Perl 6 - the future is here, just unevenly distributed

IRC log for #perl6, 2015-04-13

Perl 6 | Reference Documentation | Rakudo

| Channels | #perl6 index | Today | | Search | Google Search | Plain-Text | summary

All times shown according to UTC.

Time Nick Message
00:20 raydiak do we have a way to remove a non-empty directory without recursively emptying it one file at a time?
00:30 ugexe you could do it the way steam did it... shell out a rm -rf and hope nothing bad ends up happening >:)
00:30 TimToady shell('rm -rf dir')
00:33 ugexe then again, thats just recursively emptying the folder too
00:35 TimToady there is no other way under Unix, unless you want to damage the filesystem
00:36 raydiak guess that answers that, then (btw anyone reading along wanting to reliably and portably remove dirs could use rm_rf() from Shell::Command)
00:38 ugexe using the hyperoperator you can do it in a few lines anyway
00:40 nbdsp joined #perl6
00:40 tinyblak joined #perl6
00:40 nbdsp Greetings! Could someone advise please how to insert binary data into MySQL table with DBIish so that it wasn't interpreted as Unicode? When I try the following: my $q = "\x79\x80"; $st = $db.prepare("INSERT INTO main (hash) VALUES('$q')"); $st.execute(); I always get in the DB table 3 bytes: 0x79, 0xc2, 0x80
00:46 laouji joined #perl6
00:47 colomon joined #perl6
00:51 TimToady .encode('latin-1') or so?
00:52 TimToady or is the interface enforcing utf-8?
00:57 nbdsp I can't find in the DBIsh connection parameters relating to charset. Maybe someone uses DBIish for such kind of tasks?
00:59 simcop2387 joined #perl6
01:00 nbdsp When I try the following: my $q = "\x79\x80"; $st = $db.prepare("INSERT INTO main (hash) VALUES('$q.encode('latin-1')"); $st.execute(); there is an error "Cannot use Bud as a string, but you called Stringy method on it."
01:00 anaeem1_ joined #perl6
01:00 nbdsp "Cannot use Buf as string..."
01:00 bobkare nbdsp: have you tried passing the data via a placeholder instead?
01:01 nbdsp Yes. The same result.
01:01 aborazmeh joined #perl6
01:01 aborazmeh joined #perl6
01:16 nbdsp from mysql site: "mysql_query() cannot be used for statements that contain binary data;" and DBIish uses mysql_query. :(
01:17 gtodd aww perl8.org is gone
01:22 laouji joined #perl6
01:24 BenGoldberg joined #perl6
01:24 vendethiel joined #perl6
01:29 yeahnoob joined #perl6
01:30 cognominal r: say []^.roles
01:30 camelia rakudo-moar 8e96e4: OUTPUT«one(, (Any))␤»
01:30 camelia ..rakudo-jvm 8e96e4: OUTPUT«cannot connect to eval server: Connection refused␤»
01:31 cognominal r: say .WHAT for  []^.roles
01:31 camelia rakudo-moar 8e96e4: OUTPUT«(Junction)␤»
01:31 camelia ..rakudo-jvm 8e96e4: OUTPUT«cannot connect to eval server: Connection refused␤»
01:45 gtodd left #perl6
01:59 elf_ joined #perl6
02:13 chenryn joined #perl6
02:13 noganex joined #perl6
02:26 SHODAN joined #perl6
02:34 aborazmeh joined #perl6
02:34 aborazmeh joined #perl6
02:42 chenryn joined #perl6
03:50 tony-o moritz: that event::emitter is available on panda now if you were looking to use it - if not thanks for the heads up that i boogered up the meta
04:09 vendethiel joined #perl6
04:17 anaeem1 joined #perl6
04:20 anaeem1_ joined #perl6
04:30 kaare_ joined #perl6
04:44 chenryn joined #perl6
05:01 moritz tony-o: thanks for fixing it; I noticed it went missing when updating https://github.com/moritz/perl6-all-modules
05:06 elf_ left #perl6
05:13 chenryn joined #perl6
05:23 KCL_ joined #perl6
05:35 virtualsue joined #perl6
05:44 espadrine joined #perl6
05:45 Psyche^ joined #perl6
05:46 chenryn joined #perl6
06:09 mr-foobar joined #perl6
06:13 jnthn morning o/
06:15 Woodi joined #perl6
06:17 rindolf joined #perl6
06:33 labster \o morning jnthn
06:39 dalek roast: e6f9b8c | usev6++ | S02-literals/heredocs.t:
06:39 dalek roast: Add reference to RT #117853
06:39 dalek roast: review: https://github.com/perl6/roast/commit/e6f9b8c0e1
06:39 synbot6 Link:  https://rt.perl.org/rt3//Public/Bug/Display.html?id=117853
06:45 masak morning, #perl6
06:46 moritz \o
06:48 labster m: say Str.chars
06:48 camelia rakudo-moar 8e96e4: OUTPUT«use of uninitialized value of type Str in string context  in block <unit> at /tmp/HEdB8nACCY:1␤␤0␤»
06:49 RabidGravy joined #perl6
06:50 labster I suppose that seems like the right behavior.
06:51 masak yes.
06:51 masak the Str type object has 0 characters :)
06:51 masak and a warning, because don't do that
06:52 labster I was thinking of adding a method chars to Str.pm, because I think I can get 4% faster without the unboxing.
06:53 labster which is not exactly LHF, but I still feel like .chars is a bit too slow
06:53 chenryn joined #perl6
06:55 jnthn Yes, Str.chars should exist
06:56 jnthn Probably as a multi with :D and :U on the invocant
06:56 jnthn Then the :D one is just nqp::chars(self)
06:56 jnthn Or nqp::chars(nqp::unbox_s(self))
06:58 awwaiid joined #perl6
06:58 labster :U is the confusing one.  multi method chars (Str:U) { self.Str; 0 }  ?
06:59 labster well, that might happen automatically with Cool's version.  I'll give it a shot.
06:59 chenryn joined #perl6
06:59 moritz labster: I'd guess that the self.Str is there just to generate the warning
06:59 labster that was my intent
06:59 jnthn It's a fair enough way to generate it.
07:00 moritz but please add a comment to it
07:00 dalek roast: d561d49 | usev6++ | S04-declarations/my.t:
07:00 dalek roast: Add test for RT #115916
07:00 dalek roast: review: https://github.com/perl6/roast/commit/d561d4977c
07:00 synbot6 Link:  https://rt.perl.org/rt3//Public/Bug/Display.html?id=115916
07:02 * jnthn notes we're down to 670 RTs
07:02 laouji_ joined #perl6
07:02 masak \o/
07:02 jnthn I hope we can get down to 650 and below during April
07:02 bartolin \o/ jnthn++
07:02 masak it's been way higher than 670, hasn't it?
07:02 moritz bartolin++ # relentless RT work
07:02 masak have we ever pushed 1000?
07:02 Ven joined #perl6
07:02 jnthn It's been over 700 before, for sure
07:02 masak oh, ok.
07:02 masak bartolin++ indeed
07:02 labster \o/ indeed
07:03 bartolin it has been near 800 last autumn :-)
07:03 jnthn Wow :)
07:03 vendethiel joined #perl6
07:03 * jnthn wonders if there's a way to get RT to spit out stats on that :)
07:03 masak bartolin: ok, you push it to 0, I push it to Inf. first one to get there wins :P
07:03 bartolin *g*
07:04 bartolin would be a ... tie?
07:04 labster bartolin++ keeps filling up my inbox with RT closures
07:05 bartolin yeah, at times I think I'm a bit noisy at perl6-compiler
07:05 * masak .oO( it's because of all these RT closures that we have to allocate stuff on the heap! ) :P
07:05 masak bartolin: it's good noise.
07:05 masak I don't mind p6c being noisy with RT stuff.
07:06 zakharyas joined #perl6
07:06 masak for each RT ticket, there's at least one person who cares about the ticket.
07:06 moritz and if we do start to consider it noise, we'll simply have RT not send those mails to p6c
07:07 Alina-malina joined #perl6
07:07 bartolin moritz: I'm Cc:'ing p6c most of the time manually, though
07:08 jnthn I thought they went to p6c automatically?
07:09 * moritz too
07:13 jnthn hm, seems my blog post didn't make it to pl6anet.org
07:14 Ven joined #perl6
07:15 jnthn itz_: Is pl6anet yours? Any chance of getting https://6guts.wordpress.com/ there, now I'm actively Perl 6 blogging again? :)
07:15 moritz jnthn++ # https://6guts.wordpress.com/2015/04/12/this-week-unicode-normalization-many-rts/
07:17 RabidGravy m: say homedir('.')
07:17 camelia rakudo-moar 8e96e4: OUTPUT«Dynamic variable $*HOME not found␤  in method <anon> at src/gen/m-CORE.setting:16089␤  in any find_method_fallback at src/gen/m-Metamodel.nqp:2908␤  in any find_method at src/gen/m-Metamodel.nqp:1052␤  in sub homedir at src/gen/m-CORE.setting:19088…»
07:18 RabidGravy where would $*HOME be best implemented?  In Process.pm with most of the dynvars?
07:22 Rounin joined #perl6
07:29 FROGGS joined #perl6
07:29 bjz joined #perl6
07:34 awwaiid joined #perl6
07:41 nwc10 OMG bloggage
07:42 abraxxa joined #perl6
07:46 Ven joined #perl6
07:47 darutoko joined #perl6
07:55 yvan1 joined #perl6
07:57 yakudza joined #perl6
08:05 chenryn joined #perl6
08:12 FROGGS .schnickschnack
08:12 yoleaux 11 Apr 2015 16:25Z <jnthn> FROGGS: Please undo removing MODE_UNBOX_STR_NOT_EMPTY_OR_ZERO; it's probably useful for, for example, v5.
08:12 yoleaux 11 Apr 2015 16:33Z <nwc10> FROGGS: Ooops you're innocent. Please try harder next time :-)
08:12 yoleaux 11 Apr 2015 16:35Z <nwc10> FROGGS: Guess what, Problem *still* Exists Between (my) Keyboard And Chair.
08:12 espadrine joined #perl6
08:12 FROGGS is the jvm build still borken? and if so, how can I reproduce it?
08:13 cschwenz joined #perl6
08:15 nwc10 FROGGS: I believe that jnthn was able to fix it
08:15 FROGGS ahh cool
08:15 FROGGS need to backlog the entire weekend...
08:16 nwc10 Friday's party was *that* good? :-)
08:16 moritz python, python
08:16 moritz ImportError: cannot import name 'OperationalLevel'
08:16 FROGGS nwc10: I had to build a fence, almost all on my own :/
08:16 moritz how about telling me *why* you can't import it?
08:16 nwc10 where "Almost" was a LTA thing, because you had small assistants?
08:17 [ptc] moritz: PYTHONPATH?
08:17 FROGGS I did not even know until then where one can have aching muscles
08:17 FROGGS nwc10: no, the owner of the fence helped
08:17 ab5tract_ joined #perl6
08:17 FROGGS but there are things that only one person can do
08:17 FROGGS and sadly that was me
08:18 moritz [ptc]: I guess the symbol that I want to import simply isn't there in the file
08:18 moritz [ptc]: but if that's the case, python could *tell me that*
08:18 FROGGS and now I see who broke the jvm build :/
08:18 [ptc] moritz: I feel your pain!  Python is my $dayjob
08:18 moritz python's error message simply aren't written from a user's perspective
08:19 [ptc] torture the user on behalf of the implementor
08:20 FROGGS moritz: that's often the case
08:20 FROGGS moritz: look at visual basic for example
08:20 dalek roast: 223db17 | usev6++ | S14-roles/parameterized-type.t:
08:20 dalek roast: Fix test for RT #68136
08:20 dalek roast: ... Mouq++ for pointing out the correct syntax
08:20 dalek roast: review: https://github.com/perl6/roast/commit/223db17d80
08:20 synbot6 Link:  https://rt.perl.org/rt3//Public/Bug/Display.html?id=68136
08:28 dalek rakudo/nom: e28b43d | labster++ | src/core/Str.pm:
08:28 dalek rakudo/nom: Add Str.chars instead of inheriting Cool.chars
08:28 dalek rakudo/nom:
08:28 dalek rakudo/nom: .chars on strings now runs 41% faster
08:28 dalek rakudo/nom: review: https://github.com/rakudo/rakudo/commit/e28b43dc6a
08:28 bjz joined #perl6
08:29 labster I think that helped out the optimizer quite a bit more than I expected.
08:29 dakkar joined #perl6
08:29 azawawi joined #perl6
08:29 azawawi hi
08:30 azawawi What are the recommended Perl 6 syntax highlighter(s) these days?
08:31 Ven joined #perl6
08:36 [ptc] azawawi: the vim syntax highlighter is good, although you'll need to use the very latest version (https://github.com/vim-perl/vim-perl/blob/master/syntax/perl6.vim) since the one shipped with e.g. Debian is very slow
08:37 azawawi i see thx
08:40 laouji joined #perl6
08:41 maddingue joined #perl6
08:42 [ptc] azawawi: if you just want to do syntax highlighting (in HTML, say), then pygments (pygments.org) provides a Perl6 highlighting module
08:42 ab5tract_ wasn't someone getting a 'textmate' coloring going? which would theoretically give us our github coloring back?
08:43 RabidGravy now a few more of you have woken up, where should $*HOME be implemented?
08:44 RabidGravy m: homedir('.')
08:44 camelia rakudo-moar 8e96e4: OUTPUT«Dynamic variable $*HOME not found␤  in method <anon> at src/gen/m-CORE.setting:16089␤  in any find_method_fallback at src/gen/m-Metamodel.nqp:2908␤  in any find_method at src/gen/m-Metamodel.nqp:1052␤  in sub homedir at src/gen/m-CORE.setting:19088…»
08:44 telex joined #perl6
09:02 bjz joined #perl6
09:07 pecastro joined #perl6
09:10 Woodi joined #perl6
09:22 vendethiel joined #perl6
09:31 cognominal joined #perl6
09:34 mr-foobar joined #perl6
09:36 dalek Heuristic branch merge: pushed 22 commits to rakudo/newio by lizmat
09:41 pecastro joined #perl6
09:44 rurban_ joined #perl6
09:45 yqt joined #perl6
09:46 chenryn joined #perl6
09:50 cognominal joined #perl6
09:59 yqt joined #perl6
09:59 laouji_ joined #perl6
10:26 tadzik joined #perl6
10:27 tadzik joined #perl6
10:30 kinslayer joined #perl6
10:32 cognominal joined #perl6
10:32 thilp joined #perl6
10:38 silug joined #perl6
10:38 cschwenz joined #perl6
10:41 yqt joined #perl6
10:47 raiph joined #perl6
10:48 Ven joined #perl6
10:52 * tadzik looks at https://github.com/tadzik/panda/issues/148, puzzled
10:52 tadzik ...how did it ever work? Was it always that way?
10:53 tadzik oh, withp6lib :)
10:57 tadzik thanks to ugexe++, rakudobrew now works on windows :)
11:00 colomon ooooo
11:00 colomon what does it need to build?
11:00 colomon (or rather, maybe, what does Rakudo need to build?)
11:01 tadzik Perl only, I think
11:02 tadzik er, wait
11:02 tadzik yeah, think so
11:02 tadzik try it out, I don't have windows :)
11:02 |Tux| tadzik, Text::CSV does not run the tests under panda and then does not install
11:02 tadzik [Tux]: ok, looking
11:02 |Tux| if you can spot the reason why, I owe you another beer
11:03 tadzik :)
11:03 tadzik [Tux]: is that related to the prove -Ilib ticket you opened?
11:03 |Tux| yep
11:03 tadzik and fixed by your patch?
11:03 |Tux| that is why I came here
11:03 tadzik [Tux]: it just tested and installed correctly for me
11:03 |Tux| it fixed the panda install Text::CSV, but I saw core dumps later
11:03 tadzik aside from the fact that it does not have provides
11:04 |Tux| but I have no idea if that was related
11:04 tadzik so it cannot be loaded properly
11:04 tadzik let's see if I can fix that
11:04 |Tux| it *does* have provides!
11:04 tadzik oh?
11:04 tadzik oh, it does! Very weird
11:05 |Tux| https://github.com/Tux/CSV/blob/master/META.info#L13
11:05 tadzik yeah
11:05 tadzik wtf
11:05 tadzik also, [Tux]++ for a nice, proper META.info :)
11:05 diana_olhovik_ joined #perl6
11:06 tadzik ok, this is crazy. When I install it with 'panda install .', it doesn't test anyore
11:06 * |Tux| is unsure if the test-depends also requires the entries in depends
11:06 tadzik 'panda install Text::CSV' works fine except it doesn't see "provides"
11:06 chenryn joined #perl6
11:06 tadzik [Tux]: I think test-depends is a superset of depends :)
11:06 tadzik by definition
11:07 |Tux| but you understand why I lost track
11:07 tadzik clearly
11:07 tadzik I'll set up my coffee and get on it
11:12 dalek rakudo-star-daily: 7faca23 | coke++ | log/ (2 files):
11:12 dalek rakudo-star-daily: today (automated commit)
11:12 dalek rakudo-star-daily: review: https://github.com/coke/rakudo-star-daily/commit/7faca231bc
11:12 dalek perl6-roast-data: ce236c5 | coke++ | / (9 files):
11:12 dalek perl6-roast-data: today (automated commit)
11:12 dalek perl6-roast-data: review: https://github.com/coke/perl6-roast-data/commit/ce236c5104
11:14 raiph .tell RabidGravy http://irclog.perlgeek.de/perl6/2015-02-14#i_10117483
11:14 yoleaux raiph: I'll pass your message to RabidGravy.
11:14 [Coke] 2 files failing on nqp parrot now.
11:15 [Coke] integration/advent2012-day04.t failing on all rakudo backends
11:15 [Coke] S12-attributes/native.t failing on rakudo-nojit
11:16 [Coke] S17 flappers still present.
11:16 thilp left #perl6
11:16 [Coke] lots more failures on jvm still.
11:17 tadzik grr, I should test PRs before merging them
11:25 tadzik [Tux]: ok, this is quite interesting. Your tests segfault rakudo, but crash the debugger
11:25 tadzik (perl6 debugger)
11:25 tadzik with a mysterious yet scary "Unhandled exception: ctxlexpad needs an MVMContext"
11:25 |Tux| yeah, that is where I stopped too
11:25 |Tux| waaaaay out of my comfort zone
11:26 tadzik haha
11:26 tadzik at least it consistently crashes on Text::CSV.new
11:29 [Coke] (rt, p6c - no, that's not automatic. you have to opt in every time you comment on a ticket via the webterface)
11:31 tadzik |Tux|: it gets better: I can't even precompile Text::CSV
11:31 tadzik perl6 --target=mbc -o blib/lib/Text/CSV.mbc lib/Text/CSV.pm says: Cannot dump this object; no dump method
11:31 tadzik good news is, Text::CSV works perfectly fine when not precompiled
11:31 * |Tux| looks over his shoulder, feeling haunted
11:31 tadzik so it is *just* a precompilation bug
11:32 tadzik I'm sure jnthn'll love it :P
11:32 |Tux| LOL
11:32 fhelmberger joined #perl6
11:35 Ven joined #perl6
11:39 masak I suspected precompilation might be involved somehow.
11:39 masak nice detective work, |Tux|++ and tadzik++
11:40 masak can you also manage to golf the precompilation thing, preferably with as few module dependencies as possible? kthx
11:41 * masak doesn't even care all that much, he just likes it when things are golfed :>
11:41 tadzik ah, masak, asking the hard questions...
11:41 masak life is the hard teacher. I'm just a TA. :P
11:42 tadzik TeAcher? :P
11:42 masak teaching assistant
11:42 tadzik . o O ( You can't have teacher without tea )
11:42 masak tadzik: maybe "amanuensis" or something like that in your quaint language? :P
11:42 tadzik :o
11:42 tadzik An amanuensis (/əˌmænjuːˈɛnsɪs/) is a person employed to write or type what another dictates
11:43 tadzik A student!
11:43 masak haha
11:43 tadzik too often in the school system I'm used too
11:43 tadzik s:2nd/too/to/
11:44 tadzik "Do you want some beer?" "Yeah, give me an IPA" "Do you want some /bɪə(ɹ)/?"
11:45 masak \(ɹ)/
11:45 tadzik ᕕ( ᐛ )ᕗ
11:46 masak tadzik: you brighten my day by just being you. :)
11:46 tadzik ᕕ( ᐛ )ᕗ
11:46 tadzik oops
11:46 tadzik http://media.tumblr.com/tumblr_m7n45immio1r6eqtv.jpg
11:46 masak T.Haha
11:46 tadzik also, look what I found in another window: Stage parse      : Internal error: invalid thread ID in GC work pass
11:47 sergot it's my window, isn't it? :P
11:47 sergot I mean, a windown with a chat with me
11:47 aborazmeh joined #perl6
11:47 aborazmeh joined #perl6
11:48 sergot tadzik: any ideas? :)
11:48 tadzik no, it looks crazy
11:49 tadzik and yes, it's from your window :)
11:49 masak internal error (c) sergot 2015
11:50 sergot :D
11:52 sergot I can't get my rakudo working :'(
11:53 nbdsp joined #perl6
11:54 sergot .seen FROGGS
11:54 yoleaux I saw FROGGS 08:20Z in #perl6: <FROGGS> moritz: look at visual basic for example
11:54 FROGGS wassup?
11:55 sergot are you aware of anything that could have broken IO::Socket::SSL?
11:55 sergot tadzik's just got this: https://gist.github.com/tadzik/22999b3fbb06069ab994
11:58 FROGGS I can imagine what this is about, yeah
11:58 FROGGS I guess one passes an Int to a subroutine that expects a native int
11:59 FROGGS we should make MoarVM smarter about that
11:59 FROGGS I mean, we can pass Ints to native numeric params as long as these params are not 'is rw'
12:00 sergot thanks FROGGS I will fix it in IO::Socket::SSL as soon as I get my rakodu working
12:01 sergot or, tadzik you can fix it on your own if you're in hurry :)
12:04 abraxxa FROGGS: i'd like to see such a type coercion too
12:04 tadzik sergot: I'll put it on my todo for today :)
12:05 domidumont joined #perl6
12:07 domidumont Hello. I trying to build nqp for Debian using a pre-installed moar. The command "perl Configure.pl --backends=moar --with-moar=/usr/bin/moar" works fine (well, it does not complain). But make fails with:
12:07 domidumont make: /home/domi/debian-dev/perl6-stuff/nqp/install/bin/moar: Command not found
12:07 domidumont Any clue ?
12:08 sergot tadzik++ I will try to fix my rakudo anyway
12:09 FROGGS abraxxa: aye
12:10 yqt joined #perl6
12:10 FROGGS domidumont: I'll check that, though I cannot do that now sadly :o(
12:11 FROGGS domidumont: can you please open a gh issue? it's easier to track progress there
12:11 domidumont no problem, I have to go back to $work also. What's gh ?
12:12 FROGGS github
12:12 domidumont heh :-)
12:12 FROGGS :o)
12:12 FROGGS domidumont: can I install that moar package somehow? (on my ubuntu box)
12:13 domidumont Best bet is to download Debian source package (still in new queue) and build/install it on your system.
12:15 domidumont well, using debian's git repo is probably faster: git://git.debian.org/git/pkg-rakudo/pkg-moarvm.git
12:16 FROGGS ahh, that's probably what I am looking for... thanks
12:19 rmgk joined #perl6
12:19 domidumont FROGGS: https://github.com/perl6/nqp/issues/231
12:35 vendethiel joined #perl6
12:52 tadzik FROGGS: I wonder if the concepts of panda's --prefix and DESTDIR are even valid now that eleven is out
12:53 tadzik or they just need to work a little bit differently
12:55 FROGGS tadzik: I dunno, really :o(
12:55 tadzik :)
12:55 tadzik something to talk about on perla
12:55 tadzik perlqa*
12:55 FROGGS my head is somewhere else right now, and I wish it were not
12:55 FROGGS yeah
12:59 tadzik ohh, another precomp bug, this time in jvm %)
13:04 vendethiel joined #perl6
13:12 RabidGravy joined #perl6
13:14 Irinaioana joined #perl6
13:19 Ven joined #perl6
13:31 aborazmeh joined #perl6
13:31 aborazmeh joined #perl6
13:33 vendethiel joined #perl6
13:34 RabidGravy is there a way of making for a module to see if any deprecations have been "raised" ?
13:34 yoleaux 11:14Z <raiph> RabidGravy: http://irclog.perlgeek.de/perl6/2015-02-14#i_10117483
13:36 RabidGravy right
13:48 RabidGravy I think I'll carry on with what I am doing in Process.pm anyway as it's only small and it'll make homedir() work and can get backed out when/if newio gets merged
13:49 skids joined #perl6
13:51 RabidGravy is there a general injunction against using NativeCall in the rakudo core library? or is that no-one has got round to doing so?
13:52 RabidGravy I think any implementation of $*HOME would enormously benefit from getpwuid() ;-)
13:53 arnsholt I think using NativeCall in the core is probably a smell
13:53 arnsholt If there's something you'd need NativeCall for, we're gonna want to support for it in NQP
13:53 moritz +1
13:54 mr-foobar joined #perl6
13:54 RabidGravy that was my suspicion
13:58 lolisa joined #perl6
13:59 RabidGravy Anyhoo I have:
13:59 RabidGravy [jonathan@coriolanus rakudo]$ ./perl6
13:59 RabidGravy > say $*HOME;
13:59 RabidGravy "/home/jonathan".IO
13:59 RabidGravy > homedir("/tmp");
13:59 RabidGravy "/tmp".IO
13:59 RabidGravy versus
14:00 RabidGravy > say $*HOME;
14:00 RabidGravy Dynamic variable $*HOME not found
14:00 RabidGravy in method gist at src/gen/m-CORE.setting:15985
14:00 RabidGravy in sub say at src/gen/m-CORE.setting:18823
14:00 RabidGravy in block <unit> at <unknown file>:1
14:00 RabidGravy > homedir("/tmp");
14:00 RabidGravy Dynamic variable $*HOME not found
14:00 RabidGravy in method <anon> at src/gen/m-CORE.setting:15989
14:00 RabidGravy in any find_method_fallback at src/gen/m-Metamodel.nqp:2908
14:00 RabidGravy in any find_method at src/gen/m-Metamodel.nqp:1052
14:00 RabidGravy in sub homedir at src/gen/m-CORE.setting:18988
14:00 RabidGravy in block <unit> at <unknown file>:1
14:01 Alina-malina joined #perl6
14:06 [ptc] does anyone use the --with-moar option when configuring nqp?
14:07 mr-foobar joined #perl6
14:08 RabidGravy just PR'd that with https://github.com/rakudo/rakudo/pull/406
14:13 lizmat RabidGravy: there's $*DISTRO.is-win
14:14 lizmat also, I would drop the check on definedness...now that "0" is true
14:14 lizmat if %*ENV<HOME> -> $home { $HOME = $home } elsif $*DISTRO.is-win { $HOME = .... }
14:15 * sjn is looking forward to the qa hackathon, to hack on install/deploy stuff with lizmat and FROGGS and tadzik and anyone else :)
14:16 pecastro joined #perl6
14:16 RabidGravy lizmat++ # cheers
14:18 lizmat afk again&
14:21 RabidGravy I'll redo that PR after I've tested
14:21 RabidGravy but y'know MacGyver is on and all
14:25 RabidGravy done
14:28 eternaleye joined #perl6
14:35 caymanboy joined #perl6
14:41 espadrine joined #perl6
14:45 vendethiel joined #perl6
14:54 molaf_ joined #perl6
14:56 fhelmberger joined #perl6
14:58 * [Coke] wonders if any of the regulars have Diabetes.2, like Coke.
15:01 domidumont [ptc]: I've tried --with-moar and filled https://github.com/perl6/nqp/issues/231
15:02 nornagest joined #perl6
15:03 nornagest hi
15:05 tadzik hello nornagest
15:05 [ptc] domidumont: ah, that's good to know.  I've managed to work out what's going wrong, but it'll be a while before I can make the relevant PR
15:06 [ptc] domidumont: the issue is that the variable set in --with-moar isn't used to set any Makefile variables...  Hence even though --with-moar sets a configuration variable (and correctly), it still has no effect
15:07 grondilu joined #perl6
15:08 [ptc] domidumont: after having set the variable in the relevant Makefile.in file, NQP still can't compile since the Makefile.in assumes a path to an installed MoarVM and thus can't find files such as lib/MAST/Ops.nqp
15:08 [ptc] domidumont: thus there's a bit of yak shaving to do before --with-moar will actually work
15:09 * grondilu was looking at http://rosettacode.org/wiki/SHA-256#Perl_6 and tried to replace @Z = @Z Z[m+] @h by @Z [Z[m+]]= @h.  Amazingly enough, it worked.
15:09 aborazmeh joined #perl6
15:09 aborazmeh joined #perl6
15:09 domidumont [ptc]: right... unfortunately, this feature is required to provide nqp with moar for Debian. So an update of nqp package on Debian will be delayed until this problem is fixed
15:11 FROGGS [ptc] / domidumont: what do we want here? additional option like --moar-libs= or MOARVM_LIBS env vars?
15:11 FROGGS options*
15:11 domidumont FROGGS: options are easier to handle in debian/rules
15:11 FROGGS because I guess /usr/lib/MAST/Ops.nqp is not what you want
15:11 [ptc] domidumont: hrm, that's strange as I've got nqp, moar, rakudo working for debian
15:12 [ptc] domidumont: have a look in the os-build project for the (currently still very rough) control files
15:12 domidumont [ptc]: you have the old versions (~ 2014.07) based on parrot. I'm working on the next version based on moar and nqp 2015.03 (as dod@debian.org)
15:13 [ptc] FROGGS: --moar-libs could help; I got things to build and test by hacking Makefile-Moar.in, editing the Makefile after it was built, and running Configure.pl as well as make with an updated LD_LIBRARY_PATH
15:13 [ptc] domidumont: no I don't have the old versions; I've updated them to use moar and nqp on 2015.03, just like you :-)
15:13 [ptc] I've even got .debs floating around, however noone seems to want to test them for me
15:13 FROGGS domidumont: so we need to get the install locations straight first for moarvm, before we can tackle nqp...
15:14 domidumont ah ok.
15:14 [ptc] ... well other than the tests that I've done
15:14 domidumont FROGGS: so you mean that moar package sitting in debian new queue will need an update ?
15:15 [ptc] domidumont: do you want any help, or any information from what I've learned?  It doesn't making sense with several people reinventing the wheel
15:15 FROGGS domidumont: depends... where does it install install lib/MAST/Ops.nqp to?
15:15 FROGGS s/install//
15:15 Ven joined #perl6
15:15 domidumont FROGGS: /usr/share/nqp/lib/MAST/Ops.nqp
15:16 [ptc] domidumont: I managed to get everything to install under /usr in the debs that I built, however one big problem was a weird /usr/languages directory which gets installed.
15:16 [ptc] domidumont: I'm pretty sure Debian isn't going to like that hanging around...
15:17 domidumont [ptc]: indeed :-)
15:18 FROGGS [ptc]: that has changed recently due to moritz++
15:18 cognominal joined #perl6
15:18 [ptc] domidumont: btw: the plan with the os-build repo (under the perl6/ project on github) was to collect all os-releated build scripts in one location
15:18 dalek rakudo/nom: 0b1e4f1 | peschwa++ | t/02-rakudo/repl.t:
15:18 dalek rakudo/nom: Add a few REPL tests.
15:18 dalek rakudo/nom: review: https://github.com/rakudo/rakudo/commit/0b1e4f171c
15:19 psch hi #perl6 \o
15:19 [ptc] domidumont: we already have files to build arch linux stuff, and things are in place to get Fedora RPMs working as well
15:19 [ptc] domidumont: I just haven't had time recently to focus on that
15:19 FROGGS hi psch
15:19 [ptc] FROGGS: ah, cool, that's good to know :-)
15:19 [ptc] afk
15:20 vendethiel joined #perl6
15:22 domidumont [ptc]: well, having the same repo for upstream and Debian package file is a work in progress. Some debian tools still need some work for that.
15:23 domidumont [ptc] there's some tricky setup also: nqp master branch from your point of view must end up in upstream branch from my point of view. master is used for debian activities...
15:24 domidumont [ptc] I'll think about it. I must get up-to-speed with the current state of the art regarding sharing a repo between upstream and Debian
15:25 chenryn joined #perl6
15:30 Ven joined #perl6
15:32 dalek nqp: 16df8ce | paultcochrane++ | t/nqp/ (2 files):
15:32 dalek nqp: Unlink temporary file after having used it in tests
15:32 dalek nqp:
15:32 dalek nqp: This prevents a dirty working directory after having run the test suite.
15:32 dalek nqp: review: https://github.com/perl6/nqp/commit/16df8ce14d
15:32 dalek nqp: bbeb540 | paultcochrane++ | t/nqp/ (2 files):
15:32 dalek nqp: Convert tmp_file variable to kebab-case
15:32 dalek nqp:
15:32 dalek nqp: This is merely to be consistent with other variable names in the same files.
15:32 dalek nqp: review: https://github.com/perl6/nqp/commit/bbeb540a31
15:32 dalek nqp: 8128fa0 | moritz++ | t/nqp/ (2 files):
15:32 dalek nqp: Merge pull request #232 from paultcochrane/pr/unlink_tmp_file_after_tests
15:32 dalek nqp:
15:32 dalek nqp: Unlink tmp file after tests
15:32 dalek nqp: review: https://github.com/perl6/nqp/commit/8128fa0f1e
15:38 tinyblak joined #perl6
15:38 hoelzro o/ #perl6
15:40 domidumont [ptc]: err, forget what I said about shared repo, os-build repo does not fit this kind of setup.
15:40 [ptc] domidumont: I'm also not sure about how to share things between upstream and Debian
15:41 [ptc] domidumont: there was a discussion on irc a few weeks ago about where to put this kind of stuff and we came up with the os-build repo
15:41 TimToady for fixing ([] ... []) I found a workaround to the **@lol problem
15:42 TimToady m: sub foo (**@lol) { say nqp::iscont(@lol[0]) }; foo((1,2,3),(4,5,6))
15:42 camelia rakudo-moar e28b43: OUTPUT«1␤»
15:42 [ptc] domidumont: that might not be the best idea, however at least we were able to update the Debian control files ourselves and start to bring things back up to date with Rakudo, Moar etc
15:42 TimToady m: sub foo (|lol) { my @lol := lol.list; say nqp::iscont(@lol[0]) }; foo((1,2,3),(4,5,6))
15:42 camelia rakudo-moar e28b43: OUTPUT«0␤»
15:42 domidumont [ptc] it's a good idea, but I'm worried about the work required to keep everything up to date
15:42 [ptc] domidumont: definitely something one needs to think about carefully
15:42 [Coke] NYC, NY, US May 2nd perl hackathon; any sixers in the area, please sign up and come down!
15:42 TimToady but we'll need to figure out whether the **@lol notation should continue to itemize its sublists like that
15:43 domidumont [ptc] Be sure that I will pay attention at your work for the next versions of nqp rakudo for Debian
15:43 TimToady I know how to fix that, but lots of other things break when we do, so probably needs to be thunk along with the GLR
15:43 domidumont [ptc] At least add a readme to link the packaging repo you're getting your info from
15:44 TimToady so for now, the |lol workaround...
15:44 domidumont [ptc] for instance git://anonscm.debian.org/pkg-rakudo/pkg-rakudo.git
15:44 domidumont [ptc] gtg
15:45 [ptc] domidumont: ok, thanks!
15:45 ugexe m: my IO::Path $path = "/home/user".IO; $path = $path.resolve; # method resolve(--> IO::Path)
15:45 camelia rakudo-moar e28b43: OUTPUT«Type check failed in assignment to '$path'; expected 'IO::Path' but got 'IO::Path'␤  in block <unit> at /tmp/cPpXnJtHii:1␤␤»
15:46 ugexe am i doing something wrong here?
15:48 psch m: my IO::Path $path = "/home/user".IO;
15:48 camelia rakudo-moar e28b43: OUTPUT«Type check failed in assignment to '$path'; expected 'IO::Path' but got 'IO::Path'␤  in block <unit> at /tmp/7_lBLhwynk:1␤␤»
15:48 ugexe if i do it via perl6 -e with a perl from yesterday it says 'expected IO::Path but got Str'
15:50 TimToady m: my IO::Path $path = "/home/user".IO;
15:50 camelia rakudo-moar 0b1e4f: OUTPUT«Type check failed in assignment to '$path'; expected 'IO::Path' but got 'IO::Path'␤  in block <unit> at /tmp/WryCe03411:1␤␤»
15:50 TimToady that's the part that's failing
15:50 telex joined #perl6
15:51 TimToady m: say "/home/user".IO.WHAT;
15:51 camelia rakudo-moar 0b1e4f: OUTPUT«(IO::Path)␤»
15:51 Perl6_newbee joined #perl6
15:51 TimToady m: my IO::Path() $path = "/home/user".IO;
15:51 camelia rakudo-moar 0b1e4f: OUTPUT«5===SORRY!5=== Error while compiling /tmp/KgxcLTzd66␤IO::Path(Any) cannot be used as a type on a variable␤at /tmp/KgxcLTzd66:1␤------> 3my IO::Path() $path7⏏5 = "/home/user".IO;␤    expecting any of:␤        constraint␤»
15:51 ugexe m: sub p(IO::Path $path) { say $path.WHAT; }; p("/home/user".IO);
15:51 camelia rakudo-moar 0b1e4f: OUTPUT«Type check failed in binding $path; expected 'IO::Path' but got 'IO::Path'␤  in sub p at /tmp/XYk3HA2sGK:1␤  in block <unit> at /tmp/XYk3HA2sGK:1␤␤»
15:51 ugexe ^ that works on my perl6 from yesterday
15:52 FROGGS star-m: sub p(IO::Path $path) { say $path.WHAT; }; p("/home/user".IO);
15:52 camelia star-m 2013.03: OUTPUT«Type check failed in binding $path; expected 'IO::Path' but got 'IO::Path'␤  in sub p at /tmp/mY019NWIw9:1␤  in block <unit> at /tmp/mY019NWIw9:1␤␤»
15:52 ugexe if i add the $path = $path.resolve inside the sub p though, i get the type error
15:53 moritz oh, I'm glad we're not the only ones that fail handling of recursive data structures
15:53 * TimToady is not really awake/well enough yet to help much
15:54 moritz python's unittest.TestCase method assertEqual reaches max recursion depth quickly for cycle data structures
15:54 moritz s/recursive/cyclic/
15:58 gfldex joined #perl6
15:59 ugexe perl6 -e 'sub p(IO::Path $path) { say $path.WHAT; }; p("/home/user".IO);'                 (IO::Path)
15:59 ugexe nickl@localhost:~/.rakudobrew/moar-nom$ git show
15:59 ugexe commit 2cb72b2141f2e37c0df4fb04fc2f5b68e34cdf01
15:59 vendethiel joined #perl6
16:01 ugexe there was a commit related to type errors by jnthn on april 12
16:02 ugexe still something funny going on before that though:
16:02 ugexe nickl@localhost:~/.rakudobrew/moar-nom$ perl6 -e 'sub p(IO::Path $path is copy) { say $path.WHAT; $path = $path.resolve }; p("/home/user".IO);'
16:02 ugexe (IO::Path) \n Type check failed in assignment to '$path'; expected 'IO::Path' but got 'Str'
16:05 spider-mario joined #perl6
16:07 raydiak m: '.'.IO.resolve.WHAT.say
16:07 camelia rakudo-moar 0b1e4f: OUTPUT«(Str)␤»
16:09 ugexe .resolve is supposed to return a IO::Path
16:10 raydiak yep
16:18 eli-se joined #perl6
16:21 chenryn joined #perl6
16:27 chenryn joined #perl6
16:28 aborazmeh joined #perl6
16:28 aborazmeh joined #perl6
16:38 cognominal joined #perl6
16:51 dalek perl6-roast-data: 7d82b5c | coke++ | / (9 files):
16:51 dalek perl6-roast-data: today (automated commit)
16:51 dalek perl6-roast-data: review: https://github.com/coke/perl6-roast-data/commit/7d82b5cfb2
16:52 [Coke] no change in nqp test or rakudo spectest
16:53 RabidGravy yep, the resolve method will return a String
16:56 ugexe ive regressed to march 30 and .resolve is still returning a string
16:56 ugexe anyone know if it ever returned an IO::Path?
16:58 [Coke] er, not true, some changes. S12-attributes/native.t stopped failing, added all the jnthn NF* tests.
16:59 PerlJam ugexe: looks like it was implemented at 61c59db4d5ac491d730311fbabb4ead71ec8798a and it returned a string then
16:59 ugexe heh i was looking at that
17:00 RabidGravy to be fair I can't even find the method documented
17:02 PerlJam RabidGravy: S16:399
17:02 synbot6 Link: http://design.perl6.org/S16.html#line_399
17:03 raydiak also http://design.perl6.org/S32/IO.html#.resolve
17:03 PerlJam yep
17:04 PerlJam S32:IO/951  # test
17:04 synbot6 Link: http://design.perl6.org/S32.html#IO
17:04 PerlJam oops
17:04 PerlJam S32/IO:951
17:04 synbot6 Link: http://design.perl6.org/S32/IO.html#line_951
17:05 PerlJam that's some dyslexia I have with symbols :)
17:05 raydiak S32/IO:.resolve ?
17:05 raydiak guess not :)
17:06 RabidGravy got it
17:07 synbot6 joined #perl6
17:07 PerlJam S32/IO:.resolve
17:07 synbot6 Link: http://design.perl6.org/S32/IO.html#.resolve
17:08 TimToady .oO(the 6th voyage of synbot)
17:08 timotimo ooooh a synbot <3
17:08 PerlJam S32/IO:move()
17:08 synbot6 Link: http://design.perl6.org/S32/IO.html#move()
17:08 timotimo meh, that doesn't turn into a single link in my terminal
17:08 timotimo only the ( is linkified, not the )
17:08 PerlJam timotimo: yeah, I just discovered that here to
17:08 PerlJam er, too
17:09 PerlJam but you can still copy and paste it as needed
17:10 beastd joined #perl6
17:10 timotimo sure
17:10 timotimo i just put the ) in the url manually
17:14 mohij joined #perl6
17:18 n0vacane joined #perl6
17:19 n0vacane left #perl6
17:22 vendethiel m: List[::T]{}; role Functor[::T where ???] {}; # how do I assure that List is a higher-order type? can I do that?
17:22 camelia rakudo-moar 0b1e4f: OUTPUT«===SORRY!===␤Could not locate compile-time value for symbol T␤»
17:23 KCL joined #perl6
17:24 vendethiel m: List[::T]{submethod BUILD{say 'hi'~T.WHAT.perl;}}; role Invoke[::F, ::T] { method new { F[T].new; } }; Invoke[List, Int].new;
17:24 camelia rakudo-moar 0b1e4f: OUTPUT«===SORRY!===␤Could not locate compile-time value for symbol T␤»
17:24 vendethiel aw :[
17:24 moritz vendethiel: are you trying to declare role List?
17:24 moritz vendethiel: or are you trying to use it?
17:24 mr-foobar joined #perl6
17:25 vendethiel m: role List[::T]{submethod BUILD{say 'hi'~T.WHAT.perl;}}; role Invoke[::F, ::T] { method new { F[T].new; } }; Invoke[List, Int].new;
17:25 camelia rakudo-moar 0b1e4f: OUTPUT«===SORRY!===␤Cannot find method 'parameterize'␤»
17:25 vendethiel moritz: thanks!
17:26 tinyblak joined #perl6
17:27 moritz vendethiel: and, uhm, don't name a role List; that's a built-in type that the compiler references, and "funny" things are bound to happen when you override that without adhering to the core List API
17:27 vendethiel m: role C[::T]{submethod BUILD{say 'hi'~T.WHAT.perl;}}; role Invoke[::F, ::T] { method new { F[T].new; } }; Invoke[C, Int].new;
17:27 camelia rakudo-moar 0b1e4f: OUTPUT«===SORRY!===␤Cannot find method 'parameterize'␤»
17:35 * masak .oO( don't name it X, don't name it List... the compiler gets all the good names! ) :P
17:40 * PerlJam looks into the future and see that as the subject of an anti-p6 rant
17:40 PerlJam s/see/sees/
17:40 lucasb joined #perl6
17:41 * masak .oO( oh yes, we luuurv lexical scoping. except for package-like things, which are global names )
17:41 TimToady you can still declare it 'my' if you like
17:42 TimToady m: my role List[::T]{submethod BUILD{say 'hi'~T.WHAT.perl;}}; my role Invoke[::F, ::T] { method new { F[T].new; } }; Invoke[List, Int].new;
17:42 camelia rakudo-moar 0b1e4f: OUTPUT«===SORRY!===␤Cannot find method 'parameterize'␤»
17:42 TimToady though I guess it doesn't help much here
17:43 PerlJam All programming languages have a bias that they impose upon the programmer.  In Perl, the bias is that the programmer must *think* about what they are doing.   ;-)
17:44 itz_ jnthn: pl6anet.org is fixed (wordpress uses https now) .. sorry for delay I was behind corporate wall from hell
17:52 ugexe https://github.com/ugexe/rakudo/commit/4330f86cc81a7b452ad9cd861798adc3ed0eebfc this seems like it fixes the .resolve stuff
17:54 ugexe probably overlooking something, but at least it looks easy enough to fix
17:59 raydiak ugexe++
18:00 ugexe im mostly unsure if i should be using absolute or not
18:01 Ven joined #perl6
18:08 masak jnthn: on consideration, I suspect you're right about RT #123894
18:08 synbot6 Link:  https://rt.perl.org/rt3//Public/Bug/Display.html?id=123894
18:08 masak synbot6! \o/
18:09 masak jnthn: S04 says "C<FIRST>, C<NEXT>, and C<LAST> are meaningful only within the lexical scope of a loop, and may occur only at the top level of such a loop block."
18:09 masak which could (easily) be taken to apply here, and actually make the ENTER { ...; NEXT { ... } } nesting illegal.
18:09 masak (which is too bad, 'cus it was cute. but at least it's spec.)
18:13 Foxcool_ joined #perl6
18:18 FROGGS joined #perl6
18:18 vendethiel ENTER { my Lock $lock .= new; NEXT { $lock.release } } #*g*
18:19 yqt joined #perl6
18:19 masak yeah, that's the essence of what I was doing.
18:25 yvan1 joined #perl6
18:26 ugexe m: my $io = "foo/./bar/..".IO.resolve; say $io;
18:26 camelia rakudo-moar 0b1e4f: OUTPUT«/home/camelia/foo/bar/..␤»
18:27 ugexe is that correct?
18:29 lucasb Does resolve makes absolute? If not, then I would expect just "foo", right?
18:29 masak "foo/"
18:29 lucasb foo o/
18:29 lucasb :)
18:29 ugexe # bar is a symlink pointing to "/baz" my $io = "foo/./bar/..".IO.resolve;  # now "/" (the parent of "/baz")
18:30 ugexe actually creating the symlinks doesn't change anything from camelias output
18:30 masak m: say "f{"o" x $_} o/" for 0..*
18:30 FROGGS masak: $_ in there is buggy
18:30 camelia rakudo-moar 0b1e4f: OUTPUT«(timeout)f o/␤f o/␤f o/␤f o/␤f o/␤f o/␤f o/␤f o/␤f o/␤f o/␤f o/␤f o/␤f o/␤f o/␤f o/␤f o/␤f o/␤f o/␤f o/␤f o/␤f o/␤f o/␤f o/␤f o/␤f o/␤f o/␤f o/␤f o/␤f o/␤f o/␤f o/␤f o/␤f o/␤f o/␤f o/␤f …»
18:30 masak so it is.
18:30 masak hm.
18:31 masak m: say "f$("o" x $_) o/" for 0..*
18:31 camelia rakudo-moar 0b1e4f: OUTPUT«(signal XFSZ)f o/␤fo o/␤foo o/␤fooo o/␤foooo o/␤fooooo o/␤foooooo o/␤fooooooo o/␤foooooooo o/␤fooooooooo o/␤foooooooooo o/␤fooooooooooo o/␤foooooooooooo o/␤fooooooooooooo o/␤foooooooooooooo o/␤fooooooooooooooo o/␤foooooooooo…»
18:31 domidumont joined #perl6
18:31 masak \o/
18:31 masak only known good use of $() :)
18:32 FROGGS $(:-)
18:33 masak that dollar hat suits you
18:33 cognominal perljam: in most languages, there is a good way to do things that may be excellent but get in the way if you are not proficient enough. Very often, in Perl, my first idea is good enough meaning perfect for a kleenex script.
18:35 bobkare I've been playing around with some parsing, but I keep stumbling on something I assume is pretty basic and simple. I can't seem to get an array to flatten out so I can pass it to a constructor, as seen in method TOP here: http://pastebin.com/KtbXKj2M
18:41 lichtkind joined #perl6
18:42 PerlJam bobkare: looks like you're making an array of strings rather than an array of Test::Foo  (as you've declared @!foos to be)
18:44 bobkare PerlJam: Huh? the debug print output of Foo AST: [[Test::Foo.new(foo => "bar "), Test::Foo.new(foo => "baz")]]<> makes me think it isn't strings
18:44 TimToady does it work any better with .list.flat instead of just .flat?
18:44 PerlJam yeah, I'm apparently not fully awake right now, sorry
18:45 bobkare TimToady: nope, exactly same results
18:46 TimToady well, seems like that constructor can't take a list because it's itemizing the array or so then
18:47 TimToady either that or it's passing in an AoA somehow, but I don't see how
18:49 bobkare It looks like it's becoming an AoA, if I replace the assignment to @a with this: my @a = [Test::Foo.new(foo => "bar "), Test::Foo.new(foo => "baz")].flat; it works. Add the extra set of brackets from the debug output and it fails.
18:49 TimToady otoh, maybe it's just our typing of arrays difficulty
18:49 TimToady and it wants Array[Test::Foo], not Array
18:50 TimToady if you have access to the class code, take off the array type and see if it works
18:51 TimToady oh, there it is, right there
18:51 TimToady so change it to has Test::Foo @.foos;
18:51 TimToady er
18:51 TimToady has @.foos;
18:51 TimToady or hash @.foos where all(Test::Foo) or some such
18:52 TimToady since jnthn++ just fixed that (or was that in a branch?)
18:52 bobkare That kinda works, but leaves foos as AoA which isn't what I want
18:54 ugexe my @a = $<foo>».made>>.flat; ?
18:54 ugexe Test::Result.new(foos => Array[Test::Foo].new(Test::Foo.new(foo => "bar "), Test::Foo.new(foo => "baz")))
18:56 ugexe s/>>//
18:56 bobkare Yes, finally something that works! ugexe++
18:57 bobkare not without the extra » though, then it continues to throw the type check failure at me
18:58 vendethiel m: $(4) ==> my $a; say $a # masak <- another *great* use of $().
18:58 camelia rakudo-moar 0b1e4f: OUTPUT«4␤»
18:59 PerlJam I guess that makes sense.   foo is plural in the regex, so you get an array of whatever each .foo creates, which is an Array of Test::Foo;
18:59 eli-se joined #perl6
19:02 perl666 joined #perl6
19:03 bobkare Oh, right, so I had » making a list of arrays and .flat needed to be applied to all the inner arrays of Test::Foo. I get it now.
19:04 eli-se hi!!!
19:04 * bobkare goes back to fix that in his real test script, which at this point is a rats mess with debug statements and extra .flat and .list calls added at random all over the place
19:05 PerlJam bobkare: maybe printing   $<foo>.perl  and $<foo>».made.perl  (before the assignment to @a) would have helped?
19:05 FROGGS hi eli-se
19:07 PerlJam bobkare: and/or put more than one foo-thing in your string to parse.  :)
19:08 vendethiel o/ eli-se
19:08 timotimo TimToady: do you know how big the performance improvement from $sawcaps was?
19:09 virtualsue joined #perl6
19:09 FROGGS ohh, it's Monday again :o)
19:09 * moritz guesses timotimo++ will want to write about http://examples.perl6.org/
19:11 timotimo oh, cool!
19:11 timotimo huh, yahtzee.p6 seems to have the source code twice
19:12 timotimo same thing for wizard.pl
19:13 skids joined #perl6
19:15 timotimo TimToady: is it just "a few %s" as stated in the very first commit regarding it?
19:17 TimToady I was picking up about 5% +/- 2% on a somewhat noisy system
19:17 TimToady that's on stage parse, but that's been a rather noisy number lately for some reason
19:18 * TimToady didn't bother turning off temp throttling, for instance
19:18 TimToady but I ran it several times
19:21 timotimo OK
19:23 timotimo my desktop computer is not assembled/set up at the moment, so i'll have to wait a bit until i can run proper benchmarks again
19:24 timotimo especially since the spec tests are a very bad place to test my recent reprop devirtualization work with
19:28 Possum joined #perl6
19:45 Sqirrel joined #perl6
19:46 timotimo m: say <4+3i>.WHAT
19:46 camelia rakudo-moar 0b1e4f: OUTPUT«(Complex)␤»
19:49 rurban_ joined #perl6
19:51 masak m: say <+4+3i>.WHAT
19:51 camelia rakudo-moar 0b1e4f: OUTPUT«(Complex)␤»
19:56 dalek rakudo/nom: 872d74c | TimToady++ | src/Perl6/Actions.nqp:
19:56 dalek rakudo/nom: don't cache Rat literals
19:56 dalek rakudo/nom: review: https://github.com/rakudo/rakudo/commit/872d74cd2c
19:56 dalek rakudo/nom: af9a7ca | TimToady++ | src/core/operators.pm:
19:56 dalek rakudo/nom: regularize list of list semantics of infix:<...>
19:56 dalek rakudo/nom:
19:56 dalek rakudo/nom: All confusion between iscont and decont values is fixed, I hope.
19:56 dalek rakudo/nom: []...[] now works and produces a list containing a single [].
19:56 dalek rakudo/nom: () is now disallowed as either a start value or an endpoint.
19:56 dalek rakudo/nom: review: https://github.com/rakudo/rakudo/commit/af9a7caa1d
19:56 dalek rakudo/nom: 49f168e | TimToady++ | src/core/operators.pm:
19:56 dalek rakudo/nom: oops, don't use .elems to test for non-empty
19:56 dalek rakudo/nom:
19:56 dalek rakudo/nom: (especially on lists that might not know their infinite)
19:56 dalek rakudo/nom: review: https://github.com/rakudo/rakudo/commit/49f168e42d
19:56 vendethiel (s/their/they're/ :P)
19:57 DrForr joined #perl6
19:57 TimToady s/infinite/infinitenesslessness/
19:58 vendethiel fun word!
20:00 DrForr eval: "foo" ~~ /<[\x{0041}..\x{007f}]>/
20:01 DrForr Or is it 'm:' to invoke the bot, I've lost backlog here.
20:01 vendethiel it's m for rakudo-moar :)
20:01 DrForr m: "foo" ~~ /<[\x{0041}..\x{007f}]>/
20:01 camelia rakudo-moar 0b1e4f: OUTPUT«5===SORRY!5===␤Illegal reversed character range in regex: }..\x␤at /tmp/13KfIro5ta:1␤------> 3"foo" ~~ /<[\x{0041}..\x{007f}]7⏏5>/␤Other potential difficulties:␤    Repeated character ({) unexpectedly found in character class␤    at /…»
20:01 vendethiel so you should go for that, unless you want to try other backends (which probably don't work nowadays anyway)
20:02 DrForr Nod. Do I need to do something special to match ranges of Unicode characters? I'll dig in the synopses anyway.
20:02 TimToady m: "foo" ~~ /<[\x[0041]..\x[007f]]>/
20:02 camelia rakudo-moar 0b1e4f: ( no output )
20:03 TimToady m: say "foo" ~~ /<[\x[0041]..\x[007f]]>/
20:03 camelia rakudo-moar 0b1e4f: OUTPUT«「f」␤»
20:03 FROGGS "Repeated character ({)" is a good hint
20:03 timotimo can you use this link for previewing my weekly post? https://p6weekly.wordpress.com/?p=480&amp;preview=true&amp;preview_id=480&amp;preview_nonce=5023d4a89f
20:04 FROGGS timotimo: "
20:04 FROGGS Oops! That page can’t be found.
20:04 FROGGS "
20:04 timotimo god fucking damnit wordpress
20:04 TimToady DrForr: in P6 we try to avoid use of {} in spots where it doesn't mean a closure, and [] is the New Grouper in P6 regexen anyway
20:04 DrForr Thanks. I'm back to delving in grammars again.
20:05 FROGGS tadzik: "==> Successfully installed OpenSSL"... how do I reproduce this? https://gist.github.com/tadzik/22999b3fbb06069ab994
20:05 timotimo it seems like they threw out the classic editor
20:05 timotimo and there's no functionality any more to give out a link that lets people provide feedback on a draft post
20:07 ugexe FROGGS: that looks like its from an old rakudo
20:07 ugexe i had that problem last week? and you fixed it
20:08 FROGGS ugexe: well, I'd guessed that tadzik would upgrade before posting it...
20:08 timotimo https://gist.githubusercontent.com/timo/f4e54c2e93703a8584ba/raw/dcab8f99f83c2dbbbc54629a7be57d70931aa71e/gistfile1.txt - feedback on my weekly post pls
20:09 * timotimo will be AFK for a bit
20:09 dalek roast: d962576 | TimToady++ | S03-sequence/ (2 files):
20:09 dalek roast: tweaks for recent sequence changes
20:09 dalek roast: review: https://github.com/perl6/roast/commit/d962576dd4
20:09 ugexe well, ive had situations where rakudobrew messed up updating silently. so at least there is an excuse
20:11 DrForr m: say "\x[236a]"
20:11 camelia rakudo-moar 0b1e4f: OUTPUT«⍪␤»
20:12 ugexe i wonder if it would be useful to put the perl -v in the Failure Summary
20:12 DrForr m: say "\x{236a}"
20:12 camelia rakudo-moar 0b1e4f: OUTPUT«5===SORRY!5===␤Unrecognized backslash sequence: '\x'␤at /tmp/UVzqPRUPEu:1␤------> 3say "\7⏏5x{236a}"␤Missing block␤at /tmp/UVzqPRUPEu:1␤------> 3say "\x{2367⏏5a}"␤    expecting any of:␤        whitespace␤␤»
20:14 TimToady "Missing block" is rather wrong
20:14 psch could be an artifact of my bad cheat for moreinput i think
20:15 TimToady it's probably worth a .obs to catch this particular p5 carryover
20:16 DrForr Also in passing, 'repeated character', while correct, isn't terribly edifying, though at this late hour I can't come up with a good wording.
20:17 FROGGS timotimo: you could perhaps mention the perl6-gdb-m and perl6-valgrind-m runners...
20:17 nwc10 about 16 minutes, I think, http://www.spacex.com/webcast/
20:17 masak nwc10++
20:18 FROGGS but timotimo++ besides that :oO)
20:18 FROGGS err, -O
20:18 TimToady DrForr: the message usually shows up when someone says <[ "foo" | "bar" ]> or so
20:18 RabidGravy nwc10++ # had forgotten
20:19 DrForr Aha.
20:19 TimToady m: /<["foo"]>/
20:19 camelia rakudo-moar 0b1e4f: OUTPUT«Potential difficulties:␤    Repeated character (") unexpectedly found in character class␤    at /tmp/IIOoj23v9M:1␤    ------> 3/<7⏏5["foo"]>/␤    Repeated character (o) unexpectedly found in character class␤    at /tmp/IIOoj23v9M:1␤    --…»
20:19 DrForr Incidentally I'll be damned, the ECMAscript parser actually compiles now.
20:20 TimToady well, you're in transylvania, what did you expect?
20:20 DrForr More transvestites.
20:21 DrForr (/Rocky Horror Picture Show/ reference, if that wasn't clear :) )
20:23 Ven joined #perl6
20:23 masak I don't think I'll ever tire of launches.
20:23 dalek rakudo/nom: cc3e654 | peschwa++ | src/Perl6/Grammar.nqp:
20:23 dalek rakudo/nom: Remove a few moreinput related bits.
20:23 dalek rakudo/nom:
20:23 dalek rakudo/nom: It's not working well enough to be worth the troubles it shows.  Add to that
20:23 dalek rakudo/nom: the fact that there shouldn't be any troubles and it should go back to
20:23 dalek rakudo/nom: something that at least works for all other use-cases, if only limited for
20:23 dalek rakudo/nom: multiline REPL input.
20:23 dalek rakudo/nom: review: https://github.com/rakudo/rakudo/commit/cc3e65405c
20:24 psch eh
20:24 psch i wanted to spectest first but got distracted :/
20:24 psch i'll do that now
20:34 masak aww
20:34 RabidGravy nwc10: scrubbed due to weather :-(
20:34 masak you will not go to space today
20:34 nwc10 try again tomorrow
20:34 masak yeah
20:34 masak ok, let me cheer you up with
20:34 RabidGravy wanted to see the landing attempt
20:34 masak today's mini-challenge :)
20:34 nwc10 me too
20:34 masak solve http://www.theguardian.com/science/alexs-adventures-in-numberland/2015/apr/13/can-you-solve-the-singapore-primary-maths-question-that-went-viral using Perl 6 ;)
20:35 ugexe hmmm Digest is failing tests on JVM today
20:35 ugexe sub infix:<⊞>(\x, \y)    {  (x + y) % 2**32 } seems like unicode infix stuff
20:37 ugexe or Nil related
20:39 FROGGS ugexe: Digest always failed on JVM: http://testers.perl6.org/dist/D/Digest/Lucien%20Grondin.html
20:41 ugexe ah. i could have swore i had openssl build on jvm sometime recently. too much caffeine this month apparently
20:42 FROGGS m: my @a = <foo bar baz>; map( { s/a/A/ }, @a ); say @a;
20:42 camelia rakudo-moar 49f168: OUTPUT«foo bar baz␤»
20:42 FROGGS is that still supposed to change the content of the array? RT #116731
20:42 synbot6 Link:  https://rt.perl.org/rt3//Public/Bug/Display.html?id=116731
20:42 [Coke] ugexe: you take that back.
20:42 timotimo FROGGS: nothing like "VALGRINDARGS" or so for the launchers?
20:42 [Coke] . o O (too much caffeine? heresy!)
20:43 ugexe [Coke]: not till tomorrow morning
20:43 FROGGS timotimo: I don't know... did you add it?
20:43 timotimo no
20:43 FROGGS s/did/didn't/
20:43 FROGGS me neither...
20:43 timotimo OK
20:43 timotimo may be a thing to add later on
20:44 FROGGS also the --cleanup-on-exit option or what it is called
20:44 FROGGS I really thought you wanted to do it
20:44 timotimo oh
20:45 timotimo moi? :)
20:45 FROGGS like with my wife... where "We should do X." means "DO X NAO!"
20:46 tadzik FROGGS: huh, let me see. I thought I have it all up-to-date...
20:58 timotimo well, at least the post is published now
21:00 psch timotimo++ # great title
21:01 FROGGS timotimo++
21:01 dalek rakudo/nom: 8b639ba | TimToady++ | src/Perl6/Grammar.nqp:
21:01 dalek rakudo/nom: catch use of braces on escapes
21:01 dalek rakudo/nom: review: https://github.com/rakudo/rakudo/commit/8b639ba6f0
21:04 TimToady timotimo: there's no h in jnthn's name, except for the second one
21:05 TimToady also, nothing gets flipped when the length is known
21:06 tadzik FROGGS: ...ok, it actually works
21:06 tadzik must be me being stupid
21:10 FROGGS tadzik: that's actually the better option :o)
21:13 masak 'night, #perl6
21:14 tadzik :)
21:14 tadzik gnight masak
21:18 cognominal joined #perl6
21:20 * flussence grumbles at browsers defaulting to latin-1 when so many .txt files I see these days are utf-8...
21:20 gfldex timotimo++ for telling the truth
21:21 cognominal r: < a b >.map: { return }  # this is probably stupid, but the error msg does not help
21:21 camelia rakudo-jvm 8b639b: OUTPUT«cannot connect to eval server: Connection refused␤»
21:21 camelia ..rakudo-moar 8b639b: OUTPUT«Unhandled exception: Cannot look up attributes in a type object␤   at src/gen/m-CORE.setting:9642  (/home/camelia/rakudo-inst-1/share/perl6/runtime/CORE.setting.moarvm::116)␤ from src/gen/m-CORE.setting:9632  (/home/camelia/rakudo-inst-1/share/perl6/ru…»
21:27 pecastro joined #perl6
21:29 tinyblak joined #perl6
21:30 telex joined #perl6
21:32 timotimo TimToady: oh? so flipping already didn't happen when length was known before you wrote that patch?
21:35 torbjorn_ joined #perl6
21:36 jdv79_ joined #perl6
21:36 orevdiabl joined #perl6
21:38 agentzh joined #perl6
21:38 agentzh joined #perl6
21:49 orafu joined #perl6
21:55 RabidGravy cognominal, actually I think there is some underlying weirdness around that:
21:55 RabidGravy m:  < a b >.map: { return $_ }
21:55 camelia rakudo-moar 8b639b: OUTPUT«Unhandled exception: P6opaque: no such attribute '$!storage'␤   at src/gen/m-CORE.setting:9642  (/home/camelia/rakudo-inst-2/share/perl6/runtime/CORE.setting.moarvm::116)␤ from src/gen/m-CORE.setting:9632  (/home/camelia/rakudo-inst-2/share/perl6/runti…»
22:00 RabidGravy m:  < a b >.map: { say $_.WHAT }
22:00 camelia rakudo-moar 8b639b: OUTPUT«(Str)␤(Str)␤»
22:00 RabidGravy not sure what it is though
22:06 cognominal me neither
22:07 ugexe it should just say cannot return outside of a routine
22:09 rhr joined #perl6
22:29 TimToady timotimo: no, the part where you said "only part gets flipped now"; if the length is known, nothing gets flipped, it just turns it into a <?before> with an offset
22:32 nbdsp joined #perl6
22:39 nbdsp Greetings! Could someone advise please how to use nested enumerations? If I have:  class EnClass { my enum En <Sun Mon Tue>; } - how to refer to En members?
22:40 nbdsp m: class EnClass { my enum En <Sun Mon Tue>; }; say EnClass::En::Sun;
22:40 camelia rakudo-moar 8b639b: OUTPUT«Could not find symbol '&Sun'␤  in method <anon> at src/gen/m-CORE.setting:16099␤  in any find_method_fallback at src/gen/m-Metamodel.nqp:2908␤  in any find_method at src/gen/m-Metamodel.nqp:1052␤  in block <unit> at /tmp/KrBxEESH0H:1␤␤»
22:40 skids joined #perl6
22:40 TimToady m: class EnClass { enum En <Sun Mon Tue>; }; say EnClass::En::Sun;
22:40 camelia rakudo-moar 8b639b: OUTPUT«Sun␤»
22:41 TimToady m: class EnClass { my enum En is export <Sun Mon Tue>; }; import EnClass; say En::Sun;
22:41 camelia rakudo-moar 8b639b: OUTPUT«Sun␤»
22:41 TimToady m: class EnClass { my enum En is export <Sun Mon Tue>; }; import EnClass; say Sun;
22:41 camelia rakudo-moar 8b639b: OUTPUT«Sun␤»
22:41 nbdsp oh, without 'my' or with 'export'. thanks!
22:52 BenGoldberg joined #perl6
23:26 grondilu sub f(uint32 $n) { $n + 2**32 }; say f 7;
23:26 grondilu m: sub f(uint32 $n) { $n + 2**32 }; say f 7;
23:26 camelia rakudo-moar 8b639b: OUTPUT«4294967303␤»
23:26 grondilu ^not sure if that's ok
23:27 grondilu m: say 4294967303.base(2)
23:27 camelia rakudo-moar 8b639b: OUTPUT«100000000000000000000000000000111␤»
23:28 gfldex m: sub f(uint32 $n) { $n + 2**32 }; say (f 7).WHAT;
23:28 camelia rakudo-moar 8b639b: OUTPUT«(Int)␤»
23:28 gfldex magic o.0
23:28 grondilu m: sub f(uint32 $n --> uint32) { my uint32 $ = $n + 2**32 }; say f 7
23:28 camelia rakudo-moar 8b639b: OUTPUT«4294967303␤»
23:29 grondilu m: say my uint32 $ = 2**32
23:29 camelia rakudo-moar 8b639b: OUTPUT«4294967296␤»
23:29 grondilu ^well that can't be right.
23:29 TimToady m: say (my uint32 $ = 2**32).WHAT
23:29 camelia rakudo-moar 8b639b: OUTPUT«(Int)␤»
23:30 grondilu m: say my uint32 $ = 2**100
23:30 camelia rakudo-moar 8b639b: OUTPUT«0␤»
23:30 grondilu m: say my uint32 $ = 2**60
23:30 camelia rakudo-moar 8b639b: OUTPUT«1152921504606846976␤»
23:30 TimToady it would appear to be returning the RHS, not the LHS
23:30 grondilu m: say my uint32 $ = 2**64
23:30 camelia rakudo-moar 8b639b: OUTPUT«0␤»
23:30 grondilu m: say my uint32 $ = 2**63
23:30 camelia rakudo-moar 8b639b: OUTPUT«-9223372036854775808␤»
23:31 gfldex it
23:31 * grondilu has no idea what's going on there
23:31 gfldex it's confusing in a odd way
23:31 grondilu the number of bits that seems critical is 64, when it should be 32.
23:32 grondilu so it behaves like int64 instead of uint32?
23:32 gfldex looks like it. Please rakudobug it.
23:33 grondilu sure?
23:37 grondilu also, in the REPL if I enter 'my uint32 $ = 1;' twice, I get:
23:37 grondilu Unhandled lexical type in lexprimspec for '$ANON_VAR__1'
23:37 grondilu but not with 'my uint $ = 1;', weirdly enough.
23:40 aborazmeh joined #perl6
23:40 aborazmeh joined #perl6
23:42 * grondilu submitted the bug
23:42 grondilu FYI, I found this while considering using native types for http://rosettacode.org/wiki/SHA-256#Perl_6
23:50 tinyblak joined #perl6

| Channels | #perl6 index | Today | | Search | Google Search | Plain-Text | summary

Perl 6 | Reference Documentation | Rakudo