pugscode.org | temporary feather address: perl6.wall.org | sial.org/pbot/perl6 | ?eval [~] <m oo se> | We do Haskell, too | > reverse (show (scanl (*) 1 [1..] !! 4))
Set by allbery_b on 24 March 2007.
dduncan fyi, I just did another smoke, my first of March afaik ... normal Haskell backend, snapshot/darwin, r15788 ... all but 554 tests pass ... interestingly, the smoke completed considerably faster than the last time I smoked ... it was 63 minutes this time, vs 90 or 135 (forget) minutes last time ... maybe the new way that tasks are handed out to concurrent smoker threads has something to do with it? 00:02
nothing about my environment really changed since the last smoke ... same hardware and GHC etc 00:03
moritz dduncan: did you compile with GHC 6.7? 00:04
dduncan no
last time and this, it is 6.6.0
tene dduncan: pugs startup time has decreated quite a bit. 00:05
moritz anyway, audreyt talked about speeding up startup by un-atomifying things
tene decreased
dduncan and it is the pre-packaged 6.6.0 at that
in any event, it will be good, whenever that happens, that Perl 6 performs no slower than Perl 5 on the same hardware with equivalent code 00:07
geoffb dduncan: It will be a great while before we reach that point. 00:08
Pugs is currently about 1000x slower than Perl 5 by my measurement 00:09
dduncan sure ... but "good enough" is also good enough
moritz probably it will only be reached when parrot is used as backend and parrot optimizes well ;-)
dduncan eg, if one can run an application written in it regularly and not have it feel slower to the user
this includes web apps 00:10
geoffb dduncan: Oh, I know that measure well -- I just meant, we're a ways from that point. I *do* think we will get there.
dduncan for my purposes, if a web program compiles and finishes a page request in less than 1 clock second with a load of 1 user, that's good enough to begin using for production 00:12
that's considerably less demand than a full server
geoffb nodnod
moritz dduncan: it's the same for me, but pray to god that you'll never have to scale ;-) 00:13
dduncan the thing is, if the system becomes fast enough to tolerate with a small number of users, then its actual use should increase exponentially, and then we have a much larger base that can focus on optimizing it for scaling 00:14
s/users/concurrent users/
s/base/developer base/ 00:15
FYI, a revised GPLv3 process time table has finally been announced, via the mailing list, [email@hidden.address] ... draft 3 will come out in 2 days, with 60 day discussion, then last call draft, then 30 days, then final publication ... so we should be able to start using GPLv3 in about 92 days! 00:23
I'm happy to see that revised timetable.
00:23 dmpk2k joined
moritz hey, cool 00:24
dduncan the web site isn't updated yet, though
or not the front page
moritz and I hope sun will release opensolaris, especially ZFS under GPLv3
dduncan indeed
and barring any major problem, unlikely that is, I will be using GPLv3 for all my important stuff 00:25
afaik, the Perl Foundation plans to make the Perl 6 license the disjunction of GPL v3 and Artistic v2, which would be good
... or later 00:26
moritz still likes the WTFPL ;-)
dduncan that sounds like the BF of licenses 00:27
moritz "WTFPL - Do What The Fuck You Want To Public License" 00:28
00:28 H0LE joined
moritz just another name for public domain ;-) 00:28
00:37 justatheory joined
offby1 "Do What The F*** Thou Wilt" shall be the whole of the law 00:39
dmpk2k It'd certainly make things... exciting. 00:41
00:44 encryptio joined 00:59 Khisanth joined
specbot6 r14354 | larry++ | Suggestions from TheDamian++ and Juerd++ and others. 01:04
r14354 | larry++ | All punctuation is now treated as potentially meta.
r14354 | larry++ | <'foo'> form is gone; just use 'foo'.
r14354 | larry++ | Conjectural syntax positive and negative submatches of isolated substring. 01:05
r14354 | larry++ | Conjectural syntax for recursive calls to anonymous substructures.
r14354 | larry++ | < a b c > is now just a list of strings.
01:11 Psyche^ joined
specbot6 r14355 | larry++ | typo from dduncan++ 01:14
01:18 Psyche^_ joined 01:19 Alias_ joined 01:21 amnesiac joined 01:24 justatheory_ joined 01:25 Psyche^_ is now known as Patterner 01:27 nekokak joined 01:29 bonesss joined 01:37 nperez joined 01:38 stevan_ joined 02:37 justatheory joined 02:56 ruoso joined 03:08 xern joined 03:09 xern joined 03:15 esstone joined 03:20 mako132_ joined 03:58 ProperNoun joined 04:04 rofl joined, rofl is now known as mathlord 04:18 Debolaz2 joined 04:20 kavelot joined
pasteling "kavelot" at 200.158.87.162 pasted "how can "A" be printed but "B" not (the script is, for some reason, stopping on second line)" (3 lines, 55B) at sial.org/pbot/23806 04:20
04:24 Debolaz joined 04:29 prly joined 05:07 esstone joined 05:18 prly_ joined 05:25 prly__ joined, Psyche^ joined 05:30 kanru joined 05:31 baest joined, prly joined 05:36 penk joined 05:38 Psyche^ is now known as Patterner 05:41 prly_ joined 05:44 rfordinal joined, amnesiac joined 05:46 prly__ joined 05:49 justatheory joined 05:55 kanru joined 05:57 beppu joined 06:07 carson joined 06:14 BooK joined 06:19 devogon joined 06:24 nxu7 joined 06:30 carson_ joined 07:04 buetow joined 07:05 Psyche^ joined 07:11 iblechbot joined 07:19 Psyche^ is now known as Patterner
gaal I like spec:r14354 :-) 07:21
07:25 jisom joined 07:41 Psyche^ joined, xinming_ joined 07:44 nxu7 left 07:54 jisom_ joined 07:55 Psyche^ is now known as Patterner, dduncan left 08:03 ntgrl joined 08:04 ntgrl is now known as integral 08:14 the_dormant joined 08:32 buetow joined 08:37 elmex joined, the_dormant joined 08:38 the_dormant joined 08:41 offby1 joined 08:47 chris2 joined 09:09 buetow joined 09:36 IllvilJa joined 10:32 ayrnieu joined 10:35 Psyche^ joined 10:51 autin joined 10:52 Psyche^ is now known as Patterner 11:10 lisppaste3 joined 11:16 autin joined 11:21 thorat joined 11:41 mako132_ joined 11:42 Psyche^ joined 11:45 xinming_ is now known as xinming 11:54 Psyche^ is now known as Patterner 11:56 Khisanth joined 11:58 JeeBee joined, JeeBee left
svnbot6 r15789 | fglock++ | PCR - fixed Match deref (perl5.9.5 is more strict) 12:01
12:01 the_dormant joined 12:04 buetow joined 12:14 iblechbot joined 12:15 Limbic_Region joined 12:20 fglock joined 12:38 rfordinal_ joined 12:39 marmic joined 12:52 charsbar_ joined 12:53 Psyche^ joined 13:04 autin joined 13:08 Psyche^ is now known as Patterner 13:09 [particle] joined 13:15 buetow joined 13:41 ofer0 joined 13:44 VanilleBert joined 13:48 ruoso joined, DarkWolf84 joined 13:49 drupek12178 joined 13:59 vel joined
DarkWolf84 hi there 14:06
:)
moritz hi DarkWolf84 ;) 14:07
DarkWolf84 is there any way to convert perl6 code to binary or byte code 14:08
moritz DarkWolf84: there is pugs with parrot backend if you mean that
DarkWolf84 yes
but i just compiled parrot and pugs 14:09
and pugs gives me an error
moritz there is a bit of bit rot in the parrot backend :(
DarkWolf84 when i'm converting program it ends with an error 14:10
moritz and don't even know if there is somebody actively working on the parrot backend 14:12
DarkWolf84 maybe that's the point
nvm pugs is running fine :) 14:13
moritz DarkWolf84: do you have a commit bit?
DarkWolf84 what is commit bit
may be i haven't 14:14
moritz DarkWolf84: it allows you to commit changes to the pugs svn repository
DarkWolf84: if you /msg me your email address I can invite you
DarkWolf84 is it included in svn version 14:15
ok
hm know only perl(original perl) for now and little perl6 14:17
svnbot6 r15790 | lanny++ | * from_perl6_rules
r15790 | lanny++ | + turn fails into todo's. There are so many todo's already
r15790 | lanny++ | in existence I /think/ this is the intention.
moritz btw on freenode you need to be registered to be allowed to /msg :(
DarkWolf84 i've registered yet 14:18
my english is not good because i'm not english
moritz then you're in good company, the perl6/pugs people come from all over the world 14:19
DarkWolf84 sorry 14:20
moritz I'm from Germany, but we have hackers from the US, Taiwan, Isreal etc
DarkWolf84 i forgot to identify myself
i think now it's done 14:21
i send the my email two times 14:22
14:22 Jedai joined
DarkWolf84 ok 14:22
moritz DarkWolf84: got it now 14:24
DarkWolf84: email is on its way
DarkWolf84 fine :)
moritz DarkWolf84: it's custom to add yourself to the AUTHORS file on first commit as a test if commit works
DarkWolf84 how work this 14:26
i'm not familiar whit it
with it* 14:27
moritz DarkWolf84: www.google.com/search?hl=en&q=svn+tutorial 14:28
lambdabot Title: svn tutorial - Google Search
DarkWolf84 tu
14:58 bonesss joined 15:00 amnesiac joined, weinig|bbl is now known as weinig
avar ?eval 1 15:07
15:10 perlmonkey2 joined 15:11 kanru joined
moritz avar: the answer is 1 (and Bool::True) ;-) 15:13
avar return "The answer to life, the universe, and everything" but 42 15:14
DarkWolf84 :)
15:26 chris2 joined, VanilleBert left 15:31 marmic joined 15:36 seano joined 15:37 prly joined 15:39 justatheory joined 16:05 weinig is now known as weinig|away 16:24 fglock joined 16:41 buetow joined
svnbot6 r15791 | fglock++ | PCR - version 0.22 in CPAN 16:50
16:51 diakopter joined 16:56 evalbot_r15790 joined 17:14 vel joined, falesca joined 17:15 vel6608 joined 17:18 lichtkind joined 17:19 vel6608 joined
svnbot6 r15792 | fglock++ | v6.pm - fixed precedence of ';' separator 17:24
17:24 justatheory joined 17:48 offby1 joined 17:50 bernhard joined 17:51 falesca joined 17:54 diakopter joined
lichtkind Leibsle: soso das leibsle 17:54
17:54 offby1 joined 18:00 justatheory joined 18:25 justatheory_ joined 18:28 jisom joined 18:34 chris2 joined 18:36 dduncan joined 18:38 jisom_ joined
kolibrie I am trying to add values to a match object as a named capture, as a side effect 18:44
18:44 justatheory_ is now known as justatheory
kolibrie I did not see any examples of this in the spec, but think it should look something like: 18:44
$<numeral> := <roman_numeral> { $<notation> := 'roman' }
pasteling "kolibrie" at 66.166.118.50 pasted "attempting to add values to match object as a side-effect" (82 lines, 1.4K) at sial.org/pbot/23818
18:47 frobnitz joined
kolibrie is there a canonical way of doing this? 18:48
19:13 geoffb joined 19:14 perlmonkey2 left 19:24 fglock joined
kolibrie fglock: any insights into sial.org/pbot/23818 ? 19:24
lambdabot Title: Paste #23818 from "kolibrie" at 66.166.118.50 19:25
fglock kolibrie: looking 19:27
kolibrie fglock: not sure if it is specced, but it seemed like it might work 19:28
19:33 buubot joined
fglock kolibrie: I think it is specced; but Pugs::Runtime::Match->hash returns a copy of the hash - assigning to the hash doesn't change the Match object, unfortunately 19:37
kolibrie fglock: so the syntax is okay, but not currently implemented in PCR? 19:38
fglock yes 19:39
kolibrie will try to find another way, then
(unless fglock has a solution already in mind) 19:40
fglock hmm - we could use tie, if it weren't slow 19:41
19:41 geoffb joined
fglock otoh, everything is slow these days 19:41
19:43 prism joined
kolibrie chose Perl regexes mostly because the syntax is so nice - was not expecting speed for a little while 19:44
fglock the solution I'm thinking about is to make a flexible emitter, that does some cost-benefit analysis, based on the kp6 compiler
DarkWolf84 i like much regexp too :) 19:45
maybe because i haven't learned the new ones 19:46
kolibrie DarkWolf84: I am using Perl 6 regexes, which are a great improvement over Perl 5 regexes 19:47
though the spec and implementation are not finished yet
DarkWolf84 I said i haven't learned them yet
19:49 riffraff joined
DarkWolf84 but there is a lot of time to leran new things 19:49
:)
kolibrie a lifetime 19:50
fglock kolibrie: tieing the result of 'sub hash' in Pugs::Runtime::Match is not too bad
DarkWolf84 life is a big school for me
kolibrie fglock: cool 19:51
fglock kolibrie: done 20:08
svnbot6 r15793 | fglock++ | PCR - minor hack to make %($/) rw - kolibrie++
kolibrie fglock: thank you
kolibrie pulls
fglock didn't need tie :)
kolibrie nice
20:12 Aankhen`` joined
kolibrie fglock: fewer lines of code, too 20:15
svnbot6 r15794 | fglock++ | PCR - ChangeLog update 20:17
20:22 Psyche^ joined
DarkWolf84 strange 20:24
let vars don't work
ingy seen audreyt 20:25
DarkWolf84 it gives me no such subroutine
is it removed 20:26
moritz @seen audreyt 20:27
lambdabot audreyt is in #haskell-soc, #haskell and #perl6. I don't know when audreyt last spoke.
moritz ingy: the sigils are very important here ;-)
ingy @gracias
lambdabot Unknown command, try @list
kolibrie ?eval my $x; for 1..4 { let $x = $_ } say $x; 20:28
20:28 evalbot_r15790 is now known as evalbot_r15794
evalbot_r15794 Error: No such subroutine: &let 20:28
DarkWolf84 yes
the same error
moritz what should 'let' be good vor?
kolibrie DarkWolf84: don't know what happened, it still works for me on an older pugs 20:29
DarkWolf84 i tyed it because the tutorial on www.programmersheaven.com/2/Perl6-FAQ-Variables
kolibrie moritz: hypothetical assignment, rolls back if block is false
lambdabot Title: Perl6-FAQ-Variables at Programmers Heaven
moritz kolibrie: thanks ;) 20:30
DarkWolf84 i saw it there
so it's remouved lately 20:31
removed*
moritz DarkWolf84: I wouldn't conclude that unless it's remove from the specs
DarkWolf84 that's realy strange 20:32
kolibrie fglock: that fix works here, thank you. fglock++
fglock kolibrie: np :) 20:33
20:34 Psyche^ is now known as Patterner
DarkWolf84 maybe that's a bug in the new vesrion 20:34
fglock kolibrie: how about adding a test for it in pugs/t/ ? 20:37
kolibrie fglock: like from my nopaste example, perhaps? 20:38
DarkWolf84 test for let?
fglock yes
DarkWolf84 why not
:)
fglock DarkWolf84: that too :)
kolibrie DarkWolf84: for match value assignment from a side effect 20:39
DarkWolf84 i'll write some program for test
20:40 diotalevi joined
DarkWolf84 the subversion is new thing for me but I found tutorial for it 20:40
moritz DarkWolf84: you should read docs/Pugs/Doc/Hack.pod as well ;-) 20:42
20:51 miyagawa joined, theorbtwo joined 20:53 fglock left 20:57 Psyche^ joined 21:09 Psyche^ is now known as Patterner
TimToady @tell fglock we need to start switching our regex metacharacters over to simple '...' for quoting. Step 1) enable parsing of '...' in addition to <'...'>. Step 2) make all punctuation "meta" and deprecate <'...'>. Fix whatever breaks. Step 3), remove <'...'>. 21:10
lambdabot Consider it noted.
21:10 polettix joined
kolibrie TimToady: is that recent? 21:11
Juerd kolibrie: Steaming fresh.
Though discussion about it isn't quite all that new :)
kolibrie thought he didn't recognize it
Juerd I would look up when I first suggested it, but my mail archive is broken and only lists things more recent than June 2006 21:13
svnbot6 r15795 | kolibrie++ | binding subrules to aliases, in the regex and as a side-effect
Juerd I have no idea where the rest went
TimToady was several years ago, anyway... 21:14
Juerd Probably around 2002-ish, when I worked on my XML grammar :) 21:15
21:18 justatheory joined
Aankhen`` Wowie. That should shake things up a bit. 21:19
TimToady not as much as you'd think--2/3 of the punctuation was already meta 21:21
Aankhen`` Oh. I never realized. 21:22
The '...' bit will be very useful.
[particle] timtoady\+\+
TimToady and '=>' is a lot easier to read than =\>
moritz my "make smoke" dies after t/builtins/strings/uc.t without any message 21:23
TimToady Well, the '..' part of it was originally 'Juerd++'s idea.
but it was a small jump from there to realizing we could get rid of another arbitrary list that the user has to memorize. 21:24
I'd just got done bragging about such 'list reduction' on perlmonks, and then realized we still had this arbitrary list... 21:25
Aankhen`` Heh.
TimToady the list of punctuation chars that weren't meta before was just: ~ ` ; ' " , - 21:26
occasionally meta are: ! = / 21:27
and these were always meta: # ^ $ @ % & | \ * + ? : . ( ) < > [ ] { } « »
Aankhen`` What about „? :-D
Yeh, I see what you mean now. 21:28
TimToady and by the time we pull out ' and " it doesn't leave much else 21:29
so we decided to make it a simple rule.
Aankhen`` Yeah.
TimToady also makes the regex grammar rather more extensible.
Juerd Extensible. I usually call it forward compatible :) 21:31
Extensions as a possibility versus as a necessity :)
In any case, I'm very happy with the change. 21:32
(Which isn't that big a change anyway)
TimToady yes, that was the pleasantly surprising part
bit of a pain to switch, but worth it. 21:33
and certainly better to do it now than later
dduncan seen mugwump 21:39
geoffb TimToady: link to list reduction brag on PM?
I'm going to a .pm meeting tonight, and that would be good fodder ....
21:41 CardinalNumber joined
TimToady geoffb: www.perlmonks.org/?node_id=606369 21:47
lambdabot Title: Re^2: Builtin functions defaulting to $_
geoffb TimToady: thanks 21:48
TimToady: out of curiosity, do you keep a (mental, electronic, or physical) list of the things Perl 6 users will be expected to memorize? I know you have been trying to shrink the cognitive load, but is this done on a "low-hanging-fruit" basis, or is there some master list you're working to minimize? 21:53
(one of those "view into the designer's mind" questions)
DarkWolf84 btw how to report a bug 21:56
geoffb Oh, and: if this list is written down somewhere, can we see it's current state? 21:57
DarkWolf84: Add a test in t/
TimToady there's no list. it's more like whack-a-mole. 21:59
geoffb I wouldn't have been surprised either way. :-) 22:00
Is $*OUT autoflush default on or off for Perl 6? And how do you get the other behavior? 22:01
svnbot6 r15796 | lwall++ | Some whackage on the <'...'> to '...' transition, mostly tests. 22:06
DarkWolf84 btw I found let test in t/
i gives the same result 22:07
22:07 meppl joined
TimToady let is just a form of temp (p5 local) that might or might not restore. so it won't unbreak the binding if the binding is broken. :) 22:08
DarkWolf84 it should be better to be there if that's not a problem 22:10
moritz I just hate the "let" keywords which reminds me of visualbasic agony ;-) 22:12
DarkWolf84 basic oh no
0_0
:)
TimToady we ignored that and borrowed it from the mathematicians to mean "provisionally" 22:13
because it was shorter than "suppose". :)
buu This is .. scary in whole new ways.
moritz DarkWolf84: I had to code a bit in VBA 6.0 for $prev_job, and I remember some braindamage having to use "let" for assignemnt to references
DarkWolf84 basik is to lame for some tasks 22:14
moritz had to write FFT in VB ;(
so insane 22:15
DarkWolf84 that was my first programing language after all
moritz DarkWolf84: same for me, startet with QBasic on win 3.11/dos 6.1
geoffb What happened to that new VB version that audreyt was all excited about because it had been cross-pollinated with Haskell? Is it out yet?
TimToady buu: I don't know why you think it's new... "let" as been in the language since 2002. 22:16
DarkWolf84 before 5-6 years we have only 8-bit apple][ clonings
TimToady *has
DarkWolf84 the only way to programing them was basic
imean in our contry 22:17
i was just 6-th grade
:)
btw there the word let there was optional 22:18
Juerd If enough people hate 'let', we'll automatically get a "use let", like "use self" ;) 22:19
DarkWolf84 %( my english
after all it's not removed from definitions 22:21
not yet
actually i haven't find way to use let 22:24
found* 22:25
I don't really need it. 22:26
22:27 ludan joined
svnbot6 r15797 | lwall++ | STD.pm now believes in '...' (oddly, it never had a rule to parse <'...'>) 22:36
TimToady DarkWolf84: essentially, you're using 'let' semantics all the time whenever a regex captures, because the semantics are that the old binding is restored if the capture is backtracked over (fails). 22:38
22:40 Limbic_Region joined
DarkWolf84 is that means the old regex will not work? 22:41
TimToady just because something is backtracked doesn't mean that it won't succeed some other way. 22:42
DarkWolf84 may be it's used different way 22:43
TimToady p5 already has 'let' within its closures, except that it overloaded 'local' for that.
we decided to keep 'let' and 'temp' separate in p6 22:44
and that lets you use 'let' outside of regex as well
DarkWolf84 well 22:45
TimToady where throwing an exception or returning undef is equivalent to backtracking.
good for hypothetical programming like Prolog
though we'd really need a unification matcher to do that right. 22:46
DarkWolf84 but now it's not defined 22:47
not in my version
TimToady what isn't defined?
DarkWolf84 let
TimToady might not be implemented yet.
DarkWolf84 why it works for the old versions then 22:48
TimToady ?eval my $x; do { let $x = 1 }; say $x
22:48 evalbot_r15794 is now known as evalbot_r15797
evalbot_r15797 Error: No such subroutine: &let 22:48
TimToady ?eval my $x; do { temp $x = 1 }; say $x
evalbot_r15797 Error: No such subroutine: &temp
TimToady looks like temp is broken too...
DarkWolf84 i haven't tryed temp yet 22:49
TimToady they were recently demoted from declarators, I think, so probably got broken then
?eval my $x; do { temp $x; $x = 1 }; say $x
evalbot_r15797 Error: No such subroutine: &temp
TimToady ?eval my $x; do { $x.temp; $x = 1 }; say $x 22:50
evalbot_r15797 Error: No such method in class Scalar: &temp
TimToady yeah, it's borken.
22:51 [particle] joined
TimToady probably interacts with the recent type system change badly, and will get picked up again after things settle down under the new type system. 22:51
23:00 vel joined 23:11 offby1` joined 23:13 offby1` is now known as offby1 23:16 Tene joined 23:19 justatheory joined, Daveman joined 23:22 DarkWolf84 joined
DarkWolf84 i'm back 23:23
net problems
23:24 ozo_ joined 23:36 rashakil_ joined 23:38 zgh joined 23:39 snoop0x7b joined
obra lambdabot: @pl /x -> x + 1 23:39
lambdabot (line 1, column 1):
unexpected "/"
expecting white space, natural, identifier, lambda abstraction or expression
obra lambdabot: @pl \x -> x + 1
lambdabot (1 +)
23:53 diakopter joined

Powered by Cro and the Raku® Programming Language.

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

Thank you!