaux 1 Apr 2015 15:22Z <FROGGS> lizmat: We need to talk about CUR::Installation and the way that non-existent dirs trigger that CURs do not appear in @*INC... (I "fixed" it in a rather scary way and want to fix it properly)
16:47 lizmat left
PerlJam TimToady: What's an "official distribution"? (I seem to be having difficulty with the word "official" lately) Is it just one that claims to be a "Perl 6 distribution"? 16:47
jercos The official distribution is the one that gets handed around at the office, right? :p 16:48
TimToady it's official if it has any authority :) 16:49
16:49 RabidGravy left
dalek rl6-roast-data: 43afcd6 | coke++ | / (9 files):
today (automated commit)
16:51
PerlJam It makes it sound like there's some Perl6 godhead from which officialdom springs forth. (and that interpretation only makes sense to me for the "Perl 6 standard")
(and yes, this can be construed as a problem within my brain, but I'm slightly worried about the other brains that would agree) 16:52
16:53 lizmat joined
tony-o arnsholt: the 'act' method is in S17 under supplies 16:53
TimToady maybe we should ask the French Academy to nail down the official meaning of all these words...
lizmat maybe we should find a better wording instead of ridiculing an effort to make sense of the mess that nobody else dared to touch 16:54
16:54 lizmat left
ab5tract maybe it would be helpful to move from "official" to "reference" distribution? 16:54
TimToady all words are squishy
and I don't think there's any ambiguity at the top end, really, since official distributions generally know they're official; it's the local site-wide stuff that might get stuck in PERL6LIB-land 16:55
ab5tract TimToady: indeed. All the more reason, though, to heed that tale about castles in the swamp. 16:56
TimToady the important thing is not so much the haggling about the words, but actually having and conveying an accurate notion of the boundary where identity becomes critical to correctness, and when a module crosses that boundary 16:58
and certainly this is something to be worked out over time, since it's probably impossible to define accurately in advance
16:59 mr-foobar joined
TimToady in a sense, git errs on the side of giving everything an exact identity right from the start, even dev stuff 16:59
[Coke] finds "standard perl 6 library" confusing wording since that can be taken to mean "what is shipped with a p6 implementation"
[Coke] is probably slightly behind backlog, apologies. 17:00
TimToady hence the parenthetical remark to expand
17:01 molaf joined
PerlJam maybe S11:668 needs a similar parenthetical, or a link to a more comprehensive explanation that has yet to be written :) 17:01
TimToady at some point it all becomes tl;dr without achieving much more clarity...
17:02 yqt joined
TimToady the intent is to push the immutable identity boundary as close to development as possible, though we probably can't take the git approach of driving all the way down to the floor 17:03
ab5tract TimToady: reading your the clarification of "standard Perl 6 library", I get the more of a sense that you mean 'everything coded in Perl 6' -- . o ( call it the 'Perliverse' ? )
where 'everything coded in Perl 6' necessarily includes not only the local/site environment, but anything which might be conceivably added to that environment 17:04
dalek kudo/nom: d17c404 | lizmat++ | docs/release_guide.pod:
It's all tl;dr, won't have time
TimToady anything that might be shared and then reshared must have an appropriate identity to prevent accidental semantic action at a distance 17:05
PerlJam I'm mostly just worried about the meme that "official" means "Larry Wall said it was so" that seems to continue to pervade that populace. To me, "official" should mean 'conforms to an agreed upon interface' or something like that.
ab5tract PerlJam: then in a sense you could say 'conformant' or 'idempotent' ... 17:07
PerlJam (and clearly, identity would be part of the "interface" :)
TimToady I can't think of a good word that means "blessed by the authority in question" other than "official", but maybe it just requires more coffee 17:08
17:08 RabidGravy joined 17:10 Hor|zon left, espadrine_ left 17:11 spider-mario left
TimToady the problem is that it's a circular definition; what we're trying to do is to get authorities to authorize things when they ought to, and not send ambiguous things out into the wide world, because that's a big source of accidential genericity 17:12
ab5tract TimToady: but what is 'the authority' WRT the DarkPAN? Perl 6 is pupating in highly distributed times.
isn't the 'authorities who authorize' essentially the module writers who conform to the interface?
PerlJam ab5tract: there's no 'the authority', there's many of them.
ab5tract PerlJam: it was intended as a leading question :) 17:13
TimToady my parenthetical remark was intended to indicate that dark authorities still should consider this issue of accidentally mutating the language as well
it's like a contract, yes 17:14
ab5tract *module writers who choose to release their module into/according to this idempotent system
17:14 spider-mario joined
TimToady maybe there should just be a footnote on "official" linking back to this conversation :) 17:15
ab5tract so "official" really is circular here: "official because official therefore official"
I think a more appropriate phrasing would be 'participating'
TimToady circles the wagons :) 17:16
funny how that's also ambiguous, you can circle the wagons if you're a settler, or you can circle the wagons if you're attacking the settlers 17:17
PerlJam I think I'd elide the words "official" and "standard" and just say: Whenever an authority (such as a CPAN author or a company) posts a compilation unit as part of a Perl 6 distribution, or enters it into any Perl 6 library, the module is required to declare its full name so that installations can know its unique, immutable identity, ... 17:18
lichtkind jnthn: for my talk 2 weeks ago i used your picture with our inofficial slogan : all VM belong to us
PerlJam i.e. "Perl 6" is what makes it "standard" or "official"
(or conformant or idempotent) 17:19
TimToady ss/Perl 6 distribution/distribution of Perl 6 code/
PerlJam yeah, that too
17:21 rindolf joined
PerlJam This feels like one of those concepts that it's *really* important to make easily understood by the widest audience and so should be explained such that a child could understand 17:21
17:22 eli-se left
PerlJam (or, perhaps like many introverts, I'm *way* over-thinking it) 17:24
(or, perhaps I'm prone to hyperbole ;)
PerlJam lunch &
raydiak good morning #perl6 17:25
TimToady well, these are very slippery concepts, and I don't blame lizmat for blaming me for sniping at her when she's been trying to do something about it all along
if it's any help, in my head I was sniping at the words, not at lizmat++ 17:26
17:27 geever joined 17:32 yqt left
TimToady m: my @a; say @a.of; my $a = []; say $a.of; my %h; say %h.of; my $h = {}; say $h.of; 17:33
camelia rakudo-moar d17c40: OUTPUT«(Mu)␤(Any)␤(Mu)␤(Any)␤»
TimToady jnthn: do you see any particular reason not to make all of these Mu? (other than that .^name then returns Str<324189768>[Mu] instead of Hash?) 17:34
oddly, there's no trouble with .^name and Array, which still returns Array
TimToady assumes the Str<324189768>[Mu] problem is some kind of circularity and/or mixin cache issue 17:45
or somewhere it's assuming that Hash means Hash[Any] 17:46
ab5tract TimToady: I think you should consider the potential benefits of leaving the wagons be and throwing a collaborative pow wow instead. I wish it had happened in history more often, that's for sure. 17:50
17:50 coffee` left 17:52 joffrey joined 17:53 kjs_ joined 17:54 joffrey left, coffee` joined, Caymanboy joined 17:55 coffee` left
TimToady for good or ill, this is the pow wow 17:55
17:56 abraxxa left 17:57 edward left 17:58 coffee` joined 18:07 Hor|zon joined
ab5tract Tim 18:07
TimToady: gist.github.com/ab5tract/4dc7deb90ae8a75ecbcc
some potential phrasings for the pow wow :) 18:08
TimToady ab5tract++ :D
18:11 Hor|zon left
ab5tract my personal favorite is: The CompUnit is said to be 'representing' if and only if 18:12
because hip hop :) 18:13
18:15 johnjohn101 joined, yqt joined, Caymanboy left 18:17 Hor|zon joined
tony-o are you guys talking in reference to an article or a synopsis? 18:21
18:22 kjs_ left
johnjohn101 hi perl6 people 18:22
skids o/ 18:24
18:25 spider-mario left, eli-se joined
TimToady tony-o: mostly about S11 and how to manage modules out in the wide world 18:26
timotimo ah it's tony-o :) 18:27
eli-se I wonder what this Easter thing is.
18:28 spider-mario joined
tony-o TimToady: i interpreted 'authoritive' to mean that the module was uploaded to that authority. IE if i upload my module to panda: and cpan: wants to make that module available, then cpan: should defer to panda: (or the module author should upload the module to both places) 18:29
so if i uploaded CSV::Parser to panda: - panda: is the authoritive because, as the author, i chose it to be my one true authoritive module provider 18:30
TimToady modules are uploaded *by* authorities, not *to* authorities
you name might be conferred by CPAN, but you are the authority, not CPAN
tony-o then i should go reread s11 with new context 18:31
18:31 spider-mario left
tony-o last time i read it was like 8 months ago 18:31
[Tux] tony-o, /me would welcome generic feedback on github.com/Tux/CSV if you'd care :)
TimToady I guess I'd say that CPAN is an authority, but on user identities, not module identities 18:32
CPAN (or reasonable facsimile) does have to convey module identity properly, but that's also different from being a module name authority 18:33
tony-o is it expected, then, that CPAN share its modules with other apps, panda and the like
moritz yes
TimToady it doesn't matter, as long as module identity is maintained
tony-o so CSV::Parser:auth<tony-o>
is available from Panda, CPAN, etc.
moritz I fully expect panda to be able to use CPAN, and github, and other sources
TimToady what we don't want is different repos to offer different modules that appear to have the same identity 18:34
module identities need to be universal, like a URI
tony-o so you don't want CPAN to offer CSV::Parser:auth<tony-o> which differs (from a code perspective) from PANDA CSV::Parser:auth<tony-o> 18:35
TimToady not unless they are different versions :)
18:35 Aaron`J joined
tony-o both :vers<1.0> :-) 18:35
TimToady correct
and if there's a checksum to enforce it, so much the better 18:36
timotimo having gotten the user name "timo" on github gave me the doubtful honor of discovering random projects where people named timo are involved somehow
TimToady if it's a good enough checksum, you can use it as a proxy for the identity, as git does 18:37
timotimo i'd expect :auth<timo> would have to be prefixed with a name-authority in every case ... %)
TimToady where's your sense of self-authority? :) 18:38
tony-o how do you handle me uploading CSV::Parser, changing a couple of lines and uploading it to panda with same version? is the 'authority' expected to be different as timotimo is saying about the prefix?
*should* you be able to do panda upload CSV::Parser:auth<cpan:tony-o>:vers<1.0> ?
and have panda go get the information from cpan ? 18:39
TimToady that's why it really needs a hash of some sort to remind you to change the version 18:40
tony-o i'd think cpan should complain about duplicate versions the way npm does
TimToady if you're going to put it multiple places, we probably need a staging area locally from which it is then uploaded to other places 18:41
presumably any two repos out in the world should be able to exchange metadata and cross check it, and raise an alarm if two supposedly identical entries are not identical 18:42
because one of them could well be a trojan horse if that were true 18:43
it's a big security issue, as well as a know-what-language-you're-running issue
moritz why does "let's sign our packages" immediately comes to mind?
tony-o lol 18:44
18:44 rurban_ left
moritz speaking of which, I recently added instructions to the rakudo release guide that we at least sign our release tags 18:44
TimToady moritz++
moritz next up is probably signing the tarballs
and doing a key signing party at the next big p6 hackathon, with all the hilarity that involves 18:45
18:45 spider-mario joined
tony-o i don't know, it doesn't seem like a bad idea to have a package manager defer to another authority for a download (or upload) and error out the user that the repo has changed from what it expected upon download 18:46
18:47 qman_ joined
tony-o i guess that gets hairy too though 18:47
moritz tony-o: think of the package manager a bit like a bittorrent client
18:47 yqt left
tony-o moritz: that was what i was trying to convey - 18:47
moritz except that it doesn't necessarily need to talk bittorrent :-) 18:48
TimToady I guess my big point is that collective authority out in the world or in a distro's package manager does not absolve us of the duty to enforce identity and the level of modules
tony-o and doing a poor job, sinus infection season
TimToady *at the level
18:49 yvan1 joined
eli-se What's the recommended way to add something to @*INC? 18:50
18:50 Aaron`J is now known as amdj
moritz eli-se: depends 18:50
eli-se: why do you need it?
eli-se in Perl 5 I do this: github.com/mill-lang/mill/blob/dev...illc#L2-L5 18:51
So I don't have to set PERL5LIB all the time while developing (and I don't want it in .zshrc).
tony-o eli-se: you can do just: use lib '../lib'
eli-se oh cool
tony-o that works in p5 too iirc 18:52
moritz yes
as long as you execute it always from the same dir
which is fine for development, usually
for for installed things, you, well, install the modules too
eli-se nice, thanks
dalek kudo/nom: 5e842d6 | moritz++ | docs/release_guide.pod:
Instructions for signing the release tarballs
tony-o didn't notice it was a bin
eli-se well with __file__ preferably actually 18:53
sometimes I run it from the VM source dir
dalek p: 2641c50 | moritz++ | docs/release_guide.pod:
Instructions for signing the release tarballs
18:55
18:57 qman___ joined
tony-o eli-se: for the __file__ thing you could: use lib "{$?FILE.IO.dirname}/../lib"; 18:59
eli-se awesome 19:00
TimToady wow, spectests get really upset if you forget to itemize circumfix:<{ }> :)
tony-o eli-se: check out the bottom of this page for some other cool variables doc.perl6.org/language/variables 19:01
moritz TimToady: I'm surprised the setting even builds 19:03
usually when I try to change *anything* about how list iterations work, setting compilation hangs
19:05 muraiki joined 19:08 qman___ left, rindolf left, yqt joined 19:09 Qman___ joined 19:11 Qman___ is now known as nurumaik 19:12 spider-mario left
dalek kudo/nom: a4fd3da | TimToady++ | src/core/List.pm:
.invert must return a flat list of pairs
19:13
rakudo/nom: 639f55d | TimToady++ | src/core/List.pm:
rakudo/nom: start undistinguishing Parcel and List
19:13 dalek left
TimToady killed dalek! 19:13
19:13 dalek joined, ChanServ sets mode: +v dalek
TimToady maybe it'll report the rest? 19:14
maybe not?
19:14 spider-mario joined
[Coke] moritz: I don't have a gpg key, so I feel I shoud drop the next release. 19:21
dalek ast: 493af7f | TimToady++ | S (3 files):
keep up with .perl and []/{} changes
19:25
19:30 telex left 19:32 telex joined
masak thinks, not for the first time, that he would like to set up a date between dalek and rosettacode.org/wiki/Find_Common_Directory_Path 19:32
TimToady [1,2,3][] and {:foo, :bar}{} look a little funny as a way to call nqp::decont though, so maybe we could recruit .<> as a standard way to decont anything 19:33
then .perl could output [1,2,3]<> and {:foo, :bar}<> instead
and we could also say $foo<> for a decont
masak is not immediately a fan 19:34
TimToady a prefix would also work, if we could find one 19:35
except that a postfix is usually where you want it 19:36
and .decont is just too ugly
and just about as uninformative as <>
masak .decont is ugly mostly because it's just a made-up word
TimToady my internal bikeshedder has yet to come up with a good short word that means that 19:37
.contents is too long
and I'm trying to Omit Needless Words in .perl output anyway
masak *nod* 19:38
TimToady that was part of the motivation for :{}
PerlJam [Coke]: is creating a gpg too onerus (or maintaining one) because you don't ever use gpg ?
*gpg key
masak I think my initial not-fan reaction to <> is because to me, <> is sugar for {}, so it makes no sense for it to sometimes mean []
TimToady at the moment we're cheating and not putting a .list on List to distinguish it from Parcel anymore
but those are converging anyway
well, S09 mixes {} and [] on the same entity anyway 19:39
but both of those Zen slices are simply implemented as nqp::decont
m: {:foo}<> 19:40
camelia ( no output )
TimToady std: {:foo}<>
camelia std 28329a7: OUTPUT«ok 00:00 136m␤»
TimToady std: <>
camelia std 28329a7: OUTPUT«5===SORRY!5===␤Unsupported use of <>; in Perl 6 please use lines() to read input,␤ or ('') to represent the null string,␤ or () to represent the empty list at /tmp/JbzV0ygVs5 line 1:␤------> 3<7⏏5>␤Parse failed␤FAILED 00:00 134m␤»
masak sometimes I'm thinking "why don't other languages appear to have these kind of problems with distinguishing container stuff from decontained stuff?"
PerlJam masak: other languages have painted themselves into different corners :)
masak I never see people fretting over these issues in the Python or Ruby communities 19:41
neither on the implementor or end user side
PerlJam: yeah, I guess that's it
most languages do not offer both = and :=
and so the distinction is not so obvious, or maybe not even there at all
geekosaur in haskell and java, boxing/unboxing comes up
TimToady it all comes from wanting push @foo, @bar; to work right :)
masak geekosaur: I'm not talking about primitives versus wrapper types 19:42
TimToady similar principle
masak geekosaur: I'm talking about values versus containers
TimToady you're talking about optional levels of indirection
most languages don't give you the choice
19:43 FROGGS[mobile] joined
masak anyway, I fully agree with the underlying goal here: making things seem effortless :) 19:43
for example in the .perl output
19:43 LonelyGM joined
TimToady has been tormenting himself the last couple days on behalf of the .perl user :) 19:44
masak :P
FROGGS[mobile] :D
TimToady wasn't expecting my @ or % to have a different .of type than [] or {}...
masak .oO( come for the compiler lore, stay for the convoluted in-jokes and in-puns based on in-memes ) :P
TimToady the road to purgatory is paved with melting snowclones 19:45
19:45 raiph left
masak read that as "melting semicolons" 19:45
come to think about it, semicolons kinds look like melting colons 19:46
TimToady sorta like semisweet chocolate
PerlJam TimToady: no circumfix for decont? (kind of as irony :)
19:46 darutoko left, raiph joined
TimToady PerlJam: thought about that, or a decoration on existing circumfixes, but decided something orthogonal was preferable 19:47
19:47 johnjohn101 left
TimToady whether the generic version turns out to be <> or something else 19:47
PerlJam The only marginally constructive thing I have to say about <> is that, like masak, I'm not a fan. (yet) 19:48
masak I could've added "yet". TimToady can be persuasive.
but I don't buy the S09 argument. :) 19:49
I don't know that this is the right quesiton to ask, but it may push us towards a better optimum: 19:50
what would Perl 5 do?
if this design question somehow turned up in Perl 5, what syntax would they arrive at?
PerlJam
.oO( the wrong thing )
masak well, yeah.
but what, in particular?
[Coke] PerlJam: I have no gpg key, correct. I suppose I could get one solely for issue rakudo releases. 19:51
Juerd You should get one, and once you have it, use it for a lot of things. But one thing is enough reason to create it. 19:52
masak isn't something like @$array_ref basically Perl 5's decont already?
but that's on variables only, I guess. not on literals like we're discussing now.
PerlJam [Coke]: I just made one. It didn't take too long (generating enough entropy seemed to take a while). But now I need to learn how to "port" it to the machines where I'm likely to use it (I assume I can just copy ~/.gnupg around) 19:53
TimToady thing is, in P5 @ guarantees you get an array, while it only guarantees Positional in P6 19:54
masak not that we use that a lot yet 19:55
lucasb m: say '!' x !$_ ~ 'foo' for True, False 19:56
camelia rakudo-moar 80afb7: OUTPUT«foo␤!foo␤»
masak lucasb: numifying Bool considered a code smell :) 19:57
PerlJam still cute though
lucasb :D
I learned this from a few commits ago
masak I think it was RandallSchwartz who said True numifying to 1 is damaging, and it should be 42 or something just so people don't do stuff like that
PerlJam masak: ->*[] could be considered a P5 decont perhaps 19:58
masak eeks
ok, forget I asked what P5 would do
PerlJam masak: and I think it was merlyn who said "Randal only has one ell"
TimToady so while we could use %() for the bare Hash, @() doesn't work for the bare Array
masak PerlJam: oops. you're right.
TimToady darn microwave wifi lag... 20:00
dalek kudo/nom: 1602e1d | coke++ | docs/release_guide.pod:
Drop release.
20:03
[Coke] we could use a release manager for April. (note the new signing requirement) 20:04
20:05 mr-foobar left
FROGGS[mobile] what does he has to sing? 20:05
masak FROGGS[mobile]: anything, as long as it's opera. 20:06
FROGGS[mobile] :S
vendethiel- so, [..][] and {..}{} "decont" now 20:08
20:08 amdj left
TimToady they have for quite some time now 20:08
m: say [1,2,3].perl; say [1,2,3][].perl
camelia rakudo-moar 80afb7: OUTPUT«[1, 2, 3]␤[1, 2, 3][]␤»
TimToady star-m: say [1,2,3].perl; say [1,2,3][].perl 20:09
camelia star-m 2015.02: OUTPUT«[1, 2, 3]␤Array.new(1, 2, 3)␤»
TimToady but the repetition of the new form is suboptimal
so it'd be nice to have a standard decont postfix of some sort
m: say 42.item[].perl 20:10
camelia rakudo-moar 80afb7: OUTPUT«42␤»
TimToady m: say 42.item.perl
camelia rakudo-moar 80afb7: OUTPUT«42␤»
TimToady heh
m: say 42.item<>.perl
camelia rakudo-moar 80afb7: OUTPUT«42␤»
TimToady hmm 20:11
m: say []<>.perl
camelia rakudo-moar 80afb7: OUTPUT«[][]␤»
TimToady heh, already works :D
that proves we should make it work...er... 20:12
20:12 raiph left
vendethiel- 's not sure what it means to be "decont". 20:12
PerlJam Right now, I don't think I'd mind an ugly, but explicit .decont
vendethiel- it means I can't "foo = x"?
and call .VAR, and so on so forth
20:13 qman_ left
El_Che TimToady: I suspect a python/java/.net conspiracy: your fosdem talk is still not online :( 20:13
20:13 kjs_ joined
vendethiel- El_Che: are you sure it wasn't lost? 20:13
PerlJam vendethiel-: it's taking your presents out of the box and throwing the box away :)
El_Che I'll mail them tomorrow to check the status
video was pretty problematic this edition
FROGGS[mobile] El_Che: Mine is also missing :/
vendethiel- El_Che: it really was :[. 100+ videos missing! 20:14
masak I agree with PerlJam about ugly-but-explicit .decont -- it seems the sanest choice right now, even if it's a bit too explicit for Perl ;)
vendethiel- does .decont warrant such a "huffmanized" postfix?
El_Che FROGGS[mobile]: sucks. Most talks were very nice
FROGGS[mobile] Also +1 for. decont from me
El_Che I saw all the perl talks :)
20:15 nurumaik left
vendethiel- TimToady: do you have other usages for postfix [] and {} in mind? 20:15
or does it look like perl6 code will regularly need decont?
vendethiel- hasn't needed it much, but might be doing something "incorrectly" 20:16
TimToady we've had Zen slices for interpolation for about 13 years now...
vendethiel-
.oO( surely you had sigil variance in perl for longer :P )
masak I do like the Zen slices. very Perl 6. 20:17
TimToady is opposed to .decont because it's too much noise in .perl, and because it's no more informative than something shorter that already works
PerlJam It's slightly more informative to the humans.
TimToady and it's quite visually distinct
20:18 Qman___ joined
FROGGS[mobile] .perl is not quite for humans 20:18
PerlJam If they have to read its output it is. 20:19
vendethiel- TimToady: it seems like I have to revisit how zen slices are useful once in a while :). (except for interpolating)
FROGGS[mobile] I also very often read HTML output 20:20
TimToady .[] is really a postfix form of @, and {} of %
jnthn TimToady: (make all of them Mu) I suspect that them coming out as Any is accidental rather than intended, so yeah, no problem wiht that. 20:24
TimToady vendethiel-: actually, tomorrow it will be 13 years and 11 months since A02 came out with the idea 20:28
vendethiel- TimToady: can I read a RFC on that, maybe?
TimToady jnthn: good, cause I already checked it in
www.perl6.org/archive/doc/design/apo/A02.html 20:29
oh wait, that was in an Update:, so it was some later than that 20:30
20:31 colomon left
TimToady but today is the 14th anniversary of A01, which does talk about sigil invariance 20:33
jnthn TimToady: heh 20:34
jnthn finally caught up with backlog at least
vendethiel- :D 20:36
dalek kudo/nom: d4fd249 | TimToady++ | src/core/ (2 files):
switch to .<> as more visually distinct decont
20:38
ast: ad98ada | TimToady++ | S32- (2 files):
fix .perl roundtrip tests that use decont
20:39
20:43 Rounin left
TimToady m: my @a; say @a.of; my $a = []; say $a.of; my %h; say %h.of; my $h = {}; say $h.of; 20:44
camelia rakudo-moar 1602e1: OUTPUT«(Mu)␤(Mu)␤(Mu)␤(Mu)␤»
TimToady yay for consistency
nap & 20:46
20:48 xfix left 20:50 Sqirrel joined
dalek ecs: 3eeab76 | PerlJam++ | S11-modules.pod:
Excise "official" wording
20:52
bartolin_ [Coke]: you said earlier we'd need tests for reporting line numbers for failing tests. I tried something with shelling out and looking of the output. since I'm not sure about that approach I didn't commit yet, but created a PR (56) for roast 20:55
21:01 kurahaupo1 joined
vendethiel- feels sad soooo many videos are missing from fosdem :( 21:02
eli-se fosdamn 21:05
[Coke] bartolin_: ISTR test::util has a runner that could be used. 21:17
masak .oO( lostdem ) 21:19
21:20 raiph joined
vendethiel- hah! 21:20
21:20 Qman___ left, FROGGS[mobile] left
bartolin_ [Coke]: do you mean "is_run"? I don't know if it's possible to run test files with that. will have to look at it -- but not today ... 21:21
21:21 fhelmberger joined 21:26 fhelmberger left 21:33 skids left, kurahaupo1 left 21:34 eternaleye left, hoelzro left 21:35 eternaleye joined
literal is there something up-to-date that has info similar to rakudo.git/docs/compiler_overview.pod ? that one doesn't mention MoarVM at all 21:35
eli-se Is there a more direct Perl 6 equivalent of shell {a..z}{a..z} other than $_ X~ $_ given 'a'..'z'? 21:36
21:37 mattp__ joined
eli-se (order is insignificant) 21:37
21:37 raydiak left, rhr_ joined 21:38 xinming_ joined, hobbs left, mattp_ left, xinming left, pnu left, rhr left, f3ew left 21:39 f3ew joined, pnu joined, raydiak joined 21:40 hobbs joined
tony-o you want a random a-z length 2 string? 21:42
m: (aa..zz).pick(1).say;
camelia rakudo-moar d4fd24: OUTPUT«5===SORRY!5=== Error while compiling /tmp/Ymm3Wn4H5O␤Undeclared routines:␤ aa used at line 1␤ zz used at line 1␤␤»
tony-o m: ('aa'..'zz').pick(1).say;
camelia rakudo-moar d4fd24: OUTPUT«xk␤»
tony-o m: ('aa'..'zz').pick(1).say; 21:43
camelia rakudo-moar d4fd24: OUTPUT«bs␤»
tony-o bs indeed perlbot
21:43 hoelzro joined
literal m: ('a'..'z').combinations(2).perl 21:43
camelia ( no output )
literal output too long?
eli-se: anyway, that will give you all 2-element combinations
tony-o literal: you didn't "say" it 21:44
literal oh
m: say ('a'..'z').combinations(2).perl
camelia rakudo-moar d4fd24: OUTPUT«(["a", "b"], ["a", "c"], ["a", "d"], ["a", "e"], ["a", "f"], ["a", "g"], ["a", "h"], ["a", "i"], ["a", "j"], ["a", "k"], ["a", "l"], ["a", "m"], ["a", "n"], ["a", "o"], ["a", "p"], ["a", "q"], ["a", "r"], ["a", "s"], ["a", "t"], ["a", "u"], ["a", "v"], ["a…»
eli-se cool
tony-o if you want just one you can .pick
or if you wanted 5 you could .pick(5)
eli-se awesome! 21:46
literal hm, actually, combinations() won't give you aa and bb, etc, though
so tony-o's solution sounds better 21:47
b2gills m: say ('a'..'z').combinations(2)>>.join.pick(5); 21:52
camelia rakudo-moar d4fd24: OUTPUT«px lz tv dx ho␤»
21:52 eli-se left
tony-o m: (('a'..'z') x (0..9)).pick(2).say 21:53
camelia rakudo-moar d4fd24: OUTPUT«a b c d e f g h i j k l m n o p q r s t u v w x y za b c d e f g h i j k l m n o p q r s t u v w x y za b c d e f g h i j k l m n o p q r s t u v w x y za b c d e f g h i j k l m n o p q r s t u v w x y za b c d e f g h i j k l m n o p q r s t u v w x y za…»
21:54 muraiki left 21:56 colomon joined 21:58 eli-se joined
lucasb m: say ('a'..'z')[0..5].perl 22:01
camelia rakudo-moar d4fd24: OUTPUT«("a", "b", "c", "d", "e", "f")␤»
lucasb ^^ What is a more idiomatic way of getting just the first N elements from an list/array?
Also, how can I do a take, drop, take-while, drop-while in Perl 6?
literal m: say ('a'..'z')[^5] 22:02
camelia rakudo-moar d4fd24: OUTPUT«a b c d e␤»
lucasb literal: oh, that's better 22:03
masak lucasb: yeah, `.[^5]` is our `take 5`
I guess `.[5..*]` would be our `drop 5`
`take-while` and `drop-while` are just a function definition away, each. 22:04
lucasb m: say ('a'..'z')[20..*]
camelia rakudo-moar d4fd24: OUTPUT«u v w x y z␤»
lucasb literal, masak: thanks! 22:05
masak m: sub take-while(&cond, @list) { @list.grep: { &cond($_) ?? True !! last } }; say take-while * < 10, 3..Inf 22:07
camelia rakudo-moar d4fd24: OUTPUT«3 4 5 6 7 8 9␤»
masak seems to work.
funny, my mind designed it with .map, but my fingers wrote it with .grep :) 22:08
lucasb that "last" will shortcut the grep? 22:09
I guess I mean "short-circuit" 22:10
:)
22:13 geever left
masak yeah, a .grep is implemented with a loop, which responds to a `last`. 22:13
timotimo so, are these the first commits towards GLR? 22:22
22:25 raiph left 22:27 raiph joined 22:38 BenGoldberg joined 22:44 lucasb left 22:47 Foxcool left 22:48 RabidGravy left
eli-se gonna write PCRE library for Perl 6 22:51
22:52 yqt left
tony-o you can use p5 regex in p6 22:52
eli-se omg
how
masak one engine, two syntaxes. 22:53
tony-o m: say "a" ~~ :p5/a/;
camelia rakudo-moar d4fd24: OUTPUT«5===SORRY!5=== Error while compiling /tmp/Vz2X3uL27G␤Missing required term after infix␤at /tmp/Vz2X3uL27G:1␤------> 3say "a" ~~ :p5/a/7⏏5;␤ expecting any of:␤ prefix␤ term␤»
eli-se absolutely terrific! 22:54
masak the adverb needs to go on something
Juerd That's a relief. I can still pester everyone with unreadable regexes in Perl 6! :-)
literal m: say "a" ~~ m:p5/a/;
camelia rakudo-moar d4fd24: OUTPUT«5===SORRY!5=== Error while compiling /tmp/KKeZYNtSG5␤Adverb p5 not allowed on m␤at /tmp/KKeZYNtSG5:1␤------> 3say "a" ~~ m:p5/a/7⏏5;␤»
tony-o m: say "a" ~~ /a/:p5;
camelia rakudo-moar d4fd24: OUTPUT«5===SORRY!5=== Error while compiling /tmp/3bbfrIKeeL␤You can't adverb that␤at /tmp/3bbfrIKeeL:1␤------> 3say "a" ~~ /a/:p57⏏5;␤ expecting any of:␤ pair value␤»
masak m: say "a" ~~ rx:p5/a/
camelia rakudo-moar d4fd24: OUTPUT«5===SORRY!5=== Error while compiling /tmp/WURr9EMVw9␤Adverb p5 not allowed on rx␤at /tmp/WURr9EMVw9:1␤------> 3say "a" ~~ rx:p5/a/7⏏5<EOL>␤»
masak huh.
Juerd Who doesn't love a nice nested stack of (?>...)
tony-o it's something like that
masak m: say "a" ~~ /:p5 a/
literal m: say "a" ~~ m:P5/a/;
camelia rakudo-moar d4fd24: OUTPUT«5===SORRY!5=== Error while compiling /tmp/m102tZTNjD␤Unrecognized regex modifier :p␤at /tmp/m102tZTNjD:1␤------> 3say "a" ~~ /:p7⏏055 a/␤»
rakudo-moar d4fd24: OUTPUT«「a」␤»
Juerd Silly colons, you don't nest! :P 22:55
tony-o m: say 'a' ~~ m:P5/./;
camelia rakudo-moar d4fd24: OUTPUT«「a」␤»
masak oh, capital P
tony-o duh
Juerd masak: Yes, small p would be perl, the executable. It's just the language, so that's Perl... (Probably not the real reason, but it fits.)
BenGoldberg p56: say 'a' =~ /./;
yoleaux 30 Mar 2015 02:23 EDT <TimToady> BenGoldberg: see irclog.perlgeek.de/perl6/2015-03-30#i_10360515 and following
camelia p5-to-p6 : OUTPUT«say('a' ~~ m:P5!.!)␤»
masak Juerd: I buy that. 22:56
eli-se m: say '#(foo #(bar) baz)' ~~ m:P5:s/[ \n]|#(\((?:(?-1)|.)*?\))|#.*?\n/
tony-o the perl6 regex is much nicer after having forced myself to use it for a while
camelia rakudo-moar d4fd24: OUTPUT«「#(foo #(bar)」␤ 0 => 「(foo #(bar)」␤»
Juerd m: say "foo".Perl; # bug!
camelia rakudo-moar d4fd24: OUTPUT«No such method 'Perl' for invocant of type 'Str'␤ in block <unit> at /tmp/2qvds03oC9:1␤␤»
eli-se meh
m: '#(foo #(bar) baz)' ~~ m:P5:s/[ \n]|#(\((?:(?-1)|.)*?\))|#.*?\n/; say $<>
camelia rakudo-moar d4fd24: OUTPUT«「#(foo #(bar)」␤ 0 => 「(foo #(bar)」␤»
Juerd looks up (? ...) 22:57
Argh
(?-...)
eli-se recursive backreference
to relative group -1
Juerd Ah
BenGoldberg TimToady++
.ping 22:58
yoleaux There is no ping command; nor can this be construed as a response.
Juerd I think I've even used it once.
BenGoldberg .tell TimToady TimToady++
yoleaux BenGoldberg: I'll pass your message to TimToady.
eli-se Can you use all Perl 5 modifiers with m:P5? such as /s and /x and /g
tony-o iirc you have to :x:g them 22:59
eli-se nice
tony-o m: say "a" ~~ m:P5:g/./;
camelia rakudo-moar d4fd24: OUTPUT«「a」␤»
Juerd eli-se: I don't know if you can use them as modifiers, but if not, there's (?...:...) for most of them. 23:01
eli-se right
23:04 Sqirrel left
eli-se brb; gonna sleep 23:06
23:07 Sqirrel joined
jdv79 p5 p6 perf diff on my parse heavy code seems more like 3 orders of mag now. did something slow down? p5 is ~44ms and p6 is ~3.2s 23:08
23:08 eli-se left
tony-o [Tux]: what kind of feedback are you looking for on the csv bench? 23:09
23:11 Sqirrel left 23:12 Sqirrel joined, skids joined 23:13 raiph left 23:16 kjs_ left 23:21 pecastro left
tony-o jdv79: ouch 23:22
jdv79 i'm waiting for the glr. idk. 23:24
:(
23:25 KCL_ joined
tony-o glr? 23:25
jdv79 most of the time seems to be spent in listy things
tony-o oh
jdv79 gimme, reify, etc...
the code is pretty simple
my code i mean
timotimo jdv79: we've recently lost inlining of many blocks at the moarvm level, but it shouldn't have caused a gigantic slowdown 23:26
skids is the SVG ecosystem module panda install broken for other people, or just me?
timotimo goes towards bed 23:27
jdv79 ok
23:27 KCL left
tony-o panda install SVG 23:34
oops
skids: built and installed for me
skids Hrm. I get a "no STable" and it tries to blame XML::Writer for being old, but panda install XML::Writer works fine, and doing so does not fix things. 23:39
tony-o what does your $PATH look like 23:40
skids Oh, I;ve got rakudobrew's path in there, forgot I was in that subshell, retrying. 23:41
nope same result. 23:42
23:43 hobbs left
skids Hrm maybe I have to get a new panda into my local install dir. 23:43
tony-o yea you might have a manual perl6 and rakudobrew panda or other way around from your path 23:44
which perl6 ; which panda ;
that might give some hints 23:45
skids Yeah panda is weeks old. Because rakudobrew doesn't support installing it anywhere but in-tree
masak 'night, #perl6 23:46
skids o/
23:49 Sqirrel left 23:56 aborazmeh joined, aborazmeh left, aborazmeh joined 23:58 raiph joined 23:59 Sqirrel joined