»ö« Welcome to Perl 6! | perl6.org/ | evalbot usage: 'perl6: say 3;' or rakudo:, niecza:, std:, or /msg camelia perl6: ... | irclog: irc.perl6.org | UTF-8 is our friend!
Set by sorear on 25 June 2013.
timotimo FROGGS: i'll give moar-support the missing commit 00:12
dalek kudo/moar-support: 93caa93 | (Timo Paulssen)++ | tools/build/NQP_REVISION:
bump NQP revision to get the regex optimizer.
kudo/moar-support: ffb923d | (Timo Paulssen)++ | src/Perl6/Optimizer.nqp:
run the QRegex optimizer from rakudo's optimizer.
timotimo wait, that didn't make sense 00:14
timotimo you were building nqp 00:14
in that case, all you need is to re-run configure.pl
dalek kudo-star-daily: 1a43d69 | coke++ | log/ (5 files):
today (automated commit)
 00:24
rl6-roast-data: a5139a1 | coke++ | / (6 files):
today (automated commit)
rjbs was re-reading this today. Possibly of interest to some here: 02:53
inform7.com/learn/documents/WhitePaper.pdf
rjbs whitepaper on the natural-language-like Inform 7 programming language 02:53
lue Fun I7 exercise: replace "The Lounge is east of the Kitchen" with "The Lounge is to the east of the Kitchen" 02:59
rjbs :)
BenGoldberg p: say (e ** 3).perl, " ", exp(3).perl; # unsurprising 05:59
camelia rakudo-parrot dc5a1f: OUTPUT«20.0855369231877e0 20.0855369231877e0␤»
BenGoldberg j: say (e ** 3).perl, " ", exp(3).perl; # surprising? 06:00
camelia rakudo-jvm dc5a1f: OUTPUT«20.085536923187675e0 20.085536923187668e0␤»
FROGGS moritz: are the double quotes in the header intentional? irclog.perlgeek.de/moarvm/2013-12-16 07:49
timotimo and at the very bottom?( 07:53
moritz FROGGS: "not really" 08:03
timotimo "" 08:04
FROGGS .oO( "Lazer" )
moritz will hopefully be fixed as soon as the template cache expires 08:05
that is, in 10 minutes
timotimo cache expiry: one of the harder problems in computer science 08:06
moritz nah, already fixed
FROGGS++
timotimo has an idea brewing in his head for a cool perl6-related project 08:07
i should get a github paid account so that i can make it private and tease you about it from time to time :P
timotimo .o(or ... just put it on bitbucket) 08:08
BenGoldberg Does perl6-related mean written in perl6, or merely inspired by perl6?
timotimo the first 08:10
timotimo but also related to the community and stuff 08:12
timotimo there's no database i can use on the jvm port of rakudo right now, no? 08:21
except maybe if i use the jvm interop for JDBC?
moritz please do that, and patch it into DBIish 08:35
timotimo ... yikes 08:36
seems pretty daunting
though the payoff is probably quite nice
dalek kudo/nom: 3264e9f | (Timo Paulssen)++ | src/core/CallFrame.pm:
give non-parrot's CallFrame the $.level, too.
 08:45
timotimo is amazed he found that :P
moritz nqp::bindattr($self, CallFrame, '%.level', $l); 08:48
timotimo whoops
thanks!
moritz it's good that you noticed the dubiousity before I could put it into words :-)
dalek kudo/nom: 01e150a | (Timo Paulssen)++ | src/core/CallFrame.pm:
copypasto, spotted by moritz++
 08:49
moritz still not convinced
shouldn't it be $!level?
timotimo WHY DO I DO THIS
moritz the low-level attribute always has a !
moritz knows exactly how timotimo feels now, having been there before 08:50
dalek kudo/nom: 4136508 | (Timo Paulssen)++ | src/core/CallFrame.pm:
there are no words to explain this blunder.
timotimo hm, so a state var in method new can be used to implement singletons, right? 08:51
though at that point i might as well have a package with "my" vars and subs instead of methods 08:52
FROGGS thanks woolfy 09:04
advent calender for today published
moritz FROGGS: could you maybe state more clearly that this is not in mainline rakudo right now? 09:08
FROGGS++ # advent calendar
FROGGS moritz: mainline rakudo? what do you mean? 09:09
ahh, that you need to recompile it to use it?
moritz FROGGS: it's in your fork of rakudo, right?
also 09:11
mlenz@mlenz-workstation:~/p6/rakudo$ ./perl6-p -e 'my $x = "123abc"; $x++; say $x'
123abd
doesn't fail
FROGGS ahh true 09:13
moritz: it is not a fork 09:14
moritz FROGGS: is it a branch? or is it in rakudo/rakudo branch nom? 09:16
FROGGS it is nom + a tiny patch that needs to be applied (that is shown in v5's README) 09:18
but yes, I will make clear how one gets v5 09:19
p: say "123abc" + 1 09:20
camelia rakudo-parrot dc5a1f: OUTPUT«Cannot convert string to number: trailing characters after number in '123⏏abc' (indicated by ⏏)␤ in method Numeric at gen/parrot/CORE.setting:12058␤ in sub infix:<+> at gen/parrot/CORE.setting:3964␤ in sub infix:<+> at gen/parrot/CORE.settin…»
FROGGS moritz: fixed the example code and added a paragraph near the bottom 09:24
moritz FROGGS++ 09:25
FROGGS thank you :o) 09:26
arnsholt o/ 09:32
FROGGS hi arnsholt 09:33
hoelzro good morning #perl6! 10:30
brrt hi hoelzro 10:31
hoelzro dost mine eyes deceive me? 11:04
I have a perl6-m?
\o/
everyone++
masak g'day, #perl6 11:10
to... to what extent do we have a perl6-m? :) 11:11
hoelzro it exists 11:12
I don't think it does much (yet(
but that's a huge step!
masak ok :) 11:17
& 11:18
FROGGS it passes17 of 23 sanity tests 11:19
jnthn wonders if that can be 23 out of 23 before he takes Christmas vacation :) 11:44
go|dfish t 11:45
jnthn t/00-parrot and t/01-sanity, to be precise... :P 11:47
If anybody feels like merging those into a single t/01-sanity, that may make sense :)
FROGGS except the pir tests?
jnthn aye, that can go in a t/parrot or so 11:48
dalek kudo/moar-support: 40a022a | (Tobias Leich)++ | t/0 (40 files):
move backend independent tests to t/01-sanity
 12:01
clkao_ win 4 12:09
moritz lose 8
FROGGS :o( 12:12
[Coke] reviewingly sees Inform 7. It was fun to play with, but lacked solid documentation, as I recall. 13:25
dalek rl6-roast-data: c660f54 | coke++ | bin/rakudo.moar.sh:
use a nice'd perl6-m
 13:30
rl6-roast-data: 314a738 | coke++ | cull:
add moar to the daily csv

  (as of this commit, still not passing any spectests)
woolfy FROGGS: perl6advent.wordpress.com/2013/12/1...16-slangs/ is an excellent blog. Even though, I found some typos and I got a question. I just sent you an email about it. 14:29
timotimo oh wow, what a nap :| 14:34
should have been working on my advent post instead :|
FROGGS woolfy: thanks, got it 14:35
timotimo FROGGS++ # advent calendar post :3 14:45
moritz wow, the advent schedule is full up to the 22nd 14:52
timotimo how should i write "Perl6" in the post? 14:52
Perl 6, Perl6 or perl6?
moritz Perl 6 14:53
timotimo okay. and Perl 5, too? 14:53
moritz if you want with a non-breaking space :-)
yes
woolfy I would write "Perl 6" (including the space)
timotimo my keyboard layout has a key for non-breaking spaces! :D 14:53
moritz breaks it 14:54
timotimo nooooo :(
tadzik :D
timotimo i was using that!
moritz it's Ctrl+k NS for me (in vim)
moritz wonders if compose <blank> <blank> works
aye, it does 14:55
raydiak good * #perl6 15:27
moritz good *, raydiak 15:41
timotimo my article is turning into a "report of what i did" rather than "how you can be a cool new contributor like me!" article :\ 15:49
moritz that's OK, IMHO 15:50
timotimo i was really hoping for the latter rather than the former 15:50
FROGGS can you describe how to get from "ohh, there is a bug / a missing thing" through "that might be a patch" to "that even works!" ? 15:51
timotimo that would be nicer, yes 15:52
FROGGS if that is possible to describe :/ 15:54
moritz providing examples helps 15:55
jnthn decommute & 16:02
masak FROGGS++ # perl6advent.wordpress.com/2013/12/1...16-slangs/ 16:49
timotimo i'm surprised he didn't use many slang terms in that post :P 16:49
moritz_ yo man, wadcha mean by dad? 16:52
[Coke] finds a perl blog post that mentions five vs. six performance and is forced to rebuild his local copy of perl6 so he can time their p6 vs. more idiomatic perl6 and ---- oh look, where did the day go. 16:53
[Coke] thinks he's going to setup a cron job on his mactop to wipe & rebuild perl6 once a week so he always has a fresh copy to code with. :| 16:53
timotimo [Coke]: i'd be interested in the link :) 17:02
timotimo huh. indirect object syntax 17:10
FROGGS[mobile] :o) 17:32
FROGGS[mobile] timotimo: that is about turning "bar Foo" into "Foo->bar" 17:33
[Coke] timotimo: blogs.perl.org/users/jarich/2013/12...and-5.html
timotimo FROGGS[mobile]: yes, i saw that 17:34
FROGGS[mobile] like: print $fh "text"
timotimo not sure if i like that. oh well.
i think there has to be a : after the $fh, though?
FROGGS[mobile] no, please don't like that
no, a space 17:35
timotimo OK
what is perl5i?
FROGGS[mobile] Perl 5 that is
tadzik perl5i is fun :)
FROGGS[mobile] Where does the I come from?
tadzik imaginary
www.youtube.com/watch?v=4X4BNqE-mIE 17:36
erm
metacpan.org/pod/perl5i
FROGGS[mobile] Indirect object syntax is the good old perl
tadzik *relief*, good that it's _you_tube...
FROGGS[mobile] Hehe 17:36
Would you turn *red* otherwise? 17:37
It is a shame I cant write actions on that phone client
tadzik :> 17:38
well, not the worst thing I accidentally pasted :)
timotimo [Coke]: getting only the numbers up to 20 seems like bad idea for comparing times, tbh 17:39
humppaa, saatanat!
[Coke] timotimo: that url is already outdated, whoops: 17:41
blogs.perl.org/users/jarich/2013/12...ise-1.html
timotimo already saw it :) 17:43
was about to ask you if you had seen that, too
[Coke] I like his new p6 version better than mine. ah well.
still no reason for that to take 8s
timotimo but it's much slower ;)
you could tell him that he could use @multiples[] instead of {@multiples}
and ..^ $max instead of .. $max - 1 17:44
bbi10
[Coke] I prefer {@multiples} 17:44
Ulti isnt the 8s coming from JVM startup time 17:47
[Coke] quite possibly. 17:48
Ulti nearly everything I do with the JVM version of Rakudo takes 8s big or small :) its an oracle that has constant time for hard or easy problems with sufficiently small input size
I posted a reply with my time which was 1s on September R* 17:54
Ulti lol I just noticed I ran without the '30' parameter so the time is actually even faster 18:03
doh
I guess 0.1s is a big deal for people worried about times like this :S
an 8 second overhead on startup is basically meaningless for me when my Perl scripts run for days or weeks
timotimo well, with jvm they would rather take days than weeks :) 18:04
Ulti no I mean Perl 5! lol
timotimo oh, ok 18:05
Ulti considering perl 6 in that example is ~20x slower if we had autothreading you could argue Perl 6 could be as fast on my desktop since it has 24 cores 18:06
Ulti finally gets around to doing his ELM regex task with Util's translated regex library :D 18:20
lame the bioperl6 module fails to build
japhb_ Ulti, sounds like an excellent opportunity to earn yourself some good karma. :-) 18:21
FROGGS[mobile] ++Ulti 18:22
dwarring ADD 1 TO FROGGS 18:32
er I mean FROGGS++ (advent post) 18:33
FROGGS[mobile] hehe 18:34
TimToady COMPUTE FROGGS = FROGGS + 1
gotta love Grace Hopper, but she didn't grok huffman coding...
Ulti I hope I get some karma for this epic patch I'm making to bioperl6 just to get it to compile on Rakudo... looks a lot like someone translated the Perl 5 by just reading the spec a decade ago :S 18:36
I think modules.perl6.org needs an extra medal for "compiles without error"
FROGGS that is what colomon++ does 18:37
Ulti orly?
as in thats a thing displayed on the module list page, or colomon is testing the ecosystem?
FROGGS he said a few days ago that he should make his smoke results available properly
he's testing all modules 18:38
Ulti yeah that would be really worthwhile
FROGGS but there is no site where you can see it
Ulti is the source for modules.perl6.org not commitable?
FROGGS it is 18:39
github/perl6/modules.perl6.org or so
he could upload his json results... yeah
colomon 75.134.170.198:3001/report 18:42
Ulti yeah I guess you just patch in the extra info in here github.com/perl6/modules.perl6.org...ct-list.pl before it builds the P6Project object
colomon nicer looking graphics for the page broke sometime recently, haven't looked into why.
Ulti neat!
colomon and ooo, five new failures overnight. 18:43
including panda?
Ulti yay I have all ticks
Ulti I only have one test though which basically just checks to see if you can instantiate the class :S 18:44
colomon and the module author field isn't working either?!
clearly I need to fiddle with things.
Ulti there are more Bs than I imagined there would be 18:45
guess a lot has changed quickly
colomon Ulti: I think most of those B failures have been there at least six months. 18:52
timotimo oh no! ADT fails! 18:57
that shines a bad light on me
Ulti colomon I have a feeling some of them are abandonware, bioperl6 feels like it anyway 19:01
colomon I'm sure that's it.
Ulti %data{"$!min_$pos"} assuming $!min is a private member of the class how can you put that in for string interpolation?
$!{min} isnt right 19:02
colomon probably we should have a concerted clean up campaign, fixing the modules that can be fixed and removing the rest from the ecosystem.
timotimo no, it would be {$!min}
Ulti aha
colomon but it doesn't fit in my schedule right now……
timotimo what you put between { } is full-fledged perl6 code
retupmoca colomon: does that page list the actual test failures? 19:03
colomon retupmoca: which test fails? no 19:03
Ulti oh lordie there is no attribute $!min no idea what this code is meant to be doing
japhb TimToady: Grace Hopper produced the first compiler the same year Huffman published his coding paper, so she might not have had the *opportunity* to grok it. 19:04
:-)
lue Hm. Something tells me valgrind wasn't the best tool for discovering why the r-m setting fails :) 19:07
timotimo dwarring: if i don't succeed in making my post publishable today, is your blog post ready to be published?
japhb lue: valgrind v. hand-written GC: And ... FIGHT! 19:08
timotimo :D
lue stage mast took 6971.933 seconds :) 19:09
timotimo oh yikes
lue the segfault happened in UV code it seems. I can gist the output from valgrind if you like, I'll try running it through gdb next.
bibifuc lue: may I suggest you Address Sanitizer (from GCC or CLang) 19:14
the overhead is only 100%
lue bibifuc: I'll look into it, though it means recompiling MoarVM :) 19:16
arnsholt Yeah, it's usually worthwhile to run a segfault through gdb first
bibifuc I agree
arnsholt A lot cheaper than valgrind, and sometimes the bug is reasonably localised
timotimo r: 'abc' ~~ /a $<a>=[\w+]/; say :$<a>.value ~~ Match 19:29
camelia rakudo-parrot 413650, rakudo-jvm 413650: OUTPUT«True␤»
BenGoldberg r: 'abc' ~~ /a $<a>=[\w+]/; say :$<a>.value.perl 19:30
camelia rakudo-parrot 413650: OUTPUT«Match.new(orig => "abc", from => 1, to => 3, ast => Any, list => ().list, hash => EnumMap.new())␤»
..rakudo-jvm 413650: OUTPUT«Match.new(to => 3, hash => EnumMap.new(), ast => Any, list => ().list, orig => "abc", from => 1)␤»
lue Umm... how do I gdb? "bt" is coming up with "No stack", so I've no clue :/ . 19:34
TimToady did you run it? 19:36
.oO("Have you tried turning it off and on?")
 19:37
lue yep. I got control of gdb upon "[Inferior 1 (process 9733) exited normally]"
Maybe I have to do `gdb make` ... 19:38
TimToady you'd use gdb on the actual command, which means you might have to look through the wrapper shell scripts first to find the real command 19:39
lue I put the command right in front of the $(M_RUN_PERL6) --setting=NULL ... line. /me is trying gdb make now 19:40
TimToady though if you're running install/bin/moar, that seems to be an executable
dalek kudo-star-daily: 3260919 | coke++ | log/ (5 files):
today (automated commit)
 20:05
: 5f6c746 | (Nick Patch)++ | docs/feather/syn_index.html:
[feather] S15 is a draft
: 53bd7fb | (Nick Patch)++ | / (9 files):
change github http urls to https

since the official urls are all https in github now.
: a695ef5 | (Nick Patch)++ | / (2 files):
update dead github links
: e840845 | (Nick Patch)++ | util/Text-SmartLinks/lib/Text/SmartLinks.pm:
[Text::SmartLinks] fix html

the specs link was broken: s/href"/href="/
timotimo i think my post is ready, +/- proofreading 20:09
docs.google.com/document/d/1wqauGT...Nq8B0/edit ← if you'd like to help, spread comments :) 20:10
timotimo feather.perl6.org:3000 is unreachable again? 20:13
[Coke] timotimo: instead of putting it in a try/eval, I'd put it in a block, then impl. can skip the block if needed. 20:14
also, add a comment mentioning the RT when adding testneeded tests. 20:15
timotimo++
timotimo i'd rather suggest putting a block around the whole thing
because if there's no try/eval it'll blow up at compile-time 20:16
[Coke] not if you skip the block. 20:18
(that's what fudging is for)
just my 2¢. 20:19
tadzik timotimo: should be better now
timotimo ah, that's how you meant it
lue timotimo++ 20:20
hm. After running the setting compilation command through gdb, it somehow works now? (And gets stuck on Test.pm) 20:21
timotimo try setting the ulimit for core dump files up and attach gdb to the already-dead process?
timotimo [Coke]: what do you think about replacing try eval q{ with do { ...? 20:22
lue (not that I can run perl6-m, it merely finishes CORE.setting)
[Coke] timotimo: the standard at the moment seems to be ala S32-list/first.t 's setup - small blocks of self contained tests that you can #?skip if you want. 20:30
s/want/need/
timotimo right
that's not what i meant, though 20:31
[Coke] what advantage does using a do block give? 20:31
timotimo it makes a lexical scope so that the $/ doesn't leak around and grabs the last value inside it for our return value 20:32
r: my $res = { 1; 2; } say $res;
camelia rakudo-jvm 413650: OUTPUT«===SORRY!=== Error while compiling /tmp/IEbnaOTctL␤Two terms in a row␤at /tmp/IEbnaOTctL:1␤------> my $res = { 1; 2; } ⏏say $res;␤ expecting any of:␤ postfix␤ statement end␤ statem…»
..rakudo-parrot 413650: OUTPUT«===SORRY!=== Error while compiling /tmp/13l7dgRpyh␤Two terms in a row␤at /tmp/13l7dgRpyh:1␤------> my $res = { 1; 2; } ⏏say $res;␤ expecting any of:␤ postfix␤ statement end␤ sta…»
timotimo r: my $res = do { 1; 2; } say $res;
camelia rakudo-parrot 413650: OUTPUT«===SORRY!=== Error while compiling /tmp/39UbQcYvLK␤Two terms in a row␤at /tmp/39UbQcYvLK:1␤------> my $res = do { 1; 2; } ⏏say $res;␤ expecting any of:␤ postfix␤ statement end␤ …»
..rakudo-jvm 413650: OUTPUT«===SORRY!=== Error while compiling /tmp/OEOyYtxy8y␤Two terms in a row␤at /tmp/OEOyYtxy8y:1␤------> my $res = do { 1; 2; } ⏏say $res;␤ expecting any of:␤ postfix␤ statement end␤ sta…»
timotimo er, oops
r: my $res = do { 1; 2; }; say $res;
camelia rakudo-parrot 413650, rakudo-jvm 413650: OUTPUT«2␤»
timotimo r: my $res = { 1; 2; }; say $res;
[Coke] that doesn't deal with the #?skip issue.
camelia rakudo-parrot 413650, rakudo-jvm 413650: OUTPUT«Block.new()␤» 20:33
[Coke] and the bare {} will scope.
timotimo right. the #?skip issue is dealth with by having the whole thing surrounded by { }
[Coke] Yes.
and it sounds like you're arguing against that for this other thing.
timotimo oh, sorry about that
that wasn't what i meant
[Coke] I figured I was confused. :) 20:34
timotimo dwarring: in case you backlog, you don't need to publish your post today, i finished mine and it's okay 20:37
dwarring timotimo++ 20:44
BenGoldberg : $res = prog1 { 1; 2; }; say $res; 20:46
timotimo did you mean $res = sub prog1 { 1; 2; }; say $res;? 20:47
BenGoldberg No, sorry, I've got lisp on the brain :)
In lisp, (prog1 foo bar blatch) evaluates each of foo bar and blotch, and it's return value is whatever foo evaluated as 20:48
lue has had it with panda and its brokenness. 20:49
BenGoldberg Kinda like, in perl5, (do { foo() }, do { bar() }, do { blatch })[0]
tadzik lue: I know the feeling 20:50
lue Honestly, it's much faster/better/* to use a (slightly modified) ufo, even if lacking the warm and fuzzy insulation of a package manager :) . 20:51
lue will hesitantly try panda once FROGGS' S11 work is completed. 20:56
FROGGS yeah
need to discuss with jnthn and tadzik, and then we need to test every possible scenario :o) 20:57
lue :) I just hope the S11 changes de-break panda, though it seems to be an issue with parrot magically not doing things sometimes. :/ 20:58
FROGGS hmmm 20:59
the only thing my S11 changes to panda do is that it delegated the work of the .install method 21:00
lue I noticed that the random build failures seem to be a result of parrot *not even trying the command*, judging by how much faster I got "False" results v. the "True" results.
FROGGS a proper test case would help 21:01
but sounds like there is none :/
lue FROGGS: a non-random failure would help with that. e.g. just before I deleted my panda repo out of frustration, I got a totally brand-new failure on Shell::Command, which I'd not seen previously. 21:02
lue (my blaming of parrot comes from previously-seen parrotland errors.) 21:03
FROGGS if it is reproduceable then pull request that failing test or make an issue please
lue FROGGS: I'm failing the (re)bootstrap.pl script during panda's Builder usage. I'm not sure of any relevant test cases. 21:04
(And the failures are reproducible to a point; but like I mentioned it has a nasty habit of changing behavior somehow.)
lue hmm, I want to see what happens if I replace all instances of shell() with qqx... 21:09
lue *sigh* and now I've got a failure on JSON/Tiny.pm >.< 21:20
I'm still placing the blame on parrot, 'cos I've yet to get a hint pointing elsewhere :/ 21:23
FROGGS lue: what is your PERL6LIB btw? 21:24
and your @*INC ?
because recompiling panda's deps requires a special path ordering
lue FROGGS: $ echo $PERL6LIB 21:25
$ perl6 -e 'say @*INC'
</usr/local/lib/parrot stuff> /home/matthew/.perl6/2013.10-315-ga6eeb0b/lib
FROGGS maybe not soo special: the directory it wants install to must come before the panda/ext/JSON__Tiny or so
lue FROGGS: this is the *build* step that's failing, the one that generates the .pir 21:26
FROGGS it does that in the bootstrap script, but maybe that fails?
lue FROGGS: gist.github.com/lue/01c2f45754f97f98460a
(the install dirs should not have an effect on the build process) 21:27
FROGGS can you insert a --ll-exception after the $*EXE_NAME? gist.github.com/lue/01c2f45754f97f...-above-L13 21:28
lue tries that 21:29
FROGGS I had expected that it fails building something that has deps
lue FROGGS: I should let you know that the "git diff", if gen'd now, would look different from that, as well as what in bootstrap.pl fails. I only gave you that link to give you a representative example of the issue. 21:30
FROGGS k
does this command there build the .pir? gist.github.com/lue/01c2f45754f97f...command-L1
if not then you can put the --ll-exeption there 21:31
lue FROGGS: whenever I do a manual run of a failing file, I get a real file out of it.
FROGGS --ll-exception*
k
I had that too when fiddling with S11
FROGGS IIRC it failed because of "out of memory" 21:32
(according to parrot)
lue FROGGS: irclog.perlgeek.de/perl6/2013-12-15#i_8004264 :) 21:33
FROGGS :S
if I could remember what fixed it for me... 21:34
lue considers writing a parrot detonation script, to be more dramatic about the day he switches to MoarVM ...
FROGGS .oO( CTRL+a CTRL+d ) 21:36
lue I don't get anything by adding --ll-exception, so it truly seems to be an issue with the norwegian blue executing commands.
timotimo how do i format in-line code snippets in wordpress?
FROGGS I use <pre> >.< 21:37
timotimo that's not in-line
apparently i can use <code>
FROGGS ahh, yes
code
lue <code> is in-line, <pre> is block-formatting, and [code] is block-formatting+line numbers 21:38
timotimo yikes, the source code looks frightful 21:39
FROGGS compile it to something else then :P 21:43
timotimo :)
lue FROGGS: do you have any ideas as to how you stopped parrot from breaking panda? 21:44
timotimo perl6advent.wordpress.com/?p=2112&a...eview=true
FROGGS lue: no 21:44
[Coke] eyerolls at the constant parrot bashing. 21:45
lue will wait for MoarVM before trying panda
timotimo quite bashful, no?
lue [Coke]: when Parrot is preventing me from using a pretty important piece of software, I think it quite deserves it.
FROGGS lue: prove it that it is parrots fault 21:47
lue FROGGS: the one and only singular error message I have ever gotten comes from the parrot source code. Seems to me, a little bit, that Parrot is to blame. 21:48
takadonet Ulti: ping 21:48
[Coke] I blame nqp for misusing parrot. 21:52
timotimo :)
[Coke] I do get that it's frustrating, honest.
lue here's the stack trace from that one error message: gist.github.com/lue/bf883d13e7917067e765 21:53
[Coke] is there a ticket somewhere that shows the issue from beginning to end 21:54
?
lue [Coke]: not that I'm aware of. This issue pops up semi-randomly, so I'm not sure what use a ticket is until I figure out the root cause. 21:55
takadonet phenny, tell Ulti that bioperl6 is not abandonware just myself and cjfields were waiting for some elements of the spec to stop changing. We did start the port a few ago so some code looks very old. Hoping to start working on it asap.
... did not think that worked
[Coke] try a colon. 21:56
lue tries running the bootstrap itself with --ll-exception
[Coke] jnthn: using perl6-j, using the example on perl6advent.wordpress.com/2013/12/0...-the-jvm/, I get this error: 21:57
No such method 'method/update/(B)V' for invocant of type 'java.util.zip.CRC32'
... oh. we can't call back into perl6 code from java yet, making me doubt we could implement an interface. 21:58
timotimo no, we can implement interfaces
at least some ...
timotimo there really isn't an update with B, just with I or B[] 21:58
is that how you spell it?
byte[] is the type in the signature
[Coke] the error and the callback were separate questions, btw.
lue [Coke]: I see now that it's probably NQP to blame, but I'm still blaming parrot until I know for sure what's happening. How often do parrot errors bleed through and assume the guise of a higher-level error? 21:59
[Coke] lue: nevermind, I cancel my end of this conversation. 22:00
timotimo meeeeh, my java_interop_dispatchers branch stopped working altogether :( 22:01
[Coke] timotimo: using (I) gives me This representation can not unbox to a native int
timotimo right, because you have a whole buffer
rather than just a single integer 22:02
you could write for @$_ at the end, though :P
but yeah, the example needs fixed
lue [Coke]: I was honestly asking how often the errors bleed through. It's why I still place blame on Parrot, really. Because I don't know if that bleed-through is normal or the fault of parrot messing up. 22:03
(adding ll-exception for the script itself revealed nothing :/) 22:04
[Coke] sticks with straight java for now. (sadface) 22:05
lue Come to think of it, how would NQP cause Parrot to claim it can't allocate memory? ... 22:08
If I'm reading the stacktrace correctly, it's not actually showing me to where in the qqx/shell/run/* execution the error pops up. 22:10
timotimo jnthn: did you see coke's concerns with the java interop post? 22:11
jnthn timotimo: Well, I only talk code I had run successfully at some point...didn't verify it on latest, and assumed it'd keep on working. Odd that it woulda broke, though. 22:17
*took code
timotimo strange, yeah 22:18
jnthn Though, the whole really needs some tests writing.
Otherwise we'll never know when it breaks. 22:19
timotimo yeah. like my build is just completely breaking and i don't understand it at all
timotimo Unhandled exception: java.lang.NoSuchMethodError: org.perl6.nqp.runtime.Ops.printfh(Lorg/perl6/nqp/sixmodel/SixModelObject;Ljava/lang/String;Lorg/perl6/nqp/runtime/ThreadContext;)Ljava/lang/String; 22:20
when doing --output=gen/jvm/stage2/NQPHLL.jar gen/jvm/stage2/NQPHLL.nqp
lue does rakudo register its own nqp::open? Because I can make no sense of control.pm:202 22:23
TimToady timotimo: it still mentions 'eval' 22:25
timotimo oh! good thing i didn't hit publish yet 22:26
ah yes, i only changed the text below the code, not the text above it
jnthn timotimo: Grr. That's 'cus the stage0 depends on an op that's no longer there.
timotimo ah damnit :( 22:27
jnthn timotimo: If we do a NQP JVM rebootstrap at some point, we'll clear this up.
Feel free to do it. It's better than somebody hitting this issue once every week or two.
timotimo edit'd 22:29
lue Finding a commonality between shell and qx to figure out my woes is not going so well :/ (I've followed shell down to the nqp ops level so far, and qx down to the parrot ops level.)
lue &
[Coke] It's been so long since I wrote XMPP code, I figured things had to be better by now. :P 22:44
tadzik oh, xmpp is fun :) 22:45
[Coke] the java library I'm using has completely different interfaces for chat vs. multi-user-chat, and, better, you can't run both listeners simultaneously. 22:46
tadzik doh 22:47
the only client I ever wrote didn't have support for muc
FROGGS timotimo++ # advent post :o) 22:51
timotimo it'll be published 0100, whenever that is in the blog's timezone 22:53
if it doesn't pop up at the right time, feel free to publish it manually for me :) 22:54
timotimo maybe i'll do the stage0 regen tomorrow 23:00
i'm kinda tired. but maybe i'll play a vidya game before i go to bed
lue I can't seem to find a common thread between qx and shell. This is frustrating. 23:10
woolfy timotimo++ for the advent calendar blog. Interesting stuff, a bit over my head, but I got the gist. And I have just two remarks, which I will send separately. 23:43
timotimo : it looks like you wrote the beginning of what can be a primer on how to make tests for Perl 6. Or does that already exist? 23:45

Powered by Cro and the Raku® Programming Language.

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

Thank you!