»ö« Welcome to Perl 6! | perl6.org/ | evalbot usage: 'perl6: say 3;' or rakudo:, niecza:, std:, or /msg camelia perl6: ... | irclog: irc.perl6.org | UTF-8 is our friend!
Set by sorear on 25 June 2013.
TimToady and the Victorians would have been shocked to see a 'leg' operator :) 00:00
00:01 rurban1 left
TimToady (but then they even covered the legs of their pianos) 00:02
(and invented the terms 'light meat' and 'dark meat' so as not to be forced into discussing legs and thighs, or even, horrors, breasts) 00:03
linguists kinda hate the concept of taboo words... 00:04
well, the concept is fun, but the practice is evil
dalek rl6-roast-data: f05ea48 | coke++ | / (4 files):
today (automated commit)
 00:12
00:28 ssutch left 00:38 betterworld left, ivanshmakov left 00:39 ivanshmakov joined 00:40 slavik joined 00:43 betterworld joined 00:45 bloonix left 00:48 cognominal left, cognominal joined 00:50 dayangkun joined 00:57 clsn_ joined
clsn_ perl6: say ("Z" ~~ /<[A..MZ]>/).perl 00:59
camelia rakudo-parrot 0bf3de, rakudo-jvm 0bf3de: OUTPUT«Nil␤»
..niecza v24-108-g17d73e4: OUTPUT«(timeout)[auto-compiling setting]␤»
TimToady huh, looks like a rakudobug 01:00
clsn_ That should be a match though, shouldn't it? It used to match in older perl6s I think.
TimToady r: say ('Z' ~~ /<[ Z A..M ]>/).perl 01:01
camelia rakudo-parrot 0bf3de, rakudo-jvm 0bf3de: OUTPUT«Nil␤»
TimToady r: say ('Z' ~~ /<[ Z A..M ]>/).WHAT
camelia rakudo-parrot 0bf3de, rakudo-jvm 0bf3de: OUTPUT«Nil␤»
TimToady r: say ('Z' ~~ /<[ Z A..Z ]>/)
colomon niecza> say ("Z" ~~ /<[A..MZ]>/).perl
#<match from(0) to(1) text(Z) pos([].list) named({}.hash)>
(locally)
camelia rakudo-parrot 0bf3de, rakudo-jvm 0bf3de: OUTPUT«｢Z｣␤␤»
TimToady r: say ('Z' ~~ /<[ N..Z A..M ]>/)
camelia rakudo-parrot 0bf3de, rakudo-jvm 0bf3de: OUTPUT«｢Z｣␤␤»
TimToady r: say ('Z' ~~ /<[ Z..Z A..M ]>/) 01:02
camelia rakudo-parrot 0bf3de, rakudo-jvm 0bf3de: OUTPUT«｢Z｣␤␤»
clsn_ Doesn't like mixing ranges with straight chars I think.
TimToady looks like the range version is discarding the singletons
r: say ('Z' ~~ /<[ A..M Z..Z ]>/)
camelia rakudo-parrot 0bf3de, rakudo-jvm 0bf3de: OUTPUT«｢Z｣␤␤»
TimToady r: say ('Y' ~~ /<[ A..M Y ]>/) 01:03
camelia rakudo-parrot 0bf3de, rakudo-jvm 0bf3de: OUTPUT«Nil␤»
01:03 beastd left
TimToady so feel free to report to rakudobug, or fire up the masakbot to do so :) 01:04
clsn_ [email@hidden.address] 01:05
TimToady [email@hidden.address] I believe 01:06
clsn_ Thanks.
colomon just realized he would almost rather try to fix Rakudo bugs than report them.
TimToady clsn_, meet colomon. colomon, meet clsn_. :)
colomon clsn_: howdy 01:07
TimToady a match made in limbo, as it were
clsn_ Hello!
colomon does not have time to tackle another Rakudo bug tonight, however. ;)
TimToady oh, it might just be [email@hidden.address] 01:09
lue This is another opportunity to bring up RT's new web-based submission feature. :) 01:10
clsn_ Yeah, found it. Thanks.
01:11 raiph joined
TimToady lue: been poking through your S15ish, and don't like the notion that Str has a .storage 01:12
colomon lue! \o
TimToady thinks Str just means NFG always, and Uni is codepoints, and we have real types for NFD, NFC, etc, not fake types under Str 01:13
01:13 thou left
lue TimToady: tbh I came up with it when I realized that I was writing a bunch of functions that needed to know the storage of the Str, but the user themself couldn't access this :) 01:13
TimToady we shouldn't reinvent the type system poorly 01:14
lue So, Str is NFG, Uni is one of the other non-NFG forms, and NFD, NFC, etc. are more specific forms of Uni?
(i.e. NFC ~~ Uni[NFC] ?) 01:15
TimToady I'd say your functions are not sufficiently polymorphic instead
yeah
much like utf8 is a specific form of blob8
the polymorphism is over the Stringy role, not the Str type 01:16
is how I've been imagining it
lue So Stringy[NFG], Stringy[NFC], etc.?
TimToady well, maybe Uni is a role too
but yeah, basically 01:17
lue
.oO(Stringy[NFC] ~~ Uni[NFC] ~~ NFC)
TimToady: I'd like to think that Uni is for Unicode storage, and Stringy is a superset of that and other user-defined things, so Stringy[EBCDIC] but not Uni[EBCDIC] 01:18
TimToady in any case, NFC ~~ Uni and NFD ~~ Uni should be true
01:18 telex left
TimToady mebbe 01:18
01:18 japhb_ joined
TimToady in a sense, Str is also outside of Uni because NFG is not defined by the consortium 01:18
so I think Str ~~ Uni is false 01:19
and almost everything we get in from other languages will be some kind of Uni
01:20 telex joined
[Coke] gets to sit in on a lovely rehearsal of christmas music at a real theatre. 01:20
TimToady encoded in some kind of UTF
[Coke] nice backdrop for coding. :)
lue TimToady: I'm starting to see how locale-aware string types could form possibly. Things like Turkish dotted/dotless I things would be handled by it.
colomon is watching the Big Ten Championship tonight, just three days late.
TimToady we've always seen language-aware strings as a level higher than Str even 01:21
since it can change the notion of graphemes even further than Str does 01:22
language-aware strings might even use their own synthetics tables in preference to NFG's 01:23
lue
.oO(class ẞStr is Str ...)
TimToady and maybe failover to NFG's for things outside the language
lue if those l10n'd types inherited from Str, I think it would.
TimToady but we'll be doing good to get to the NFG level first 01:24
TimToady is not sure inheritance is right
lue Yeah, I imagine the l10n stuff should be handled in modules somehow, like there are perhaps too many locales to account for :)
TimToady NFG is a bit of a native type, so inheriting from it can't change the memory layout if so 01:25
inheritance of native types being restricted that way
but they can at least cargo-cult the ideas from NFG, even if the relationship is not one of inheritance 01:26
so FrenchStr ~~ Stringy, but maybe not FrenchStr ~~ Str 01:27
mabye the role of class Uni is really Unicodey 01:28
lue Depends on what FrenchStr does to accomplish its ideas of characters, methinks. FrenchStr removing features makes FS ~~ Str wrong, but just adding features should allow FS ~~ Str
TimToady well, if you have to substitute different tables, it's just something else 01:29
lue
.oO( role Uny )
I'm kinda stuck on how an l10n Str would modify various rules, e.g. change "ß".uc -> "SS" to "ß".uc -> "ẞ". Changing the .uc method wouldn't be enough, because /:i ß/ 01:31
lue imagines the l10n Str would intercept any queries of Unicode properties and send back different values when appropriate, kinda like a Gameshark 01:32
TimToady thinks this will be for every language to work out, and we only need a vague notion of keeping the types separate for now
delegation may be the answer here
lue TimToady: Yeah, I'm more concerned with what P6 needs to provide to let language strings do that stuff than with how those Strs would go about it specifically. 01:33
TimToady not inventing a second-class type system is the most important thing in my estimation
lue
.oO(I'll quickly add a waitaminute note to my gist, I fear I haven't quite enough time to switch to S15-editing mode ☺)
 01:34
TimToady there isn't a LocaleStr class, since such a thing is both impossible and unneeded 01:35
we don't need a type that is generic across all languages
most people only speak a few languages locally
thanks; we should probably move this over to a real S15 soonish 01:36
lue TimToady: you're welcome. And I kinda want to write a blog post about my S15ish soon, before moving S15 to official. However, moving it first doesn't break any masterplans :) 01:37
lue might not bother converting the Pod6 file to POD, because we're Perl *6* ;) 01:38
TimToady all of the synopses started out rough, so there's no shame in checking something into git and then changing it
we supposedly have the Pod6 technology already
if anything we should be converting the others to it 01:39
lue I'd like to start a Pod6ify branch soon, probably after moving S15ish.
diakopter we have the technology
lue In any case, 30-odd Pod6 Synopses are likely to attract more rendering improvements than 1 or 2 Pod6 Synopses :)
TimToady dogfood, and all that 01:40
timotimo gmake[1]: *** read jobs pipe: Invalid argument. Stop. - what does that even mean!?
lue When I have time in ~1h, I'll add the Str, Uni, etc. stuff into S15ish accordingly, and then possibly move it. I've been waiting until it gets to the point where an iron grip on the contents of the spec becomes detrimental rather than beneficial.
01:40 SevenWolf joined
TimToady timotimo: it means, "Then don't do that." :P 01:41
iron grips are overrated
timotimo yeah, but ... all i'm doing is "make all install" :( 01:42
lue
.oO(maybe just make some of the installs then.) Have you tried running make and make install separately?
 01:43
timotimo let's see ...
okh! 01:44
forgot to checkout the subbmodules i bet!
nah
01:49 jeffreykegler joined
lue can't wait to be able to edit the synopses without pulling up perlpod and making sure even the most basic tenets of Pod6 are from POD 01:52
raydiak this talk of pod reminds me of a vaguely related question...is there anything particular that a person with web development proficiency could be doing to further Perl 6? 01:53
jeffreykegler Re NFG: I have wondered if anyone but me noticed the pun. 01:54
TimToady wonder no more :)
jeffreykegler Although it is a very obvious followon to NFC, NFD, etc., so the question is raised why it is not in the Unicode standards. :-)
timotimo raydiak: you can try to figure out how to work PSGI with our rakudo-jvm and make it do requests in parallel :)
you can write a tutorial on how to deploy a Bailador application on heroku 01:55
i wanted to do that once, but i was overwhelmed by not having a single clue what anything was ... ever
jeffreykegler In any case, as followers of the Unix tradition, fearlessness in the face of treacherous puns is our heritage.
timotimo i don't understand the NFG pun :(
TimToady jeffreykegler: well, they're trying, but when you start writing out longer and longer phrases for what should be simpler concepts, it slows you down :) 01:56
01:56 dayangkun left
TimToady timotimo: No F*cking Good in the vernacular 01:56
an intensive form of NG, for No Good
timotimo i see! 01:57
that's amusing indeed :)
TimToady I think I first heard "That's NG!" maybe three or four decades ago
but maybe it predates that 01:58
raydiak timotimo: thanks, I made a note for a more long-term list :) I guess I was thinking more bite-sized to start, like P6-related websites that need help, or pod6-to-html tools that aren't done
TimToady does not recommend that anyone trust his memories in this regard...
timotimo ah
[Coke] next F***ing generation
raydiak iow, lots of people here know perl and perl 6 as well or better than I, but I think my html and css might be a useful addition to the talent pool 01:59
(if needed anywhere)
jeffreykegler Helmut Wollmersdorfer on the Marpa mailing list told me that the Unicode standards call the rough equivalent of NFG, Grapheme Cluster boundaries extended ... 02:00
timotimo gcbe? doesn't sound as nice as nfg
jeffreykegler which is not a great name, but is pretty pun-proof
[Coke] my lon gterm goal is to get a mojo 6 running on the jvm.
02:06 raiph left 02:07 jeffreykegler left
raydiak github.com/perl6/doc/blob/master/README says "write a tool that renders the documentation as HTML" ...what's the status of that? 02:08
02:08 aindilis` left, raiph joined, jeffreykegler joined 02:11 aindilis` joined
[Coke] raydiak: github.com/perl6/Pod-To-HTML/, maybe? 02:11
raydiak thanks [Coke]++ 02:13
02:19 raiph left
raydiak I suppose basic html and css are very common skills, and more advanced stuff like being able to vertically center variable-length content and stretch columns to equal height without javascript isn't very important to perl 6 in any immediately obvious way...*shrug* 02:23
02:25 logie_ left 02:27 raiph joined 02:28 benabik joined 02:34 stevan_ left 02:38 ponbiki_ is now known as ponbiki 02:39 salv0 left 02:40 btyler joined 02:41 logie joined, salv0 joined 02:42 raiph left 02:59 jeffreykegler left 03:03 raiph joined 03:05 jeff_s1 left 03:08 dayangkun joined 03:15 araujo left 03:16 araujo joined
lue TimToady: You know, if we separate various string storages into various types, perhaps String.length could be reintroduced :P 03:33
timotimo so ... compiling on my laptop isn't exactly super fast ... but it's definitely better than on my old one 03:35
TimToady lue: 'length' is too long :P 03:38
lue How about .ln then ... wait. 03:39
timotimo gonna time my stage parse to figure out where my laptop stands in comparison to my desktop 03:40
lue TimToady: if Uni is a role, would you prefer enums to define the NFD, NFC, etc. classes, e.g. Uni[NF::C] or something, or would those NF* classes simply C<does> a non-parametric Uni 03:41
(or is Uni a synonym for NFC, and there's a higher Unicodey role?) 03:42
TimToady I suspect it's non-parametric 03:44
Uni doesn't know whether it has NFC or NFD
or some mixture
it's just a list of codepoints
much like our current Str type
03:47 thou joined 03:48 xinming joined
lue And then the NF* types decide which codepoints to convert to the local religion. The last thing I'm wondering if Uni as a class eqv. to NFC is a useful "I don't care about encoding, just Unicode please" synonym, or if worrying about using Unicode loses you the right to not care :) 03:48
(i.e. should we or shouldn't we make Uni === NFC, and spell Uni the role as "Unicodey") 03:49
TimToady no, we need a physical type for mixtures/unknowns
lue so Uni is that 'any semantics' you occasionally mention, a "freedom of religion" type, to extend the awkward metaphor :) 03:50
timotimo huh, jvm stage parse takes 200 seconds?
TimToady Uni is the closes we get to a round-trippable Unicode type that isn't a UTF
timotimo i must be doing something terribly wrong
TimToady Uni is a you-get-what-you-paid-for type, and you chose not to pay for normalization 03:51
03:52 thou left
lue
.oO(too bad NF* isn't a valid type name :P)
 03:52
TimToady No F*ing Idea :)
lue So like I said, "freedom of religion", i.e. You can have a􏿽xCC􏿽x88 in one part of the text, and ä in another. The Uni village doesn't care. 03:53
diakopter so doge
well, lots of the villagers do care 03:54
TimToady what we want to do, however, is to encourage people to handle normalization at the same spot that we're already checking for UTF decoding, so we don't have to make two passes on the data
the point at which we already have a codepoint in the register is the right point to check to see if it's a Mark character 03:55
so going straight from, say UTF-8 to NFG is going to be much more efficient than from UTF-8 to Uni to NFC to NFG
and that's where we want to be insane about efficiency 03:56
and design the language so that direct translation is default
and we don't want the overhead of plunging down multiple layers of function call on every such character translation either 03:57
it's a very good place to cheat
lue [ Uni ~~ all(NFC,NFD,NFKC,NFKD), Str does Stringy, Stringy does Unicodey(?), Uni does Unicodey, NF* is Uni is what I have in my head right now. ] 03:58
TimToady Stringy does not do Unicodey
Unicodey might do Stringy
lue TimToady: so ideally you'd care about NF when doing Uni $data = slurp($filename, :encoding<utf-8>, :nf<c>), and then not bother anymore.
TimToady in that case the real type of the returned string should be NFC, not Uni 03:59
lue is specifically excluding an NFG type, because Str is that already.
TimToady that's fine, since the name NFG is NFG. :) 04:00
timotimo did i get a ultra-low-voltage cpu in my laptop by accident? >_<
lue We could s:g/Str/NFG/, but not unless you want to watch how many tests fail the next day :)
TimToady NFG is not a user-friendly name :)
people think they know what "strings" are, so we are taking psychological advantage of that thinkage 04:01
lue TimToady: I'd like to be able to do Uni $uni = $NFCtypedStringThing , as a way of saying "I have this NFC string, but I don't care about that anymore when putting it into $uni" 04:02
TimToady well, Uni will pass the type check, if that's all you want, but the object itself will still know it's NFC
lue (aka. Uni ⊆ NFC, NFD, NFKC, NFKD) 04:03
eh, s/⊆/⊇/ 04:04
TimToady well, not sure how sets come into it really, but I take your meaning
lue I wasn't using them as Perl 6 ops, but convenient shorthand :) 04:06
timotimo my desktop does stage parse 3x faster :\
tomorrow i'll measure my old laptop against that, and it better be even slower.
TimToady we'll have to decide what NFC ~ NFD does
one can argue for a Uni result, or an NFC result 04:07
lue TimToady: my gut reaction says "that amorphous Uni type"
Though I can see NFC. And maaaaybe NFD too.
TimToady yes, but then you probably force someone downstream to renormalize a substring that's already NFC 04:08
unless the bits of rope keep different normalization info internally
lue suspects one of the NFK* types should always be preferred when present, because those things seem to be special-use-ish. 04:09
04:09 salv0 left
TimToady so concat really results in strand(NFC),strnad(NFD), and then a normalization can skip normalizing the strands that are right already 04:09
04:09 preflex left
TimToady alternately, we just blow up on type mismatch, and force the user to be explicit 04:09
lue I can see either Uni or Unicodey being ropes. I can also see refusing to play nice :) . 04:10
TimToady welcome to the world of language design :)
the trick is knowing when to be sure of things, and not being sure before then...
04:11 preflex joined, ChanServ sets mode: +v preflex
lue NF* does Strandy, Uni does Ropey (formery Unicodey), Str does Stringy. :) 04:11
TimToady we need to see how diakopter++'s ropes implementation works out in practice
lue in MVM or Rakudo?
TimToady MVM, so far, but not yet merged 04:12
04:12 salv0 joined
TimToady (that I know of) 04:12
lue I suspect we'll have a web of String-based types as complex as the Number-based types before we're done :D 04:14
TimToady now imagines D~ and C~ operators
mostly for [D~] @list 04:15
lue oh yeah, this makes the ~ prefix a bit interesting too.
TimToady but that's probably a metaop two far.
lue I think the preliminary decision should be "we don't allow that kind of mismatch", though that's not a very good final decision. 04:17
TimToady well, one wants to see what the workarounds are in each case; picking left type means we see [~] ''.NFC, @list 04:18
lue (I see NFG > all other NF* and NFK. > NF. , but getting those two relations to mesh is a bit weird atm) 04:19
TimToady picking no type means we see [~] @list».NFC
lue ».NFC feels less hacky to me
TimToady is really not all that inclined to make life easy for codepointers though :) 04:20
TimToady has much more sympathy with the graphemers
lue I'm trying to recall what I know of type promotion in numberland, but the waters are muddied with separate types that are on equal footing :/
TimToady so yeah, the » feels cleaner 04:21
lue at least, I don't think I've seen any sort of [<op>] <constant>.<meth> @list in the wild before, but I've seen plenty of [<op>] @list».<meth> 04:22
TimToady we could, I suppose, have special listops, nfc @list and nfd @list that concat
those might be frequent enough and readabler enough to warrant existence 04:23
04:24 ribasushi left, Vlavv_ left
lue Could be. Too bad we don't know how often that would be needed :) 04:24
04:25 Vlavv_ joined, ribasushi joined
lue r: say '[~] @list».NFC'.chars - 'nfc @list'.chars; 04:25
camelia rakudo-parrot 0bf3de, rakudo-jvm 0bf3de: OUTPUT«5␤»
lue Yeah, I'm thinking less and less that you can say with certainty which of NFC and NFD are preferred, and our options are: 04:27
1) prefer LHS
2) Go to Uni 04:28
3) die
(how Uni stores its amorphous collection of variously-NF'd parts is up for later discussion methinks) 04:29
04:30 wbill left
TimToady anime & 04:31
04:44 Woodi left
lue
.oO(A lot of the specifics in my S15 needs changing to fit the new multi-type setup we've sketched out.)
 04:44
TimToady: on the mismatch thing, what about $nfdStr ~= "Not NFD string, but Str string" ? Would that be solved by prefer-LHS? Or would separating rules between ~= and ~ be a viable option? 05:04
05:08 xinming left 05:10 xinming joined
lue (or would that ~ result in a Uni, and the = part of ~= does the conversion to NFD?) 05:10
05:17 raiph left 05:18 btyler left, raiph joined 05:28 [Sno] left 05:37 xenoterracide joined
lue updates to the S15ish: gist.github.com/lue/7761244 05:43
05:43 sftp_ joined
lue The framework of multiple string types is there, but sorely lacking. I suspect moving this to the specs repo soon would be the best solution to expanding that. 05:44
05:44 sftp left 05:48 sftp_ left 05:49 sftp_ joined 06:02 xenoterracide left 06:25 xinming left 06:28 [Sno] joined 06:32 xinming joined
raiph preflex: tell raydiak questhub.io/realm/perl/quest/527bca...1d4c00004c 06:34
preflex Consider it noted.
06:37 xinming left 06:49 SamuraiJack_ joined 06:59 dmol joined 07:07 FROGGS joined 07:08 Mouq left 07:20 dmol left, kaleem joined 07:27 raiph left
FROGGS o/ 07:29
07:50 xinming joined 08:02 xinming_ joined 08:06 xinming left 08:09 zakharyas joined 08:11 dmol joined 08:15 dmol left 08:17 dmol joined 08:18 xinming joined 08:20 xinming_ left, dmol left 08:21 dmol joined 08:22 xinming left 08:23 xinming joined 08:30 xinming left 08:35 dmol left, dmol joined 08:36 dmol left 08:40 xinming joined 08:46 xinming_ joined 08:48 xinming_ left, xinming left 08:52 Mouq joined 08:56 Mouq left, xinming joined
dalek : 9b19624 | (Tobias Leich)++ | t/test_summary:
TEST_JOBS environment variable
 09:09
: 44d7af5 | (Tobias Leich)++ | / (2 files):
re-apply 27b050ef74464d845e8bd4779eba51fb0d53c040
09:12 darutoko joined 09:19 denis_boyun joined 09:26 sqirrel_ joined 09:39 daniel-s joined 09:43 dprelec joined 09:46 rindolf joined 09:48 filly left 09:51 vaclavstrachon joined 09:56 finanalyst joined 10:02 fhelmberger joined
finanalyst hi from Hong Kong. Are bugs in perl6 still being handled via Bugzilla? I found that running fmt on an Inf hangs perl6 10:03
10:03 daniel-s left
moritz finanalyst: rt, not bugzilla 10:07
finanalyst: [email@hidden.address]
finanalyst moritz: thanx. Long time been out of the conversation. Hope to rejoin soon. 10:08
moritz p: say Inf.fmt('%d')
camelia rakudo-parrot 0bf3de: OUTPUT«get_boxed_ref could not unbox for the given representation␤ in any directive:sym<d> at gen/parrot/stage2/NQPHLL.nqp:2204␤ in any !reduce at gen/parrot/stage2/QRegex.nqp:724␤ in any !cursor_pass at gen/parrot/stage2/QRegex.nqp:686␤ in any dire…»
moritz p: say 1e1.fmt('%d') 10:09
camelia rakudo-parrot 0bf3de: OUTPUT«10␤»
finanalyst p: say Inf.fmt('%f')
camelia rakudo-parrot 0bf3de: OUTPUT«(timeout)»
finanalyst p: my $x= Inf; my $y = 5; say ($x + $y); say ($x + $y).fmt('%f') 10:10
camelia rakudo-parrot 0bf3de: OUTPUT«(timeout)Inf␤» 10:11
finanalyst p: my $x= Inf; my $y = 5; say ($x + $y);
10:11 SamuraiJack_ left
camelia rakudo-parrot 0bf3de: OUTPUT«Inf␤» 10:11
FROGGS p: printf('%s', Inf)
camelia rakudo-parrot 0bf3de: OUTPUT«Inf»
FROGGS p: printf('%d', Inf)
camelia rakudo-parrot 0bf3de: OUTPUT«get_boxed_ref could not unbox for the given representation␤ in any directive:sym<d> at gen/parrot/stage2/NQPHLL.nqp:2204␤ in any !reduce at gen/parrot/stage2/QRegex.nqp:724␤ in any !cursor_pass at gen/parrot/stage2/QRegex.nqp:686␤ in any dire…»
finanalyst Inf is handled OK by rakudo, but not fmt. I think this is because fmt is a wrapper around C code, which does not handle Inf 10:12
FROGGS finanalyst: it is nqp code...
finanalyst: here github.com/perl6/nqp/blob/master/s...f.nqp#L162
finanalyst specs say its C code. 10:13
10:13 finanalyst left
FROGGS specs say something about the actual implementation? 10:14
specs/S32-setting-library/Str.pod:434:This function is mostly identical to the C library sprintf function.
10:21 dayangkun left 10:35 vaclavstrachon left 10:41 mtk left 10:42 tgt joined 10:44 mtk joined 10:46 mtk left 10:47 SevenWolf left 10:49 mtk joined
FROGGS std: multi postcircumfix:<P5[ ]>($a) { $a } 11:00
camelia std 3b262af: OUTPUT«ok 00:01 129m␤»
FROGGS p: multi postcircumfix:<P5[ ]>($a) { $a }
camelia rakudo-parrot 0bf3de: OUTPUT«===SORRY!=== Error while compiling /tmp/vEEjoGRu8F␤Cannot add tokens of category 'postcircumfix'␤at /tmp/vEEjoGRu8F:1␤------> multi postcircumfix:<P5[ ]>⏏($a) { $a }␤ expecting any of:␤ colon pai…»
FROGGS postcircumfix has methodcall precedence? 11:04
11:07 Fatalnix left 11:09 Fatalnix joined 11:20 denis_boyun left
lizmat good *, #perl6! 11:21
FROGGS hi lizmat
lizmat postcircumfix parses as sub calls nowadays
maybe std: isn't up to date in that respect? 11:22
FROGGS why? it parses it, while rakudo bails out
it is just that I need the precedence level of postcircumfix and methodcall seems appropriate
this has not much todo to what @a[1] delegates to 11:25
it is about declaring your own sub postcircumfix:<P5[ ]>( )
lizmat no? does the parser know when to generate AST's with subcalls and when with method calls? 11:26
last time I looked at that code, it didn't
FROGGS the actions know what to do, but that is not the point here
lizmat ok, then I'm misunderstanding
FROGGS p: multi postcircumfix:<P5[ ]>($a) { $a } # <--- it bails out but it should not 11:27
camelia rakudo-parrot 0bf3de: OUTPUT«===SORRY!=== Error while compiling /tmp/DIyTbKWDOd␤Cannot add tokens of category 'postcircumfix'␤at /tmp/DIyTbKWDOd:1␤------> multi postcircumfix:<P5[ ]>⏏($a) { $a } # <--- it bails out but it s␤ expectin…»
11:31 Mouq joined 11:32 denis_boyun joined
lizmat
.oO( needs more tea )
 11:33
FROGGS lizmat: btw, I am doing the method2sub in v5 now too, but to do that I need to be able to declare such a subroutine 11:40
before I augmented Any, Array and List, and added postcircumfix:<P5[ ]>( ) as methods
finally I don't have to augment these types, which is a nice extra 11:41
lizmat FROGGS++ 11:44
rindolf lizmat: hi. 11:50
lizmat hi rindolf!
rindolf lizmat: what's new?
lizmat a method for creating objects ?
:-)
I think most people are waiting for parl6-m to pass a significant amount of spectests 11:51
slightly less new: S17 and implementation thereof in perl6-j 11:52
did I mention perl6advent.wordpress.com ?
and you? 11:53
rindolf lizmat: there's some code I want to write, but I'm being lazy. 11:59
lizmat: and I solved today’s binarypuzzle.com/ .
12:00 Mouq left
lizmat nice brainteaser 12:01
12:06 aborazmeh joined
diakopter preflex: help 12:12
preflex try 'help help' or see 'list' for available commands
diakopter list
preflex: list
preflex Botsnack: [botsnack]; Cdecl: [cdecl]; 8ball: [8ball]; excuses: [excuse]; Factoid: [+, -, ., ?, delete, get, store]; Help: [help, list]; Karma: [++, --, karma, karmabot, karmatop]; Nickometer: [nickometer]; Nickr: [nickr]; PlokiRE: [re]; Seen: [seen]; Sixst: [6st, ordinal]; Tell: [ask, clear-messages, messages, tell]; Rot13: [rot13]; Quote: [be, quote, remember]; WCalc: [calc, wcalc]; Version:
[version]; XSeen: [xseen]; ZCode: [zdec, zenc]
diakopter gah
preflex: tell TimToady new syntax damieng.com/blog/2013/12/09/probabl...llustrated 12:13
preflex Consider it noted.
FROGGS diakopter: he backlogs anyway
diakopter I knew that, but this bumps it to the front of the line if he says something before finishing backlogging :p 12:15
FROGGS p: sub foo(\self) { self } # dunno if this should be bugreported 12:21
camelia rakudo-parrot 0bf3de: OUTPUT«===SORRY!=== Error while compiling /tmp/wIbQq3sIJk␤'self' used where no object is available␤at /tmp/wIbQq3sIJk:1␤------> sub foo(\self) { ⏏self } # dunno if this should be bugrepo␤ expecting any of:␤ …»
12:25 atroxaper joined 12:28 filly joined 12:31 tgt left 12:33 tgt joined 12:40 SamuraiJack_ joined 12:44 xinming left 12:46 aborazmeh left 13:06 benabik left
bbkr github.com/bbkr/GeoIPerl6/blob/mas....pm#L3-L23 - should GeoIP::City::GeoIPRecord also be mentioned in "provides" section of META.info? 13:16
FROGGS only if one ever is going to "use" it 13:18
13:19 darutoko left, dayangkun joined
FROGGS but since one could do "use GeoIP::City" and will get it exported, all is fine 13:19
bbkr: look at it like an enum, if you have a file that only contains that enum and which should be used somehow, you need to add it to provides 13:20
bbkr i thought this section purpose was to index all provided name spaces. thanks for clarification
13:20 darutoko joined
FROGGS it is more like a compilation unit lookup hint 13:21
13:22 atroxaper left, Mouq joined 13:23 atroxaper joined, rindolf left 13:27 Mouq left 13:32 atroxaper left, atroxaper joined 13:33 xinming joined
lizmat FROGGS: I thought it is more authoritative information 13:34
13:37 rindolf joined 13:43 sqirrel_ left 13:47 kaleem left
dalek rl6-roast-data: a6f9384 | coke++ | bin/rakudo.moar.sh:
Cargo cult config option from TimToady++
 13:51
13:53 kaare__ joined
[Coke] sees advent postings are going up at the beginning of the day, not the end of the day. whoops. 13:53
[Coke] voice memod a bunch of notes on the drive in which will hopefully coalesce into an appropriately-word-counted article by this evening. 13:54
timotimo oh my, my article is due soon-ish 13:55
13:59 btyler joined
timotimo FROGGS: i need to add a provides entry for all modules in my package even if they are not supposed to be used outside of it, right? because the use statement inside the distribution itself would also need to query the database? 13:59
14:03 hummeleB1 joined 14:04 raiph joined 14:05 btyler left 14:10 ajr joined, ajr is now known as ajr_ 14:16 PacoAir joined, bluescreen10 joined
FROGGS timotimo: correct 14:26
lizmat: it is because we can't guess that information 14:27
14:33 xinming_ joined
timotimo updated a few random modules 14:33
can we get an extra logo to modules.perl6.org that shows if there's a provides? like a panda with a little "S11" in the corner?
FROGGS nice idea... dunno how to do that yet 14:34
14:35 kaleem joined
timotimo where do i have to look to find the code? 14:35
i think i can make that patch
14:35 xinming left
FROGGS timotimo: github.com/perl6/modules.perl6.org/ 14:36
14:37 FROGGS left 14:40 xinming joined
timotimo surprisingly obvious :) 14:42
14:42 xinming_ left, sftp_ left
timotimo oh, this is perl5 :\ 14:42
14:42 araujo left 14:43 araujo joined
FROGGS[mobile] hehe 14:43
timotimo $projects->{$name}->{'badge_panda'} = defined $json->{'source-url'};
seems straightforward to add to that
FROGGS[mobile] yeah
the hardest thing should be the png 14:44
14:44 hummeleB1 left 14:45 kaleem left
timotimo hah, that's not that hard 14:46
i suppose it could make badge_panda an integer instead of a boolean
so ... (defined $json->{'source-url'}) + (defined $json->{'provides'} && defined $json->{'provides'}) 14:47
that doesn't look elegant :P
so maybe a ternary instead
FROGGS[mobile] *g*
Ohh, that is fine
timotimo maybe the original badge should get a "not s11 compliant" symbol of some sort, instead of adding "s11" to the pandas of the modules that have it 14:51
because we're expecting s11 to be the norm soon, and not being s11 should give a tiny warning, sort of
14:54 sftp joined 14:55 hummeleB1 joined, btyler joined
FROGGS[mobile] +1 14:58
timotimo oh, HTML::Template requires me to make it two bools 15:04
FROGGS[mobile] Why that? 15:06
dalek href="https://modules.perl6.org/s11:">modules.perl6.org/s11: d166faf | (Timo Paulssen)++ | web/ (4 files):
display conformity to the new S11

  ... by giving all other modules a red strike-through "S11"
on the panda's head.		 15:07
timotimo because there's TMPL_IF, but not TMPL_IF_== ;) 15:08
what do you think?
oh, i forgot to add the alt to the image 15:09
dalek href="https://modules.perl6.org/s11:">modules.perl6.org/s11: 56bfafe | (Timo Paulssen)++ | web/index.tmpl:
supply an alt text for the image, too.
 15:10
15:10 Mouq joined
timotimo PR'd, because i'm not certain of my p5 skills and didn't test-run it 15:11
15:15 Mouq left 15:20 araujo left 15:21 araujo joined 15:24 thou joined
timotimo FROGGS[mobile]: github.com/perl6/ecosystem/blob/master/spec.pod <- when is the right moment to update that? 15:25
putting provides into "mandatory fields"? or perhaps first in "optional fields" and then moving it over?
FROGGS[mobile] Dunno 15:26
Let's wait a few days for comments on the post... 15:27
timotimo OK :)
FROGGS[mobile] But then I'd make it mandatory and review the spec
15:29 bluescreen10 left 15:34 Fatalnix left, Fatalnix joined 15:37 xinming_ joined 15:39 xinming left 15:43 eternaleye left 15:45 SamuraiJack_ left 15:47 eternaleye joined 15:56 dprelec left 15:59 Psyche^_ joined 16:01 benabik joined 16:02 Psyche^ left 16:11 ajr joined, ajr_ left 16:12 ajr is now known as Guest7442, Guest7442 is now known as ajr_ 16:25 vytas joined 16:26 [Sno] left 16:29 FROGGS joined 16:33 xinming joined 16:35 xinming_ left 16:42 jnap joined
FROGGS timotimo: I'd propose this: froggs.de/perl6/panda_nos11.png 16:44
timotimo i like that better 16:45
16:46 zakharyas left
timotimo i'll integrate it into my pull request 16:46
dalek href="https://modules.perl6.org/s11:">modules.perl6.org/s11: 0557ce9 | (Timo Paulssen)++ | web/panda_nos11.png:
FROGGS++ design for the nos11 icon
href="https://modules.perl6.org:">modules.perl6.org: d166faf | (Timo Paulssen)++ | web/ (4 files):
display conformity to the new S11

  ... by giving all other modules a red strike-through "S11"
on the panda's head.		 16:47
href="https://modules.perl6.org:">modules.perl6.org: 56bfafe | (Timo Paulssen)++ | web/index.tmpl:
supply an alt text for the image, too.
href="https://modules.perl6.org:">modules.perl6.org: 0557ce9 | (Timo Paulssen)++ | web/panda_nos11.png:
FROGGS++ design for the nos11 icon
href="https://modules.perl6.org:">modules.perl6.org: 4806ddf | (Tobias Leich)++ | web/ (4 files):
Merge pull request #3 from perl6/s11

S11
timotimo cool :)
how long until the website gets updated? can we poke it manually?
FROGGS I don't think there could went something wrong :o) 16:48
16:50 Celelaptop left
FROGGS[mobile] No idea about the update 16:51
16:51 kaleem joined
timotimo it'll happen soon enough 16:51
16:52 Celelaptop joined
jnthn FROGGS: From backlog - I don't think Rakudo knows how to do custom post-circumfixes yet 16:55
16:58 Mouq joined 16:59 Celelaptop left, nnunley joined
FROGGS[mobile] jnthn: true, I found out that I just need to tweak v5's grammar 17:02
just so I am able to use that from the actions 17:03
17:03 Mouq left
FROGGS[mobile] rakudo can stay unchanged 17:03
17:03 denis_boyun left
jnthn btw, my two Build Stuff talks were recorded, and I signed a video disclaimer, so I guess they'll appear online sometime soon :) 17:03
timotimo cool :)
FROGGS[mobile] cool
:o) 17:04
jnthn: can I get a comment on my advent post from you?
jnthn In the meantime, slides are at jnthn.net/papers/2013-bs-invoke-dynamic.pdf and jnthn.net/papers/2013-bs-secret-life-of-gc.pdf
FROGGS[mobile] (about the implementation)
jnthn FROGGS[mobile]: yes, though I have to head out in 5 mins... 17:05
link?
FROGGS[mobile] Todays post
17:06 dmol joined
FROGGS jnthn: here perl6advent.wordpress.com/2013/12/1...g-modules/ 17:06
jnthn: read whenever you brane has some capacity :o)
FROGGS[mobile] your* 17:07
btw, hacking v5 is pretty fun atm :o) 17:09
17:09 cognominal left, rindolf left 17:11 spider-mario joined
jnthn FROGGS[mobile]: Dinner now, will read it when I return :) 17:13
bbiab
17:14 Celelaptop joined
FROGGS umm, is that valid json? github.com/timo/ADT/blob/master/META.info#L8-L9 17:14
17:17 atroxaper left
moritz looks right to me 17:19
FROGGS star: use JSON::Tiny;
camelia ( no output )
FROGGS star: use JSON::Tiny; say from-json '{ "a": 42 }' 17:20
camelia star 2013-09: OUTPUT«("a" => 42).hash␤»
FROGGS k
was it YAML that required whitespace before the colon?
moritz no, YAML doesn't require that
17:24 sqirrel_ joined
FROGGS weird 17:25
timotimo: it does not show the <s>S11</s> icon... modules.perl6.org/ 17:26
benabik Is the S11 icon a panda? I see it. 17:28
FROGGS benabik: do you see a panda for the ABC module?
benabik FROGGS: Not for ABC, but for ADT 17:29
FROGGS benabik: right, and it should show this one here github.com/perl6/modules.perl6.org..._nos11.png
(for AB)
ABC*
benabik No, I see the standard no metal here thing. 17:30
FROGGS correct, this is what I am complaining about :o)
17:32 filly left 17:33 filly joined
moritz I just watched class.coursera.org/compilers/lecture/4 17:34
there are many points I agree with, but which I haven't thought very clearly about before 17:35
class.coursera.org/compilers/lecture/preview for the table of contents
FROGGS nice!
17:41 kaleem left
dalek href="https://modules.perl6.org:">modules.perl6.org: 522594e | (Tobias Leich)++ | web/lib/P6Project/ (2 files):
this hopefully fixes displaying the no-S11 badge
 17:42
FROGGS looks like we have to wait 40 minutes 17:43
17:44 denis_boyun joined, [Sno] joined
TimToady good time to read jnthn++ slides :) 17:46
17:54 baest left
TimToady or ponder a bunch of relatively low-wattage syntax sugars for C# 17:55
FROGGS[mobile] I smiled when spotting their "using" :o) 17:58
retupmoca modules.perl6.org/panda_nos11.png --> 404 Not Found :( 18:01
lots of "Bronze Panda badge" text on the modules page, though 18:03
FROGGS retupmoca: seen it too right now 18:05
moritz FROGGS[mobile]: you have to add the new .png to @MEDALS
FROGGS[mobile]: then web/build-project-list.pl will copy it to the destination dir
retupmoca so when I try to use a specific version of a module in rakudo, is the META.info information the only thing it looks at? 18:06
or will it look at pod stuff or the class Name:ver<1.1> stuff as well? 18:07
dalek href="https://modules.perl6.org:">modules.perl6.org: 8f0deff | (Tobias Leich)++ | web/build-project-list.pl:
added panda_nos11.png to @MEDALS, moritz++ and retupmoca++
FROGGS retupmoca: it takes only the META.info into account
retupmoca: maybe this META.info will be generated for you out of information available in pod, we'll see 18:08
retupmoca ok. What does it do for modules that haven't been panda-installed?
FROGGS k, 14 minutes to wait
retupmoca (as in I'm creating something new and do a use lib 'lib'; in a repl)
FROGGS retupmoca: these will takes, whatever version they have
that is what I call "development modules" 18:09
but these should vanish over time
18:09 ajr_ left
FROGGS so only real dev modules should be in such a path some day 18:10
retupmoca right
does version "*" in the META.info do anything special? Or should I put a real number in?
FROGGS when installed it will be treated as version 0
retupmoca simple enough 18:11
FROGGS otherwise it would beat every future version of that module
because:
p: say * > 42
camelia rakudo-parrot 0bf3de: OUTPUT«WhateverCode.new()␤»
FROGGS p: say so * > 42
camelia rakudo-parrot 0bf3de: OUTPUT«WhateverCode.new()␤»
FROGGS gah
18:14 gdey joined 18:15 sqirrel joined, jnap left 18:16 sqirrel left, sqirrel joined
FROGGS p: say so Version.new('*') > v42 18:16
camelia rakudo-parrot 0bf3de: OUTPUT«===SORRY!=== Error while compiling /tmp/mgMJpCHKry␤Undeclared routine:␤ v42 used at line 1␤␤»
FROGGS p: say so Version.new('*') cmp v42 18:17
camelia rakudo-parrot 0bf3de: OUTPUT«===SORRY!=== Error while compiling /tmp/MqSO6cwkmG␤Undeclared routine:␤ v42 used at line 1␤␤»
FROGGS p: say so Version.new('*') cmp v42.0.0
camelia rakudo-parrot 0bf3de: OUTPUT«===SORRY!=== Error while compiling /tmp/8OfYJf3ghl␤Confused␤at /tmp/8OfYJf3ghl:1␤------> say so Version.new('*') cmp v42.⏏0.0␤ expecting any of:␤ argument list␤ postfix␤ dotty m…»
FROGGS p: say so Version.new('*') cmp Version.new(42)
camelia rakudo-parrot 0bf3de: OUTPUT«Default constructor for 'Version' only takes named arguments␤ in method new at gen/parrot/CORE.setting:800␤ in method new at gen/parrot/CORE.setting:795␤ in block at /tmp/B9TSe4LEQA:1␤ in any at /tmp/B9TSe4LEQA:1␤ in any at gen/parrot/s…»
FROGGS p: say Version.new('*') cmp Version.new('42') # I am going nuts 18:18
camelia rakudo-parrot 0bf3de: OUTPUT«More␤»
FROGGS it is not easy to use Version
18:22 jnap joined
FROGGS modules.perl6.org/ <--- works now! :o) 18:22
timotimo++ # that is really a nice thing :o) 18:23
18:23 beastd joined
benabik The difference between good and bad is really subtle. 18:23
18:24 sqirrel_ left
Fatalnix oh boy, who what the hell is with all of the colors 18:25
FROGGS Fatalnix: colors? what colors? 18:27
18:28 sqirrel left
[Coke] modules.perl6.org/fame-and-profit - should that show the new S11 badge? 18:31
benabik It does, for me.
timotimo hmm
it would be cool to get a daily report of ecosystem activity
FROGGS timotimo: are an adhoc update to a separate (logged) channel 18:32
timotimo "$n modules updated their master, $nb modules changed their META.info, $m modules had changes in branches, $x new modules, ..."
FROGGS [Coke]: I hope we don't need the S11 badge for long... (in fact, I hope we can drop it in january already) 18:34
18:36 darutoko left 18:39 ajr joined, ajr is now known as ajr_ 18:41 rindolf joined
Fatalnix FROGGS: heh. 18:43
[Coke] FROGGS: ok. in the meantime, I have no idea what it means.
benabik: for me, it shows the panda, not the branded panda. 18:44
timotimo you can mouse over the icon to get a short explanation
but yeah, i can write a tiny bit about it in the other page if you'd like
benabik [Coke]: The branded panda is the bad one. ('tis a bit confusing)
[Coke] benabik: yes. which is why having docs would help. 18:46
[Coke] is happy enough with the mouseovers for now.
18:47 Mouq joined, jnap1 joined, jnap left
timotimo it would be much more amusing if the sign was black on yellow and said "S21" 18:47
18:51 Mouq left 18:54 gdey left 18:55 fhelmberger_ joined
FROGGS hehe 18:56
18:57 fhelmberger left 19:03 fhelmberger_ left 19:04 fhelmberger joined 19:05 sqirrel joined 19:08 fhelmberger left 19:10 sqirrel left, sqirrel joined 19:12 sqirrel_ joined, sqirrel left 19:13 AndChat-372521 joined, sqirrel_ left
lue hello world o/ 19:14
19:15 AndChat-372521 left, ssutch joined, ssutch left
FROGGS hi lue 19:16
19:16 ssutch joined
timotimo "$foo modules lost their freshness badge today" 19:17
19:22 ssutch left 19:30 ssutch joined, Shozan left 19:32 sqirrel_ joined, SHODAN joined 19:46 jnap1 left 19:55 aindilis` left 19:57 aindilis joined 19:58 rindolf left
timotimo jnthn: questhub.io/realm/perl/quest/52a5fc...b87a000056 - can you make sense of the dramatically small difference between those measurements? 19:59
benabik timotimo: That's like 20% 20:01
timotimo yeah. i was expecting more like 80%
benabik I don't think calls are THAT inefficient compared to string operations. 20:02
timotimo the whatever code example supposedly does a clone of a WhateverCode object and calls that for each operator or . in that thingie
benabik What backend?
timotimo this is parrot 20:03
.o(no, this is patrick!)
benabik: you were the one i talked to about the find_cclass caching thing, weren't you? 20:04
benabik timotimo: Sounds familiar
timotimo is there an efficient and simple way to tell (while inside an op in parrot) if i'm being executed by the first thread?
benabik No idea. 20:05
timotimo OK
lizmat suddenly realizes she has no time to write an advent blogpost tomorrow 20:12
which means I need to write it tonight :-)
FROGGS lizmat: good that there is always a plan B :o) 20:13
timotimo i may end up with a working (but quite ugly) prototype for working chained WhateverCode some time later tonight 20:15
FROGGS working++
timotimo more like ++working actually ;)
benabik Well, working++, but ++timotimo :-) 20:16
timotimo hehe
FROGGS exactly :o)
lizmat [Coke]: I also just realized that *if* you're strapped for time for you blog post, we could do a swap 20:33
so that would give you some more time?
20:35 Mouq joined
jnthn back 20:36
lizmat jnthn o/
nwc10 post-beer?
jnthn nwc10: Yes, went to a place with 18 different Lithuanian beers on tap and some nice pub food.
jnthn looks at the FROGGS post 20:37
FROGGS :o) 20:38
20:39 Mouq left
FROGGS /o\ de.webfail.com/7fe994aa327 20:42
20:43 araujo left 20:44 benabik left, araujo joined 20:45 benabik joined
jnthn FROGGS: In the perl6 -ICompUnitRepo::Local::File:prio[10]=/home/peter/project-a:/home/peter/project-b how does the priority stuff work? 20:48
jnthn is struggling to guess what the 10 means :)
FROGGS the default priority is 0, which the CompUnitRepo::Local::Installation (aka %*CUSTOM_LIB< home site vendor perl >) have 20:49
the higher the priority the earlier they appear in the list of candidate groups 20:50
all repos of one priority are considered a group, and the best compunit wins
if there is no candidate for the first group, go the next group (== lower prio) 20:51
lizmat which effectively makes @INC a hash without order ?
FROGGS lizmat: it is not in @INC yet
lizmat or would you use the priority to insert the CompUnitRepo at the right place in INC ?
FROGGS CompUnitRepo itself keeps a @repos atm, wich could because @INC 20:52
lizmat: yes, sort of
the default is that there is a CompUnitRepo::Local::File:prio[1] and four CompUnitRepo::Local::Installation:prio[0] for home, site, vendor...
the File:: thingy points to the current install/lib/parrot/.../lib and install/lib/parrot/.../site/lib 20:53
::File thingy*
jnthn: you might want to read the diff of the eleven branch, it is not too muh to read 20:55
though, I need to add comments
lizmat #120739 # abuse rtbot
synopsebot Link: rt.perl.org/rt3//Public/Bug/Displa...?id=120739
benabik lizmat: Isn't that less an abuse and more the intended purpose? 20:56
lizmat well, I'm just interested in the link, to be copied into my blog post 20:57
not to have everybody's attention about that bug
hmmm... I guess I should just privmsg synopsebot :-) 20:58
21:00 jnap joined
FROGGS lizmat: how are these variables called? ${^TAINT}, ${^UNICODE}, ${^UTF8CACHE}, ${^UTF8LOCALE} 21:01
lizmat FROGGS: I have no idea 21:02
FROGGS k
21:02 dayangkun left 21:04 dmol left 21:06 dmol joined 21:18 sqirrel_ left 21:19 dayangkun joined 21:21 shinobicl joined
shinobicl hi! 21:21
21:21 denis_boyun left
FROGGS hi 21:21
shinobicl in modules.perl6.org/... 21:22
21:22 Mouq joined
shinobicl what does the panda icon with S11 means? 21:22
FROGGS yeah
shinobicl: S11 is about synopsis 11, and we added a requirement in the META.info for this synopsis 21:23
shinobicl is a warning or an error, or deprecated?
ahhhh ok thanks!!!
FROGGS shinobicl: perl6advent.wordpress.com/2013/12/1...g-modules/
see the last paragraph there
21:23 jnap left
FROGGS shinobicl: it is about a new way to install modules, and for making that work we need that extra bit of information in the META.info 21:24
21:27 denis_boyun joined, jnap joined
shinobicl thanks FROGGS, i've updated mi META.info :) 21:29
s/mi/my/;
lizmat timotimo: just an idea for a blogpost: something about the RaNIW, and the slides and lovely excercise material that came from it? 21:30
FROGGS shinobicl++ # thank you!
:o) 21:31
lizmat or is that something you wanted to mention in your Day 17 blog post?
21:35 raiph left
timotimo lizmat: why me? :P 21:38
did you even look at my writing? :D 21:39
21:39 raiph joined
lizmat timotimo: did I miss something? 21:42
jnthn time for some sleep...'night 21:43
lizmat gnight jnthn, sweet dreams
21:44 denis_boyun left
lizmat timotimo: I guess I thought about you because you were there and since then has done lots of internals stuff ? 21:44
21:46 spider-mario left 21:52 kivutar joined
timotimo hehe 21:59
you missed that i don't write very well :)
lizmat well, if you don't practice, it will never get better 22:00
PerlJam timotimo: you're on channel full of potential editors to help you improve the writing.
timotimo hehe 22:01
PerlJam (even if we occasionally drop a word or two :)
22:02 kaare__ left
lue
.oO( (even if we occasionally a word or two :) )
 22:03
lizmat licorice! 22:06
.oO( only understandable for the native Dutch amongst us )
 22:07
22:08 BenGoldberg joined, DrEeevil joined 22:09 bonsaikitten left 22:11 jnap left 22:16 pochi joined 22:17 labster_ joined 22:19 benabik left 22:20 pochi_ left, labster left, labster_ is now known as labster 22:22 cxreg2 is now known as cxreg, BenGoldberg left 22:24 denisboyun joined, denisboyun left 22:25 BenGoldberg joined 22:31 BenGoldberg left, Ben_Goldberg joined 22:32 Ben_Goldberg is now known as BenGoldberg
[Coke] lizmat: I was going to say you shouldn't enable me, but my boss decalred he's taking the team out for drinks and dinner this evening, which makes a date switch much more attractive. 22:32
lizmat then by all means, go out and have drinks and dinner 22:33
[Coke] I can still make one in time for tomorrow, but would appreciate the day of slack.
lizmat gives [Coke] a day of slack
[Coke] \o/
you are, as always, the best. Danke. 22:34
dalek : c77798b | (Elizabeth Mattijsen)++ | misc/perl6advent-2013/schedule:
Switching with [Coke]
 22:35
lizmat yw
:-)
22:38 ajr_ left 22:39 ssutch left
[Coke] lizmat++ 22:43
oh, crap, I also owe a blog post to TPF. 22:44
tomorrow is a twofer.
I need to start using obra++'s hiveminder again
22:45 PacoAir left 22:46 beastd left
lizmat [Coke]: I'm afraid I won't be able to give you some more slack tomorrow :-( 22:48
maybe jnthn or masak may be able to move up theirs 22:49
*their next one
work in progress so far: perl6advent.wordpress.com/?p=2040&a...eview=true
BenGoldberg Doesn't show up for me 22:51
404 not found
lizmat I guess you don't have a login
BenGoldberg Correct
lizmat gist.github.com/lizmat/7919978/raw...tfile1.txt
should be public, although not rendered 22:52
BenGoldberg Yup, works
lue r: "aaaaa" ~~ /(a)+/; for $/[0..*-1] { say $_ } # today I learned this is apparently crazy-talk behavior for repeated capture groups in the rest of the regexverse.
camelia rakudo-parrot 0bf3de, rakudo-jvm 0bf3de: OUTPUT«｢a｣␤␤｢a｣␤␤｢a｣␤␤｢a｣␤␤｢a｣␤␤»
22:53 jnap joined 22:54 ssutch joined
BenGoldberg In perl5, each successive match would replace the previous value of $1 22:57
BenGoldberg imagines extending perl5, so that a regex like that filled in @1
BenGoldberg . o O ( Too much work ... that's what perl6 is for! ) 22:58
22:59 ssutch left
lue BenGoldberg: that explains why Boost is being a jerk about doing the right thing :/ . At least there's seemingly an experimental feature thing I can turn on to get the right behavior. 23:00
lizmat: good post so far, the very last codeblock was formatted differently from the previous ones (<pre>/<code> issue methinks)
lizmat [Coke]: make it a blog post postable on both perl6 advent and TPF ?
I mean, if Hollywood can do that with cross-fertilizing TV series and cutting on production costs like that 23:01
why can't we ?
(assuming the TPF one is to be about Perl 5? )
lue: that was me testing: the <code> renders way too large 23:02
lue lizmat: for me, the <pre> renders waaaaaaaaaaay too small :) . [I still used it in my posts because it was the norm though] 23:03
(of course, setting my monospace font to the size of the normal fonts would fix that, but then the monospace font's too big for me everywhere else on the web.) 23:04
23:07 kivutar left
lue
.oO(nvm, I'd have to recompile Boost. *shakes fist at Perl 5*)
 23:08
23:15 jeffreykegler joined
timotimo oh wow 23:24
i was wondering where the hell the call to the old whatevercode came from
turns out i had my $new := QAST::Op.new( [...], $old )
yeah, that's going to add the old stuff in there, too
i was looking at *every* line in the function except that 23:25
blergh, i feel i was dumb now :| 23:26
23:26 BenGoldberg left 23:28 raiph left
TimToady lizmat: line with say %h:!exists seems to be missing a <a b> maybe? 23:34
or some such
timotimo "does the hash not exist?"
23:35 ssutch joined
lizmat argh.... <a b c> is eaten by wordpress as some invalid HTML 23:35
timotimo yeah, i think it won't if you put it in <pre> or something
what do other people use to make in-line code segments?
TimToady well, it's just one True so maybe <a> 23:36
lue perl6++ for not being dumb about repeated capture groups by the way :) 23:38
TimToady you mean perl5-- for being stupid about them :) 23:39
lizmat TimToady: fixed, thanks!
TimToady well, perl1 really
lue yeah, but p6++ for fixing it :)
TimToady it seemed a pretty obvious thing to fix, what, more than decade ago now... 23:40
lue I really wish Perl 6 regexes would start replacing P5-inspired ones real soon. Stuff like (?: ) feels incredibly wrong, for starters.
lue doesn't quite buy Boost's "repeating captures is a massive performance hit, so you have to recompile all of Boost to get them", seeing as Rakudo on Parrot can run pretty fast with this magical ability. 23:41
p: "abcabcabcabc" ~~ / (abc)+ /; say now - BEGIN now, " is just too long!"; 23:42
camelia rakudo-parrot 0bf3de: OUTPUT«Instant:1386805375.631707␤»
lue p: "abcabcabcabc" ~~ / (abc)+ /; say (now - BEGIN now), " is just too long!";
camelia rakudo-parrot 0bf3de: OUTPUT«0.0846417 is just too long!␤»
23:43 xenoterracide joined
lue TimToady: It almost makes me want to write a libp6regex, because Perl 6's regexes are *so much* better :) 23:43
23:44 xinming left, xinming joined
lizmat maybe diakopter could be persuaded to allow for Petl 6 regexen in Perl 5 code in the MoarVM perl 5 interop ? 23:45
timotimo wow, i've derped the hell out of this code
lue lizmat: that sounds like a nice idea, so long as diakopter has the tuits for it :) 23:46
Won't benefit me, but P6 regexes are too nice to keep to ourselves :)
timotimo that's true
23:46 jnap left
lizmat maybe it's on his list already, but probably on the "nice to have" list 23:46
timotimo could we also get lexical $_ and $/ and stuff? :P
lue wonders how a C++ p6regex library would implement <subrules>... 23:47
lizmat but since P5 now has installable regex engines (afaik), that could make things simpler
23:47 jnap joined
timotimo with a register that's kinda like a hash? 23:47
lue timotimo: perhaps. The "regex" part of Perl 6 regexes is dead simple; the "grammar" aspect is the interesting part :) 23:48
timotimo yeah
especially deriving and stuff
we won't be able to write them directly into the code like in p6 anyway except with a crazy amount of preprocessoring 23:49
lue I imagine the worst part about writing libp6regex would be the lack of regexes to parse regex strings :P
timotimo :D
it should, of course, be self-hosting!! ;)
lue Sure, I could use a regex library, but the C++11 one on my system is buggy atm, and as mentioned I'm underwhelmed with Boost. 23:53
lue might just do libp6regex after finishing his DLX solver in C++. 23:54
timotimo wow. i finally hvae that prototype i talked about!
lue \o/ 23:55
23:55 benabik joined, benabik left
lue (or rather, 🙌) 23:55
timotimo gist.github.com/timo/13090bda08ab4824b515
added the AST output, too 23:56
\o/
but the code is a *mess*
a royal mess
lue timotimo: I have no idea if that's good or not, aside from the lack of backtraces :) 23:57
timotimo r: say "$_, {(10 < * < 20)($_)}" for ^30
camelia rakudo-jvm 0bf3de: OUTPUT«use of uninitialized value of type Nil in numeric context␤0, True␤use of uninitialized value of type Nil in numeric context␤1, True␤use of uninitialized value of type Nil in numeric context␤2, True␤use of uninitialized value of type Nil in numer…»
..rakudo-parrot 0bf3de: OUTPUT«0, True␤1, True␤2, True␤3, True␤4, True␤5, True␤6, True␤7, True␤8, True␤9, True␤10, True␤11, True␤12, True␤13, True␤14, True␤15, True␤16, True␤17, True␤18, True␤19, True␤20, True␤21, True␤22, True␤23, True␤…»
timotimo compare the outputs :)
23:59 xenoterracide left

Powered by Cro and the Raku® Programming Language.

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

Thank you!