feather's all normal now :)
Set by Juerd on 21 May 2007.
00:02 Limbic_Region joined
Limbic_Region salutations all 00:05
00:06 kaether joined
moritz hi Limbic_Region ;) 00:06
00:12 rlb3_work_ joined
dduncan hi ho 00:13
stevan_ polettix: your welcome, you can always ask on #moose over on irc.perl.org as well 00:14
polettix thanks 00:15
stevan_ polettix: the #moose crowd is pretty helpful :) 00:18
polettix some time ago I tried to start a new project with moose, but I was quite in hurry and I eventually falled back to Object::InsideOut
but with some nasty feeling that I was missing a good chance 00:19
also considering that (for what I know) moose could be a soft bridge towards Perl 6 00:20
BTW, is there a best practice to write Perl 6? Should it be a single word "Perl6" or two separate tokens? 00:21
polettix confesses not to have RTFM regarding this
moritz polettix: Perl 6 00:23
00:25 kaether joined
polettix ok, thanks for bearing with my non-constructive lazyness. Good night for those that live where it's night :) 00:25
stevan_ night
moritz goes to sleep as well 00:26
night
00:29 weinigLap_ joined 00:33 weinigLap_ joined 00:35 irc joined, cognominal joined, JarJarBinks joined 00:36 weinigLap joined 00:38 amnesiac joined 00:57 [particle1 joined 01:02 nipotaway is now known as nipotan 01:04 SubStack joined
bsb dduncan: I'm looking at sql ast possibilities for $work and many paths link to you. I'd appreciate your advice on what to check out 01:07
currently it's perl5, but the build stages could potentially involve other things 01:09
dduncan bsb, my AST is for relational databases in general, and isn't specific to SQL, but you can parse SQL to it and generate SQL from it, either with minimal or more indirection depending what SQL features you use 01:11
I also made an AST once that was specific to SQL, which was on CPAN as Rosetta::Model, but I abandoned that in favor of QDRDBMS::AST 01:12
regardless, I'm open to discussing your needs and helping as I can 01:13
for your $work, do you have something in house you started, or are you trying to save yourself from making your own?
bsb Thanks, I'll check out Rosetta and QDRDBMS 01:14
the later
dduncan just keep in mind that Rosetta is officially aborted
QDRDBMS is a rewrite
and Rosetta never actually worked, per se
bsb :) 01:15
my dream is a declarative validation framework, but that's a long way above the gutter of sql 01:16
dduncan depending on your time table, you could help enhance QDRDBMS to meet your need
it is intended to be what SQL DBMSs could have been were they not led astray 01:17
01:17 trunix joined 01:18 [particle1 is now known as [particle]
dduncan right then, lets talk later ... meanwhile, I suggest you check out mm.DarrenDuncan.net/mailman/listinfo 01:18
lambdabot Title: mm.darrenduncan.net Mailing Lists
dduncan I set up lists specific to QDRDBMS there
that might be better fora for discussions 01:19
bsb will do, thanks again
01:23 irc joined, cognominal joined, JarJarBinks joined
dduncan dinner & 01:28
01:28 JarJarBinks|meta joined 01:34 vel joined 01:36 aukjan|gone joined 01:39 stevan__ joined 01:58 japhb joined 02:20 miyagawa joined 02:30 drbean joined 02:57 fridim joined 03:09 REPLeffect joined
PerlJam Anyone awake and happen to remember the reasons why **{} looks the way it does? 03:17
03:26 stevan__ joined 03:29 avarab joined 03:38 Alias_ joined 03:46 mncharity joined
mncharity observers that python is generally good at guiding/confining people in the direction of clean code. with one major caveat. dsl's are not part of the vision. so when you really really want a dsl, eg, to abstract away gui hair, python... leaves you in a nest of gui hair. :( 03:48
pining for p6.... 03:49
Tene mncharity: Why not start using it now? 03:51
mncharity it's not xmass yet. 03:52
03:52 mncharity is now known as putter
Tene huh. Apparently the IRC bot I left running ran out of memory. 03:52
although, on reflection, that's probably my mistake, not pugs' 03:53
putter g'night & 03:56
04:18 devogon joined 04:21 anatolyv joined 04:23 thoughtpolice joined 04:26 zperl joined 04:27 drupek12 joined
zperl hi 04:27
?eval my $N is constant = 5; $N = 10; say $N; 04:29
04:29 evalbot_r16488 is now known as evalbot_r16493
evalbot_r16493 OUTPUT[10ā¤] Bool::True 04:29
zperl hmmm how can i make $N constant and readonly?
anyone? 04:32
TimToady ?eval constant $N = 5; $N = 10; say $N 04:34
evalbot_r16493 Error: Can't modify constant item: VInt 5
TimToady "is constant" is a fossil
zperl cool ty larry 04:35
TimToady there is now an "is readonly" which applies readonlyness to something that is intrinsically writable
that's what parameters to functions are by default
but the constant declarator lets you declare a compile-time constant
(scopes like my) 04:36
04:37 toku_away is now known as tokuhirom_
zperl ?eval my constant $N = 5; $N = 10; say $N; 04:38
evalbot_r16493 OUTPUT[10ā¤] Bool::True
zperl interesting 'constant' is ignored by pugs? 04:39
TimToady when used as a type, yes
zperl so is it a bug or a feature?
TimToady fossil, going away
zperl ok, cool
TimToady gotta go pick up my son from the airport... 04:40
zperl any place where fossils are documented so we can avoid?
TimToady eh, not really
zperl ok larry, cya and ty.
TimToady just memorize the synopses and you'll be fine
:)
afk &
zperl lol
04:40 kunwon1 joined 05:03 chylli joined 05:04 chylli left 05:07 BooK_ joined 05:25 skrit joined
Aankhen`` ?eval constant $N = 5; $N = 10; $N.say 05:25
evalbot_r16493 Error: Can't modify constant item: VInt 5
05:33 jisom joined
Aankhen`` Couple of questions about grammars. Can grammars export regexes? Is the syntax for matching from a specific grammar always <My::Awesome::Grammar.foo> to use the `foo` subrule from My::Awesome::Grammar? 05:34
(My understanding is that the answer is "yes" to both.)
05:43 skrit left 05:45 kanru joined 05:48 rfordinal joined, rashakil_ left
thoughtpolice Aankhen``: a grammar is a class, too. it just inherits from Rule, so you're basically working the same way. 05:56
05:57 rfordinal joined
Aankhen`` Okay. 05:58
06:23 renormalist joined 06:25 Averell joined 06:34 skrit joined 06:35 skrit left 06:36 IllvilJa joined
xinming what the related name for ->meta->make_immutable in moose is in perl 6 please? 06:46
06:47 weinigLap joined 06:54 iblechbot joined 06:56 kane_ joined 06:57 kane_ left 06:58 kane-xs joined
renormalist ?eval my $in = "y" ; given $in { when "y" { say "yes" } }; 06:58
evalbot_r16493 OUTPUT[yesā¤] Bool::True
renormalist ?eval my $in = "y" ; given $in { when ({ $_ ~~ "y"}) { say "yes" } };
evalbot_r16493 OUTPUT[yesā¤] Bool::True
renormalist ?eval my $in = "y" ; given $in { when $_ ~~ "y" { say "yes" } };
evalbot_r16493 undef 06:59
07:00 kane-xs joined
renormalist Why do I have to put ({ }) around an explicit smart match expression, as seen in my second eval. Without the paran/brace (3rd eval) it doesn't work. 07:01
07:02 franck__ joined 07:03 obvio171_ joined 07:11 justatheory joined 07:12 justatheory joined, franck___ joined 07:16 japhb joined 07:22 beppu joined 07:24 kane-xs is now known as kane_
svnbot6 r16494 | Aankhen++ | * added examples/rules/Grammar-URI-RFC1738.pm 07:30
07:39 dduncan left 07:51 LaserQ joined
LaserQ What's the eqiv to sort ( $hash{$b} <=> $hash{a} ) in perl6? 07:52
Aankhen`` @foo.sort:{ $^a <=> $^b }, most likely.
lambdabot Unknown command, try @list
LaserQ Ah, twigils. Thanks 07:53
Aankhen`` NP.
You don't necessarily have to use placeholder variables like that.
Tene @foo.sort: { %hash{$^a} <=> %hash{$^b} }
lambdabot Unknown command, try @list
07:53 dmq joined
Aankhen`` You can specify any block that takes two variables. 07:53
Er, yeah, with %hash too.
Aankhen`` blushes.
@foo.sort: -> $a, $b { %hash{$a} <=> %hash{$b} } 07:54
lambdabot Unknown command, try @list
Aankhen`` @foo.sort: sub ($a, $b) { %hash{$a} <=> %hash{$b} }
lambdabot Unknown command, try @list
Aankhen`` To the best of my knowledge, any of those should suffice.
Tene There was a proposal to have sort() inspect the arity of its code block and automagically parse blocks with single arguments only into sort: { &code($^a) <=> &code($^b) }
Aankhen`` Say what?
Ah.
Hmm.
Tene Although perhaps that should have a different keyword 07:55
sort_by, ocelot
Aankhen`` So basically you take the result of the_block($^a) and the_block($^b) and compare those?
gaal mooses
Tene Right.
Aankhen`` capybaras gaal.
Ah.
Sounds neat.
LaserQ %tally.keys.reverse.sort: { %tally{$^b} <=> %tally{$^a} } gives me Undeclared variable: "$^b" 07:56
svnbot6 r16495 | yiyihu++ | r16572@HomePc: xinming | 2007-05-18 06:11:39 +0800
Aankhen`` Okay, I think I've ported all the easy Regexp::Common modules.
svnbot6 r16495 | yiyihu++ | Renamed from_moose_0.21 to from_moose
Aankhen`` All three of them.
LaserQ Can yo only do that kind of sort with a .kv?
Aankhen`` Although I did end up writing a grammar for RFC 1738 in the process.
LaserQ: Er, where is this giving you an error?
Tene There are lots of convenient little operators like sort_by that would be nice to have builtin, but are only a line or two to implement.
Aankhen`` And why do you reverse the keys when you're sorting them anyway? 07:57
Tene: Fair enough.
LaserQ That's a very good question. Ops. Still get the eror with that bit removed though
Aankhen`` ?eval my %foo = (a => 'c', b => 'b', c => 'a'); %foo.keys.sort:{ %foo{$^a} <=> %foo{$^b} } 07:58
07:58 evalbot_r16493 is now known as evalbot_r16494
evalbot_r16494 ("a", "b", "c") 07:58
Aankhen`` Uhm.
Tene LaserQ: pugs doesn't handle auto-magic parameters properly yet. It handles them in the order they appear in the code, not in alphabetical order.
07:58 kane_ joined
Aankhen`` ?eval my %foo = (a => 'c', b => 'b', c => 'a'); %foo.keys.sort:{ %foo{$^a} cmp %foo{$^b} } 07:58
evalbot_r16494 ("c", "b", "a")
Aankhen`` There we go.
svnbot6 r16496 | yiyihu++ | r16665@HomePc: xinming | 2007-05-22 15:56:04 +0800
r16496 | yiyihu++ | An attempt to translate 000_load and 001_recipe moose tests in to perl 6.
Aankhen`` And what Tene said.
'Cept s/auto-magic/placeholder/. :-P 07:59
Tene Yeah. Couldn't remember the term. Thanks.
Aankhen``++
Aankhen`` Oh, please. I'm always happy to correct other people. ;-) 08:00
LaserQ Aankhen`` - your sort worked so I'm made a mistake somewhere. I'll have a dig, thanks for the pointers
svnbot6 r16497 | Aankhen++ | * added examples/rules/Regexp-Common.pm (only contains `list`, `ws` and `lingua` patterns so far).
Aankhen`` LaserQ: <Tene> LaserQ: pugs doesn't handle auto-magic parameters properly yet. It handles them in the order they appear in the code, not in alphabetical order.
LaserQ oohh 08:01
Aankhen`` LaserQ: You didn't make a mistake, Pugs just doesn't handle placeholder variables correctly as yet.
You can work around it for now by reversing $^a and $^b. And presumably altering your algorithm accordingly.
@quote ghc
lambdabot ghc says: Illegal constraint
Aankhen`` @quote fortune 08:02
lambdabot No quotes match. Are you on drugs?
Aankhen`` :-(
Tene @quote tene
lambdabot Cale says: desrt: Did you hear about the trivial religion that I came up with? Its tenets consist of a single statement: "Believing in this statement will make you happier". It's like a terminal
object in the category of religions and theomorphisms.
Aankhen`` Yeh, just realied it, heh.
Realized, even.
And LMAO at the quote. 08:03
LaserQ Thank you both.
renormalist @list 08:06
lambdabot www.cse.unsw.edu.au/~dons/lambdabot/COMMANDS
renormalist @help
lambdabot help <command>. Ask for help for <command>. Try 'list' for all commands
renormalist @help karma 08:07
lambdabot karma <polynick>. Return a person's karma value
renormalist @karma renormalist
lambdabot You have a karma of 11
renormalist @karma audreyt
lambdabot audreyt has a karma of 815
renormalist @karma TimToady
lambdabot TimToady has a karma of 36
meppl good morning 08:10
renormalist @help quote
lambdabot quote <nick>
remember <nick> <quote>
Quote somebody, a random person, or save a memorable quote
renormalist @quote
lambdabot AdamPeacock says: Once I looked at the source code, 25000 lines of ASP, I reckoned it would be easier to rewrite it in a real language.
renormalist @quote 08:11
lambdabot Syzygy- says: Why do I get the feeling that the arrow code was written during "Speak like a pirate day"?
08:15 kane_ joined
renormalist a wording question: Does one use the word "twigil" also for the parameters of functions, eg sub quicksort(*$x, *@tail) 08:20
or are twigils just what starts with $, %, @, like $.accessor?
s/parameters of functions/modifiers of parameters of functions/ 08:25
Aankhen`` Hi meppl. :-) 08:26
renormalist: A twigil is a secondary sigil, so only things like "*" in $*FOO or "^" in $^foo are twigils. 08:27
Or "." in $.accessor, since you mentioned it.
renormalist k, thx 08:28
Aankhen`` Happy to help.
Aankhen`` wonders if eric256 has abandoned examples/rules/basic.pl 08:29
(Most likely because of vastly incomplete implementations.)
His last commit was r10195, so I'm guessing yes.
moritz good morning ;) 08:31
08:31 VanilleBert joined
Aankhen`` Eek, it's moritz. Run. 08:32
Tene He killed audreyt!
Aankhen`` Oh the huge manatee~! 08:33
s/~//
Stupid ~ key keeps sneaking in there.
renormalist Aankhen``: you have a special Perl6 keyboard? :-) 08:34
Aankhen`` Shh, don't tell anyone. ;-)
renormalist as we just speak about Perl6: can someone explain me what the pipe symbol in the quicksort example (examples/algorithms/quicksort.pl) do: return(quicksort(|@pre), $x, quicksort(|@post)); 08:36
Aankhen`` Well, |@pre is a Capture, like |@post. Unfortunately, I'm not sure I understand them completely. 08:38
I *think* what's happening there is that it's sort of "flattening" @pre to fill the argument list of quicksort. 08:39
Although "flattening" might be a particularly bad word to use.
08:42 offby1` joined
Aankhen`` I could have sworn there was a POD document on captures. 08:42
Ah, there's a docs/Perl6/FAQ/Capture.pod 08:43
Though I have no idea how up-to-date that is. 08:44
S02 says: "Capture Function call arguments (right-hand side of a binding)"
Actually, just read S02 from "An argument list may be captured into an object with backslashed parens", that might give you some idea.
08:45 rho joined 08:46 franck__ joined
renormalist hm, not sure I understand, maybe it's indeed due to the use of := binding of @pre/@post. that I also don't understand, why it's done. Or generally when I should use binding. 08:49
Aankhen`` I'm afraid I don't really know. :-( 08:50
renormalist m'kay 08:51
09:04 rho joined
renormalist ?eval 3 ~~ 1 .. 10 09:11
09:11 evalbot_r16494 is now known as evalbot_r16497
evalbot_r16497 Bool::False 09:11
moritz ?eval grep {$_ == 3}, 1 .. 10 09:13
evalbot_r16497 (3,)
moritz ?eval 3.grep(1..10) 09:14
evalbot_r16497 Error: Cannot cast from VList [VInt 1,VInt 2,VInt 3,VInt 4,VInt 5,VInt 6,VInt 7,VInt 8,VInt 9,VInt 10] to Pugs.AST.Internals.VCode (VCode)
09:25 riffraff joined
riffraff hi 09:26
moritz hi riffraff ;)
09:28 buetow joined
riffraff too bad that $x does role doesn't work yet :/ 09:29
renormalist how can I count how many tests the testsuite contains without running it? 09:41
or can someone tell me current numbers
moritz renormalist: you can look into the smokes on smoke.pugscode.org 09:42
renormalist: there is a count at the very bottom
renormalist indeed, thx 09:43
is it really around 18.000 tests? sounds good :-)
moritz I had 20.000 in mind 09:45
I think a few of the test files don't compile currently, so the plan $n; is never executed...
riffraff renormalist, well, not all opf them pass though 09:46
moritz and they don't count in that statistic
renormalist can i sum the numbers in "17954 ok, 0 failed, 4638 todo, 747 skipped " to 17954+4638+747==23339 ? 10:00
10:01 audreyt joined
riffraff I remember more than 17954 10:02
there a re a lot of new TODOs, I believe
moritz it explicitly says "18590 test cases" 10:03
riffraff completely OT: what does pile-drive means?
moritz or something 10:04
renormalist: just take that number
riffraff is been doing make smoke for a while now
10:11 lampus_ joined 10:22 yves_ joined 10:30 LCamel joined 10:38 trunix joined
svnbot6 r16498 | audreyt++ | * Pugs.Compat, Pugs.Embed.Parrot: Change the PUGS_HAVE_POSIX 10:40
r16498 | audreyt++ | check in .hsc from reading #include <../pugs_config.h> into
r16498 | audreyt++ | using a _PUGS_HAVE_POSIX variable, because relative paths
r16498 | audreyt++ | under .hsc is not reliable (in GHC-HEAD's Cabal, .hsc files
r16498 | audreyt++ | generates dist/build/Pugs/Embed/Parrot.hs rather than the
r16498 | audreyt++ | assumed src/Pugs/Embed/Parrot.hs).
r16499 | audreyt++ | * UTF8.hs - GHC.Prim is not visible in GHC-HEAD anymore; 10:46
r16499 | audreyt++ | change import to GHC.Base.
11:29 mugwump left 11:32 VanilleBert left 11:33 avarab is now known as avar 11:35 rgs joined 11:38 rfordinal_ joined
svnbot6 r16500 | audreyt++ | * Move src/RRegex into third-party/hsregex/, both to simplify the 11:41
r16500 | audreyt++ | building process (no more mucking with src/pcre/pcre.o etc), and
r16500 | audreyt++ | to continue the move of third-party projects (see README) for
r16500 | audreyt++ | clarifying licensing/redistribution.
11:41 ggoebel joined 11:42 deq` joined
svnbot6 r16501 | audreyt++ | * Remove pcre/re-override-*; they are thoroughly replaced by 11:42
r16501 | audreyt++ | the re-engine family of modules on CPAN.
11:43 mako132_ joined 11:44 prism joined
svnbot6 r16502 | audreyt++ | * Thoroughly remove src/pcre and src/RRegex in favor of 11:48
r16502 | audreyt++ | third-party/hsregex/.
r16503 | audreyt++ | * gen_prelude.pl: Do not precompile a module more than once..
r16504 | audreyt++ | * Pugs.Meta.Str: The "split" method should be named "split". :) 11:51
r16505 | audreyt++ | * docs/notes/mo-bootstrap-plan.pod: Spellchecking. 11:54
11:57 chris2 joined 12:00 nipotan is now known as nipotaway
svnbot6 r16506 | audreyt++ | * Unbreak the build. 12:09
moritz fears for the debian packages ;) 12:12
but since I'm not too keen on learning for my exams, I'll spend some time fixing it if something breaks ;) 12:13
12:13 iblechbot joined, the_dormant joined 12:19 Alias_ joined, rgs is now known as args 12:23 rodi joined
avar ew @ re::override:) 12:27
12:28 f0rth joined 12:42 offby1` is now known as offby1 12:45 bloonix joined
bloonix salute 12:46
12:46 f0rth joined
moritz hi bloonix ;) 12:46
bloonix hey moritz, what up? :) nice to meet you again
moritz bloonix: I'm thinking about two dimensional regexes 12:47
bloonix *hehe*
moritz to extend regex matching to image data
but so far I'm not very successfull ;)
args hmm image spam detection
moritz not only spam
other features as well 12:48
bloonix use bugs as a feature moritz ;)
moritz bloonix: that's your job ;) 12:52
bloonix i plead guilty 12:54
moritz executes bloonix
13:11 stevan__ joined 13:22 [particle1 joined 13:23 audreyt joined 13:27 buetow joined 13:29 pbuetow joined, rfordinal_ is now known as rfordinal 13:34 buetow__ joined 13:35 vel joined 13:38 yves_ joined
moritz feather.perl6.nl and perlcabal.org don't have a favicon 13:42
maybe we should use the same as pugscode.org?
masak moritz: I've been thinking at times about 2d regexes in relation to Hex: en.wikipedia.org/wiki/Hex_(board_game) 13:49
it carries additional difficulties due to the non-rectangular layout of the board...
but I'd need something like that to do efficient pattern matching on a board 13:50
13:50 [particle1 is now known as [particle] 13:55 kaether_ joined 14:03 VanilleBert joined 14:06 devogon joined 14:15 rhandom joined 14:18 renormalist joined 14:34 REPLeffect joined
renormalist What are the most important URLs that people should know of, when they hear from Pugs in an article? 14:35
moritz www.pugscorde.org and the svn url ;)
VanilleBert run.pugscode.org/ 14:36
if it would work
moritz if it doesn't work, report to andara++ (iirc) 14:37
VanilleBert it redirects me to feather.perl6.nl:8080/runpugs/ and no connection can be made 14:38
moritz that's propably due to feather's reboot 14:39
renormalist: what are you writing about pugs? 14:44
renormalist: maybe you could link audrey's talk on pugs+perl6 at google video
renormalist moritz: I'm working on an article in a magazine. Google video might be cool. 14:48
moritz renormalist: which magazine? ix? 14:49
renormalist y
what's the url of audreys talk I already had it but forgot
moritz video.google.de/videoplay?docid=-38...6103839772 14:51
lambdabot Title: Audrey Tang - Perl 6 Today - Google Video
14:55 mikfire joined, mikfire is now known as mikfire09 14:57 rindolf joined
renormalist moritz: thx, I made it short: xrl.us/perl6today 15:04
lambdabot Title: Audrey Tang - Perl 6 Today - Google Video
15:09 amnesiac joined
mikfire09 My pugs install seems to be missing HsBridge.pm. Could somebody tell me how to fix this? I want to play with the new match syntax but cannot get past this. 15:10
15:13 the_dormant joined
rho is away: (ice cream time) 15:20
15:20 _bernhard joined
TimToady mikfire09: do you have a file in ./blib6/pugs/perl5/lib/Pugs/Runtime/Match/HsBridge.pm ? 15:21
(we don't generally try to install pugs because it tends to interfere with future development, with the consequence that it tends not to be maintained much either...) 15:22
mikfire09 That is a fair answer 15:23
I may have to back what I did out and try again the old fashioned way.
TimToady some of us just symlink /usr/bin/pugs into the build directory 15:24
mikfire09 I can set my PATH easier. Hmmm. Thanks Tim, I think that was the answer I needed.
15:27 moritz_ joined
Juerd Is there a place where one can pull up to date versions of the synopses from? 15:28
POD preferrably
moritz pugscode.org | Think twice before installing pugs 15:28
kolibrie Juerd: svn.perl.org/perl6/doc/trunk 15:31
lambdabot Title: Revision 14400: /doc/trunk
Juerd kolibrie: Thanks 15:32
kolibrie np
renormalist @help seen 15:34
lambdabot seen <user>. Report if a user has been seen by the bot
renormalist @seen rafl
lambdabot I haven't seen rafl.
renormalist oh
15:35 SubStack joined 15:39 myren joined
masak @seen lambdabot 15:44
lambdabot Yes, I'm here. I'm in #friendly-coders, #dreamlinux-es, #xmonad, #unicycling, #perl6, #parrot, #oasis, #jtiger, #haskell-soc, #haskell-overflow, #haskell-blah, #haskell, #ghc, #gentoo-uy, #gentoo-
haskell, #darcs and #scannedinavian
masak nice.
[particle] can somebody kick lambdabot from #parrot--that channel isn't used on freenode 15:45
will the real #parrot please stand up?
allbery_b @tell dons
also remind him to rejoin #haskell on irc.perl.org :) 15:46
moritz are '@tell's scoped to channels?
allbery_b nope
they're scoped to nicks 15:47
moritz cool ;)
[particle] @tell dons lambdabot is on #parrot, but that channel isn't used on freenode (the real #parrot is on irc.perl.org) 15:49
lambdabot Consider it noted.
15:50 weinigLap joined 16:03 kanru joined 16:08 iblechbot joined 16:29 rodi left
rindolf Hi all. 16:35
16:36 ribasushi joined
rindolf Happy Shavuoth everybody. 16:36
ribasushi hi
anyone here uses/used Petal? Although the documentation specifically states it works with HTML as well as XHTML, it can not parse validated HTML 4.01 16:37
nothingmuch: rindolf suggested you are the one that might know
16:40 the_dormant joined
rindolf ribasushi: maybe try #petal on irc.perl.org, assuming this channel exists. 16:40
ribasushi rindolf: nope 16:41
I am still milking google
nothingmuch ribasushi: i don't like petal 16:43
i used to
but then I realized it's only useful for the first version of a templat
e
because it takes too much work to tweak small things
nowadays I use TT if it's klunky
or HTML::Element if it's clean 16:44
YMMV
it's nice to have a "100% guaranteed to be valid" template
but you can get that with HTML::Element, Template::Declare or HTML::Seamstress ( basically HTML::Element on crack) 16:45
it also has issues with unicode
ribasushi nothingmuch: I am into petal because of the wysiwyg friendlines
nothingmuch yeah, i thought that pipe dream was worth it
i'm not sure now
HTML::Seamstress should get you the same friendliness
basically you parse the HTML 16:46
and run replacement patterns on it
ribasushi reading... 16:48
can you tell me anything on my question though?
nothingmuch i don't think it can
it's an XML parser 16:49
ribasushi so the input/output options are leftovers from the pre-xml versions I suppose?
16:58 buetow joined, damog joined, renormalist joined 17:07 Psyche^ joined, Psyche^ is now known as Patterner 17:14 prism joined 17:15 elmex joined
renormalist is it correct to use the term "Perl6" (without space) in an article, or should I always put a space before the 6? Opinions? 17:16
rhandom On the conversation of template engines, I'm releasing a new version of a year old module today that keeps its TT2 and TT3 support and adds HTML::Template and HTML::Template::Expr support - now for the pugs/perl6 connection - the engine is grammar based and I really can't wait to rewrite it in perl6 - I guess its time to start :). 17:17
I am hoping I can define my grammar, refine the AST, and play it natively in different runtimes. 17:18
myren is parrot still seen as "the" engine which will ultimately host perl6? 17:19
listening to mr audrey tang's presentations it started to seem in ways that perl6 was the perl6 runtime 17:20
[particle] renormalist: i think "Perl 6" is more appropriate
myren: parrot as "an" engine for perl6 17:21
TimToady renormalist: we tend to prefer "Perl 6" in order to indicate that it's version 6 of the language "Perl" 17:23
otherwise people think "Perl6" is the name of the language 17:24
avar mp6 and co are looking much more viable than parrot
TimToady however, I'm not as picky as some people... :) 17:25
avar: different levels--they can be complementary
avar they can, but mp6 emitting perl5, java bytecode etc. would allow it to target different vm's 17:26
17:27 Arathorn joined 17:28 BuRn-X joined, BuRn-X left
renormalist TimToady: ok, I will use "Perl 6" (with space). the same when referring to Perl 5 17:29
rho is back 17:30
17:33 franck__ joined 17:34 rodi joined
renormalist brb 17:36
PerlJam TimToady: why is **{} writ that way? (as opposed to *{} for instance) 17:37
TimToady because otherwise it requires a space to distinguish * {} from *{}, and I thought it would be better to keep it visually distinct as well
PerlJam TimToady: I was reading S05 last night and **{} still seems like an oddball. Specifically the difference between procedural vs. declarative. Everywhere else it seems procedural has doubled chars, and declarative has single chars. 17:38
17:38 rafl joined 17:39 renormalist joined
TimToady well, we might re-evaluate now that all non-alpha are considered meta by default 17:40
17:41 _bernhard joined
PerlJam okay, don't let it impinge upon your brain too much, but a re-eval would be nice :-) 17:42
17:42 justatheory joined 17:46 dduncan joined 17:57 kaether_ is now known as kaether, the_dormant joined
TimToady kaether: greetin's 17:58
would you like a commit bit? :)
kaether Can I have two? 18:00
Tene You going to pass the second one on to someone else? 18:02
kaether I was thinking about ebay...
moritz ;-)
Tene Heh.
TimToady gimme two email addresses and you can have two bits... 18:03
18:03 SubStack joined
[particle] hums "shave and a haircut" 18:03
PerlJam I could use both of those 18:05
kaether I was thinking that maybe the second bit would be immutable. 18:06
TimToady if your second email address is immutable, shouldn't be a problem... 18:07
actually, I know your immutable email address, it's the mutable one I don't know. :)
[particle] i have a mutable email address, with gmail 18:10
TimToady yeah, well, heisenberg and all that... 18:11
[particle] [email@hidden.address] 18:12
TimToady commitbit seems to be hosed 18:16
spewing bare templates at me...
Tene I remember audreyt saying that she was turning it off because it was having too many issues.
TimToady I don't suppose there's a commandline interface that can be used on feather... 18:17
obra TimToady: let me look. I have admin bits now 18:18
TimToady yeah, I remember her turning it off too 18:19
even said where to turn it back on at the time
18:19 japhb joined
obra TimToady: this was after I last messed with it? 18:19
[particle] yes, this was yesterday 18:20
obra I see her running a current rsync of commitbit
which leads me to not want to touch it
(like 1 minute ago) 18:21
TimToady moritz.faui2k3.org/irclog/out.pl?ch...21#id_l728 18:22
lambdabot Title: IRC log for #perl6, tinyurl.com/yspw57
moritz how can I tell irssi not to hilight my nick if it's followod by a dot '.'?
renormalist Perl-6.0.0-STD.pm currently doesn't compile with pugs -c. The log says it worked some time ago. Is it currently meant to be parseable by pugs or is this not a current goal? 18:23
TimToady audrey does not appear to be logged on to feather
sure it wasn't a cron job?
obra audreyt 26872 0.2 0.0 2980 1484 ? Ss 20:20 0:00 rsync --server --sender -vulogDtprz . /data/svn/commitbit
18:23 jisom joined
obra Not sure it wasn't a cronjob 18:23
I may have just had good timing
TimToady oops, now she's there
moritz renormalist: I think you have to pipe it through "cheat"
TimToady yes, currently you have to use "cheat" 18:25
as pugs's parser improves, we can remove transformations. when "cheat" turns into "cat", we'll have made progress. 18:26
diakopter s/he// 18:27
TimToady will the intermediate state be ceat or chat?
"Yes! Both! Neither!" says QM. 18:28
18:29 prly joined
[particle] c[h|e]at 18:29
obra I'd really like to get commitbit off this snapshot from november 18:30
renormalist moritz, TimToady: cheat? what's that?
ah, cat 18:31
moritz renormalist: a p5 script that transforms some syntactic sugar into something that is easier to parse
but not so nice to read
/src/perl6/cheat in the pugs repository 18:32
renormalist moritz: thx I found it. works with that 18:34
TimToady smile when you say "works"
18:36 REPLeffe1t joined, prly_ joined
Tene moritz: irssi has regex as an option to specify hilights. 18:37
moritz: /help hilight
moritz Tene: I know it has regexes, but it's a bit... unusual (compared to PCRE) 18:39
18:41 prly__ joined
moritz damned, `man 7 regex` doesn't seem to mention lookahead assertions 18:41
obra commitbit Database vacuumed and indexed and tested with the standalone server. I expect it to be back in a bit. 18:42
(renabled in httpd.conf)
moritz does commitbit do anything else except handling of svn access?
it seems to cause a lot of trouble for such a seemingly simple task 18:43
obra moritz: indeed it does. We built it for code.bestpractical.com and haven't seen any of these issues there.
But pugscode is running an ancient snapshot, so I don't actually know if actively my fault or passively my fault ;) 18:44
18:44 chris2 joined
moritz obra: ;) 18:44
obra Part of it is that I can't understand why feather is So. Slow. 18:45
TimToady: want to try that commit bit? 18:46
TimToady loadavg went up to 10 :(
moritz it feels faster now 18:47
the "People" page no longer lists all authors
obra TimToady: most of the load was svk and irssi (It was there before commitbit was up) 18:48
moritz: who's missing?
TimToady people doesn't list any people
obra commitbit.pugscode.org/project/Pugs looks ~rasonable to me
moritz obra: everybody ;) 18:49
obra I'm going to spend my afternoon trying to get commitbit up to the latest (this involves fixing a bug in the latest) so that I can _know_ this is my fault. 18:51
18:53 isaacd joined
renormalist Does the word "hyperoperator" only mean the meta operator "Ā«", or is an hyperoperator the constructed result from the meta with the base operator, like Ā»+Ā« ? 18:59
19:00 toshinori joined
TimToady it's the resulting operator 19:00
"hyper map" will also be a hyper operator 19:01
hyper means "run to completion but I don't care what order"
wolverian can hypers short circuit in boolean context? 19:02
TimToady that would depend on what type the hyper is returning, and how that type behaves in boolean context
moritz I guess in a reduce [...] they can 19:03
wolverian I'm wondering about the naive if any(@foo) == 2 { ... }
er
TimToady that's not a hyper
moritz you could 'grep' or 'first' for it
wolverian TimToady, yeah. sorry. I don't know what's wrong with me. :) 19:04
TimToady junctions are meant to parallelize in boolean context
(and short-circuit)
moritz ?eval 'x'.HOW.methods 19:10
19:10 evalbot_r16497 is now known as evalbot_r16506
evalbot_r16506 ["","reverse","chop","index","quotemeta","join","chars","bytes","LIST","ITEM","_split_str","bless"] 19:10
moritz what's that empty method?
'x'. or 'x'."" 19:11
or a bug?
19:17 ozo joined
audreyt moritz: not a bug; it's currently MO's way to get to the metaobject from an object, since all other method names must resolve "normally" 19:19
maybe I should hide it from the .methods list 19:20
19:20 ozo_ joined
audreyt but it's more of an exposed implementation detail 19:20
moritz if you can't call 'x'.'', it should be hidden IMHO ;)
audreyt agreed. (afk)
19:26 Tene is now known as tenen
renormalist Can someone explain me why binding with := is used in examples/algorithms/quicksort.pl ? Is it ok to use just = (it seems so) 19:27
audreyt no real reason except it avoids creation of a mutable array structure 19:28
svnbot6 r16507 | audreyt++ | * Pugs.Class: Mask the "" magic-backdoor-method in the .^methods list 19:31
r16507 | audreyt++ | as requested by moritz++
dduncan speaking of MO, what is the major stuff still pending? 19:36
audreyt dduncan: having .methods actually list stuffs in Pugs.Prim 19:37
you'll note that a string can't actually only have those methods...
?eval "x".^methods
evalbot_r16506 Error: ā¤Unexpected "^"ā¤expecting ".", "\187", ">>", "=", operator name, qualified identifier, variable name, "...", "--", "++", "i", array subscript, hash subscript or code subscript 19:38
audreyt er, hm, it works here
?eval "x".HOW.methods
evalbot_r16506 ["","reverse","chop","index","quotemeta","join","chars","bytes","LIST","ITEM","_split_str","bless"]
dduncan okay
audreyt but in doing so all kinds of misassumptions and bad-multi-dispatching and model corner cases got turned up 19:39
but I'm making progress... just need some continuous hours of cycles.
dduncan good fortune 19:43
19:46 lampus_ is now known as lampus 19:49 awwaiid joined 20:03 obvio171_ joined, obvio171_ is now known as obvio171
moritz S02 has an example: if $x ~~ 1..* {...} # if 1 <= $x <= +Inf 20:06
does this work because 1..* is a range rather than a list? 20:07
TimToady Range object can be used as a matcher
moritz ok
TimToady pugs does not yet implement Range objects
hence tends to get unhappy about 1..* 20:08
moritz I noticed that ;)
audreyt Juerd: 20:12
$ time perl -e 'print "x" x 10485768' > z
real 0m0.221s
user 0m2.320s
$ time perl -e '$_ = "z"; open _; <_>'
real 0m10.120s
user 0m4.608s
on feather
hypothesis: software raid going crazy?
er, 20:13
real 0m10.120s
for read, and
real 0m5.972s
for write
reading 10m shouldn't be 2x slower than writing 10m, and either way they shouldn't take many seconds...
@tell Juerd (time perl -e 'print "x" x 10485768' > z) and (time perl -e '$_ = "z"; open _; <_>') are showing weird perf losses and that's what blocked commitbit 20:14
lambdabot Consider it noted.
Juerd audreyt: Could be anything 20:15
lambdabot Juerd: You have 3 new messages. '/msg lambdabot @messages' to read them.
Juerd audreyt: But I have nothing to diagnose it further than I have, and everything I did showed that it was okay
@messages
lambdabot Tene said 8d 11h 55m 24s ago: tracd uses a literal 'md5crypt' to check passwords in the htpasswd file. Patching tracd to recognize the '{SHA}' sequence in the htpasswd file is an option?
Tene said 8d 11h 53m 14s ago: python-support/trac/trac/web/auth.py:268
audreyt said 1m 12s ago: (time perl -e 'print "x" x 10485768' > z) and (time perl -e '$_ = "z"; open _; <_>') are showing weird perf losses and that's what blocked commitbit
moritz is DMA enabled? 20:16
Juerd moritz: It's S-ATA
moritz ok
audreyt k, but the fix is not going to be at commitbit level -- I upped to latest jifty and tuned db a bit
all it does though is reducing 10mn per page to 3min
commitbit is re-enabled 20:17
Juerd 104857600 bytes (105 MB) gekopieerd, 0,688841 seconden, 152 MB/s
audreyt and I need to crash into sleepland :)
moritz night ;)
Juerd audreyt: I understand it's feather, but I don't know what to do.
audreyt: RAID seems to perform just right
152 OTOH, this could be cache
20:17 polettix joined 20:18 fglock joined
Juerd 1073741824 bytes (1,1 GB) gekopieerd, 15,4734 seconden, 69,4 MB/s 20:19
That's certainly not cached I think
70 MB/s is fine.
1;0 root@feather:~# time dd if=/dev/zero of=tmp/1024mb.bin count=1048576 bs=1024
So I think the RAID is not guilty
The question remains... what is...
fglock would pugs be able to use a Range class implemented in Perl 6? 20:20
Juerd 1;0 root@feather:~/tmp# time dd if=1024mb.bin of=/dev/null
2097152+0 records in
2097152+0 records uit
1073741824 bytes (1,1 GB) gekopieerd, 9,73728 seconden, 110 MB/s
Also great.
(Double speed, because it can read from both drives of the mirror set) 20:21
moritz really misteriously 20:22
I once had a vim instance that lost its terminal and produced 100% CPU usage without showing up in top... 20:23
but theses cases are extremely rare
so I can't see how anything like that could happen regularly
20:27 larsen_ joined
renormalist Juerd: did you already consider that some malicious thing might doing that? like a root kit that sniffs local network with flood attacks (just an example). They can feel like high load. 20:31
Juerd renormalist: Yes, considered that. 20:32
renormalist: But chkrootkit couldn't find anything - do you know more ways to diagnose this? 20:33
moritz Juerd: rkhunter
or maybe sniff traffic at the router
Juerd Have sniffed already 20:34
Nothing out of the ordinary
japhb Juerd: can you paste a line or two of 'vmstat 5'?
Juerd japhb: Sure
procs -----------memory---------- ---swap-- -----io---- -system-- ----cpu---- r b swpd free buff cache si so bi bo in cs us sy id wa
19 0 0 1183088 65336 266596 0 0 382 149 2 48 68 2 25 4 20:35
That's during an apt-get install
16 0 0 1203192 65356 266484 0 0 1 132 300 231 99 1 0 0
10 0 0 1200040 65400 266692 0 0 18 50 342 284 98 2 0 0
r b swpd free buff cache si so bi bo in cs us sy id wa
20:35 baest joined 20:36 lisppaste3 joined
Juerd I'm installing rkhunter now 20:36
20:37 ggoebel joined
japhb hmmm ... so user CPU% and io blocks are probably both the apt-get ... 20:37
... the interrupts looks normal for clock tick + NIC irq, assuming you're not running a tickless kernel.
and the context switches is in line with the interrupts 20:38
diakopter /home/agentz/autosmoke has taken all available cpu for the past while... 20:39
20:39 the_dormant joined
Juerd diakopter: I assume it was niced? 20:39
diakopter err
/home/agentzh/auto-smoke I mean
I dunno. I was just looking at htop
Juerd Agent said he'd niced everything, so that shouldn't be a problem 20:40
renormalist Juerd: don't get panicked with 2.6 kernels when rkhunter/chkrootkit find LKM rootkits. They are often false positives. Unfortunately the "false" you never know ...
20:40 rhandom left
Juerd renormalist: Same with chkrootkit 20:41
But that's why I disabled module support on this public box :)
1;0 root@feather:~# lsmod
Module Size Used by Not tainted
lsmod: QM_MODULES: Function not implemented
20:42 zamolxes joined
integral It's okay, you can still load modules via /dev/kmem :-) 20:43
Juerd Those aren't technically LKMs :)
integral I guess no one's written something that'll load real modules via that
20:44 riffraff joined
riffraff hi 20:45
moritz hi riffraff ;)
riffraff err.. trying to run make smoke yesterday, but it seems that something in t/xx-unspecced/ loops forever
20:45 SubStack joined
riffraff xx-uncategorized, sorry 20:46
file svn.pugscode.org/pugs/t/xx-uncatego...f_doesnt.t
[particle] should probably move that test to xx-infinite ;)
riffraff sorrty, no svn and compiler atm, but it would be nice if someone could confirm that :)
moritz it terminates within a second for me 20:47
r16436
[particle] riffraff: platform?
riffraff linux, ubuntu 7
(now on win32)
usua x86 cpu (athlon) 20:48
20:55 weinigLap_ joined
diakopter wolverian: are you around? 20:58
wolverian yes 20:59
well, let's make sure: diakopter, a highlighted yes. 21:00
21:00 deq` joined
diakopter wolverian: :) 21:01
wolverian :)
obra Juerd: are there any other builds of perl5 on feather? 21:02
Juerd obra: Nafaik 21:03
obra *nod*
TimToady maybe trac installed a vm rootkit... 21:05
and now the python gang are spying on every keystroke...
obra Just for the hell of it, I'm going to build a 5.8.8 in my ~ to see if I get the same weird io performance
wolverian hrm. any ideas what irssi scripts could be taking lots of cpu? 21:06
TimToady Paranoids ŠÆ Us
obra Ya.
TimToady :) 21:07
Juerd obra++ # "ya" 21:08
diakopter Juerd: anyone know why cron is set to renice all irssiz to -1 ? It looks like someone tried to filter screen'd irssiz out, but failed. 21:09
Juerd There's this store "HEŠ˜DRIKSEN" that I keep reading as Heidriksen. 21:10
obra wow. ./configure.gnu is ... very slow
Juerd diakopter: To keep this ultra-interactive process at priority 21:11
diakopter: This has never been a problem, and shouldn't be, because irssi typically is done very quickly. It's just to prevent lag.
rkhunter is finished and found nothing special
diakopter ok... thanks.
wolverian Juerd, I'm running irssi at 10 now and don't notice a difference. then again feather isn't terribly loaded right now. 21:12
Juerd Hm, I saw htop running in top
Nice tool.
wolverian (only 11...)
Juerd Who was running that? Thanks.
diakopter that was me...
Juerd diakopter++ 21:13
wolverian I ran it a week or ago, though ;)
Juerd wolverian: Its load is at 12.
wolverian s/or/or so/
Juerd, yeah. then I guess: irssi++ # still interactive
diakopter obra's configuring all of gnu
wolverian (or maybe the linux scheduler)
Juerd Wow, htop takes even more resources than top :)
diakopter yeah it's heavy
obra running perl's configure script...shouldn't be doing this 21:14
diakopter htop's tree view is nice
21:14 ruz joined
Juerd diakopter: I specifially like the nice- and nice+ :) 21:16
diakopter svk info is using *tons* of cpu every so often, when called from examples/network/evalbot/evalbot-loop, but I don't see any !eval lately. could people be messaging it directly or is that a normal part of the loop? 21:17
moritz that's... nice *SCNR*
Juerd I hae no idea
moritz diakopter: the looks is propably to ensure it's rebuild every so often 21:18
s/looks/loop/
tenen yes, it's to check if it needs to rebuild pugs
Juerd wolverian: What kinds of irssi scripts do you run, by the way?
tenen perhaps you could commit a method of checking that used less cpu
diakopter it could be switched to an svn commit-hook (with a forced wait-after-rebuild, and perhaps a !rebuildYourself command to force it), now that the repo is on the same machine. 21:19
wolverian Juerd, I use hilightwin, go, wlstat, people, scriptassist, keepnick, dccstat, thistory, title, hideauth, bitlbee_tab_completion, usercount, cleanlinkchan, buf, splitlong
I wonder if there's an irssi profiling plugin 21:20
Juerd bitlbee_tab_completion is only active on tab presses, buf is only active on /upgrade, splitlong only for outgoing stuff
scriptassist is safe, keepnick ought to be safe
The rest I don't know
If you're in a lot of active channels, things that process incoming output can damage performance 21:21
Especially if brain dead regexes are used
wolverian I'll look into it. thanks.
diakopter filters out wolverian's irssi from the cron renicer temporarily 21:22
wolverian feels fat
21:22 IllvilJa joined
obra Juerd: just for testing, would you be up for renicing my ./configure.gnu to a high priority? 21:23
I'm curious to see if I'm just low in the IO queue or what.
(I'm seeing about 3s per syscall check from configure. it's Impressive)
Juerd obra: Go ahead 21:24
obra About the same 21:26
Juerd Then restore its priority please
irssi is noticeably lagging 21:27
wolverian hrm. none of my scripts really parse input in large amounts
obra ok
Juerd wolverian: Anything that registers signal "message public"?
obra restored
21:28 SamB joined
wolverian Juerd, a few. looking 21:29
Juerd wolverian: Quite often, those have inefficient (.*) regexes and re-emit themselves.
wolverian ah. hm. I think it might be people 21:30
Juerd Don't blame the machine. Blame people!
wolverian the script ;)
wolverian disables it
Juerd Holy crap that script is huge 21:31
wolverian it's.. useful
though it has features that have nothing to do with a people system
japhb wolverian: what does it do?
wolverian (like nick colours..)
Juerd It better be, otherwise its usefulness/lines ratio would be very low... Like Brainfuck :)
wolverian japhb, userlist with auto-opping, etc.. and lots of vaguely relevant things 21:32
anyway, it's disabled for now.
vim++ # vimgrep
Juerd It does rather heavy things
Like retrieving multiple settings per incoming line. Understandable, though. 21:33
obra Juerd: next time you reboot, will you try audrey's trivial line-of-perl tests?
wolverian I don't really see any way around it
Juerd obra: How is that related to rebooting?
wolverian it's not an essential script, so no worries if I can't use it
Juerd obra: I'm sitting right next to feather. I can reboot it right now if you think that'll help
wolverian: Could you /upgrade now to reset your CPU counter? :)
wolverian sure :) 21:34
Juerd Thanks
buf.pl is the only Irssi script that I wrote and still use.
Jmax what does it do?
Juerd (Speaking of slow scripts...)
obra Juerd: mostly, I'm curious to know if the issue is load related
Jmax oh, neat
Juerd Jmax: Store scroll buffers, and restore them after irssi's /upgrade has exec()ed a new irssi.
obra: If I reboot now, your perl compile is aborted 21:35
obra I don't want to inconvenience folks
21:35 weinigLap joined
Juerd obra: It needs reboots once in a while anyway 21:35
obra Juerd: I'm ok with that. the ./configure being slow strongly suggests taht it's not in perl
Juerd obra: And it doesn't matter much at which time of the day you do it, there's always people using feather 21:36
obra Well, it's only been up a day ;)
21:36 isaacd joined
Juerd This is one of the reasons I want to make it a xen box by the way 21:36
wolverian that was a slow upgrade indeed
Juerd Then it can just suspend and resume in between reboots
wolverian: /upgrade+buf.pl on a slow box can cause a ping timeout... 21:37
wolverian ouch..
Juerd wolverian: I used to run irssi on a Pentium 100
obra heh. *nod*
Juerd buf.pl should be implemented in irssi's core, not in a script.
A few function calls PER LINE OF THE SCROLL BUFFER is way too much to do in Perl space :) 21:38
Scroll buffers tend to get large, and typically people have a dozen windows at least.
buf.pl was a proof of concept. But as it goes, the proof of concept worked and nobody wanted to spend time rewriting it in C. 21:39
Most people run irssi on dedicated fast idle boxes :)
wolverian: Hm, apparently exec doesn't replace the CPU counter. Ah well 21:40
21:41 rindolf joined
wolverian Juerd, aw. I can restart irssi. 21:42
Juerd wolverian: No, never mind. 21:43
meppl good night 21:47
21:49 perlDreamer joined
wolverian load is down now.. I hope it was obra's configure, not my irssi.. :) 21:50
perlDreamer ?eval {a => 1} eqv {a => 2} 21:51
21:51 evalbot_r16506 is now known as evalbot_r16507
evalbot_r16507 Error: ā¤Unexpected end of inputā¤Closing hash curly may not terminate a line;ā¤please add a comma or a semicolon to disambiguate 21:51
perlDreamer ?eval {a => 1} eqv {a => 2};
evalbot_r16507 Bool::False
Juerd wolverian: Hehe 21:52
perlDreamer ?eval !({a => 1} eqv {a => 2});
evalbot_r16507 Bool::True
Juerd is configuring an old bleadperl just to see the load
perlDreamer ?eval ok(!({a => 1} eqv {a => 2}), "eqv on anonymous hash references (1)"); 21:53
evalbot_r16507 Error: No such subroutine: &ok 21:54
Juerd Nah, load still goes up very quickly 21:55
obra Juerd: is it possible to add a non-mirrored disk into the mix?
even a small one
Juerd obra: I can add USB drives while it's running 21:56
obra: Do you need anything specific?
wolverian ?eval {a => 1} !eqv {b => 1};
Juerd obra: Note that I could write a GB at 67 MB/s and read it back at 110 GB/s. That suggests to me that disk I/O is not the problem.
evalbot_r16507 Bool::True
renormalist does something need to be restarted to make rnu.pugscode.org work again?
wolverian oo, that's implemented.
obra I just want to try my syscalls against something not raided.
perlDreamer wolverian: I was testing some things that are crashing pugs in a test 21:57
renormalist s/rnu/run/
perlDreamer but evalbot seems pretty happy with them
wolverian perlDreamer, strange
perlDreamer I'm going to update and rebuild
Juerd obra: Wait, I'll attach a small USB disk
I wonder if the kernel has usb support even :) 21:59
obra :)
Juerd It does
japhb Juerd: what about soft errors and other such problems on the physical drives ...? Your test may not hit the problem areas of the disk, while something that gets used frequently *does* happen to be hitting the failing sectors ...
21:59 xinming_ joined
Juerd obra: /mnt/for-jesse-only 21:59
obra: Feel free to empty it if needed
japhb: Could be, but it doesn't report any of it. Do you know how to examine this? 22:00
obra ok :)
22:00 Jedai joined
japhb Juerd: I dunno, but smartmontools appeared in an apt-cache search. That looks promising 22:00
Juerd It's really nice to have servers at a few yards away instead of a 1h30m drive
japhb: I was just installing that. I don't think it'll work on sata drives though 22:01
obra Juerd: yeah. that has the same horrible IO performance
japhb Hmmm, I would think if it supports both ATA and SCSI it's got a decent chance 22:02
obra Juerd: done. thanks.
"now I have no idea"
[particle] what's the magic eight ball say?
Juerd I've started smart short tests
s/smart/SMART/
00:03 <@Juerd> purl: excuse for slow feather? 22:03
00:03 <+purl> Juerd: Recursivity. Call back if it happens again.
wolverian purl++
Juerd # 1 Short offline Completed without error 00% 17438 - 22:05
(that's sda)
mikfire09 Juerd, smartmon does work on SATA drives. I have it installed on my machine using SATA and it works fine.
Juerd # 1 Short offline Completed without error 00% 17436 -
(that's sdb)
?eval 17438/24
evalbot_r16507 8719/12
Juerd I hate that. 22:06
726.583333333333
2 years, that's about right.
22:06 drupek12 joined
wolverian completed without error: 00% 22:06
(cough)
Juerd Num Test_Description Status Remaining LifeTime(hours) LBA_of_first_error
wolverian: status: completed without error, remaining: 00%
wolverian ah.
japhb ?eval Num(17438/24) 22:07
Juerd Doing long test on sda.
Don't be surprised if disk access is slow now.
evalbot_r16507 Error: No such subroutine: &Num
japhb ?eval my Num $foo = 17438/24; $foo
evalbot_r16507 \8719/12
wolverian isn't Num the generic "any number type"
what's Float called? Float? :) 22:08
Juerd eh, Float?
japhb ?eval my num $foo = 17438/24; $foo
evalbot_r16507 \8719/12
Juerd Why would floats have a special type? That's just Num.
mikfire09 ?eval +(17438/24);
Juerd japhb: Pugs doesn't do native types afaik
?eval my aoeuaoeu $foo = 17438/24; $foo 22:09
japhb Juerd: yeah, sigh.
evalbot_r16507 8719/12
\8719/12
Juerd It doesn't really do *types*, apparently :)
wolverian Juerd, they're different in haskell... ;)
japhb ?eval my num $foo = (17438/24) ** 1.0; $foo
Juerd wolverian: Perl's not haskell, so that doesn't answer the question :)
evalbot_r16507 \8719/12
wolverian Juerd, to differentiate between rationals and irrationals I suppose
Juerd ?eval 17438.0/24.0
japhb ?eval my num $foo = ((17438/24) ** 0.5) ** 2; $foo 22:10
evalbot_r16507 8719/12
wolverian the foo/bars are even called Rats
but I don't think we have Irrats :)
evalbot_r16507 \899466600210769342731597381841/1237940039285380274899124224
Juerd Mice
japhb DAMMIT
Juerd japhb: Hahaha
22:10 thoughtpolice joined 22:12 ribasushi joined
japhb ?eval my num $foo = 2 ** (log(17438/24) / log(10)); $foo 22:12
oops
evalbot_r16507 \7.266624873097622
japhb ?eval my num $foo = 2 ** (log(17438/24) / log(2)); $foo
evalbot_r16507 \726.5833333333331
japhb HAH!
And it only ended up wrong by ~2 ulp 22:13
wolverian Juerd, anyway, you're right. I don't see a float type. I was thinking Num was a role, with integer, rational and irrational implementations (among others) 22:14
due to much haskelling lately, probably 22:15
diakopter slow feather? Put it in a vacuum. That will reduce drag by quite a bit.
22:20 SubStack joined
GeJ Hello all. 22:21
is everything right with svn.pugscode.org? I can't seem to be able to connect. 22:22
hum, well, now I can 22:23
sorry, please forgive me for the noise. 22:24
22:25 lichtkind joined
lichtkind ?eval @a = 1, 2; say @a; 22:26
evalbot_r16507 OUTPUT[12ā¤] Bool::True
lichtkind ?eval @a = 1, 2; say @a.perl;
evalbot_r16507 OUTPUT[[1, 2]ā¤] Bool::True
22:32 REPLeffect joined
TimToady by the way \1/2 is incorrect output, since it would mean (\1)/2 by precedence 22:35
Jedai ?eval my @array := [1, 2, 3]; @array.elems
evalbot_r16507 1
masak ouch.
TimToady woops 22:36
?eval my @array := [1,2,3]; say +@array
evalbot_r16507 OUTPUT[1ā¤] Bool::True
masak that's a misfeature, mehopes
TimToady indeed 22:37
Jedai Well Audrey say it will be fixed before the next release
I was just checking
masak ah.
it does seem pretty fix-worthy
lichtkind i only tried to figure out what this .perl does 22:40
moritz lichtkind: output of perl code, like Data::Dumper does 22:43
lichtkind thanks 22:44
perlDreamer weird
This test passes
ok {a => 1} === {a => 2}, "on anonymous hash references (1)";
but prints no test comment after it
and it also shouldn't pass
is comma very binding? 22:45
lichtkind moritz i currently proofread the IX perl6 article from renormalist, there i saw it the first time
moritz lichtkind: ok ;)
perlDreamer and if you parenthesize the boolean, then pugs pukes
moritz I might proofread as well if he wants to, but currently he's not on the channel 22:46
lichtkind moritz: yeah he told me not to give that away since its for a big magazine 22:47
moritz lichtkind: quite understandable ;) 22:48
lichtkind: do you read IX regularly? 22:49
lichtkind moritz: no its boring 22:50
moritz most of it, yes ;)
lichtkind too much buzzhyping
moritz: but if you see something good id be interested 22:56
22:56 Aankhen`` joined
lichtkind good night 23:00
Aankhen`` Er, shouldn't the topic be a little more specific?
As in, think twice before running `make install` rather than just installing Pugs?
moritz Aankhen``: it should... go ahead and improve it
Aankhen`` pugscode.org | Beware of `make install` and Pugs, it will do bad things to your system 23:01
Aankhen`` So is the whole coerce:<as> syntax dead? 23:14
Seems a shame.
Unless there's something similar which I missed.
moritz can't you force coercion by forcing a context? 23:15
?eval ~("foobar" ~~ m/oo/)
evalbot_r16507 "oo"
Aankhen`` Yes, but that depends on, e.g. redefining prefix:<~> for your classes.
moritz and why do you want something else? 23:16
Aankhen`` Doesn't seem as universal as multi coerce:<as> (::?CLASS $self, Str $other) { ... } # or whatever the exact syntax was
wolverian I suggested 'method --> Str { ... }' .. :) 23:17
moritz why not use roles? "does Stringify"?
Aankhen`` Every object needs to `do Stringify` anyway, so why add a role for it? 23:18
wolverian why add a role, you have a type already (Str)
Aankhen`` (That is to say, any object could conceivably be interpolated into a string.)
TimToady I'm still thinking through the whole coerce/deep semantics thing. I don't like the "is deep" approach much any more. 23:28
and we have to deal with "as" vs "to" issues. 23:29
which might or might not be the same as the difference between Str($x) and Str.new($x) 23:30
this also might interact with dduncan's desire to require a readonly copy of a potentially rw argument 23:32
Aankhen`` Fun. ^_^
TimToady: I have to say, I don't envy you your job.
moritz language design is hard ;) 23:33
dduncan I would probably comment on this discussion, but I'm in the middle of another chat that is $work related
GeJ moritz: language design with continuous almost-realtime feedback from passionate users is harder. 23:38
TimToady language design in a vacuum is even harder. :) 23:43
23:49 fridim joined 23:52 ntgrl joined

Powered by Cro and the RakuĀ® Programming Language.

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

Thank you!