pugscode.org/ | nopaste: sial.org/pbot/perl6 | pugs: [~] <m oo se> (or rakudo:, kp6:, smop: etc.) || We do Haskell, too | > reverse . show $ foldl1 (*) [1..4] | irclog: irc.pugscode.org/
Set by TimToady on 25 January 2008.
00:04 IllvilJa left
mncharity my impression is the parrot folks have been working to make t/ less monolithic. does anyone know the details? 00:11
specifically, has anyone looked at "chunking" the tests? ie, groveling over the files and determining which lines are needed for each test? 00:12
peepsalot jnthn, when was that FAQ originally written? 00:22
and how much is perl 6 still changing?
pmichaud, i added a comment to RT#39930 earlier today, regarding the smoke test being broken 00:25
oops, semi-wrong-channel 00:26
00:26 Auzon left
peepsalot time for dinner. afk 00:27
00:27 qmole left
jnthn peepsalot: I think over a year ago. And, well, hard to say how much stuff is changing. Sometimes I think something won't change, and a good reason comes up to change it, etc. 00:29
Like the list assignment discussion earlier today. 00:30
jnthn afk too, sleepz
00:30 zamolxes joined 00:41 devogon left 00:49 Juerd is now known as Whreq
meppl good night 00:55
00:58 meppl left 01:07 stevan_ joined 01:16 jferrero left 01:18 icwiener left 01:31 jjore joined 01:44 blindfish left
mncharity re tests, it looks like t/spec is the development hotspot for the test suite. (ah, so *that's* what fudge is:) 01:46
01:52 luqui joined 01:55 jiing left
TimToady ah, so you found the README didja? 02:06
02:11 zamolxes left 02:16 Connel joined
monomorph mncharity: earlier today, you wondered what keeps people from contributing to perl6 02:23
mncharity: i'm new to this channel, so i have an "outsider" perspective that might be interesting (maybe not)
mncharity: this is long, so i posted it on jottit.com/r3stt/ 02:24
lambdabot Title: Perl6 Thoughts: Home
monomorph mncharity: perhaps at least food for discussion
02:28 literal left 02:33 Lion2 joined
mncharity re README, hmm, not ./README... t/spec/README ? $ find .|grep -v svn|grep README|wc -l #=> 134 ;) 02:37
02:37 ruoso left
mncharity hmm, maybe there's a t/README... yes, but no mention. so my guess is t/spec/README, right? 02:38
re food for discussion, nifty, looking...
Unmaintained websites - very much agree - big problem. 02:51
Thoughts on cause: "perl6 isn't a project". :) Assorted related projects create stuff, which then lacking a core support team, then rots. Perhaps an expectation of brokenness develops - eg, I didn't even try to figure out how to report svn.perl.org/perl6/ 's brokenness. 02:54
lambdabot Title: Revision 14497: /
02:55 aindilis joined
mncharity Thoughts on resolution: ... hmm, good question. Searching around, and creating a wiki page summarizing the assorted brokenness, might help. With the bug list concentrated like that, it might draw attention and peoples' time. 02:55
Unclear project boundaries. Interesting. It didn't strike me as odd until you mentioned it. :) Thoughts on cause: perl doesn't have a centralized community _development_ site (repo, cia, etc). So pugs and parrot both rolled their own. Pugs being... ? most open? featureful? site of the first real test suite? encouragement of other p6 related projects? whatever, 02:59
ended up being the default p6 repo. Moving t/spec off to a different repo would require recreating infrastructure. Rather than say, if perl had a gforge, simply new-project'ing it. 03:00
Oh, and the move would lose the pugs development community (by losing their commit bits). 03:01
monomorph the actual url is not the issue IMO, but it's definitely non-obvious what on pugscode belongs to pugs and what doesn't 03:02
mncharity One could imagine a perl6 super-project on source forge, with parrot and pugs and spec projects within it, sharing access rights. But, didn't happen that way.
ah, ok. yes. pugscode is now perhaps now an approximation of "most perl 6 stuff, aside from parrot, and the mailing-lists". 03:03
which would be a useful thing to say somewhere. :) 03:04
I'm afraid shifting around directories to reflect that, eg, putting pugs off in a subdirectory, would break things. And tuits are a bit short to clean up after it. But better README's seem a good and tractable goal. :) 03:06
Parsing Perl6 requires Perl6. STD.pm is a bit complex. But then, so is p6. I am unclear on it's status. My impression is it's simply spec at this point. Given the pain redsix had trying to cope before it existed, I'm delighted that it's there. Even if I kind of expect if I end up doing something which needs it, a bunch of massage will be needed. 03:08
But yes, it's not casual reading. :) 03:09
monomorph i see. but again, it's right inside the pugs source, and there's no document anywhere (that i could find...) to explain its purpose 03:10
mncharity Over-ambition. Hmm. The specific example I'm not sure is correct. The backends were spun off, at least originally, because they were easy. And while the percent of attention they got, vs Pugs haskell development, increased in time, I'm not sure to what extent that was cause, or effect. 03:11
monomorph in fact, my last three points may all be explained by me misunderstanding the project goals
mncharity s/effect/an effect of haskell development slowing down/ 03:12
re project goals, interesting. yeah, it might be nice to have something which captured those somewhere. I was looking around earlier today for an svn heat map generator, so I could more easily see what people were spending time on. 03:13
finally, re bootstrapping, I think it's an artifact of the language was (and is) still evolving, and the language is quite complex. Lua engineered for simplicity. p6, very not so much. 03:16
(though the core can be much simpler than the surface)
So writing Lua in C was a no brainer. Hmm, curiously, you might view parrot as an attempt to write p6 in C... well, sorta kinda not really. 03:18
monomorph: very neat. might the observations be summarized as "it was hard to look for a big picture description of the Perl 6 project, and it turned out one didn't exist"? or something like that? 03:20
monomorph i'm not sure if i understand the part about perl evolving... that makes bootstrapping harder, not easier
about complexity: i agree, lua and perl are very different beasts 03:21
mncharity re makes bootstrapping harder, I guess the idea is p6 complexity and size push you towards implementing non-core stuff in the language itself, and towards using as high-level a language as you can get (which then tends to blur into bootstrap). 03:23
monomorph mncharity: in a nutshell, yes. although i wonder if there's no description of the big picture because there *is* no big picture. if that makes any sense.
mncharity (I note, re non-core and high-level language, parrot took a very different approach) 03:24
03:25 Auzon joined
mncharity re *no* big picture, well, one necessarily exists, even if it's balkanized. the 'two line summaries of everything' would have made you effort easier, no? 03:26
03:26 japhb left
mncharity *your 03:27
Tene So, I could give a two-line summary of most parts of the Perl 6 effort.
So could many people here. 03:28
mncharity the glossary might be tweaked in that direction?
Tene Where should that summary be posted such that it would solve that problem?
mncharity in the face of Unmaintained websites, it's not clear. the current wiki? pugs svn? as long as it's written somewhere, folks could point to it. 03:29
one issue is how to keep it updated. 03:30
but yeah, given a browsing user, I don't know how to get it to them before they hit #perl6.
which rather suggests any website updating should emphasize #perl6, to age better. 03:31
Tene: re two lines, if the first line is what it is, and the second line is status, I could give one-line summaries for lots of things, but few two-line summaries. 03:33
What is kp6-perl5 blocked on in core p5 and cpan module bugs? What can and can't one do with STD.pm? How far did the pugs hs new ast/compiler changes get? What if anything about pil2js is rotted? 03:36
How close is kp6 to being bootstrap-able? Do we have any idea what it would take to update pugs to current versions of ghc? What, briefly, can/can't PCR currently do?' 03:38
ok, starting to get to more than two lines there. 03:39
but lots of basic questions. pugs seems to be passing fewer tests than it used to. is that just because the language and test suite has been evolving, or is there some rot in pugs itself? 03:40
what are the two lines on Moose? how close is it to the current vision/spec of p6 oo? 03:41
but perhaps all that is more like a paragraph each, rather than two lines each. 03:42
Is parrot running against the pugs t/ ? How much is passing? 03:44
03:47 alester left
mncharity same question for pugs, pugs alternate backends, kp6, mp6, redsix. 03:47
obra getting all these questions on the wiki where folks can edit in answers would probably be very useful 03:50
(and then a note to p6c and p6l encouraging people to actually do the filling in ;)
Tene Is there now a wiki that is considered authoratative for Perl 6? 03:51
obra www.perlfoundation.org/perl6/index.cgi is likely reasonable 03:52
lambdabot Title: Perl 6 / Perl 6
mncharity www.perlfoundation.org/perl6/index.cgi claims officialness. it's unfortunately a socialtext wiki. There's also PerlNet. perl.net.au/wiki/Perl_6 03:54
lambdabot Title: Perl 6 / Perl 6
mncharity so yes, but if you used to mediawiki, well, it's a much simpler data model. so, for instance, in the absence of discussion/Talk pages, it's hard to calibrate collaboration. 03:57
monomorph almost 5am here, i'll finally go to bed. thanks for the discussion, a few things are clearer now
mncharity monomorph: thanks for your work, and the resulting discussion! :) 03:58
03:59 monomorph left 04:00 justatheory left
mncharity I note www.perlfoundation.org/perl6/ points into the ~year out of date svn.perl.org. 04:01
obra ...it is a wiki, no?
mncharity or maybe not. parrot. r25454? 04:02
obra parrot and the specs are still canonically in svn.perl.org
mncharity ah, great. 04:03
"perl6 is currently in the Parrot repository in languages/perl6/, but needs to become its own repository." ?!? 04:04
stevan_ mncharity: hey :) 04:05
mncharity www.perlfoundation.org/perl6/index...._dashboard
lambdabot Title: Development Dashboard / Perl 6, tinyurl.com/276go6
mncharity hey steven_! :)
obra that means rakudo
Rakudo was called 'perl6' until recently
stevan_ mncharity: Moose is not directly tracking the p6 OO spec,.. but I every once in a while I do review it and try and keep us heading the same general direction 04:06
mncharity: also Moose is helping push some of the other p6-ish p5 based efforts such as Devel::Declare, Devel::BeginLift, B::Simple and others 04:07
04:10 sanxiyn joined
stevan_ mncharity: and mst is been hanging out over on p5p a lot too... so I wouldnt be suprised if some of these efforts end up in 5.12 too 04:11
obra well, mst is a p5 person more than a p6 person ;) 04:12
mncharity neat. :) what are the most important current objectives and challenges? 04:13
re spec, does Moose have the same "a class is an undef instance of itself" kind of flavor, or is it more p5-like with a strong class/instance distinction? 04:15
stevan_ obra: he is one of the bridge folk :) we who stand between the two (sorry watching LoTR with my daughter) 04:16
mncharity: it has the same p5 class/instance distinction
mncharity oh, np. just curious. another time perhaps.
k 04:17
stevan_ there is a point where I cant do it in pure perl
without crazy autoload hacks, etc
so thats where mst has been coming in (Devel::Declare, etc)
B::Simple is basically LISPish macros that operate on the p5 opcode level
mncharity search.cpan.org/~mstrout/Devel-Decl...Declare.pm 04:18
lambdabot Title: Devel::Declare - - search.cpan.org, tinyurl.com/2gk6zd
stevan_ yeah look at the tests... the docs are crap 04:19
search.cpan.org/~mstrout/Devel-Decl...6/MANIFEST
lambdabot Title: The CPAN Search Site - search.cpan.org
mncharity hmm, i'm not getting a cpan hit on B::Simple 04:20
www.faqs.org/docs/perl5int/x1854.html B/Simple 04:22
04:25 japhb joined 04:27 orafu joined, simcop2387 left
mncharity obra: re questions->wiki, email->p6c|p6l, ... volunteers welcome. I'm afraid it's too non-Ofun with too low a "do it anyway, it's critical" payoff for me. 04:30
04:30 luqui left
obra k 04:33
04:34 simcop2387 joined, Limbic_Region left
mncharity TimToady: re t/spec, it might be helpful if one could easily break the .t files into fragments which could be run separately. so an implementation could use the tests without having to do a lot of test file editing. or at least many of the tests. 04:37
Auzon Well, the tests can be marked to skip for a certain implementation... 04:38
04:38 pbuetow left, pbuetow joined
obra mncharity: how much have you read up on fudge? 04:38
mncharity indeed. but if, for instance, the tests guarenteed that any state which affected a test was within a surrounding block, then one could automagically break up the test files, run the tests independently, and not have to worry about maintaining a set of skip/eval/etc tags. 04:39
re read up on fudge, skimmed only.
Auzon well, then you're maintaining a list of which tasks to execute somewhere... 04:40
obra It's my understanding that fudge is timtoady's solution to the problem you're describing
I may well be wrong, though ;) 04:41
04:43 OuLouFu left
mncharity re solution, yes, but it's the same approach of pugs. tagged tests, and no information on whether those tests depend on other parts of the file (decls, mutation in other tests, etc). 04:43
which is ok. but since the tests are being touched anyway, it might be nice to add additional nature-of-the-test tagging to permit use without editing the files. 04:44
04:45 literal joined
mncharity re "you're maintaining a list of which tasks to execute somewhere", maybe, at least for regression testing, but it means everyone with a p6 implementation or mutant doesn't need to be banging on the spec .t files. 04:46
sanxiyn mncharity: You want test inclusion/exclusion list maintained outside of test file itself? 04:50
mncharity that said, it's not clear to me whether there's a nice way to do it. without going for a perhaps painfully more verbose '(test jig / optional setup / tests... )+" approach.
sanxiyn: I'm picturing being able to mechanically grovel over t/spec, and run the tests independently. so instead of "test file" being the unavoidable unit of test running, it's instead "cluster of tests which can't be run separately". 04:52
ideally it would be nice to have full dependencies. so you know test1 and test2 depend on the preceding decl, but not on each other. 04:53
but each step down that path can make the spec less readable, less maintainable. 04:54
so I'm wondering if there might be a low-hanging fruit around. like saying top level blocks in the test files are independent of the rest of the file. 04:55
the one could go through the files, yank each top level block, and run it separately.
sanxiyn You want setup/teardown? 04:56
04:56 alester joined
mncharity that would do it. 04:57
04:57 literal left
mncharity as long as the clarity cost wasn't too high. 04:57
people with mutants can always make their own copy of t/spec, to avoid having to edit the master. so perhaps it's not a big deal. it's just been on the pugs t/ wish list since forever, for a t/ rev 2, and t/spec seemed to have some t/ rev 2-ness to it, so I thought I reraise the issue for consideration. 05:00
I was also thinking earlier today, "how might one set up a test-driven-development hackathon to create a p6 implementation?". to which an immediate response is one wants test rather than file level failure granularity. (the one bit of file you can't handle failing the entire, or rest of, the file). 05:03
a test suite one could set up at the beginning. I suppose one could go through and tag all tests, and all declarations, EVAL or some such. but... not a pretty approach to TDD. 05:04
hmm, my decls are local to eval() blocks. otherwise automagically wrapping every single statement in an eval might be a solution. 05:06
perhaps it's just an observation that new implementations should build fault tolerance into their parsers, compilers and runtimes. 05:09
though the difficulty in doing that is what motivated tagging to start with.
rather than actual blocks of related tests, perhaps one could just use comments. #cluster{ ...setup... #tests ...tests... #cluster} 05:13
with a default that otherwise, the tests are independent. or something... 05:14
05:15 manfred joined
mncharity #{ ... setup ... ...first-test-terminates-setup...tests... #} 05:15
sanxiyn IMO, tests should be independent.
mncharity "tests in ...tests... are independent, and can be run on a fresh copy of ...setup..., unless bracketed together with #( ... #) " ? 05:19
use v6; #{ my $x; #{ $x = 3; is($x,3); #( is(++$x,3); is($x,4); #) is($x+2,5) #} #{ my @a; $x = +@a; is($x,0); #}} 05:25
err, no: is($x++,3) 05:28
hmm. though that set of tests _depends_ on tests being independent. which would be a non-trivial departure from the current style. the #mumbles are no longer simply comments. so... 05:31
perhaps just tag the #{ #} with whether the enclosed tests are independent or not. #{i vs #{ ? tests in #{i being declared to be sideeffect-free. or at least not to depend on each others sideeffects. ? 05:34
re _depends_, ie, $x went from 4 to 3 just because a #) was passed. which seems a no-no. 05:35
05:35 Connel left 05:36 Connel joined
mncharity so, calling them #{i and #{d for in/dependent enclosed tests, the story might be "test files can be decomposed into independent tests, and #{d clusters of tests, all with associated setup". 05:41
anyway. maybe something to consider. 05:42
end of day. g'night all. &
sanxiyn bye 05:44
05:49 mncharity left 05:50 jordilin joined 06:21 justatheory joined 06:29 Auzon left, Auzon joined 06:40 thestarslookdown joined 06:42 _newbie_ joined 06:43 _newbie_ left 06:53 jordilin left 07:01 thestarslookdown left 07:02 kanru joined
peepsalot are negative indices supported in slices? @a[0..-2] for example could get everything but the last element? 07:03
07:07 sri_work left 07:20 Connel left 07:21 Conne1 joined, Conne1 is now known as Connel 07:39 devogon joined
peepsalot pugs: %*ENV 07:45
exp_evalbot RESULT[\undef]
peepsalot pugs: my $a = 123; %*MY; 07:47
exp_evalbot RESULT[{}]
peepsalot pugs: my $a = 123; %*MY
exp_evalbot RESULT[{}]
peepsalot this page talks about symbol tables, search.cpan.org/dist/Perl6-Doc/lib/...le_scoping but I don't understand how you access them 07:48
lambdabot Title: Perl6::Overview -- a brief introduction and overview of Perl 6 - search.cpan.org, tinyurl.com/2ez6gj
peepsalot i was able to access %*ENV in pugs on my local machine, but not the other ones? 07:49
ugh, why do i keep ending statements with question marks. past my bed time 07:50
Auzon This one is running in a safe compartment or something of the sort
e.g.
pugs: sleep
exp_evalbot OUTPUT[*** ā¤ Unexpected end of inputā¤ at /tmp/jJMHcch0Cp line 2, column 1ā¤]
Auzon pugs: sleep 1;
exp_evalbot OUTPUT[*** Unsafe function 'sleep' called under safe modeā¤ at /tmp/wZpsMmsxFY line 1, column 1-8ā¤]
peepsalot Auzon, ok, i can understand that %*ENV might not be good to display here, but what about all the other symbol tables? MY OUR GLOBAL etc. ? 07:51
Auzon No idea.
peepsalot I can't figure out how to show them in my own pugs terminal 07:52
should they theoretically be accessed in the same way? with %* ?
Auzon pugs:%::
pugs: %:: 07:53
exp_evalbot OUTPUT[*** ā¤ Unexpected end of inputā¤ expecting "(", "!", "/", twigil or word characterā¤ at /tmp/rtv3cEDYKK line 1, column 4ā¤]
Auzon pugs: %*::
exp_evalbot OUTPUT[*** ā¤ Unexpected "*::"ā¤ expecting "::"ā¤ at /tmp/tSNfUrWz7Y line 1, column 2ā¤]
Auzon pugs: %::*
exp_evalbot OUTPUT[*** ā¤ Unexpected end of inputā¤ expecting word characterā¤ at /tmp/jx3ZuTIbno line 1, column 5ā¤]
Auzon Hm. There's somewhere to get it, or there should be 07:54
TimToady peepsalot: see S09:619 "Negative and differential subscripts" 07:57
peepsalot TimToady, where do I find that? i'm new to all this 08:02
nevermind, i found it 08:03
ugh, was looking in s06 instead of s09
peepsalot smacks head
08:04 Aankhen`` joined
peepsalot will read tomorrow, it's getting too late for me to function 08:04
TimToady anyway, you'll write @a[0..*-2] for that 08:07
(not implemented yet, btw) 08:08
so you'll still see a lot of @a[-1] that will eventually be @a[*-1]
Auzon What does the star signify? 08:09
TimToady * is a term that is of type Whatever, and it is intentionally context dependent 08:10
the operator that receives a * gets to decide what it means
Auzon So it's a DWIM operator?
TimToady so an infinite range is just 0..* for instance
essentially, as long as you mean what most people would mean by * in that situation :) 08:11
Auzon Fair enough.
so what should @a[-1] do? throw an error?
TimToady turns out to be useful all over the place 08:12
yes
Auzon Thank you. :) I've been bitten in Perl 5 by $arr[$index] when $index is below 0
TimToady well, we're trying to fix everything wrong with Perl 5 and then some... :)
Auzon Yeah. I cannot wait to start using Perl 6 in my freetime project. :) 08:14
TimToady unfortunetly Perl 6 is my freetime project :/
Auzon No one wants to pay you to work on Perl 6? That's unfortunate :( 08:15
TimToady I wish people could send back a little money from the future for that, but that doesn't seem to work 08:16
course, maybe that means we've made money obsolete in the future :) 08:18
Auzon I don't know if that'd be good or not. Too much to think about at this time of night... 08:19
TimToady yeah, I'm waiting for my antihistamine to hit so I can sleep... 08:20
and thinking more about possible list assignment operators 08:21
Auzon Where are captures defined in the spec? 08:23
ah, I see something now 08:24
TimToady well, sort of all over the place, S02, S03, S04, S06 08:27
Auzon or this: svn.pugscode.org/pugs/docs/Perl6/FAQ/Capture.pod
08:28 Lion2 left 08:29 Aankh|Clone joined 08:47 Aankhen`` left 09:09 iblechbot joined 09:32 blindfish joined 09:35 drrho joined 09:40 Aankh|Clone is now known as Aankhen`` 09:54 meppl joined 10:06 IllvilJa joined 10:08 rindolf joined 10:15 ispy_ left 10:17 barney joined
rakudo_svn r25461 | bernhard++ | [Rakudo] 10:28
r25461 | bernhard++ | Clean up perl6.o
10:40 ruoso joined 10:47 Aankh|Clone joined 11:05 Aankhen`` left, chris2 joined, drrho left
rakudo_svn r25462 | bernhard++ | [Rakudo] 11:06
r25462 | bernhard++ | Fiddle with perl6/Makefile.
r25462 | bernhard++ | - Use make variable $(EXE)
r25462 | bernhard++ | - Remove setup of unused commands
r25462 | bernhard++ | - Let 'make help' tell about the targets 'perl6' and 'xmas'.
r25462 | bernhard++ | - Use $(RM_F) instead of $(RM_RF).
11:44 riffraff joined
riffraff hi 11:46
12:06 manfred left 12:09 jferrero joined 12:10 IllvilJa left 12:12 monomorph joined 12:14 pdy joined
pdy hi 12:15
pugs_svn r19796 | ruoso++ | [smop] Most of the memory leaks are solved... ALL TESTS PASSING :) 12:45
12:47 cognominal_ joined 12:49 cognominal__ left 13:07 turrepurre left 13:21 manfred joined 13:48 DarkWolf84 left 13:57 chris2 left, DarkWolf84 joined 13:59 wilx joined 14:08 barney left
rakudo_svn r25464 | coke++ | [rakudo] 14:13
r25464 | coke++ |
r25464 | coke++ | Spelling fix; minor output consistentifcation.
14:25 Whreq is now known as Juerd 14:28 rff_ joined 14:34 drrho joined 14:39 riffraff left 15:07 alester left 15:18 Aankh|Clone is now known as Aankhen`` 15:45 alester joined 16:08 marmic left, meppl left 16:09 Psyche^ joined 16:11 meppl joined 16:12 marmic joined 16:14 chris2 joined
peepsalot TimToady, so, this perl6 doc has some misinformation about arrays then? I guess that's the perl 5 way of donig things, not perl 6. 16:20
search.cpan.org/dist/Perl6-Doc/lib/...le_scoping
er, its closer to this part of the page search.cpan.org/dist/Perl6-Doc/lib/...able_types under arrays of course 16:21
should I not be reading this perl 6 doc? 16:22
16:22 penk left 16:23 penk joined 16:26 Patterner left, Psyche^ is now known as Patterner, alester left 16:28 lambdabot joined
rakudo_svn r25472 | jonathan++ | [rakudo] Allow classes whose last declaration is an attribute to compile. 16:32
16:37 penk left 16:38 penk joined 16:46 pbuetow left, pbuetow joined
rakudo_svn r25473 | jonathan++ | [rakudo] Have the new method on objects initialize the attributes to Undef or the supplied value. Note we aren't passing named parameters yet, so can't really use that bit. It does, however, let us mutate attributes through the accessor (we're always is rw by default at the 16:53
..moment, though that will change).
16:54 barney joined 17:02 meppel-san joined 17:13 pmurias joined 17:19 meppl left 17:23 schmalbe joined
pugs_svn r19797 | cosimo++ | [t/spec] Minor fixes to S16 tests (pod + fudgify) 17:29
17:31 rff_ left 17:38 schmalbe left 17:40 barney left
TimToady peepsalot: the Overview directory is not canonical, and can get out of date with respect to the synopses, as it is in this case on negative subscripts. 17:43
peepsalot ok 17:48
17:54 thoughtpolice joined 17:56 pcbuetow joined
pugs_svn r19798 | lwall++ | Added "self" for jonathan++ 17:58
18:01 drrho left 18:14 avar left
pmurias ruoso: how will s1mp look like? 18:17
18:17 wilx left 18:35 icwiener joined
rakudo_svn r25477 | jonathan++ | [rakudo] Implement self. Note that we can't make it a term yet, where it belongs, so it's a noun for now. 18:43
ruoso pmurias, s1p looks like Perl 6... but it is translated to smop C runtime + sm0p code... 18:44
it doesn't have namespaces, so everything should resolve to the smop C runtime names... 18:45
it doesn't have lexical scopes also, which means that variables should be compiled to sm0p operations
www.perlfoundation.org/perl6/index....munication
lambdabot Title: SMOP Inter Continuation Communication / Perl 6, tinyurl.com/yuqcgm 18:46
ruoso that is how one sm0p frame can communicate with another to pass the values that transcend the frames.
s1p will already be completely stackless, that's why all its code is sm0p based 18:47
after I finish the stage0 details (memleaks), we'll implement p6opaque which is the default responder interface for Perl 6 objects 18:48
s1p will already be a KP6 backend, just not the final one...
18:50 jordilin joined 18:52 riffraff joined
pmurias ruoso: i see 18:54
18:54 meppel-san is now known as meppl 18:55 jordilin left
pmurias ruoso: s1p will feed source code into the kp6 the ususal way rather than via q:s1p {...} right? 19:02
peepsalot is this class Marker ... example correct? www.oreillynet.com/onlamp/blog/2007...ors_1.html 19:16
19:16 rindolf left
lambdabot Title: YAP6 Operator: Repeat Operators - O&#039;Reilly ONLamp Blog, tinyurl.com/2jzu42 19:16
moritz_ looks 19:17
19:17 meteorjay left
moritz_ it should, yes 19:17
peepsalot i guess I just don't undertand the method declaration, and how it knows the way to convert to string
it doesn't seem to run correctly in pugs in my terminal
moritz_ when you have a method with name $type, it is used to convert to that $type 19:18
ah, pugs isn't very up to date
peepsalot ok 19:19
moritz_ class A { method Num { 1 } }; my $a = A.new; say $a + 1; # should print 2
peepsalot pugs: class A { method Num { 1 } }; my $a = A.new; say $a + 1; 19:20
exp_evalbot OUTPUT[*** Cannot cast from VObject (MkObject {objType = (mkType "A"), objAttrs = <Hash:0xb714e378>, objOpaque = Nothing, objId = MkObjectId {unObjectId = 3}}) to Double (VNum)ā¤ at /tmp/AEgbJxf10x line 1, column 50-56ā¤]
peepsalot heh
moritz_ that feature was introduced to the specs after audrey got sick 19:21
and nobody cared to implement it since
peepsalot do you not need semicolons after method declarations?
19:21 Connel is now known as SomeWiktAdmin 19:22 avar joined
moritz_ no 19:22
in general you don't need any semicolons after closing } that delimits a block
in p5 you needed here one: my $a = sub { ... }; 19:23
19:49 SamB joined 19:52 Limbic_Region joined 20:02 jferrero left 20:07 aindilis left, aindilis joined
TimToady more precisely, you don't need one after a line-ending } 20:09
you do if you put more on the same line
bbl & 20:10
20:22 Aankhen`` left
peepsalot lol 20:28
This synopsis summarizes Apocalypse 10, which discusses packages despite never having been written.
moritz_ ;)
20:29 riffraff left 20:33 sri_work joined
peepsalot is Order package? (ex. Order::Decrease) 20:34
s/Order/Order a/
i'm trying to follow these micro articles, but I guess they assume knowledge of perl 5 20:37
i don't know what is the difference (as far as perl 6 is concerned) between a package, module, class, etc
was wondering if there's a list somewhere of builtin things like Order:: 20:38
moritz_ a package is basically a namespaces
classes introduces namespaces
and modules do as well
peepsalot so what is Order ? 20:40
moritz_ in p5 namespaces are not hirarchical, so you could have Order::Decrease without having Order:: 20:41
I'm not sure how that's don in p6 20:42
ruoso pmurias, yeah... s1p is plain code to be compiled by kp6... 20:43
but still generating C code, or better, sm0p code...
peepsalot moritz_, is there a way to list all the global namespaces that are part of the language like that? 20:44
or a page that lists them
moritz_ peepsalot: there will be, both 20:45
but I don't think there is a way yet
peepsalot doh
moritz_ the most important builtins are summarized in S29 though
peepsalot ko
20:50 Khisanth left
moritz_ TimToady: are you going to make the parse tree available trough a special variable? 20:59
perhaps only on request, to avoid performance impact 21:00
21:01 chris2 left 21:02 chris2_ joined 21:07 jferrero joined
pdy slaps moritz_ with a rubber band 21:10
TimToady well, if I don't make it available, someone else surely will :) 21:11
moritz_ pdy: why that? that hurts!
TimToady: I meant in terms of specs ;)
pdy moritz_: was looking for you this weekend till i found out that i can help myself via your website :-p 21:12
moritz_ pdy: ;)
pdy: you don't have a different nick in #perlde by chance? ;-)
pdy sure, i am Leibsle there
moritz_ thought so
pdy :-)
your log bot is really a good learning example 21:13
moritz_ thanks
peepsalot what website? 21:14
ist auf deutsche? 21:15
moritz_ assumes irclog.perlgeek.de/ or moritz.faui2k3.org/en/ilbot
lambdabot Title: IRC logs - Index
pdy yep, moritz.faui2k3.org/ i finally found when i followed the white rabbit in the channel topic
lambdabot Title: Moritz Lenz
pdy cause i failed to find the svn source 21:16
21:22 Khisanth joined 21:24 silug left 21:29 cosimo_ joined 21:35 Khisanth left, Khisanth joined 21:44 alester joined, cosimo_ left 21:47 SomeWiktAdmin is now known as Connel
ruoso monomorph, about your post at jottit.com/r3stt/... the problem with trying to write a something simply in ANSI C is that Perl 6 features require a very rich runtime environment. While I don't completely agree with chromatic (in the sense that I do believe that it can be simpler than parrot), it's not simple as gcc *.c, just take a look in how smop is evolving and you can have a better understanding... 21:50
brb &
monomorph i stumbled upon smop after i wrote this
i don't yet understand what it does :) 21:51
but i agree with what mncharity said yesterday on that point, lua is unique in that the language is specifically engineered to have a very simple implementation 21:52
21:56 alester left 22:01 manfred left, pmurias left 22:05 Connel left 22:07 iblechbot left 22:08 kane__ joined, kane_ left 22:17 silug joined 22:30 wknight8111 joined 22:37 sanxiyn left 22:39 pbuetow left
moritz_ is $foo[$a] the same as $foo.postcircumfix:<[ ]>($a) ? 22:40
22:41 jferrero left 22:52 alester joined, Torment joined, alester left 22:59 jferrero joined 23:02 justatheory left 23:09 Jedai left 23:14 alester joined 23:25 jferrero left 23:27 BinGOs left, BinGOs joined 23:30 jferrero joined 23:33 marmic left 23:34 blindfish left
ruoso monomorph, smop is an effort on doing Perl 6 the most pragmatic way I can possibly find out... 23:36
23:37 jeffreykegler joined, marmic joined
jeffreykegler TimToady (if you're there): you mentioned the other day that Perl 6 would be "deterministically parseable" 23:42
I take you mean by that LR(k), or parseable by a deterministic PDA
Another reading is that you mean "no rand() stuff in BEGIN blocks can affect the parse" -- a much weaker statement 23:43
Correcting the above ...
That is, you said that under certain circumstances Perl 6 would be "deterministically parseable" 23:44
I think it was in the context of the tools discussion
TimToady I mean that Perl 6 is many languages, but we'll know exactly which one we're in at any moment 23:52
jeffreykegler So both the above readings are wrong 23:53
In the parsing world "determinism" means a zillion different things 23:54
TimToady everyone keeps treating Perl 6 like it's a single language; it's really a metalanguage 23:55
well, I'm going for vigor, not rigor :)
jql I see perl6 as the language between the {} in the rules syntax
jeffreykegler In particular, then, my reading you're using the phrase in the Grune and Jacobs sense ... 23:56
to mean LR(k) is just totally wrong
23:57 pcbuetow is now known as pbuetow
Tene "Wherever I go, there I am!" 23:57

Powered by Cro and the RakuĀ® Programming Language.

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

Thank you!