»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!
Set by masak on 12 May 2015.
00:04 BenGoldberg joined 00:18 tinyblak joined 00:20 telex joined 00:25 pat_js left 00:46 laouji joined
raydiak seems to me like there'd be many existing formats and protocols using MD5 which someone might want to support in their P6 script/module/whatever, unless we're so far in the future now that such things don't really exist any more...???... 00:49
skids Heck, you even still need MD2 if you want to do PKI. 00:51
raydiak heh wow 00:52
raydiak finds himself developing a geezer attiutude wrt the "don't need old stuff" perspective, in general 01:02
01:09 khisanth__ joined 01:10 khisanth_ left 01:13 yubimusubi joined
tony-o_ raydiak++ 01:19
01:27 vendethiel joined, cognominal left 01:28 cognominal joined 01:39 aborazmeh joined, aborazmeh left, aborazmeh joined 01:43 metalman joined 01:48 b2gills joined 01:49 salva joined 01:58 metalman left 02:01 chenryn joined 02:02 khisanth__ is now known as Khisanth, salva left
jdv79 why isn't lazy implemented then? 02:10
02:15 salva joined
colomon ‘coz we’re lazy 02:17
02:30 rmgk_ joined, rmgk_ is now known as rmgk, dayangkun joined, dayangkun left 02:31 Ben_Goldberg joined 02:33 dayangkun joined 02:34 BenGoldberg left 02:39 Zoffix joined 02:42 vendethiel left 02:50 chenryn left 02:55 noganex_ joined 02:57 salva left 02:59 salva joined 03:15 colomon left 03:16 chenryn joined, adu joined 03:17 colomon joined 03:53 raiph left 04:07 adu left 04:13 diana_olhovik_ joined, chenryn left 04:25 skids left 04:33 BenGoldberg_ joined 04:34 jack_rabbit joined
tony-o_ itll be implemented when you need to use it 04:36
04:36 Ben_Goldberg left 04:39 BenGoldberg_ left 04:42 cognominal left 04:43 cognominal joined 04:47 awwaiid left 04:51 flaviusb joined 04:59 awwaiid joined 05:14 chenryn joined 05:20 jack_rabbit left 05:23 diana_olhovik_ left
dalek osystem: ffaf1fa | (Sterling Hanenkamp)++ | META.list:
Add Template-Anti to the Perl6 Ecosystem
 05:30
osystem: adef775 | (Tim Smith)++ | META.list:
Merge pull request #45 from zostay/master

Add Template-Anti to the Perl6 Ecosystem
05:31 awwaiid left 05:34 kaare__ joined 05:36 szabgab joined
szabgab Hello there! The YAPC::EU organizers would like to evaluate the level of interest for hackathons, if you'd like to have a Perl 6 hackathon, please add yourself to the wiki page: act.yapc.eu/ye2015/wiki?node=Hackathons 05:37
05:38 awwaiid joined
szabgab and if you are already there, please submit your Perl 6 related talk to have a longer list than this: act.yapc.eu/ye2015/talks/tag/perl6 05:40
05:41 szabgab left 05:42 adu joined 05:45 Psyche^ joined 05:47 aborazmeh left 05:53 rindolf joined 06:02 domidumont joined, awwaiid left 06:05 domidumont left 06:06 domidumont joined 06:07 gfldex joined 06:08 diana_olhovik_ joined, Sqirrel joined 06:11 aborazmeh joined, aborazmeh left, aborazmeh joined 06:12 quester joined 06:15 awwaiid joined 06:18 _mg_ joined 06:20 aborazmeh left 06:26 FROGGS joined
FROGGS good morning #perl6 06:26
yoleaux 20 May 2015 20:29Z <japhb> FROGGS: From the commit message for MoarVM: commit 4102a25b1acea1d453d87f4e614d328ba0af8816
20 May 2015 20:29Z <japhb> FROGGS: From the commit message for MoarVM: commit 4102a25b1acea1d453d87f4e614d328ba0af8816 it looks like the comparison is reversed (> instead of <). Am I misunderstanding?
FROGGS TODAY IS RELEASE DAYS - please add your changes to the changelog, hold back patches when you are uncertain about them, etc 06:27
and test what you can, also modules etc
.tell japhb the commit states "round up to a multiple of"... maybe it is my bad English; if the biggest elem (perhaps an inlined struct) is bigger than 8 bytes (void* on x86_64), then it should allocate a multiple of 8 bytes 06:30
yoleaux FROGGS: I'll pass your message to japhb.
FROGGS .tell japhb there is already that bit of code: github.com/MoarVM/MoarVM/blob/4102...uct.c#L242 06:31
yoleaux FROGGS: I'll pass your message to japhb.
06:32 adu left
FROGGS .ask jnthn are you around in about 12 hours to upload the moarvm release tarball? 06:32
yoleaux FROGGS: I'll pass your message to jnthn.
06:35 [Sno] left, Alina-malina joined 06:38 chenryn left, chenryn joined 06:40 gfldex left 06:43 pochi joined, ab5tract joined, breinbaas joined, arnsholt joined, freeze joined, torbjorn joined, hernan joined, eternaleye joined, BooK joined, tadzik joined 06:44 camelia joined 06:45 ChanServ sets mode: +v camelia 06:46 agentzh joined, agentzh left, agentzh joined 06:50 RabidGravy joined 06:51 ggoebel joined 06:57 al2o3-cr joined 07:08 burnersk joined, [ptc] joined 07:09 zakharyas joined 07:12 [ptc]_ joined 07:13 fhelmberger joined 07:17 fhelmberger left 07:23 mj41 joined 07:28 lizmat joined 07:30 isBEKaml_mobile joined 07:43 darutoko joined 07:49 kurahaupo1 joined, [ptc]_ left, [ptc]_ joined 07:50 fhelmberger joined 07:56 [Sno] joined
lizmat Good *, #perl6! 07:57
things not yet in ChangeLog, which I don't know quite how to word them: 07:58
FROGGS morning lizmat
lizmat + Cunion changes in NativeCall
FROGGS I'll add that of course
lizmat + Add jna.library.path to set properties for JVM runner 08:00
FROGGS I can add that too 08:01
lizmat: can you please run 'make test' in nqp and rakudo? I did that on 64 and 32bit linux and 64bit windows 08:02
will do 32bit windows in a bit too
lizmat will do 08:03
FROGGS lizmat++ 08:04
I'll disable CURLI in a bit, by only putting file# in INC and CUSTOM_LIB
this way we can have a star...
lizmat cool 08:05
08:05 vendethiel joined
FROGGS though, I'll reenable CURLI after the release of course... maybe only in a branch 08:05
vendethiel o/ #perl6!
FROGGS hi vendethiel
lizmat Files=98, Tests=3763, 9 wallclock secs ( 0.42 usr 0.15 sys + 8.38 cusr 0.80 csys = 9.75 CPU)
FROGGS did is pass? :o) 08:06
lizmat FROGGS: yes
t/04-nativecall/02-simple-args.t (Wstat: 512 Tests: 12 Failed: 2)
Failed tests: 10-11
Non-zero exit status: 2
Files=39, Tests=511, 3 wallclock secs ( 0.20 usr 0.08 sys + 17.25 cusr 2.10 csys = 19.63 CPU)
those failures I've seen for weeks now
FROGGS hmmm, this is about uint8... I thought we fixed those just recently... hmmm 08:09
08:13 [ptc]_ left
dalek kudo/nom: 7b42268 | lizmat++ | docs/ChangeLog:
Some more entries
 08:13
masak hi vendethiel 08:17
08:17 [ptc]_ joined 08:19 [ptc]_ left
dalek ast: c5b581a | lizmat++ | S10-packages/basic.t:
Fix deprecation warning for bare package
 08:24
RabidGravy woo! "Package rakudo-star.x86_64 0:0.0.2015.04-3.fc21 will be an update" 08:26
moritz 0.0.2015.04 ? 08:27
but yes, woo!
RabidGravy who knows the mystery of fedora versioning?
FROGGS \o/ 08:28
nwc10 w.0.0 surely?
dalek ast: d8d17f9 | lizmat++ | S26-documentation/module-comment.t:
Fix deprecation message for module
ast: b4149a2 | lizmat++ | integration/advent2013-day19.t:
Fix deprecation warning for Supply.more/emit
 08:30
lizmat afk for a few hours 08:31
08:34 dakkar joined 08:53 quester left
masak hm, the Angular people actually seem not-half-bad at coordinating between 1.x and 2.0 development 08:56
didn't really expect that :)
08:56 tinyblak_ joined 08:57 phdphil joined 09:00 tinyblak left 09:08 espadrine joined
timotimo how do we actually have a 2015.04 rakudo star on fedora? 09:10
FROGGS that's an interesting question :o)
timotimo or did rabid gravy just make that up? :P 09:11
RabidGravy nope it really did just update it (the previous was 2014.12) 09:21
timotimo but ... how
we didn't make a 2015.04 rakudo star
RabidGravy jonathan@coriolanus ~]$ /usr/bin/perl6 -v 09:22
This is perl6 version 2015.04 built on MoarVM version 2015.04
timotimo did the fedora packager just do the star release process on their own? 09:23
i'm confused
RabidGravy let me dig a bit
09:24 yeahnoob joined
RabidGravy www.spinics.net/lists/fedora-packag...58251.html 09:25
timotimo ah 09:26
so just an nqp update
and thus a moarvm update
that's not that bad
it passes all the spec tests, yeah?
RabidGravy pkgs.fedoraproject.org/cgit/rakudo-star.git/ 09:27
09:28 xfix joined
timotimo BuildRequires: readline-devel ← that wasn't thrown out at that point yet? 09:29
literal sad that Debian's rakudo package is still at 2014.07
09:32 yeahnoob left
timotimo oh my, the XML module is quite unhappy 09:36
probably GLR-related
masak most likely, yes 09:37
I fixed two different GLR-fallout-related things in a script of mine yesterday.
09:48 bjz joined
timotimo Net::Irc has warnings about character classes containing quotes 09:58
Net::IRC::Bot, apparently
LibraryMake also needs a touch-up for the unit keyword
and all the embedded modules in panda 09:59
huh, there's no test failures with XML here 10:04
but something's wrong with comments
10:07 telex left 10:08 telex joined
timotimo gaaaaaaaah 10:10
i was looking at the wrong line because the last digit of the line number got wrapped in my terminal
10:11 sftp left 10:15 isBEKaml_mobile left
masak timotimo: do you have line numbers sitting on the right of the text? how did they manage to get wrapped? 10:15
timotimo Use of Nil in string context in block at /home/timo/perl6/ecosystem/exemel/lib/XML.pm6:996 10:16
exactly the 6 and only the 6 went to the next line
XML is still using [0] to get at things quantified with ? in regexes 10:18
masak m: grammar G { regex TOP { <foo>? }; regex foo { foo } }; G.parse("foo"); say $<foo>.^name 10:22
camelia rakudo-moar 7b4226: OUTPUT«Match␤»
masak m: grammar G { regex TOP { <foo>? }; regex foo { foo } }; G.parse("foo"); say $<foo>[0].^name
camelia rakudo-moar 7b4226: OUTPUT«Nil␤»
masak how many times have we changed that .[0] thing back and forth now for the ? quantifier? 10:23
this is Perl 6's random hunt for the sweet spot of the Maybe type :P 10:24
timotimo good Q
masak (the closest to which we ever got was probably TimToady++'s OKness proposal, which was contested/rejected)
timotimo i didn't pay sufficient attention
10:24 mj41 left
vendethiel never witnessed such a change 10:24
OKness?
timotimo yeah, OKness was a bit tough
how different types should signal an "ok" value vs a "not ok" value 10:25
i forgot the details
masak vendethiel: not sure I'll be able to hunt it down. it was a gist by TimToady with some pretty radical changes to the type system.
timotimo er ... huh?
Template::Anti is using a rakudo from 2014-12
on its travis-ci 10:26
dwimperl.com is szabgab's?
masak vendethiel: found it: gist.github.com/TimToady/1828667
10:27 dayangkun left
vendethiel ...that sounds very confusing 10:28
10:28 chenryn left
masak I remember feeling that it was somehow "inside-out". 10:28
I'm re-reading the proposal now.
right, captured in this sentence: "'0 but True' needs to become more like 'OK[REALLY => 0]'" 10:29
that turns it inside-out.
vendethiel doesn't get what's "REALLY" or why Ok takes a type parameter that's a hash 10:30
timotimo named type parameters?
vendethiel oh, uhmm.
m: role Foo[:$x] { method x { say $x } }; say Foo[:5x].x; 10:31
camelia rakudo-moar 7b4226: OUTPUT«5␤True␤»
vendethiel :o
masak please don't use the `:5x` syntax willy-nilly :)
vendethiel just because it's simpler to type :)
masak it's meant for a very specific purpose, and using it outside of that is basically mis-use of the feature
vendethiel went as far as removing it from the learnxiny 10:32
..exactly because of what you said :-)
masak probably a good idea :)
timotimo you're speaking of "mis-use of a feature"? :)
you know people will take that as a challenge?
masak .oO( :0good_idea )
FROGGS is this #python?
vendethiel isfroggsguido 10:33
.oO( poor FROGGS )
masak just because we're not #python doesn't mean we can't point to mis-uses of features ;)
FROGGS :o)
timotimo i want to introduce a new syntax for this kind of pair
:(value)name
masak timotimo: that totally breaks the original intent
timotimo .o( i can make that a slang! )
masak of :2nd
vendethiel gets reminded of objective-c's block syntax
masak timotimo: you would be better of disallowing the syntax everywhere except in calls to .subst :) 10:34
off*
timotimo i'm not saying i'm going to use it at all, but i'd defend a programmer's desire to use :123foo whenever she wants to use it!
masak I think that's a poor use of the defense of that freedom ;) 10:35
the question "does it improve readability of the source code" should show up somewhere 10:36
timotimo of course a programmer would only want to use it where it improves readability of the source codes 10:37
that goes without sayings
FROGGS my $datetime.add( :1day ) 10:38
masak timotimo: but it doesn't go without saying. I had to remind vendethiel above in the backlog (and he immediately agreed that it was a bad use). 10:39
vendethiel wut?
masak vendethiel: the :5x thing
vendethiel I said I used it just because it was simple as an example, not because I thought it was proper use here
masak right, exactly. 10:40
I do realize that your example was a simple, throwaway one.
I just don't want to give people ideas about that being a sensible way to write named arguments with integer values *in most cases*.
timotimo :)
masak :1day is a good example of a possible good use. but it has to be viewed on a case-by-case basis like that.
timotimo m: say :0xfood 10:41
camelia rakudo-moar 7b4226: OUTPUT«Unexpected named parameter 'xfood' passed␤ in block <unit> at /tmp/YxKX6v0MI4:1␤ in any <unit-outer> at /tmp/YxKX6v0MI4:1␤␤»
masak people hate to think for themselves, I know. sweeping rules that ignore specifics are much preferred.
timotimo m: say (:0xfood)
camelia rakudo-moar 7b4226: OUTPUT«xfood => 0␤»
vendethiel masak: I'm just responding to the "had to remind". don't worry -- I'm aware of :5x's use :)
timotimo d'aaw
10:41 breinbaas left
timotimo ^- super trolly 10:41
masak m: say 0xf00d 10:42
camelia rakudo-moar 7b4226: OUTPUT«61453␤»
10:43 breinbaas joined 10:46 maddingue joined
timotimo well, at least i got rid of the warnings in XML :) 10:48
and the deprecation for unit
is panda going to wait for the release (or even star?) to put unit keywords into its dependencies?
FROGGS m: say (0xDEADBEEF)
camelia rakudo-moar 7b4226: OUTPUT«3735928559␤»
FROGGS m: say (:0xDEADBEEF)
camelia rakudo-moar 7b4226: OUTPUT«xDEADBEEF => 0␤»
FROGGS :P 10:49
(to be avoided)
timotimo agreed
10:50 lizmat left 10:51 maddingue left 10:55 lizmat joined 10:57 maddingue joined 10:58 _mg_ left
timotimo maybe there should be a "best practice" to put a test at the end of every test file that checks if any control exceptions have been raised? 10:59
lizmat maybe we should build that into Test.pm ? 11:01
timotimo kind of hard to do, as it needs to have a CONTROL block stack-wise outside the user's code 11:02
someone suggested at some point the ability to export phasers
that would help here
lizmat yup, exactly 11:03
still. I don't see why it wouldn't be possible to export a CONTROL block to the outer scope 11:04
which at that point, hasn't been fully parsed yet
timotimo yeah, i don't think it'd be impossible
11:07 chenryn joined 11:09 mj41 joined
dalek kudo/nom: e113455 | lizmat++ | docs/ChangeLog:
Mention demise of .for
 11:12
lizmat afk again& 11:13
FROGGS lizmat++
11:17 cschwenz joined, cschwenz left 11:18 _mg_ joined 11:23 bjz left
timotimo List/Supply.rotor expects a list of Pairs (elems|elems => gap) - i wonder how this could be changed to reflect the pair/value thingie 11:24
i mean, you can still pass normal single numbers, right?
or ... you don't? 11:25
+ 'earliest' now handles for remaining channels without specific handlers 11:27
is that "for" stray, or is my grammar grammar busted?
Int/Rat ** -Int now generates a Rat if possible, else a Num - and i think this "else" would read better as "otherwise"?
moritz +1 to otherwise 11:28
or else 'orelse' :-) 11:29
timotimo andthen do the release
moritz CATCH { everything that's not supposed to be released } before you lose CONTROL 11:31
FROGGS timotimo: always patience with the young horses :D
11:35 colomon_ joined 11:36 colomon left, colomon_ is now known as colomon 11:39 andreoss joined
dalek kudo/nom: 9bdeaa9 | (Carl Masak)++ | docs/ChangeLog:
[docs/ChangeLog] remove extraneous 'for'

Pointed out by timotimo++.		 11:40
11:41 pecastro joined 11:45 retupmoca left, hoelzro left 11:46 retupmoca joined, hoelzro joined 11:49 hobbs joined, hobbs left, hobbs joined 11:57 egrep joined 11:58 zakharyas left
timotimo do you people think the language/ section of the docs should be sorted manually rather than alphabetically? 11:59
putting "grammars" before "regexes" doesn't seem terribly helpful
There are predefined character classes of the form \w. Its negation is written with an upper-case letter, \W. ← just nit picking here, really; would "their negations" be more correct than "its negation" here? 12:01
m: "foo\n\r" ~~ /\s/; say (~$/).perl 12:02
camelia rakudo-moar e11345: OUTPUT«"\n"␤»
timotimo m: "foo\r\n" ~~ /\s/; say (~$/).perl
camelia rakudo-moar e11345: OUTPUT«"\r"␤»
timotimo maybe the \s section should have an explicit "\s does not match logical newlines, just single characters." 12:03
anyways, very good that the character class sections explicitly point out exotic characters that fall into that category 12:04
12:04 laouji left
dalek c: 2475551 | timo++ | lib/Language/regexes.pod:
regexes.pod - missing word
 12:07
12:09 zakharyas joined
masak timotimo: yes, ordering the language/ section manually sounds like an improvement over alphabetical order. 12:16
12:21 andreoss left
timotimo how do we implement that? give the pod files a litle heading section "ordering" where we can define what cmp-ing the pod files against each other shall give? 12:21
regexes.pod: =ordering before grammars.pod ? 12:22
i'm not very convinced by giving the files numbers at the front
dalek kudo/nom: 29e5521 | FROGGS++ | src/core/CompUnitRepo.pm:
use hash symbol everywhere in CUR include-specs
 12:27
kudo/nom: 7adde5a | FROGGS++ | src/core/CompUnitRepo/Local (2 files):
remove method "install" for CURLF
kudo-star-daily: d74d118 | coke++ | log/ (9 files):
today (automated commit)
 12:33
timotimo [Coke]: got any clue why it fails to find all these modules? 12:36
masak timotimo: hm, ordering information should preferably not be spread out over all the files. better if they're in one place, such as the table-of-contents where they're listed. 12:42
timotimo: that way, no numbers need be used either. just an ordered list.
timotimo we don't actually have a "the table-of-contents" document 12:44
do we?
masak oh, mebbe not.
moritz I think it's auto-generated
masak can it be placed as a meta-data file in some appropriate place?
then it doesn't even have to be Pod 12:45
moritz from =TITLE and =SUBTITLE attributes
masak the important thing is that the renderer finds it and uses it, right?
12:45 yogan joined
moritz right 12:45
and since it parses the pod, it should be in the pod
12:45 abraxxa joined
FROGGS tadzik: ping 12:46
masak moritz: yes, but the document it should rightfully be in is automatically generated...
moritz: ...and spreading it over all the sub-documents just because there's no ToC document feels... ugh. 12:47
(and avoiding ordinal numbers would be worthwhile, too)
moritz masak: TBH, the whole htmlify qualifies as "ugh"
patches welcome :-)
timotimo right :S 12:48
moritz we just had the mis-fortune of me being the only one hacking on it when it was originally developed
so nobody could prevent my bad ideas from the start :-)
masak heh :) don't be so hard on yourself 12:49
we have documentation rendered as HTML. moritz++
this is the crucial difference between software as effect vs software as artifact.
software as effect: does it work? yes! yay!
software as artifact: is it beautiful? no. boo! does it contain bad ideas? unfortunately. ugh :/ 12:50
both aspects are important, but it could be argued that the first one takes precedence if the goal is to quickly deliver value.
and that's exactly what happened here. so moritz++ :) 12:51
FROGGS hmmm....
Compiling lib/Facter.pm to mbc
useless use of 'is rw' on @!search_path in any at src/Perl6/World.nqp:2498
moritz rw on a private attribute looks questionable 12:52
so does the line number :-)
FROGGS error reporting bug me thinks 12:54
it should report this line instead: github.com/cosimo/perl6-facter/blo...ter.pm#L47 12:55
12:59 raiph joined
masak reproduced locally. 12:59
masak submits rakudobug
m: class T { has $!foo is rw } 13:00
camelia rakudo-moar 7adde5: OUTPUT«useless use of 'is rw' on $!foo in any at src/Perl6/World.nqp:2498␤»
vendethiel I guess it's the ".setting" regexp?
masak vendethiel: I don't... think so...? which ".setting" regex? 13:04
vendethiel warning filters stack frames using something like "%frame<line> !~~ /'.setting'/" iirc 13:05
13:07 chenryn left
moritz vendethiel: no 13:08
13:08 liztormato joined
moritz vendethiel: the problem is that the line of the file that's being parsed doesn't appear *anywhere* in the stack 13:09
liztormato Perhaps it should be a .ends-with('.nqp') ?
moritz because the line isn't calling anything; it's the parser that does something with this line
so, can't use a normal warning here
vendethiel oh :)
thanks for the clarification!
liztormato Cycling again &
13:09 liztormato left 13:12 skids joined 13:15 PZt joined
DrForr .tell lizmat #125226 was my own fault, the 'please close this bug' message never made it, apparently. 13:18
yoleaux DrForr: I'll pass your message to lizmat.
[Coke] OBTW, we're going to have to adopt some kind of branch/dev/release strategy after "release day" that isn't "everything in master^Wnom" 13:19
masak nwc10: just saw your comment in #125216 -- "are floats real numbers?"
nwc10: well, no -- strictly speaking, they don't guarantee associativity in addition, which real numbers do.
DrForr On other notes, can the Perl6 grammar be exposed within perl6 yet? I keep getting this irritating itch to write perl6tidy... 13:20
timotimo i have a little bit of interest in writing something that deparses
masak nwc10: but I think that does not detract from the fact that floats are very much there to be a goodenuf representation of real numbers.
[Coke] timotimo: (fails to find modules) - no, I just unbusted the star build process a day ago, haven't dug into this one yet. 13:21
masak nwc10: and saying that "floats are positioned on the real line" does not feel weird or wrong to me at all.
timotimo ok, thanks for your work! :)
[Coke]: perhaps star is including modules that don't have "provides" sections yet?
shall i go through the submodules in star and bump the versions up a bit? 13:22
DrForr Well, floats are just rationals with a really big denominator :)
[Coke] timotimo: I can't remember if my star tester is grabbing master/HEAD or just the revision tagged in the submodule. I think it's the former. 13:23
timotimo: but bumping those probably couldn't hurt.
FROGGS [Coke]: why do we need to adopt another release strategy? 13:24
DrForr masak: Thanks, 'twas a thinko on my part. 13:25
masak DrForr: there's an unfortunate corner there where the syntax for blocks and hashes collides. 13:26
DrForr Yeah, I was just used enough to things such as lives_ok {} to assume for a few shining minutes that subtest {} operated the same way. 13:27
FROGGS ewww, Yapsi/t/compiler.t infiniloops 13:28
DrForr Mind you, I think it probably *should*, for consistency's sake, but we've all got our crossproducts to bear.
masak FROGGS: probably GLR-related... :/ 13:30
FROGGS masak: yeah, is just not fun in a smoke run, is all :o)
masak maybe the smoker should have a 5-minute auto-kill timeout... 13:32
FROGGS .tell panda 1) please add curlies to package decls to File::Find, Shell::Command and JSON::Tiny 13:33
yoleaux FROGGS: I'll pass your message to panda.
timotimo FROGGS: release strategy? what?
[Coke] FROGGS: right now everything is "get as much as possible of whoever wants to work on stuff into the next release". After we have a "real" release, we need to have some sort of arrangement where there's some discussion about what's going in the release, new features, backcompat, timing with the actual release date...
FROGGS .tell panda 2) the 'build reverse deps feature' builds reverse deps of not yet installed modules, but it does not install their deps
yoleaux FROGGS: I'll pass your message to panda.
timotimo telling that to ... panda? :) 13:34
FROGGS ups
hehe
[Coke] we do all of this in an ad hoc fashion right now (and better than many for pay gigs I've been on.) but "everything into master" isn't viable if you have multiple releases in the wild to care about.
FROGGS .tell tadzik 1) please add curlies to package decls to File::Find, Shell::Command and JSON::Tiny
yoleaux FROGGS: I'll pass your message to tadzik.
FROGGS .tell tadzik 2) the 'build reverse deps feature' builds reverse deps of not yet installed modules, but it does not install their deps
yoleaux FROGGS: I'll pass your message to tadzik.
[Coke] (right now, our response to anyone with an existing version is "upgrade to the latest( and sorry, some stuff might break)" 13:35
FROGGS [Coke]: my only concern is that a branch that is not master might be not that often tested on different platforms
[Coke]: so instead of release branches I would propose we keep having feature and dev branches 13:36
though, the point where stuff is merged into master/nom should be discussed more then
discussed and decided*
[Coke] I'm not tied to a particular plan of attack here, just want there to be some plan. We also should consider tagging or somehow labelling branches that need testing for whatever automatic infrastructure we have (to the question from jdv79 recently about "how are we testing what and where is it") 13:40
FROGGS aye 13:41
I very much agree
13:42 JoaquinFerrero left 13:43 coffee` joined
Woodi hi today :) 13:49
I must say that has $.fh is lazy = open( $!file, :r ); is a fun :)
it's like user want's to browse web, he starts browser, access network and modem starts to dial somewhere - such futuristic automatization :) 13:51
FROGGS hehe
it should also make some funny noise :o)
Woodi but it can be seen as kind of "finalizer", a "starter" or trigger doing action... 13:52
or general mechanism for TIE but not only for storge...
actually we have/had wrappers for methods, maybe it can be reused ? 13:53
dalek ar: a55705b | timotimo++ | modules/ (16 files):
update submodules to latest masters

this fixes a whole bunch of META.info files, for example
Woodi but generally speaking thing is like wanting to some bit of small scripting into boring, structural code :) 13:54
*to have some*
FROGGS: that noise is mandatory! it's kind of essential for modem-in-action ;) 13:58
and "is lazy" is wrong name. laziness to me is connected to messing with structures... trigger maybe ? but all this thing seems like some layer above normal code sequence... 14:00
14:03 Foxcool joined 14:08 domidumont left 14:09 domidumont joined
masak Woodi: `is lazy` has pretty solid prior art in Moose, though. 14:20
Woodi: I don't share your connotation of "messing with structures", fwiw 14:21
14:21 molaf joined
DrForr Naah, laziness is just a delayed action, doesn't *have* to be with data structures. 14:35
And a far more general act than TIEing an object.
FROGGS Inline::Perl5 fails :o( 14:36
t/call_p6_function_from_p5.t .. ok
panic: MUTEX_LOCK (22) [op.c:697].
err, the failing test file is: t/callables.t
14:36 abraxxa left
DrForr @a[3200000] = 1; # If your program never uses values from @a beyond 100, why preallocate? :) 14:38
FROGGS weird, I::P5 works when the deprecation message is not printed :o( 14:39
14:41 bjz joined, bjz left 14:42 bjz joined 14:48 phdphil left
[Coke] Inline::Perl5 tests aren't being run because I can't panda in an ./install directory - could have been failing for some time. 14:50
vendethiel .u II 14:51
yoleaux U+0418 CYRILLIC CAPITAL LETTER I [Lu] (И)
U+0419 CYRILLIC CAPITAL LETTER SHORT I [Lu] (Й)
U+0438 CYRILLIC SMALL LETTER I [Ll] (и)
14:51 Klaudioh joined
FROGGS [Coke]: it works on my box when I do not pipe the result into a file :S 14:54
14:57 darutoko- joined 15:00 darutoko left
hoelzro o/ #perl6 15:01
15:02 fhelmberger left 15:06 Ven joined 15:12 isBEKaml joined, kurahaupo1 left 15:14 diana_olhovik_ left 15:17 molaf_ joined 15:20 molaf left, mj41 left 15:24 domidumont left 15:25 domidumont joined 15:26 _mg_ left 15:29 espadrine left, zakharyas left, kurahaupo1 joined 15:30 Foxcool left 15:33 FROGGS left 15:40 ssqq joined, NYC5 joined
NYC5 Hi room 15:40
FROGGS[mobile] hi 15:42
ssqq Hi NYC5 15:43
NYC5 Hi ssqq
What is this room about?
seems sorta lifeless ;-)
ssqq about Perl 6 language.
15:43 xfix left
ssqq FROGGS: Could we add token to a grammar dynamically like add method to class? 15:44
15:44 lichtkind joined
NYC5 thanks, guess I am in wrong place then 15:44
take care
thanks for response :)
ssqq NYCS, you are welcome
15:45 NYC5 left, FROGGS joined 15:47 tinyblak_ left, xfix joined, xfix left, xfix joined 15:49 diana_olhovik_ joined, _mg_ joined, _mg_ left 15:52 tinyblak joined
hoelzro FROGGS[mobile]: do you have any plans/ideas for supporting STL stuff with NativeC++all? 15:57
16:00 tinyblak left, tinyblak joined
timotimo ssqq: well, you can certainly use mixin to put in new tokens etc 16:01
FROGGS hoelzro: STL?
hoelzro FROGGS: standard template library
ex. std::string, std::vector
timotimo impossible without a C++ compiler
hoelzro a lot of C++ libraries use them as arguments
timotimo: I figured as much =/
because of the template soup
16:02 gfldex joined
FROGGS can't we map that to Perl 6 types? 16:02
hoelzro I don't know if we can; the C++ compiler generates a bunch of code when it instantiates a templated type, doesn't it?
ssqq timotimo: let me have a look 'mixin'.:) 16:03
timotimo ssqq: with "but" and "does"
FROGGS hoelzro: I guess I need an example
hoelzro FROGGS: well, std::string is a typedef for std::basic_string<char> 16:04
and, iirc, basic_string has a few other template arguments, but the default to "reasonable defaults"
when I have a template type like basic_string, the .so doesn't contain all of the code for methods of that type 16:05
16:05 tinyblak left
hoelzro any method that utilizes the type T is generated for each value provided for T 16:06
so basic_string<char> and basic_string<wchar> methods reference entirely different code
vendethiel hoelzro: "STL" is the name of the *old* library, the c++ standard library is std :)
hoelzro one thing we *could* do is provide a support library that creates some common templated types for us 16:07
hoelzro 's inexperience with C++ is showing =)
16:07 eli-se joined
vendethiel but yeah, templates have to be in headers files because the compiler needs to specialize them for your types 16:07
16:08 domidumont left
hoelzro right 16:08
eli-se vendethiel: not in C++98! 16:09
ssqq timotimo: have some sample about it?
timotimo ssqq: there is two modules out there that construct grammars dynamically 16:11
ssqq Css::Grammar? 16:12
timotimo no 16:14
i'm confused now
it seems like this one uses EVAL
the BNF one by tadzik 16:15
all it does is .^add_method before .^compose is called
github.com/tadzik/Grammar-BNF/blob...NF.pm#L249
ssqq timotimo: thanks, let me see it carefully. 16:17
16:19 KotH joined, tinyblak joined
KotH good evening! 16:19
16:20 yubimusubi left
timotimo what part of the star release process ought to grab rakudo/, nqp/ and MoarVM/? 16:21
because when i use the makefile from star's build tools with --version=daily, they are missing
hello KotH 16:24
16:26 spider-mario joined
hoelzro oh, I forgot the release was today 16:26
\o/
FROGGS hoelzro: and there is no stdlib.so that we can load which exposes these symbols? 16:27
hoelzro FROGGS: in the case of basic_string, there are no symbols
[Coke] timotimo: step 3 in the star release guide, I think.
16:27 Possum joined
hoelzro it's all in the headers 16:27
FROGGS timotimo: are you able to upload moarvm release tarballs to moarvm.org? 16:28
16:29 domidumont joined
FROGGS I'll start cutting the moarvm release in about 1.5 hours 16:30
[Coke] FROGGS: ping me if you need me to do the macports step, no worries. 16:32
FROGGS [Coke]: k
16:40 mohij joined 16:46 rindolf left 16:48 rindolf joined
timotimo no 16:48
16:55 jepeway joined 17:02 fhelmberger joined 17:03 telex left 17:04 telex joined, zakharyas joined 17:07 fhelmberger left, zakharyas left 17:10 eli-se left 17:12 andreoss joined 17:17 diana_olhovik_ left 17:23 Ven left
andreoss can i return a path to a existing file without declaring a subset? 17:23
17:25 dakkar left
FROGGS andreoss: you want a constraint on the return value? 17:26
TimToady or do you just want a bareword like in Perl 5?
17:27 zakharyas joined
TimToady the question is hard to understand... 17:27
andreoss sub xxx($a) returns -> $x { $x.IO.e } { ... } 17:28
anonymous subset may be?
vendethiel sub xxx($a where *.IO.e) {} ? 17:30
TimToady you want a type that matches only existing files?
vendethiel or subset File of Str where *.IO.e;
17:34 tinyblak left
andreoss i know that having a subset for that will work 17:36
vendethiel I gave you another solution with "where", isn't it good?
andreoss but are there other ways to put a constaing on returning value?
vendethiel: i believe it will be applied to an argument, not to value this sub returns 17:37
vendethiel aah
FROGGS it is like he wants that: 17:38
std: sub foo($a --> IO::Path where *.e) { }
camelia std 28329a7: OUTPUT«5===SORRY!5===␤Multiple prefix constraints not yet supported at /tmp/WRPNGh9JRA line 1:␤------> 3sub foo($a --> IO::Path where *.e7⏏5) { }␤Potential difficulties:␤ $a is declared but not used at /tmp/WRPNGh9JRA line 1:␤------> 3sub foo(7⏏…»
FROGGS TimToady: would that be possible?
skids I could see that being useful for type calculus, but the error messages are going to be LTA for application-specific purposes versus checking in the function itself. 17:39
TimToady that would just mean you throw an exception if it doesn't match 17:40
vendethiel maybe use POST?
TimToady likewise
17:41 Ven joined
andreoss m: proto xxx(|) { * ; say "hi "}; multi xxx(1) { say 1 }; xxx(1); 17:46
camelia rakudo-moar 7adde5: OUTPUT«hi ␤»
17:47 ssqq left
andreoss can i have real code inside proto? 17:47
tony-o_ did Proc::Async.start[...] .status.exitcode get moved?
andreoss: yes
TimToady use {*} inside, not *
m: proto xxx(|) { {*} ; say "hi "}; multi xxx(1) { say 1 }; xxx(1); 17:48
camelia rakudo-moar 7adde5: OUTPUT«1␤hi ␤»
tony-o_ m: sub x(Str $d) { my proto p { 'p'.say; } EVAL $d; }; x("d()"); d();
camelia rakudo-moar 7adde5: OUTPUT«5===SORRY!5=== Error while compiling /tmp/sr2c1ikust␤Strange text after block (missing semicolon or comma?)␤at /tmp/sr2c1ikust:1␤------> 3sub x(Str $d) { my proto p { 'p'.say; }7⏏5 EVAL $d; }; x("d()"); d();␤ expecting any of:␤ …»
tony-o_ m: sub x(Str $d) { my proto p { 'p'.say; }; EVAL $d; }; x("d()"); d();
camelia rakudo-moar 7adde5: OUTPUT«5===SORRY!5=== Error while compiling /tmp/usj6VA8l3a␤Undeclared routine:␤ d used at line 1␤␤»
tony-o_ m: sub x(Str $d) { my proto p { 'p'.say; }; EVAL $d; }; x("d()");
camelia rakudo-moar 7adde5: OUTPUT«5===SORRY!5=== Error while compiling EVAL_0␤Undeclared routine:␤ d used at line 1␤␤»
tony-o_ m: sub x(Str $d) { my proto p { 'p'.say; }; EVAL $d; }; x("p()"); 17:49
camelia rakudo-moar 7adde5: OUTPUT«p␤»
tony-o_ m: sub x(Str $d) { my proto p { 'p'.say; }; EVAL $d; }; x("p()"); p()
camelia rakudo-moar 7adde5: OUTPUT«5===SORRY!5=== Error while compiling /tmp/8627h_v8UA␤Undeclared routine:␤ p used at line 1␤␤»
andreoss m: proto foo(|) { say "multi returns " ~ {*}; } ; multi foo(1) { "path" }; foo(1) 17:50
camelia rakudo-moar 7adde5: OUTPUT«multi returns path␤»
17:59 brrt joined
brrt \o 17:59
out of interest; tomorrow i'll be 'representing' perl6 on a python conference 18:01
that is to say, wearing my perl6 badge :-)
TimToady ask them how we can get uptake for our new version :P 18:02
brrt lol
yeah, i'' do that
Ven Writing a long commit message always gives me that good-work feel. 18:04
brrt wonders just how much of our 'professional work ethic' comes down to just such feelings 18:05
Ven "was I a good boy?" :P
andreoss m: class Y { multi method BUILD(:$a, :$b) {say $b}; multi method BUILD(:$a) { say $a} } ; Y.new( a => 1).say; 18:07
camelia rakudo-moar 7adde5: OUTPUT«(Any)␤Y.new␤»
isBEKaml Ven: I think that has more to do with "Look how much I got done!" :-) 18:09
Ven :P
I think a big part of it is also the feeling you've saved your future self a lot of work, brrt
brrt oh, probably, yes
whether that is actually true, i don't know
but we're emotional creatures, and if there are things we can do to optimize our happiness in our work, why not 18:11
Ven -Ofun, right? :)
isBEKaml Ven: The only reason I write long and descriptive commits is, I bundle a *lot* files into a single commit. Like say, one commit per feature(or a changeset). So, you really have to summarize them for sane reading through the logs (later audits or bug hunt)
Ven isBEKaml: since I've switched to emacs(with evil-mode, obviously), using magit, I realized I was much more prone to splitting my commits and writing nice messages due to the tools being available (interactive committing to only take the changes I'm interested in, say) 18:12
isBEKaml Ven: unfortunately, we're still on svn 18:13
Ven aw :(
18:14 eli-se joined
isBEKaml Ven: a lot of commits on svn isn't a very good indicator of productivity, compared to git. I'd attribute that to how you can manipulate your local history while still preserving your files. 18:15
brrt now that i think of it, i definitely see the value of a proper commit log, especially in 'production' software that you haven't written all by yourself 18:16
as in
commit: 'Fixed the FooBar' isn't always really helpful
isBEKaml Ven: so, on svn - having a lot of successive commits is like "so you wasted so much time writing this feature and you had lots of commits to prove you were fighting the code?" sounds like management-speak, but really - there's some truth in it 18:17
Ven I see.
18:18 diana_olhovik_ joined 18:21 sftp joined 18:32 andreoss left 18:35 raiph left, tinyblak joined, raiph joined 18:37 mj41 joined 18:38 Ven left, sue__ joined 18:39 tinyblak left 18:44 [Sno] left 18:46 rindolf left
dalek kudo/nom: 2172535 | (Steve Mynott)++ | / (3 files):
three minor typos
 18:46
kudo/nom: b865609 | lizmat++ | / (3 files):
Merge pull request #433 from stmuk/nom

three minor typos
lizmat argh, I just realized I'm 45 minutes past FROGGS release window, instead of 15 minutes before 18:49
yoleaux 13:18Z <DrForr> lizmat: #125226 was my own fault, the 'please close this bug' message never made it, apparently.
lizmat FROGGS: sorry, please revert / ignore if too much trouble 18:50
FROGGS lizmat: no problem
lizmat: I'm not that far yet
lizmat *phew*
FROGGS :o)
lizmat FROGGS++
lizmat is going to do a JVM build and spectest to atone 18:51
FROGGS I might even do the release tomorrow when jnthn does not pup up to upload the moarvm tarball
18:53 muraiki joined
PerlJam does the moarvm tarball *need* to be there first? Aren't most people going to come at it from the other end? 18:55
(Or do we anticipate that packagers (rpm, deb, whatever) are using the tarballs to make their packages?)
timotimo hum? 18:56
FROGGS PerlJam: basically the star repo needs the tarball, and rakudo/rakudobrew fetches from git 19:02
though, I don't like the idea of doing a partial release...
we can wait a day without doing any harm
19:03 fhelmberger joined
PerlJam ah. okay 19:03
19:07 fhelmberger left 19:12 zakharyas left 19:16 brrt left 19:17 muraiki left
lizmat is investigating breakage in t/01-sanity/54-use-lib.t 19:18
on the JVM
dalek kudo/nom: c006cd1 | FROGGS++ | / (3 files):
replace 'is inlined' trait by HAS scope declarator

Attributes in CStructs and CUnions can specify whether they are 'embedded' into the surrouning structure or referenced. To say the former one will now be able to use the 'HAS' keyword, to do the latter one uses 'has' as usual.		 19:19
FROGGS lizmat: yes, that's also on my list 19:20
TimToady: you can HAS HAS now :o) 19:21
19:21 Ven joined 19:22 Ven left
FROGGS I::P5 keeps working, nice 19:23
also the P5 integration tests pass, like they did 19:24
timotimo i don't quite understand the logic behind has vs HAS
is HAS a recronym? :)
FROGGS timotimo: I just know that it was a long discussion, and that jnthn and TimToady agreed on it 19:25
no
it probably just means: I REALLY REALLY HAVE IT
timotimo :D
do we warn or err on things that make no sense to HAS?
FROGGS no
timotimo do you see a way to do it? if so, you could tell me to do it and i'll try my best 19:26
like, only allow HAS inside CStruct and CUnion REPR'd classes?
FROGGS I dunno if I get my hands on the repr at that time
19:26 Ven joined
timotimo is moar's release already cut? 19:26
FROGGS timotimo: you can try to call .package.repr on $attr
timotimo if not, we could throw inside P6Opaque's compose reprop
timotimo pulls things 19:27
FROGGS nah, let's do it as early as we can, in NativeCall.pm
timotimo agreed
HAS is declared in NativeCall?
timotimo looks at the commit 19:28
oh, i get it
HAS gets slang'd into our main language by use-ing NativeCall
i like that
dalek ast: a30ffc1 | usev6++ | S (2 files):
Change test for 'NaN ** 0'

  ... fixes RT #124450
FROGGS timotimo: I don't want to pollute Perl6::Grammar directly 19:29
timotimo aye
tony-o_ does nqp::readfh read nothing from STDIN until it's closed? 19:31
nvm - might be a problem with it in this start block 19:33
timotimo FROGGS: yes, i can get the repr there 19:34
the only ones acceptable are CStruct and CRepr?
dalek kudo/nom: 83f09e0 | lizmat++ | src/Perl6/World.nqp:
Make sure @?INC strings are properly ro
19:35 hernan left
FROGGS timotimo: that are the only ones it is implemented for 19:35
so, yes
timotimo so ... die or warn?
FROGGS die
timotimo k
X::NYI?
FROGGS hmmm, no, that might be misleading
or.. maybe yes 19:36
yes, NYI might make sense
t/spec/S32-exceptions/misc.rakudo.moar (Wstat: 0 Tests: 305 Failed: 1) 19:37
Failed test: 304
Parse errors: Tests out of sequence. Found (1) but expected (187)
does somebody else get that too? 19:38
lizmat regularly, usually for a single file in the spectest
lemme check again 19:39
tony-o_ m: my $p = Proc::Async.new("perl6", "-e", "await start \{ $*IN.read(1).say; \};"); my $r = $p.start; $p.write("0".encode); await $r;
camelia rakudo-moar b86560: OUTPUT«Cannot use a Buf as a string, but you called the Stringy method on it␤ in block <unit> at /tmp/t1CrH1Kiyd:1␤ in any <unit-outer> at /tmp/t1CrH1Kiyd:1␤␤»
FROGGS I got this in the run before too, same file, saying Found (2) but expected (187) or so 19:40
lizmat FROGGS: I get utter fails on JVM for nativecall tests now
tony-o_ m: my $p = Proc::Async.new("perl6", "-e", 'await start { $*IN.read(1).say; };'); my $r = $p.start; $p.write("0".encode); awa[C[C[C[C[C[C[C[C[C[C[C[C[C[Cit $r;
camelia rakudo-moar b86560: OUTPUT«5===SORRY!5=== Error while compiling /tmp/hZez1ZzdO1␤Bogus postfix␤at /tmp/hZez1ZzdO1:1␤------> 3$r = $p.start; $p.write("0".encode); awa7⏏5␤ expecting any of:␤ infix␤ infix stopper␤ statement end␤ …»
lizmat $ perl6-j t/04-nativecall/01-argless.t
===SORRY!===
No such method 'AT-KEY' for invocant of type 'BOOTHash'
tony-o_ m: my $p = Proc::Async.new("perl6", "-e", 'await start { $*IN.read(1).say; };'); my $r = $p.start; $p.write("0".encode); await $r; 19:41
camelia rakudo-moar b86560: OUTPUT«Proc::Async is disallowed in restricted setting␤ in sub restricted at src/RESTRICTED.setting:1␤ in method new at src/RESTRICTED.setting:32␤ in block <unit> at /tmp/owVrAL9SPU:1␤ in any <unit-outer> at /tmp/owVrAL9SPU:1␤␤»
FROGGS k, I'll investigate
itz_ some interesting Bailador failures 19:42
dalek rl6-roast-data: 8af83e5 | coke++ | / (9 files):
today (automated commit)
 19:44
timotimo FROGGS: so what exception type do we want? 19:48
FROGGS timotimo: X::NYI is fine
dalek kudo/nom: c00dded | lizmat++ | t/01-sanity/54-use-lib.t:
Handle different reporting on different backends
19:49 domidumont left
timotimo couldn't find a good one 19:49
OK
though P6opaque will never get HAS, for example
i'd almost use AdHoc instead 19:50
FROGGS well, go ahead
I'm fine with that too
lizmat FROGGS: the 54-use-lib.t on jvm is fixed now
FROGGS lizmat++
awesome
lizmat timotimo: which would be just "die", right ?
timotimo yes
lizmat FROGGS: more bad news? 19:51
perl6-j t/03-jvm/01-interop.t
===SORRY!===
Can not call method 'orig' on a null object
ah, it's in World.DEPRECATED
looking at that one now as well 19:52
FROGGS yeah, that one is new
tony-o_ submitted a bug for that proc async thing
[Coke] roast a30ffc1 - we have the same NaN test in two different files. let's kill one of them?
19:52 dolmen joined
FROGGS tony-o_++ 19:52
lizmat seems it is caused by use java::lang::String:from<java>; 19:53
I did change that the otther day
tony-o_ it happens with channels and sockets too
19:53 brrt joined
timotimo i'll go with a regular X::AdHoc for now 19:56
19:57 xfix left
dalek kudo/nom: 36a6de1 | FROGGS++ | lib/NativeCall.pm:
fix accessing %*LANG<MAIN> on jvm
 19:57
FROGGS lizmat: nativecall tests are clean now 19:58
timotimo Can only use HAS on classes with repr CStruct or CUnion, not CUnion
;(
FROGGS O.o
that's... unfortunate
how do you check it? 19:59
and it should mention that it was applied on attributes..
timotimo perl6 -e 'use NativeCall; class Foo is repr<CUnion> { HAS $.foo }'
right 20:00
if $attr.package.REPR ne 'CStruct' || $attr.package.REPR ne 'CUnion' {
die "Can only use HAS on classes with repr CStruct or CUnion, not " ~ $attr.package.REPR;
}
FROGGS Can only declare HAS scoped attributes in classes of repr CStruct and CUnion
s/||/&&/
timotimo wait
that... yes
FROGGS :P
how you such mistake? :D 20:01
this thinko can bring down entire servers
timotimo distracted
20:01 salv0 joined
timotimo one of my hands has been monopolized 20:02
FROGGS cats...
timotimo should it also warn on int and such? 20:03
FROGGS you can put attribute inside them? 20:04
timotimo hm?
FROGGS ahh, no, gotcha
timotimo use NativeCall; class Foo is repr<CUnion> { HAS num $.foo }
FROGGS it should state that this is a noop I think
like: class Foo { has $!foo is rw }
m: class Foo { has $!foo is rw }
camelia rakudo-moar 83f09e: OUTPUT«useless use of 'is rw' on $!foo in any at src/Perl6/World.nqp:2499␤» 20:05
timotimo warn, thon
then
FROGGS yes, let's do the same here
timotimo k67hhhhhhhhhhhhhhhhhhhhhh
lizmat FROGGS: feels like class Foo { has $!foo is rw } is something I broke recently, so I'll look at that if you want 20:06
FROGGS lizmat: why broke?
lizmat incorrect file/line number ?
FROGGS ahhh, yes 20:07
did not look at that :o)
20:07 ugexe joined
lizmat the "from Java" err is caused by the deprecation message trying to show file/line number :-( 20:07
FROGGS ohh
20:08 [Sno] joined
lizmat it did not do so before 20:08
changed that with 484e9f708812aaa82ca7f33b8
HLL::Compiler.lineof($/.orig, $/.from, :cache(1)) 20:09
seems to be the offender
20:09 andreoss joined
FROGGS I can attempt to fix if you want 20:09
lizmat the above line should work if we have a $/ there, doesn't it ? 20:10
FROGGS yes
lizmat ah, but I guess we don't know for sure
FROGGS umm, we certainly have no $/ there 20:11
timotimo Using a HAS scope for an attribute typed num is useless. in block at lib/NativeCall.pm:478
looks good?
lizmat ok, checking for $/ then :-)
FROGGS timotimo: sounds weird
timotimo i typed it with one hand :P
FROGGS lizmat: we cannot have $/ there, it does not get passed to the method 20:12
timotimo yeah, i'll spell it more like "useless use of"
FROGGS yeah, exactly :o)
lizmat FROGGS: aaahhh
timotimo Useless use of HAS scope on an attribute with type num. in block at lib/NativeCall.pm:478
20:13 darutoko- left
FROGGS timotimo: it is good enough for me, but you know, me English is not the best :o) 20:13
timotimo :)
you're not the best english, are you? 20:14
FROGGS lizmat: I think I know of a patch
lizmat FROGGS: me too, working on it :-)
dalek kudo/nom: 65aaeec | timotimo++ | lib/NativeCall.pm:
some checks for HAS for reprs and types
20:15 chansen_ joined
timotimo i hope this is satisfactory 20:21
FROGGS timotimo++ # it is :o) 20:22
andreoss does BUILDALL call ordinary BUILD with callsame?
timotimo let me see 20:23
it does not
it doesn't pass on the @autovivs parameter, so it calls with (self, |%attrinit) instead 20:24
andreoss so BUILDALL won't work with multiple BUILDs
dalek kudo/nom: 498ce69 | lizmat++ | src/Perl6/ (2 files):
Handle java deprecation message properly
20:24 dolmen left 20:26 isBEKaml left
andreoss timotimo: ? 20:26
brrt no i don't believe that's right 20:28
BUILD is a submethod iirc
BUILDALL is a method 20:29
timotimo BUILDALL explicitly works with multiple BUILDs
FROGGS lizmat: that's exactly what I would have done :o)
brrt doc.perl6.org/language/objects#Obje...nstruction explains it well enough :-)
lizmat FROGGS: hehe 20:30
FROGGS: btw, make test is now clean for me on JVM
PerlJam Unless my ack-fu is weak, BUILDALL doesn't appear in the synopses.
timotimo that's right, BUILDALL is a method in Mu that doesn't get replaced very often
FROGGS still rebuilds
lizmat was glad hearing the new MacBook Pro was out 20:31
lizmat was sad to see there was no change in CPU options :-(
FROGGS lizmat: what type of change were you hoping for? 20:32
lizmat 20% faster CPU
FROGGS ahh
tony-o_ lizmat i'm pretty not happy with the imac i just bought but i'm out of the 14 day window
brrt lizmat: buy an wintel machine, install linux, be happy ;-)
tony-o_ my 3 year old macbook pro is still faster/smoother 20:33
lizmat brrt: and the wintels have faster CPU's ?
timotimo lizmat: on wintels you can freely choose your CPU :)
brrt some do :-)
timotimo well, not necessarily on a laptop
lizmat at home we're a mixed Mac / Wintel->Linux place
timotimo there's probably no laptop with a i7 5960X :) 20:34
lizmat and I hear myself screaming a lot less at my machine and rebooting a lot less than the Linux user in this house
PerlJam lizmat: that might be a function of the user rather than the OS :) 20:35
brrt apple website says just '2.7Ghz quad core i7
timotimo PerlJam: to be fair, linux has its nooks and crannies that can drive you up the wall
lizmat true, and I *am* going deaf, so there may be other factors involved
anyway, it's not a good example for me :-)
geekosaur reboots linux and os x about as often. linux would be less often if it weren't limited to 4GB RAM (and hosted under OS X...)
lizmat brrt: there's a 2.9Ghz i7 option
PerlJam timotimo: indeed. Graphics cards always seem to be hit-or-miss with me on linux 20:36
brrt which would be which model exactly?
andreoss can't have proto on BUILD?
timotimo graphics cards? i've had very little issue with my nvidia cards and the intel integrated gpus for a long time now
the intel integrated gpus have superb drivers and feature sets, but they are not exactly powerful
lizmat (hmmm.. I meant 2.8Ghz i7 option) 20:37
skids timotimo: no surprise there it's the ATIs that kill you.
andreoss m: class Y { proto submethod BUILD(|) { {*} } ; multi submethod BUILD($a) {} }
camelia rakudo-moar 36a6de: OUTPUT«===SORRY!===␤Package 'Y' already has a Method 'BUILD' (did you mean to declare a multi-method?)␤»
lizmat which is what I have noew already
timotimo i've stayed away from ati for a long time
i'm not sure i ever had an ati at all
skids I went through a phase of wanting to like them. It wore off after repeated exposure.
lizmat afk for a bit& 20:38
20:38 diana_olhovik_ left
brrt see you 20:38
jercos I haven't had good luck with ATi cards, but my AMD cards have always been fine on Linux.
brrt (i still have failed to actually use the GPU now that I have a real one) 20:39
timotimo AMD has the superior open source driver scheme; has for at least 2 years i believe
andreoss i'm trying to call BUILD explicitly from BUILDALL, but hash with arguments seems to have keys in random order so the signatures don't match. 20:42
skids Maybe, but every radeon I've used has had issues with bus timing or overheat. There always was an excuse why some part of the hardware could not be openly documented, and the result was always instability.
I'm on Intel now and won;t even look at anything else until I want a gaming system again. 20:43
brrt (i can't imagine wanting >2,8Ghz on a laptop though :-))
20:43 molaf_ left
timotimo ah. i want to do a bit of gaming every now and then - also some game development hopefully at some point ;) 20:43
skids I'll be jumping back in with pure-linux steamboxen eventually. Probably nvidia. 20:44
brrt i'm sooo avoiding homework today :-)
20:45 diana_olhovik_ joined, dolmen joined
brrt nouveau driver works wel enough, it seems, but the optimus / bumblebee stuff doesn't work at al 20:45
courtesy of **** selinux
timotimo aaw
well, modern GPUs and security is kind of mutually exclusive anyway 20:46
brrt have i not root for this machine? 20:47
in selinux lands, machine has root of you
FROGGS lizmat: I can confirm that 'make test' is clean on jvm and moar 20:48
lizmat :-) 20:49
FROGGS I am finishing going through all commits of the month and update the changelog and will then rest a bit
brrt ok, i'm going to poll a bit bigger; who thinks it'd be fun/interesting if i did a talk on JIT compilation at YAPC::EU? 20:50
FROGGS +1
that will attract ppl I think
PerlJam brrt: I'm not going to be at YAPC::EU, but I would look forward to the video of your talk :)
brrt ah well. then i'll just propose something :-) 20:51
lizmat FROGGS: t/04-nativecall/02-simple-args.t still has 2 fails for me on Moar 20:52
FROGGS lizmat: yes, I looked at that today but I cannot sport anything wrong...
lizmat so it's only on OS X ?
FROGGS given it works on 32/64bit linux and 32/64bit windows 20:53
yeah
seems like
perhaps any BSD
I can test that on an openbsd in a bit
20:55 skids left
lizmat FROGGS: in CompileTestLib, there's a if $VM.name eq 'parrot' 20:55
I guess it doesn't hurt to keep it? 20:56
FROGGS you can remove it if you want
today I doubt that there will be volunteers that revive parrot 20:57
dalek kudo/nom: 1d803b8 | lizmat++ | t/04-nativecall/CompileTestLib.pm:
Remove some more feathers
lizmat FROGGS: looks like x is -2 inside TakeUint8 / TakeUint16 20:59
brrt re:talk, 20 minutes or 50? 21:01
PerlJam brrt: if you can fill 50 minutes, do it! :)
lizmat FROGGS: sorry, that seems wrong
FROGGS lizmat: yes, I expected that it somehow messes up unsigned vs signed
lizmat when I add a "printf("x = %d\n",x);"
I get 254 and 65534 respectively 21:02
FROGGS but that would be correct
brrt hmm, there is plenty to talk about for 50 minutes or longer; but that also moves up the level of detail
lizmat FROGGS: indeed, so the check seems the problem 21:03
FROGGS there is no clang warning?
brrt also, now we're on the topic of hardware. i'm looking for a new phone. budget at most E150, preferably android or firefox phone or something like that
lizmat not that I can see
FROGGS what about changing it to: if ((unsigned char)x == (unsigned char)0xFE) 21:04
lizmat will try
FROGGS tests are clean on openbsd fwiw 21:05
lizmat nope, no difference 21:06
tony-o_ firefox phone looked like it had the right idea
shame it was laggy like android 21:07
itz_ there are cheap chinese android phones but a 2nd hand Nexus 4 or similar might have more cyanogenmod etc support 21:08
FROGGS lizmat: stackoverflow says we need to cast both to unsigned int... can you try that?
lizmat trying 21:09
nah :-(
FROGGS hmpf
21:10 spider-mario left
lizmat if I put a 'printf("x = %d = %d\n",x,0xFE);' 21:11
I get: x = 254 = 254
tony-o_ the nexus 4 with cyanogenmod is pretty nice itz_ brrt
brrt firefox phone looks like it's unavailable
andreoss m: my $x = ''; say so $x.IO.e; 21:12
camelia rakudo-moar 498ce6: OUTPUT«True␤»
21:12 Ven left
FROGGS brrt: do you know why this check fails? 21:12
brrt: this github.com/rakudo/rakudo/blob/nom/...args.c#L80
itz_ you could be the one person to get a ubuntu phone :)
lizmat andreoss: looks like "" refers to the current dir 21:13
21:13 spider-mario joined
brrt FROGGS - frankly i haven't seen that yet :-) i will see 21:13
lizmat m: say "".IO
camelia rakudo-moar 498ce6: OUTPUT«"/home/camelia".IO␤»
geekosaur unix used to do that, posix made it explicitly illegal 21:14
andreoss m: say "".IO leg $*CWD
camelia rakudo-moar 498ce6: OUTPUT«Less␤»
brrt 0xfe is 254 right
FROGGS m: say 0xfe == 254
camelia rakudo-moar 498ce6: OUTPUT«True␤»
timotimo i wonder why there's no 0d for decimal? 21:15
lizmat m: say "".IO.Str; say $*CWD.Str # andreoss
camelia rakudo-moar 498ce6: OUTPUT«␤/home/camelia␤»
FROGGS m: say 0d42
camelia rakudo-moar 498ce6: OUTPUT«42␤»
21:15 diana_olhovik_ left
FROGGS timotimo: ^^ 21:15
brrt FROGGS - i can imagine a bunch of things, really
i'd need gdb to figure it out 21:16
FROGGS hmmm, I an do that too
timotimo oh
FROGGS ohh no, I cant
timotimo neat!
FROGGS it is osx specific :/
brrt hmm
lizmat FROGGS: well, clang specific, I would guess
brrt i can't check further then, today
clang moarvm does it? that's interesting 21:17
but i'm afk for now
21:17 brrt left
lizmat FROGGS: FWIW, I've run both clang calls manually, and there are no warnings of any kind 21:18
clang -c -o 02-simple-args.o -fno-omit-frame-pointer -fno-optimize-sibling-calls -O3 -DNDEBUG -D_DARWIN_USE_64_BIT_INODE=1 t/04-nativecall/02-simple-args.c
clang -dynamiclib -O3 -DNDEBUG -Wl,-rpath,/Users/liz/Github/rakudo.moar/install/lib -Wl,-rpath,/Users/liz/Github/rakudo.moar/install/share/perl6/site/lib -lpthread -o 02-simple-args.dylib 02-simple-args.o
are you seeing anything funny with that ?
21:18 mj41 left
FROGGS well, only -O3 21:19
but, I'm just guessing wildly
lizmat yeah, that was just what I was trying to remove :-)
andreoss m: say shell "echo 1 2>/dev/null"; 21:21
camelia rakudo-moar 1d803b: OUTPUT«shell is disallowed in restricted setting␤ in sub restricted at src/RESTRICTED.setting:1␤ in sub shell at src/RESTRICTED.setting:15␤ in block <unit> at /tmp/MkomvNJk6M:1␤ in any <unit-outer> at /tmp/MkomvNJk6M:1␤␤»
lizmat FROGGS: removing -O3 makes it pass 21:23
also fails with -O2 and -O1 21:24
FROGGS hmmmm, so we need to change the default in moar until we know what e.... ewww
lizmat so, shall I just remove the -O3 for all ?
FROGGS that is passed from moar's config 21:25
lizmat well, I could add a named param :noopt to compile_test_lib and pass that in the case of OS X for that file ?
FROGGS no, let us skip these tests instead 21:26
lizmat skip?
not todo ? 21:27
ah yes, skip
tadzik FROGGS: I won't be able to fix aything for.the.next few weeks, I'm far away from any computer of mine :)
yoleaux 13:34Z <FROGGS> tadzik: 1) please add curlies to package decls to File::Find, Shell::Command and JSON::Tiny
13:34Z <FROGGS> tadzik: 2) the 'build reverse deps feature' builds reverse deps of not yet installed modules, but it does not install their deps
21:28 Zoffix left
FROGGS tadzik: bah 21:28
tadzik: give me a commit bit then :o)
lizmat FROGGS: I can merge some pull requests 21:29
FROGGS lizmat: for File::Find, Shell::Command and JSON::Tiny?
tadzik oh, I can do that :)
lizmat just did Shell::Command
FROGGS ahh, cool
lizmat itz already did some PR's
FROGGS tadzik: but I'd take commit bits anyway :o) 21:30
lizmat FROGGS: skipping the test would mean commenting them out, as there is no fudging in 'make test', right ?
21:30 cognominal left 21:31 cognominal joined
FROGGS lizmat: it uses Test.pm, and we can check for $*VM.compiler.backend 21:32
m: say $*VM.compiler.backend
camelia rakudo-moar 1d803b: OUTPUT«Method 'compiler' not found for invocant of class 'VM'␤ in block <unit> at /tmp/ARy_UnEhVq:1␤ in any <unit-outer> at /tmp/ARy_UnEhVq:1␤␤»
FROGGS m: say $*PERL.compiler.backend
camelia rakudo-moar 1d803b: OUTPUT«Method 'backend' not found for invocant of class 'Compiler'␤ in block <unit> at /tmp/iO3Gv58Glj:1␤ in any <unit-outer> at /tmp/iO3Gv58Glj:1␤␤»
FROGGS m: say $*PERL.compiler.^methods
camelia rakudo-moar 1d803b: OUTPUT«BUILD Str gist <anon> <anon> <anon> <anon> <anon> <anon> <anon> <anon>␤»
FROGGS m: say $*PERL.compiler
camelia rakudo-moar 1d803b: OUTPUT«rakudo (2015.4.353.g.1.d.803.b.8)␤»
FROGGS m: say $*PERL.backend
camelia rakudo-moar 1d803b: OUTPUT«Method 'backend' not found for invocant of class 'Perl'␤ in block <unit> at /tmp/539aTddq69:1␤ in any <unit-outer> at /tmp/539aTddq69:1␤␤»
FROGGS hmpf
lizmat say $*KERNEL.name
m: say $*KERNEL.name
camelia rakudo-moar 1d803b: OUTPUT«linux␤»
lizmat ? 21:33
$ 6 'say $*DISTRO.name'
macosx
perhaps ?
FROGGS yes, that will do too
dalek kudo/nom: 2b08d6f | lizmat++ | t/04-nativecall/02-simple-args.t:
Skip 0xFE/0xFFFE tests on OS X
 21:36
21:37 vendethiel left
ugexe m: role BB { method ok { ... }; method ok2 { 2 } }; class AA does BB { method ok { 1 } }; my $c = AA.new; say $c.ok; say $c.ok2; say $c.^roles; # role introspection bug? 21:37
camelia rakudo-moar 1d803b: OUTPUT«1␤2␤Method 'ok' must be implemented by BB because it is required by a role␤ in any compose_method_table at src/gen/m-Metamodel.nqp:2606␤ in any apply at src/gen/m-Metamodel.nqp:2616␤ in any compose at src/gen/m-Metamodel.nqp:2752␤ in any ma…»
lizmat hmmm... t/04-nativecall/13-union.t just flapped on me
FROGGS: otherwise, make test is now clean for me as well on moar
FROGGS nice :o) 21:38
eww, flapped?
lizmat yeah: t/04-nativecall/13-union.t (Wstat: 65280 Tests: 26 Failed: 0) 21:39
Non-zero exit status: 255
Parse errors: Bad plan. You planned 28 tests but ran 26.
FROGGS phew, have now 114 changelog lines I need to add and dedup/rephrase 21:40
lizmat well, I hope it is less work collecting all of the 114 in the first place ? 21:41
FROGGS no, took a while
will get some rest now... 21:42
gnight lizmat
21:42 FROGGS left
timotimo good night froggs :) 21:43
lizmat gnight FROGGS 21:44
21:52 cognominal left, cognominal joined 21:59 eli-se left
lizmat good night, #perl6! 22:04
22:06 vendethiel joined 22:07 tadzik left
timotimo gnite lizmat :) 22:09
22:12 RabidGravy left
andreoss what the difference between "<->" and "->"? 22:12
*is between
timotimo <-> makes the parameter "is rw" 22:13
my @a = 1, 2, 3; for @a <-> $_ { $_ += 5 }; say @a
m: my @a = 1, 2, 3; for @a <-> $_ { $_ += 5 }; say @a
camelia rakudo-moar 2b08d6: OUTPUT«6 7 8␤»
timotimo m: my @a = 1, 2, 3; for @a -> $_ { $_ += 5 }; say @a
camelia rakudo-moar 2b08d6: OUTPUT«Cannot assign to a readonly variable or a value␤ in block <unit> at /tmp/I_BYqr_twe:1␤ in any <unit-outer> at /tmp/I_BYqr_twe:1␤␤»
timotimo note that <-> $_ is the default for for loops
m: my @a = 1, 2, 3; for @a { $_ += 5 }; say @a
camelia rakudo-moar 2b08d6: OUTPUT«6 7 8␤»
22:14 raiph left
andreoss thanks 22:14
22:18 Alina-malina left 22:21 raiph joined 22:23 Alina-malina joined 22:30 sue__ left 22:31 dolmen left 22:34 tadzik joined 22:39 skids joined 22:47 vendethiel left 22:51 bjz left 23:02 raiph left 23:08 kaare__ left 23:09 raiph joined 23:14 andreoss left 23:25 xiaomiao joined 23:32 gfldex left 23:35 telex left 23:36 telex joined 23:40 jepeway left 23:47 jepeway joined 23:50 mohij 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!