Perl 6 - the future is here, just unevenly distributed

IRC log for #perl6, 2008-04-10

Perl 6 | Reference Documentation | Rakudo

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

All times shown according to UTC.

Time Nick Message
00:06 mncharity TimToady: trait_auxiliary:is currently takes an <ident><postcircumfix>? .  so  'class A is B::C {}' doesn't parse.  intentional?
00:12 justatheory joined #perl6
00:12 SubStack joined #perl6
00:15 TimToady hmm, well, we have to be able to derive from classes with :: in the name...
00:17 TimToady so we need to parse <name> somehow in those cases
00:19 mncharity np.  can workaround it.
00:21 mncharity tnx :)
00:21 pugs_svnbot r20253 | lwall++ | trait_auxiliary:is should look for <name>, mncharity++ for noticing
00:21 pugs_svnbot diff: http://dev.pugscode.org/changeset/20253
00:21 lambdabot Title: Changeset 20253 - Pugs - Trac
00:24 muixirt2 joined #perl6
00:55 zonkers_ joined #perl6
00:57 perlbot joined #perl6
01:02 mncharity "extends 'X';" to "use base 'X';" to "{no strict;unshift(@ISA,'X')}".   blech.
01:04 mncharity to "{no strict;CORE::unshift(@ISA,'X')}".  :/
01:11 ting joined #perl6
01:12 felipe joined #perl6
01:14 mncharity and back to "use base".  the motivating problem is one wants to be able to "is X :and_use_its_methods_in_prefer​ence_to_any_we_already_have".
01:18 armagad joined #perl6
01:49 * mncharity tries to remember what the official rule for deciding whether {...} is a block or a hash...
01:52 zakame joined #perl6
01:54 thoughtpolice joined #perl6
02:06 armagad joined #perl6
02:10 FurnaceBoy joined #perl6
02:15 zakame_ joined #perl6
02:34 pugs_svnbot r20254 | putter++ | [elf_d] Multiple 'is Foo' now work (rather than each deleting the previous one).
02:34 pugs_svnbot diff: http://dev.pugscode.org/changeset/20254
02:34 lambdabot Title: Changeset 20254 - Pugs - Trac
02:37 wtgee joined #perl6
02:46 pugs_svnbot r20255 | putter++ | [elf_e] elf_e forked from elf_d.  Needed working 'is Foo'.  Expected focus is using IR analysis to clean up the emitted runtime.
02:46 pugs_svnbot diff: http://dev.pugscode.org/changeset/20255
02:46 lambdabot Title: Changeset 20255 - Pugs - Trac
02:49 xdg joined #perl6
02:56 mncharity I'm not entirely happy with having to start elf_e already, but oh well.
02:56 pugs_svnbot r20256 | putter++ | [elf_e] Two files missing from r20255.
02:56 pugs_svnbot diff: http://dev.pugscode.org/changeset/20256
02:57 lambdabot Title: Changeset 20256 - Pugs - Trac
02:58 mncharity next step, real GLOBAL's, and no more "all subs are global" kludge.
02:59 meppl joined #perl6
03:10 mncharity good night all &
03:10 pugs_svnbot r20257 | putter++ | misc/elfish_smop: Smop emitter for elf.  A very limited start, but recently neglected, and so checked in.
03:10 pugs_svnbot diff: http://dev.pugscode.org/changeset/20257
03:10 lambdabot Title: Changeset 20257 - Pugs - Trac
03:21 peepsalot joined #perl6
03:22 peepsalot joined #perl6
03:36 alester joined #perl6
04:15 iblechbot joined #perl6
04:49 alc joined #perl6
05:06 Psyche^ joined #perl6
05:07 kanru joined #perl6
05:44 iblechbot joined #perl6
06:04 wtgee left #perl6
06:13 elmex joined #perl6
06:39 sri_work joined #perl6
07:12 agentzh left #perl6
07:39 agentzh joined #perl6
07:49 c9s joined #perl6
08:02 alanhaggai joined #perl6
08:35 IllvilJa joined #perl6
08:36 qmole joined #perl6
08:38 alanhaggai__ joined #perl6
09:24 ruoso joined #perl6
09:52 Schwern joined #perl6
10:09 iblechbot joined #perl6
10:22 zac joined #perl6
10:50 syle joined #perl6
11:08 meppl joined #perl6
11:52 bbkr joined #perl6
12:00 meppl joined #perl6
12:17 alanhaggai joined #perl6
12:26 rdice joined #perl6
13:17 alanhaggai__ joined #perl6
13:53 TJCRI joined #perl6
13:55 alester joined #perl6
14:03 cmarcelo joined #perl6
14:06 chris2 joined #perl6
14:20 sscaffidi joined #perl6
14:57 IllvilJa joined #perl6
15:20 kcwu joined #perl6
15:39 peeps[work] joined #perl6
15:50 pen1 joined #perl6
16:13 armagad joined #perl6
16:17 xdg joined #perl6
16:22 wtgee joined #perl6
16:30 ludan joined #perl6
16:51 wtgee left #perl6
16:54 wtgee joined #perl6
16:55 wtgee joined #perl6
16:56 wtgee joined #perl6
17:01 wtgee joined #perl6
17:04 Schwern joined #perl6
17:04 pen1 left #perl6
17:18 pen1 joined #perl6
17:19 gbarr joined #perl6
17:20 lisppaste3 joined #perl6
17:24 rindolf joined #perl6
17:28 eternaleye_ joined #perl6
17:35 BinGOs joined #perl6
17:37 justatheory joined #perl6
17:46 wtgee left #perl6
18:03 peepsalot joined #perl6
18:04 mj41 joined #perl6
18:10 sscaffidi joined #perl6
18:22 sri_work joined #perl6
18:26 barney joined #perl6
18:46 pmurias joined #perl6
18:48 pmurias @tell mncharity you can use CORE::unshift(our @ISA,'X') instead of use strict
18:48 lambdabot Consider it noted.
18:48 pmurias @tell mncharity s/use strict/no strict/
18:48 lambdabot Consider it noted.
18:48 smtms joined #perl6
18:52 sscaffidi_ joined #perl6
18:54 rdice joined #perl6
19:02 araujo joined #perl6
19:08 pmurias @tell based on the commits you seem to have stayed with the Moose version of the perl5 backend, why? isn't the none-Moose version much faster on your machine also?
19:08 lambdabot Consider it noted.
19:14 justatheory joined #perl6
19:16 [particle1 *phew* that was a long mail to write
19:23 ofer joined #perl6
19:29 Lorn joined #perl6
19:34 ruoso joined #perl6
19:37 mncharity joined #perl6
19:37 mncharity hi pmurias
19:37 lambdabot mncharity: You have 2 new messages. '/msg lambdabot @messages' to read them.
19:38 mncharity /msg lambdabot @messages
19:38 mncharity /msg lambdabot @messages
19:39 mncharity /msg lambdabot @messages
19:40 mncharity /msg lambdabot @messages
19:40 [particle] you have to delete a leading character
19:41 mncharity getting very odd behavior.  sometimes having to remove and retype not just /  but /m else fail.  and thrice got a no nick lambdabot.  wierd.
19:43 mncharity pmurias: re our @ISA, nifty.  :)   now just have to figure out how to say it in p6. :)   for both classes and roles.
19:45 mncharity pmurias: re non-Moose speed, good question.  will check.  need to first unbreak elf_e.  a rather startling number of bugs and release errors got together to pretend to work, while not really.
19:47 pen1 left #perl6
19:51 mncharity oky, speed...
19:51 pugs_svnbot r20258 | putter++ | [elf_e] Further clean up r20255/r20256.  Now IRx1_Analysis is actually run.
19:51 pugs_svnbot diff: http://dev.pugscode.org/changeset/20258
19:51 lambdabot Title: Changeset 20258 - Pugs - Trac
19:52 pmurias mncharity: hi
19:53 pmurias mncharity: adding a superclass in p6 would use the .HOW
19:57 mncharity "adding a superclass in p6 would use the .HOW"?
19:58 pmurias mncharity: i understood you asked how to do it at runtime?
19:58 mncharity re speed, wizzzzzzy.  With cached parse, elf_d self-compiles for me in 2.5 vs 9.5 seconds with NoMoose.
19:59 pmurias sort of ::Foo.HOW.add_parent(::Bar.HOW)
19:59 mncharity re 'at runtime', ah, sorry, no.  Wondering how to say ' is Foo :and_put_it_at_the_head_not_tail_of_inhe​ritance_and_dont_complain_about_override​s_because_thats_what_Im_using_it_for;'
20:02 mncharity if   class A { is B; is C; }  results in a search path  A,B,C,  then one also wants a  class A { is B; is_first C } which results in a  A,C,B
20:03 mncharity and something similar with does(), but which doesn't complain if the role overrides the definitions in other does()ed roles.
20:03 mncharity these may be in the Snn, haven't looked yet.  todo.
20:04 pmurias class A { BEGIN { B.HOW.methods.unshift(C.HOW) } }
20:05 mncharity :)
20:05 pmurias (not directly speced) ;)
20:05 mncharity .methods ?
20:06 mncharity i'd have expected an "isa" something.
20:07 pmurias sorry
20:07 pmurias looking up more correctly
20:08 pmurias it's isa
20:08 mncharity ah, ok
20:09 pmurias i think perl6 uses C3 inheritance
20:09 mncharity that's my fuzzy recollection
20:09 mncharity *very fuzzy
20:11 pmurias what do you need it for in p6?
20:11 pugs_svnbot r20259 | putter++ | misc/elf/elf_d_src/README: Added a ElfD_NoMoose.pm bootstrap recipe.
20:11 pugs_svnbot diff: http://dev.pugscode.org/changeset/20259
20:11 lambdabot Title: Changeset 20259 - Pugs - Trac
20:11 [particle] perl 6 is C3 by default
20:12 pmurias mncharity: what do you mean by real GLOBAL?
20:13 pmurias AFAIK the stuff in the prelude just gets imported into the current lexical scope at the start of the file
20:13 mncharity $*x should emit as $GLOBAL::x  or somesuch, but doesn't at present.
20:14 pmurias ok
20:15 mncharity currently emits as $x, and the prelude says "our $x", and we cross our fingers that no one has said "my $x" in between.  kludge.
20:16 pmurias what stops it from being implemented
20:16 pmurias ?
20:17 mncharity now, nothing.  but needed to have IR analysis in place.
20:17 pmurias why?
20:18 pmurias $*x is always $GLOBAL::x
20:18 mncharity because p5 apparently doesn't have a way to define globals, so since you can't tell if the variable you are looking at is global or not... .   fine for explicit  $*x.  not so great for foo(3).  is &foo a my sub foo?  a &*foo?  can't tell
20:19 pmurias p5 dosn't have lexical subs
20:20 pmurias but if you do $x, perl will check if it's a lexical or package scoped $x
20:21 TimToady in p5 all lexical subs have to be stored in "my $subref"
20:21 mncharity when emitting   say(3), you have to emit different code if that's referring to &*say, or referring to some sub declared in package, or some lexical sub.
20:21 pmurias &*say is lexical
20:21 pmurias i mean in say "foo", &say is lexical
20:22 TimToady yes, so called "global" subs are imported into the lexical scope by the prelude
20:22 TimToady you don't have to look in *
20:23 mncharity for a prelude say, in p5 it's  "GLOBAL::say(3)".  for "class A { sub say($x){}...}"  it's  "say(3);".  for "{ my sub say($x){}; ...}" it's "$code_say->(3)".
20:24 pmurias mncharity: for class { sub say($x) {} } it's also $code_say or say will be visible as a method
20:24 mncharity so when emitting "say(3)", you have to have groveled over the IR to determine where that say is coming from.
20:25 mncharity re "visible as a method", hmm.  sigh.  point.
20:25 TimToady in class A it's just "our" instead of "my"
20:25 TimToady but still looks lexical from the call
20:30 mncharity re "you don't have to look in *",  module NotInPrelude { sub *g(){3}; } ... module SomewhereElse { ... say g(); ... }.  no?
20:30 penk joined #perl6
20:31 pmurias is it possible to declare a lexical var from within a use'ed module?
20:31 TimToady mncharity: not under "use strict"; "no strict" has a failover to GLOBAL
20:32 TimToady but that just means adding it as the multi candidate
20:32 TimToady s/the/the last/
20:34 TimToady strict code is now required to import to omit the *; why %ENV no longer works, but %*ENV does
20:35 TimToady see S11: use GLOBAL <$IN $OUT $ERR>;
20:35 mncharity re "lexical var from within a use'ed", dunno.  as long as its use'ed at compile time, it seems at least vaguely feasible
20:35 mncharity ah
20:36 TimToady it's a matter of keeping control over the exact semantics under stricture
20:36 TimToady we can't have someone changing our language by adding things to GLOBAL
20:36 * TimToady oversimplifies, of course...
20:37 mncharity re "we can't have someone changing our language by adding things to GLOBAL", err, why not? :)
20:37 TimToady because audreyt said so, so there!
20:37 mncharity lol
20:37 TimToady just one of the many revelations of the pugs effort
20:41 mncharity /me reminds meself that the mission of elf is to be a hack, one which enables others to comfortably write hackish p6 code to create a real p6 compiler, and is not(with emphasis) trying to be a real compiler itself.
20:43 * TimToady empathizes
20:47 mncharity very slippery slope.  "sinking sensation"(especially complexity induced) => "slipping down slope" => "you left the path again, didn't you?" :)
20:49 mncharity pmurias: re Moose,
20:49 mncharity I failed earlier to sketch the following for stevan_, and so may simply be confused, but
20:53 mncharity my impression is "Moose has put a lot of work into assorted features"; "as we go forward, we'll wan't more features"; "so the key question isn't what set of Moose features are being used now"; "but what set of Moose features we may want to use soon"; "and how much work is saved by using the Moose versions, rather than rolling our own"; "but I'm unclear on
20:53 mncharity what that set looks like".
20:55 mncharity for illustrative example, say the absence of default values for has() variables gets annoying.  will it be easier to use the Mooseish version (assuming we can), or to implement it?
20:56 PerlJam mncharity: the former certainly.
20:58 Foke2 joined #perl6
20:58 mncharity or say elf_e starts trying to do something like  "[] is autoboxed as package ARRAY, which claims to be class Array.  There's also a package Array, and both inherit from a non-spec  Array::Role ."  or some such.  basically, if we want to start doing roles, it might be easier to just do whatever Moose does with them, at least for elf, than trying to do something new.
20:59 mncharity so I end up torn.  speed++, easy_development++, tradeoff.
20:59 stevan_ mncharity: use ruby :P
21:00 stevan_ or better yet,.. rails
21:00 PerlJam mncharity: I thought you just said that elf wasn't trying to be a "real" compiler?  In that light, you go with easy development.  (Assuming that's what you think Moose is :)
21:00 stevan_ mncharity: you might want to look at Coat,.. it is a mini-moose,. that is mostly upwards compat
21:00 mncharity maybe refactor EmitSimple and EmitNoMoose, so everything which is shared can just be inherited, and we can keep both running
21:00 mncharity s/can/can easily/
21:03 mncharity PerlJam: re real, indeed.  perhaps I should have said "easier development because of shorter edit-test loop" vs "easier development because less code needs to be written (modulo that code may not quite do what you would ideally have preferred)".
21:04 mncharity re "Coat", lol.  looking...
21:04 pmurias speed is very important at least for me if it fall below a certain treshhold i procrastinate too much
21:04 mncharity me too
21:05 mncharity oh, I meant to add a note to README to emphasize setting STD_RED_CACHEDIR if you are doing a lot of bootstrap compilation.  much nicer not to pay any parse cost.
21:05 stevan_ careful of the whole premature optimization stuff,.. Parrot is only now just recovering from the damage it caused
21:05 stevan_ but YMMV
21:05 pmurias stevan_: kp6 demonstrates the oposite pitfall
21:05 stevan_ anyway.. Coat is pretty good, I worked with it's author to keep things reasonably close to Moose
21:05 jrockway stevan_: but now the parrot object system is ... less than desierable
21:06 mncharity re "premature optimization", "amen". :)
21:06 stevan_ pmurias: with Moose, we went for correctness first, and optimization second,.. and we are still a little slow, but the optimizations are coming in big leaps now
21:07 stevan_ like 30 lines of XS and we were 45% less slow
21:07 mncharity the caveat being, when performance is actually one of your design objectives, then it needs to be watched like any other.  and if it is a/the critical one, then it needs to be focused on early and often.
21:07 stevan_ agreed
21:07 Auzon joined #perl6
21:07 jrockway if i were implementing perl6 right now, i would not give a fuck about performance :)
21:07 jrockway that is the easy part
21:08 stevan_ mncharity: why not try and write this in OCaml or something?
21:08 PerlJam jrockway: wait ... what's wrong with parrot's object system?
21:08 stevan_ fast and nice for writing compilers with
21:08 mncharity re 30 lines, yes.  the risk is it turning out those lines happen not to exist, because of some architecture decision made long ago, and now entwined throughout the system.
21:08 TimToady on the other hand, if you're designing Perl 6, you have to give lots of fucks about all the things your implementors are going to ignore until it's too late :)
21:09 stevan_ mncharity: my personal experince has always been that those entwined issues tend to be made in the cause of optimization more than anything else
21:09 stevan_ optimizations and death-march schedules
21:10 stevan_ but again, YMMV
21:10 mncharity re 'why not write in OCaml or something', the "yet another dead p6 implementation" problem.
21:10 TimToady there are many ways to design one's self into a corner...
21:10 stevan_ :)
21:10 stevan_ TimToady: so true
21:11 PerlJam TimToady: and then you design away the corner  :)
21:11 stevan_ PerlJam: then where will you stand ?? :P
21:11 TimToady well, you at least try to spot the corners in advance, and then declare "it can be proven that...", and go back to sleep.
21:11 mncharity re "those entwined issues tend to be made in the cause of optimization more than anything else", lol. :)  hmm, not sure my experience matches.  but certainly a leading cause.
21:12 PerlJam stevan_: in the middle of perl 6 somewhere ("you're in a maze of twisty passages, all alike" :-)
21:12 stevan_ :)
21:12 mncharity that's my model for software engineering in general
21:12 stevan_ maze of twisty passages all alike?
21:13 pmurias stevan_: i thought of OCaml before
21:13 stevan_ pmurias: and?
21:13 * [particle] uses a grue language
21:13 stevan_ its like Haskell but without all the annoying Monads
21:13 TimToady he said "it can be proven" and then fell back asleep
21:13 stevan_ :)
21:13 mncharity re "maze of twisty passages all alike?", yeah.  big maze of a heat sink, glowing white as it dissipates immense societal resources.
21:14 pmurias stevan_: i'll look it after my end of highschool exams
21:14 PerlJam jrockway: so, again, ... what's wrong with parrot's object model?  If it's "less than desirable", it should be modified to be more desirable.
21:14 * stevan_ decreases the morphine drip on TimToady's IV,,. maybe you wont nod out as often :)
21:14 TimToady I feel pretty dissipated myself at times...
21:15 pmurias s/look it/look into it/
21:15 stevan_ pmurias: SML is also worth a look, using MLTon is can be faster than OCaml
21:15 elmex there are waaaaay too many metaphors used for a 'technical' discussion about compiler/language implementation techniques.
21:15 mncharity ocaml could be a neat backend target.  just not a lot of ocaml hackers around here now, so using it rather than bootstrap+p5 seemed likely to result in  pugs-on-haskell^Hocaml - far less featureful, but similarly inactive.
21:16 stevan_ well perl5 is not a language for writing compilers in (no offense TimToady )
21:16 stevan_ but ocaml is
21:16 stevan_ and ocaml is easier to learn and use then Haskell
21:16 stevan_ again, see comment about annoying monads
21:16 mncharity so is p6 :) # a language for writing compilers in
21:16 stevan_ ocaml doesnt get in your way when you want to do something simple but impure
21:17 stevan_ mncharity: true, we should use that then :P
21:17 TimToady I'm trying...
21:17 stevan_ :)
21:17 elmex isn't the general rule of thumb: the one who does the works decides what he uses?
21:17 TimToady that's fine until you try to get someone else to do your work
21:17 stevan_ always,.. but the other rule of thumb is that everyone has an opinion as to why his choice is wrong
21:18 * stevan_ is an armchair-compiler-writer :P
21:18 stevan_ never done it, but dammit, I know the best way to do it and its not how $you are doing it
21:19 * TimToady pictures white-hot armchairs...
21:19 mncharity re impure, true, but you still hit the "I really want to just do x.  it would be so simple in another language.  i'm sorry you can't type it, but... that's really not something I want to be concerned with right now.  can I pay you to ignore it by accepting lower performance?  no?  pretty please...?"
21:20 stevan_ mncharity: yes, and p6 will solve this problem,.. once you write it :P
21:20 * stevan_ looks around,.. hmm this is very familair we have been here before
21:20 mncharity "maze of twisty little passages... all alike"
21:21 stevan_ :)
21:21 TimToady you should have dropped your dagger
21:21 mncharity and repeated revisited
21:21 stevan_ meta-circular-meta-circularity++
21:21 justatheory joined #perl6
21:22 pmurias mncharity: re elf goal you want it to be the thing the first p6 version to be written in or something which could evolve into a real compiler?
21:22 TimToady postfix:<++> fails to find the appropriate multi
21:22 stevan_ TimToady: your using the wrong compiler :P
21:23 TimToady obviously some form of Lisp, or those would be underscores instead...
21:24 mncharity re dagger, wouldn't help.  the passages are full of the detritus of dead projects.  corpses, broken tools, etc.  even if it's a memorable dagger, well, everything gets covered in bit rot too.
21:24 TimToady reverse polish Lisp maybe
21:24 stevan_ TimToady: no, thats the subtraction operators,.. I overladed 'em
21:24 [particle] everyone wants the paren
21:25 * PerlJam hands [particle] a bowl of oatmeal
21:26 armagad joined #perl6
21:26 stevan_ excuse me waiter, there are fingernails in my oatmeal
21:27 TimToady have a ] too, that'll fix everything
21:28 stevan_ but you keep hogging the colon
21:29 mncharity re elf goal, there really isn't a single elf.  it's more an incremental exercise in "we have a hack p6 implementation, that we are using to compile the p6 code for a hack p6 implementation.  what aspect of hackiness is annoying us most today.  fix it.  iterate."
21:29 armagad left #perl6
21:29 PerlJam mncharity: Hacks all the way down doesn't quite work the same as if it were turtles.
21:29 mncharity s/fix it/make it no-longer-the-most-annoying-thing/
21:30 TimToady (by breaking something else worse, in my case, ususally)
21:30 mncharity logo also has the "yet another dead p6 implementation problem".
21:31 TimToady to sleep, perchance to dream
21:31 mncharity re breaking, yeah.  so far not much in the way of regressions.  but also don't have rich p6 prelude yet.
21:31 mncharity dreams of turtles
21:31 TimToady this conversation is going all the way downhill fast
21:32 mncharity don't worry, while the shells are slippery, there's another bigger turtle underneath
21:33 PerlJam TimToady: gravity prevents us from going partway downhill.
21:33 * [particle] doesn't believe in gravity
21:34 mncharity PerlJam: more seriously, "Hacks all the way down" is indeed something to avoid.  elf is attempting to address it by not having very little down.  elf_x can always compile itself, into an executable whose only dependency is the external parser, and CPAN modules.
21:34 [particle] my god, are you at elf_x already?
21:34 pmurias elf_e
21:34 mncharity LOL
21:34 [particle] :)
21:35 TimToady STD5 is also trying to have very little down, for a very large single dose of "little"
21:36 mncharity oy, good laugh.  [particle]:  danke. :)
21:36 PerlJam TimToady: broad but shallow?
21:37 pmurias mncharity: compiling into a directory of compiled modules wouldn't be much less convinient ;)
21:38 TimToady conceptually very simple.  Just do whatever it takes to machine translate STD.pm into a perfect-enough Perl 6 compiler to compile STD.pm as is.
21:38 TimToady but that implies longest token matching works right
21:38 TimToady and all the syntax category multimethod stuff
21:38 TimToady and the derived languages
21:38 pmurias sleep&
21:39 TimToady and figuring out what a derived language is anyway..
21:39 TimToady is the inside of q/.../ a different language than the inside of q!...! ?
21:40 TimToady how does the language inside know its terminators?
21:40 TimToady along with any escapes?
21:40 PerlJam TimToady: The "story of longest token matching" strikes me as something that would make a good article for one of the perl mags (bits or dead trees both), if ever you have a hankering to write about it  :)
21:40 mncharity re "much less convinient", :)  i have trouble keeping track of bootstrap dependencies when compiling is a single step which creates a single stand-alone file.  don't have the neurons for much additional complexity.
21:41 [particle] timtoady: make it an introspectable property of the compiler
21:41 PerlJam [particle]: which compiler?  :)
21:41 PerlJam [particle]: or, how do you know which compiler to ask?
21:41 TimToady it's determined by the dynamic scope of the parse
21:41 TimToady so more like a context var
21:42 TimToady but the interaction of terminators with LTM is problematic because of that dynamic nature
21:42 TimToady you can't just recalculate different token tables because the set of terminators changed
21:42 [particle] malformed input in nested languages is difficult
21:43 mncharity /me was recently puzzling over the easiest way to get a parse of the regex.  yare, or kp6, or rakudo, oh my.
21:43 mncharity oh, or STD_red of course
22:02 aindilis joined #perl6
22:08 pugs_svnbot r20260 | putter++ | elf_d_src/EmitNoMooseP5_fromIRx1.pm: Update from EmitSimpleP5_fromIRx1.pm.
22:08 pugs_svnbot diff: http://dev.pugscode.org/changeset/20260
22:08 lambdabot Title: Changeset 20260 - Pugs - Trac
22:21 stevan_ joined #perl6
22:26 Jedai joined #perl6
22:33 pasteling "mncharity" at 71.232.7.10 pasted "a Moose error on 5.8.8. Works on 5.10.0." (16 lines, 3.3K) at http://sial.org/pbot/30791
22:34 mncharity stevan_: ^^ fyi.  both are Moose 0.40.
22:37 Limbic_Region joined #perl6
22:38 mncharity but it may result from my not using extends()
22:44 mncharity no, that's not it.  :/
22:47 wknight8111 joined #perl6
22:50 mncharity ah, ok.  now have a failure case which fails on both.  progress. :)
22:51 thestarslookdown joined #perl6
23:05 wtgee joined #perl6
23:12 mncharity stevan_: fyi... the first, isa() problem is undiagnosed.  but probably my giving bogus extends() arguments.  perhaps a leading Moose::Object.  a second issue is more easily replicated:
23:12 mncharity elf_d_src]$ perl -we '{package A;use Moose;sub f{print 3}}{package B;use Moose; extends qw(A);}A->f;B->f;'  works
23:12 mncharity elf_d_src]$ perl -we '{package A;use Moose;sub f{print 3}}{package B;use Moose; extends qw(::A);}A->f;B->f;'  doesn't work.
23:12 mncharity failing with  Can't call method "can" without a package or object reference at /usr/lib/perl5/site_perl/5.8.8/Moose/Meta/Class.pm line 278.
23:20 thoughtpolice joined #perl6
23:24 mncharity oh, and  perl -we '{package A;sub f{print 3}}{package B;@ISA=qw(::A);}A->f;B->f;'  works of course.  thus the issue.
23:29 eternaleye_ joined #perl6
23:31 thoughtpolice joined #perl6
23:44 Auzon joined #perl6

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

Perl 6 | Reference Documentation | Rakudo