Perl 6 - the future is here, just unevenly distributed

IRC log for #perl6, 2008-08-02

Perl 6 | Reference Documentation | Rakudo

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

All times shown according to UTC.

Time Nick Message
00:52 lambdabot joined #perl6
00:57 alester joined #perl6
01:04 mncharity joined #perl6
01:05 mncharity joined #perl6
01:08 meppl good night
01:26 jiing joined #perl6
01:27 hercynium joined #perl6
01:32 mncharity pmurias: a while back you mentioned you had been blocked on elf string escapes (now fixed, yes?).  are you blocked on anything else?
01:32 mncharity @tell pmurias a while back you mentioned you had been blocked on elf string escapes.  are you now blocked on anything?
01:32 lambdabot Consider it noted.
01:33 * mncharity puzzles over what goals elf should have for August'08.
01:37 mncharity @tell TimToady What is your planned direction after STD parses t/ ?  Performance improvements?  Moving back to p6?  Starting a runtime?  Wrapping it up and moving to the OO core?  Something else...??
01:37 lambdabot Consider it noted.
01:41 mncharity @tell TimToady My impression is elf is no longer providing either technical or motivational assistance for your STD work, correct?  Any thoughts on how it might be of use to you over the coming month?
01:41 lambdabot Consider it noted.
01:47 mncharity Sigh.  There seems some small interest in high performance common lisp full bootstrap.  With STD maturing, it's at least a feasible goal, but a great deal of work given the low level of interest.  There's been very slight interest in writing p6 prelude, but... very slight.
02:01 Auzon TimToady: re STD appearing accepting, it seems that it works and then fails as a result of a cronjob or svn up. p6eval's code is looking for "--- !!perl" to indicate success, though, so I don't know why it's returning ok.
02:08 mncharity moritz_: a while back you asked whether elf was "limited by parsing, or by the rest of the compiler".  I answered parsing.  Mostly because a fast authoritative-AST parser would permit close coupling with the test suite.  Real concrete TDD, rather than the incompleteness and fuzziness of recent years.
02:08 Auzon mncharity: moritz is on vacation, if you didn't know.
02:09 mncharity But the real answer is neither.  elf is limited on developer resources.  STD_red could have been transliterated into p6, so frontend was boostrapped too.  STD_red could have been synced with a more recent STD.  The prelude signatures in Snn could have been copied to a file.  the elf common lisp backend could have been fleshed out into a bootstrap.
02:09 mncharity Auzon: oh, yeah, right. :/  ah well.  thanks :)
02:10 Auzon I imagine he'll backlog and grep. And it's interesting (to me at least)
02:12 mncharity punchline is, even things which would take no great skill, little time, and are repeatedly mentioned, arent't happening.  No social momentum.  That's what elf is limited on.
02:12 mncharity Auzon: re backlog, *nod*.  re interesting, :)
02:12 mncharity questions/comments welcome/encouraged.
02:17 Auzon I think it's interesting how you mention the parser being a crucial part. I've found that STD.pm is very helpful in writing tests, and while I don't know any specifics, I can see how it'd help in a number of cases. Would STD.pm be a significant improvement over STD_red?
02:22 mncharity re STD.pm vs STD_red, well, 100% vs ~70% coverage.  but...  much of the value I think is authoritativeness.  think how much harder your work would be if the t/ tests were "parrot has one set of t/, and pugs another, and elf, and...", instead of "this is the current, authoritative, best understanding of what the language is".  having a reference parser,
02:24 mncharity regardless of whether you can easily use it, moves the "and it's all a hack" boundary from between source and AST to between AST and IR.
02:25 Auzon What's your current goal with elf? It's being used on SMOP, correct?
02:26 mncharity re SMOP, I believe so.  re goal, unclear.  thus this evening's exercise.
02:29 mncharity for most of its existence, the goal might be described as "help keep STD rolling" and "see whether once people can work on compiler internals in a fast big p6 dialect, they do".  STD rolled, and they didn't.
02:31 Schwern joined #perl6
02:31 mncharity an initial hypothesis was that if an unfinished implementation with a solid foundation/walls could be constructed, people would be motivated to flesh it out.
02:32 Auzon It was somewhat attractive to me, but I had other obligations. I don't know how many free developers you could've attracted :-/
02:34 mncharity eg, say a STD parser written in p6, with a compiler written in p6, with no performance problems (eg, a fast common lisp runtime), and solid real oo.  so you could say "the foundational oo is there.  the tests for full p6 are there.  you're getting their authoritative AST's.  the development pragmatics are good.  you can work in p6.  Christmas will be here as soon as you write enough p6."
02:36 s1n is it possible to verbose test output from using localtest (without having to execute it by hand)?
02:36 mncharity re how many free developers, don't know.  there seems a background reservoir of
02:37 mncharity s1n: err, testing what against what?
02:37 s1n mncharity: using the harness and the localtest make target
02:37 Auzon Rakudo against something in t/spec.
02:37 apeiron_ joined #perl6
02:38 mncharity s1n: ah, sorry, don't know.
02:38 s1n Auzon: yeah, but i'd like to do something like "make localtest" and have the harness spit out each test result
02:38 s1n i couldn't see anything and I don't know enough about parrot's Test.pm to change it to not capture the test output
02:38 s1n or change it to capture and print it as well (i.e. verbose mode)
02:39 Auzon I usually don't use the harness which is quite verbose.
02:39 s1n yeah but i want to see each of the tests results
02:39 s1n all of the passes and where fails occur
02:40 Auzon try: ../../parrot perl6.pbc `t/spec/fudge rakudo t/spec/whatever/foo.t`
02:40 Auzon That assumes you have a bash-like shell that supports backticks
02:40 s1n i could do this all by hand (such as your example), but i was looking for a shortcut
02:41 s1n well, there's "../../parrot ./perl6.pbc --tests-from-file=t/localtest.data" which is easier, but i'm looking for something short and sweet
02:42 s1n err well, that doesn't do the fudging, but you get the idea
02:42 mncharity there seems a background reservoir of interest.  people occasionally wander by here, before being scared off by the lack of project management.  kind of hard to judge the reservoir level in the absence of probes.  a hypothesis is if we had something to give them which actually worked, and could be easily developed, they would come.  "had an active development community" might be
02:43 mncharity another requirement there, and thus a community bootstrap issue.
02:43 kanru joined #perl6
02:43 Auzon Hm, sorry s1n, I can't think of anything else. I must go &
02:44 s1n Auzon: thanks anyways
02:44 s1n Auzon: i'll probably just write a make target for it :)
02:46 mncharity but my previous hypothesis was "people worked on kp6, before getting stuck.  given a faster better less sticky version of kp6, people would work on that".  but didn't.  modulo pmurias.  so... not sure I believe the hypothesis.  not enough to burn a couple of weeks to test it, in the absence of anyone else believing it.
02:47 s1n mncharity: rakudo isn't so hard to get started with, it just has a fairly different build system than i'm used to, and i had to adjust to the harness idea. all of this could be improved if the interface to these operations (changing, building, testing) were simplified and documented
02:57 mncharity s1n: thanks for the thought.  the rakudo folks tend to be bottlenecked on "someone getting around to it" too, so you might write up your "what I wish *I* had seen documented, and would have made my life easier", and send it to them.
03:00 avar The people who work on rakudo are also the guys who wrote the subsystems it uses so they understand all the pir stuff already:)
03:01 mncharity avar, you worked on kp6 at one point, didn't you?
03:02 avar I tried:)
03:03 elmex_ joined #perl6
03:03 avar It was really fglock's project though and he disappeared in a black hole or something
03:03 mncharity hmm... so let's see...
03:05 mncharity pretty mountains, cool breeze, someone mentions a new p6 "xmas sled" project.  describes it's attributes.  you're excited to work on it.  ...  what characteristics does it have? :)
03:08 mncharity or alternately, did you ever look at elf?  if not, any particular reason?  if so, any things you liked/disliked come to mind?  any thoughts on "if it had only been/not-been X, I might have been psyced"?
03:09 mncharity "The code was unreadable.  The performance glacial.  It wasn't really p6.  Yet another soon-to-be-abandonware p6 project. ...?"
03:11 mncharity "Didn't believe the architecture.  Completely unnecessary given rakudo.  Given STD.  Given kp6.  Don't really believe in p6 anymore. ...?"
03:11 mncharity (have run into "Don't really believe in p6 anymore" a couple of times now :/ )
03:12 alanhaggai_ joined #perl6
03:13 hercynium Do you believe in p6? … If you believe clap your hands.
03:14 hercynium Every time a hacker says, 'I don't believe in p6,' there is a unit test somewhere that fails for no good reason.
03:14 mncharity ...wind whispers though the seemingly abandoned city...
03:14 mncharity lol
03:14 hercynium something about your statement made me think of tinkerbell
03:15 mncharity noticed ;)
03:15 avar mncharity: I haven't really followed perl6 at all since I stopped hacking on kp6 other than compiling parrot's perl6 every ~3-6 months
03:15 avar I looked at elf once and didn't understand it:)
03:15 mncharity :)
03:16 mncharity so "The barrier to entry on elf is too high."
03:18 avar "avar was impatient that it wasn't a full perl6 implementation and probably proceeded to watch some crappy TV show"
03:18 mncharity *lol*
03:19 avar But I don't think I'd work on any perl6 project right now, it's all a bit pie-in-the sky
03:19 avar Hopefully parrot will have something at least as useful as raw scheme for writing programs in 1-2 years
03:20 avar Or someone starts hacking pugs again
03:20 avar But well, stuff like kp6 and elf is cool also because it shifts the compiler problem elsewhere which is nice
03:23 mncharity re sky pie, sigh, indeed lots of work needed.  perhaps not that much time with lots of people, but lots of time with current population.
03:27 mncharity so "I can be using p6 in production WHEN?!?  In only <startlingly soon> n weeks?!?  It's no longer pie-in-the-sky?!?" might be the key characteristic of interested.
03:31 mncharity for myself, I most often feel the need for p6 when I have to write something in multiple languages, eg js+python+ruby, and would much rather write it once in p6, and emit the others from there.  also when I have to do ugly architectures because I don't have high performance multimethods.  or macros.
03:33 mncharity so my "p6 is real for me when.."s might be "when I can write simple libraries in p6, and have usable from js/py/rb" and "when I can write a smallish low order 10k line programs, and use multimethods with performance more like CL than python".
03:34 mncharity anyone have their own?
03:38 mncharity Auzon: re role of STD.pm, note the "it wasn't a full perl6 implementation".  having it "it's fully real from the top down to there, and from the bottom up to here, ... it's just the compiler and prelude which need more p6 written"... might, hypothetically, who knows, be more attractive.
03:41 mncharity ok.  elf goals for 2008.08, "make elf more approachable", and... whatever comes from having asked for suggestions/requests.
03:42 mncharity Auzon: avar: s1n: thanks for your help.
03:55 Auzon @tell mncharity re "I can use p6 in $small_unit_of_time", I had some time during this summer when I wanted to make an IRC bot. I'm not a huge fan of the existing p5 modules for this (either Net::IRC or Bot::BasicBot/::Pluggable), so I would've used p6, but it had no sockets. Elf supported sockets via p5, but then I may as well have used p5. (Not sure how the p5-to-p6 interfacing worked. It may have been OK.)
03:55 lambdabot Consider it noted.
03:57 Auzon @tell mncharity For me, wanting to use p6 in a project was with IRC (and required sockets as a minimum). I've heard file IO as another minimum, but I'm really not sure what a lot of people consider as the minimum requirement.
03:57 lambdabot Consider it noted.
04:18 s1n Auzon: if it's accepted, you'll be able to do a "make localtest_loud" (see rt#57522)
04:19 alester joined #perl6
04:20 Auzon s1n++ # improving the toolchain
04:23 alanhaggai__ joined #perl6
04:25 Lorn joined #perl6
04:31 josh joined #perl6
04:31 pugs_svn r21719 | lwall++ | [STD] detect obsolete use of =cut
04:31 pugs_svn r21719 | lwall++ | [t/*.t] delete obsolete uses of =cut
04:38 pugs_svn r21720 | lwall++ | [t/*] s/:perl5/:Perl5/
04:42 pugs_svn r21721 | lwall++ | [t/*] s/qn/Q/
05:00 pugs_svn r21722 | lwall++ | [t/*] fix all the rands and $*PID."... instances
05:02 pugs_svn r21723 | lwall++ | [t/xx-uncategorized/hyperop.t] postfix hyper can't have space before it
05:31 audreyt avar: yes.
05:33 gaal hey audreyt. my build is... too fast :) and almost all smoke tests were timing out
05:33 gaal 'time make' is wel under a minute and only two files are being compiled?
05:35 gaal after make clean, only Pugs.Config and Pugs.Parser[boot] seem to require building
05:35 gaal ooh, dist/build is full of objects!
05:37 clkao gaal!
05:38 gaal hey :)
05:38 clkao has my friend landed tel aviv yet ;)
05:39 gaal I think not -- last week I was told sunday or Monday
05:49 Psyche^ joined #perl6
05:54 gaal hmmm. many tests still timing out (e.g. Cypher
05:54 gaal )
05:59 audreyt gaal: yes please hack make clean to run ./Setup clean
05:59 audreyt avar: yes that is the error I'm getting
06:08 ilogger2_ joined #perl6
06:16 clkao audreyt: http://pastebin.com/m26c2f03e
06:16 buubot clkao: The paste m26c2f03e has been copied to http://erxz.com/pb/11160
06:16 clkao doesn't quite build
06:18 audreyt clkao: it's in http://svn.pugscode.org/pugs/src/Pugs/CodeGen/PIR/Prelude.hs though
06:18 clkao what should i do then? i am just running make
06:19 audreyt committing
06:20 pugs_svn r21724 | audreyt++ | * Try resolve the "file already exists" bug
06:21 audreyt ok try r21725.
06:21 clkao Could not find module `Pugs.CodeGen.PIRPrelude':
06:21 clkao but you had:
06:21 clkao A   src/Pugs/CodeGen/PIR_Prelude.hs
06:21 clkao does that seem right to you?
06:22 clkao (sorry, i've been watching firefly)
06:22 audreyt you're one rev behine
06:22 audreyt behind
06:22 audreyt need 21725
06:22 pugs_svn r21725 | audreyt++ | * Rename PIR.Prelude to PIR_Prelude
06:22 clkao tada
06:22 clkao compiling is so fast
06:23 audreyt science have progressed
06:23 clkao indeed
06:23 clkao http://xkcd.com/54/
06:23 lambdabot Title: xkcd - A webcomic of romance, sarcasm, math, and language - By Randall Munroe
06:24 clkao gcc-4.0: -E, -S, -save-temps and -M options are not allowed with multiple -arch flags
06:25 clkao i think the universal binary flag from perl extutil is messing things up
06:26 audreyt I thought I fixed it.
06:26 audreyt did you see -arch in Pugs.buildinfo?
06:26 audreyt actually, you reported it months ago
06:26 clkao -optl-arch -optli386 -optl-arch -optlppc
06:27 clkao yeah i didi
06:27 audreyt fixed
06:27 audreyt up and make again should work
06:27 clkao does it read the build info during make? or do i have to reconfigure?
06:27 pugs_svn r21726 | audreyt++ | * Remove -arch from optl as well.
06:28 clkao oh worky now
06:28 audreyt configure is part of make
06:28 clkao so you can't build universal binary with ghc then.
06:28 clkao does that seem right to you?
06:28 audreyt cabal configure is so fast, it's okay to do it over and over again
06:28 clkao mtl:pugs clkao$ time ./pugs -e 'say 1'
06:28 clkao 1
06:28 clkao real    0m0.201s
06:28 audreyt that seems right to me.
06:28 * clkao is now happ
06:28 audreyt there's a ghc ticket for it
06:28 clkao happy, even
06:29 clkao i think someone needs to refactor the mad htmlmatrix smoke thing with the new harness infrastructure
06:29 pugs_svn r21727 | audreyt++ | * Just to be sure, nuke -optc-arch as well
06:30 clkao how do i smoke just the core tests? it's failing CGI crap
06:31 audreyt env PUGS_SMOKE_EXCLUDE_EXT=1 make smoke
06:32 clkao is inc etc not working?
06:32 clkao t/01-sanity/06-use.t
06:32 clkao pugs: *** Can't locate lib.pm in @*INC (@*INC contains: blib6/lib /System/Librar
06:32 p6eval pugs: OUTPUT[*** ␤    Unexpected "***"␤    expecting program␤    at /tmp/yXLp1NJU0m line 1, column 1␤]
06:32 audreyt did you run "make"
06:32 audreyt or jsut "make pugs"
06:33 audreyt oh nvm.
06:33 audreyt mea culpa. sec
06:33 pugs_svn r21728 | audreyt++ | * Build ext/.
06:34 audreyt there.
06:34 audreyt (make ; make smoke)
06:34 clkao well but this was 06-use.t
06:34 clkao not ext
06:34 clkao uh, lib is unde ext/ too?
06:35 clkao worky now
06:35 clkao now let's just see how well the smoke is
06:39 pugs_svn r21729 | gaal++ | * "make clean": wipe dist/ too by having it call ./Setup clean
06:43 clkao t/examples/all_parse.t
06:43 clkao 304 wallclock secs ( 0.00 usr +  0.03 sys =  0.03 CPU) @ 33.33/s (n=1)
06:43 clkao this test is slow
06:44 audreyt *nod* that's because the parser is slow
06:48 audreyt TimToady: will a longest-token re::engine::TDFA be of any interest/help to you?
06:52 clkao does make smoke submit to the right place?
06:55 audreyt you need to say "make upload-smoke"
07:07 alanhaggai joined #perl6
07:08 clkao yeah, but is the smoke aggregation server still running
07:15 audreyt sure is
07:28 meppl joined #perl6
07:33 pugs_svn r21730 | gaal++ | * Makefile.PL: don't use literal tabs to construct the makefile.
07:36 gaal most of the examples and t/junction are timing out. is it just me?
07:44 meteorjay joined #perl6
07:53 fridim_ joined #perl6
08:01 hcchien joined #perl6
08:09 audreyt gaal: t/junction worksforme
08:09 audreyt gaal: are you still on colinux?
08:13 alanhaggai joined #perl6
08:36 pmurias joined #perl6
08:54 * zamanfou is now away: In the rain my tears are forever lost.
09:58 pjcj joined #perl6
10:10 masak joined #perl6
10:13 pugs_svn r21731 | pmurias++ | [smop] start of m0ld, the DSL for Mold
10:19 masak 'make' works now!
10:19 masak audreyt++
11:10 BinGOs joined #perl6
11:25 pugs_svn r21732 | fglock++ | [v6.pm] more tests  # 7100
11:39 BinGOs joined #perl6
11:59 cognominal_ joined #perl6
12:13 sri_work joined #perl6
12:50 meppl joined #perl6
13:00 BinGOs joined #perl6
13:07 BinGOs joined #perl6
13:12 wknight8111 joined #perl6
14:02 meppl joined #perl6
14:09 masak a lot of the modules in ext/ time out during make smoke
14:28 alester joined #perl6
14:46 pugs_svn r21733 | pmurias++ | [smop] start of Mold implementation
14:47 pugs_svn r21734 | lwall++ | various buglets, STD now parses 94% of t/
14:50 gaal audreyt: no, real ubuntu hardy
14:50 gaal eep, and all of t/operator is timing out too
14:51 gaal basically my smoke from ~8hrs ago is still going
14:52 gaal ghc-6.8.2
14:53 gaal hmm, what would be a good way to add profiling to pugs?
15:01 pmurias @tell ruoso what do you think of turning has_next,next and eval into next only?
15:01 lambdabot Consider it noted.
15:33 justatheory joined #perl6
15:33 TimToady my make test seems to have hung in examples/algorithms/lambda-calculus.pl
15:33 lambdabot TimToady: You have 2 new messages. '/msg lambdabot @messages' to read them.
15:33 TimToady @messages
15:33 lambdabot mncharity said 13h 56m 14s ago: What is your planned direction after STD parses t/ ? Performance improvements? Moving back to p6? Starting a runtime? Wrapping it up and moving to the OO core?
15:33 lambdabot Something else...??
15:33 lambdabot mncharity said 13h 51m 45s ago: My impression is elf is no longer providing either technical or motivational assistance for your STD work, correct? Any thoughts on how it might be of use to you over
15:33 lambdabot the coming month?
15:36 TimToady audreyt: I'm not really interested in substituting one buggy tagged dfa implementation for another :)
15:39 TimToady @tell mncharity after t/ I'm probably going to be working on mutating my TRE-prevention code into a real dfa matcher, then translate that to P6.
15:39 lambdabot Consider it noted.
15:41 alester joined #perl6
15:41 TimToady @tell mncharity but to get t/ parsing right involves adding symbol tables/pads to STD to implement user-defined operators and post-declarations
15:41 lambdabot Consider it noted.
15:42 jhorwitz joined #perl6
15:43 pmurias TimToady: how will be the lexicality of grammar modification handled?
15:43 lambdabot pmurias: You have 1 new message. '/msg lambdabot @messages' to read it.
15:43 TimToady @tell mncharity and of course, performance improvements are always in the back of my mind in some kind of repressed state, and translating Cursor to p6, and prelude, and...
15:43 lambdabot Consider it noted.
15:45 TimToady it's handled automatically by the fact that cursor_fresh starts a new language and _SUBSUME throws it away at the end of the construct, but of course that begs the question of how the new grammar represents its lexical symbols to itself that aren't new rules
15:47 pmurias but _SUBSUME has to propagate grammar changes or they won't get far
15:47 TimToady sub postfix:<!> needs a new rule, whereas my sub foo just needs a new lexpad entry
15:47 pmurias far meaning to the next statement
15:48 TimToady _SUBSUME automatically propagates the language it entered with onward
15:49 pmurias if i have a sub infix:<more_than_plus> {...};1 more_than_plus 2;
15:51 TimToady the add_macro has to figure out which language to add it to, really
15:51 pmurias TimToady: not create a new one?
15:51 TimToady at the moment it's assuming it's adding to $¢ but that might not propagate out of the current construct
15:52 pmurias it dosn't
15:52 TimToady so it probably needs to tell some higher SUBSUME to switch to the new language, maybe more than one
15:53 pmurias by changing SUBSUME, STARr and PLUSr i got it as far as too the noun
15:53 pmurias s/too/to
15:53 TimToady it probably needs to change any SUBSUME up to the statement level
15:54 TimToady since it needs to influence the rest of the declaration as well as the next statement
15:54 pmurias the autolexer blocks the propagation at noun, so it needs to be changed too
15:55 pmurias i could put a path some where which propagates a hardcoded infix:<more_than_plus>
15:57 TimToady I wonder if we could fake it by failing and reparsing
15:57 TimToady that is, fail back to the statement level, add the symbol to the language, and reparse.
15:58 TimToady probably not workable for deeply embedded declarations
16:00 pmurias the grammar changed could be effective from the next statement
16:00 TimToady anyway, I will think about how to do that, since we need it for the one failing test in t/spec
16:00 macroron joined #perl6
16:00 TimToady that's what p5 does, but that's got its own problems
16:01 TimToady which is why p6 tries to consistently introduce it right at the name, even if the definition isn't complete
16:03 TimToady but it's basically a situation that we're adding a method to a base class and expecting the autolexer to cope, I think
16:03 TimToady (which it doesn't currently)
16:05 TimToady it may be as simple as invalidating all the lexers derived from the current statement language and letting the autolexer recalculate (hopefully lazily)
16:10 pugs_svn r21735 | pmurias++ | [STD] fixed bug with autolexer hack in derived grammars
16:15 pmurias s/changed/change/
16:18 TimToady anyway, there can be more than one "next" statement when the declaring construct contains a block, and we'd like to be able to express recursive ideas, like using postfix:<!> in its own definition
16:19 TimToady a related interesting problem is that if we introduce infix:<foo> right when we see the name, we don't yet know its precedence until the traits are parsed
16:20 TimToady so the natural place to introduce a new symbol may be after its traits
16:20 iblechbot joined #perl6
16:20 TimToady this works for parameters in sigs too, since an earlier parameter is visible to a later parameter's default expression
16:21 TimToady and parameters can have traits
16:23 TimToady this is probably all magically orchestrated by the COMPILING scope somehow, since that represents the current lexpad, so maybe it keeps an autolexer valid bit somewhere that sublanguages can check
16:25 TimToady at some point we'll probably also have to think about how different autolexers can share common elements so that we don't have to regenerate everything every time we mixin a new symbol
16:26 TimToady using polymorphic tries or some such
16:27 TimToady at the moment I'm just hoping most mumble:<x> definitions get added without being used soon, so we don't have to autolex the intermediate languages
16:28 TimToady and I'm also banking on the fact that normal sub defs don't really change the language, but just add to the symbol table accessed by a \w+ match
16:28 TimToady (which is why we can have postdeclared identifiers, but not post-declared operators)
16:42 TimToady but after thinking a bit longer, the correct solution is probably to just provide a way for SUBSUME to have some way of merging interior declarations with the outer language, if there's some sane way to represent the current set of pending declarations
16:44 TimToady the list of pending declarations could even be used as an exception list of some sort in the current sublanguage, so the sublanguage could decide on its own whether to go to the trouble of autolexing a grammar that will soon be thrown away
16:44 TimToady but maybe this is isomorphic to tweaking the outer grammar in a way that is visible to inner grammars
16:45 TimToady in any case, it is an *interesting* problem  :)
16:48 pmurias &
16:50 gaal TimToady: is your smoke still running, and timing out on many subsequent tests?
16:55 jjore joined #perl6
17:25 TimToady I thought it would continue when I ^C'd the hanging test, but it stopped everything, and I didn't restart.  next time I'll try kill
17:25 TimToady and it wasn't a smoke, but a test
17:42 TimToady starting a real smoke to see how it does
17:44 TimToady ext/Cipher/t/0-cipher-oo.t seems to loop
17:45 ruoso joined #perl6
18:00 rhr I just ran a smoke, and nothing seemed to hang here
18:01 rhr however everything in t/regex/from_perl6_rules fails with ~/.cabal: createDirectory: already exists (File exists)
18:10 masak rhr: no five-minute pauzes in almost all of ext/ ?
18:11 rhr masak: the whole smoke completed in 4:59 :)
18:11 masak lucky you :)
18:12 gaal that's just a spurious warning (which I'm in the middle of fixing), not an error
18:21 araujo joined #perl6
18:28 gaal hee, looks like I was wrong and so was 9630
18:29 gaal createDirectory is failing with a alreadyExists exception, killing the caller (in this case, doMatch)
18:30 gaal unfortunately, the paths are wrong, so switching to createDirectoryIfMissing True $ foldl1 (</>) [_cabal, "share", "Pugs-" ++ versnum] only yields a perl error
18:30 gaal gotta go now :(
18:39 pmurias joined #perl6
18:40 pmurias ruoso: hi
18:40 ruoso hi pmurias
18:40 lambdabot ruoso: You have 1 new message. '/msg lambdabot @messages' to read it.
18:42 pmurias ruoso: what do you think about changing that (see @messages)?
18:43 ruoso pmurias, it might be ok... I would keep the "eval" name for that tho
18:43 ruoso but...
18:43 ruoso how do you know when to stop the loop?
18:43 ruoso has_next is still needed, isn't it
18:43 pmurias when the eval returns false
18:43 ruoso ?
18:44 ruoso what if the expected return from the eval is false
18:45 pmurias i don't understand
18:46 ruoso eval returns the value returned by that step
18:46 pmurias not in SLIME
18:47 meppl joined #perl6
18:47 ruoso hmm... indeed
18:47 pmurias and i think we should abstract away the idea of step so we don't impose a particular implementation on interpreters
18:49 ruoso that's ok... as long as we always return to the runloop instead of recursing
18:49 ruoso and also that any object may "goto" another interpreter at any time
18:51 alanhaggai joined #perl6
18:59 pugs_svn r21736 | ruoso++ | [smop] sm0p.pl now uses non-blocking IO and no longer freezes waiting for a return of the compiler.
18:59 ruoso pmurias, test/22 no longer freezes
19:01 pmurias ruoso: why did it block?
19:02 ruoso no idea ;)... but making it non-blocking solved ;)
19:05 pugs_svn r21737 | ruoso++ | [smop] test/22 passes :) :) :) :) :)
19:08 ruoso btw... that means we have attributes already
19:11 pmurias what will we do with/to slime once mold is ready?
19:11 jferrero joined #perl6
19:11 ruoso throw it away? :)
19:11 ruoso slime was meant to be replaced
19:12 pmurias good
19:18 pmurias should constants in mold be virtual ro registers or be just used for setting defaults to registers
19:18 pmurias ?
19:19 Lorn joined #perl6
19:23 pmurias constants here meaning  low level constants like 44 not high-levl constant $foo = 4
19:23 pmurias * high-level
19:48 pmichaud Aloha
19:50 pmurias pmichaud: hi
19:54 xinming joined #perl6
20:18 japhb joined #perl6
20:29 jan__ joined #perl6
20:43 cathyal joined #perl6
20:50 pugs_svn r21738 | pmichaud++ | S29-trig/trig.t:  remove skip marker from atan2 tests.
21:07 pmurias c--
21:30 lambdabot joined #perl6
21:30 ruoso joined #perl6
21:31 * ruoso back
21:35 pmurias ruoso: do the frame receive correct interpreters to their methods?
21:36 ruoso yes
21:37 pmurias are you aware that test 09 fails sometimes?
21:39 pugs_svn r21739 | pmurias++ | [smop] added Mold.Frame, it's currently a battleground with segfaults but comit early, commit often...
21:39 gaal @tell audreyt I think getDataFileName isn't working: or rather, that the modules looked for there aren't being found. this is causing RE matches to fail
21:39 lambdabot Consider it noted.
21:40 explorer__ joined #perl6
21:41 gaal @tell audreyt also, the particular path is wrong; we create a cabal_ </> Pugs-... but return _cabal </> "Pugs-" ... </> fn
21:41 lambdabot Consider it noted.
21:41 gaal @tell audreyt anyway, my ~/.cabal has none of the actual files we want.
21:41 lambdabot Consider it noted.
21:45 ruoso pmurias, I am, but I have not a single clue on why it does...
21:45 * ruoso dinner
22:29 hercynium joined #perl6
22:45 Alias__ joined #perl6
23:19 ruoso pugs: module Foo { my &foo; sub foo { }; };
23:19 p6eval pugs: RESULT[\:() {}{Prim ([Pugs.AST.Internals.Val] -> Pugs.AST.Eval.Eval Pugs.AST.Internals.Val)}]
23:19 ruoso pugs: module Foo { my &foo = sub { 1 }; sub foo { 2 }; }; say Foo::foo();
23:19 p6eval pugs: OUTPUT[*** Can't modify constant item: VStr "MkCode {isMulti = True, subName = \"&\", subType = SubPrim, subOuterPads = [], subInnerPad = MkPad (padToList []), subPackage = , subAssoc = AIrrelevantToParsing, subParams = [], subBindings = [], subSlurpLimit = [], subReturns = (mkType \"An...
23:21 ruoso do Package hold different storages for the subs and global variables?
23:22 hercynium joined #perl6
23:23 ruoso or is that just like p5, where package subs are just a special type of package variable
23:50 ruoso joined #perl6

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

Perl 6 | Reference Documentation | Rakudo