The Return of the Journal : pugs.blogs.com/ | pugscode.org | pugs.kwiki.org | paste: paste.lisp.org/new/perl6 or sial.org/pbot/perl6
Set by GammaRay on 31 December 2005.
00:07 frederico joined
rafl dduncan: Some things in parrot changed. We should make sure embedding works with both a parrot sandbox and an installed parrot. 00:12
00:16 PJF joined 00:28 scook0 joined 00:49 stennie joined 01:04 mjl69 joined 01:05 joepurl joined 01:18 joepurl joined 01:54 DaGo joined
stevan dduncan: I wouldnt worry about the release at this point 01:58
I think the focus is more on getting to 6.28.0 right now 01:59
which should happen by the end of January at the latest
01:59 joepurl joined
dduncan does that mean there will not be a 6.2.11? 02:13
stevan dduncan: I dunno 02:17
02:17 gaal_ joined
stevan audrey should be up in a few hours,.. you can ask her :) 02:17
my point was more that if we did have 6.2.11, that 6.28.0 would probably be out not too much longer after that 02:18
and as far as your work is concerned with it's heavy use of OO features,.. 6.28.0 is probably more interesting 02:19
dduncan I'm not concerned about having anything working before 6.28.0 ...
stevan ok, then what are you looking to get into 6.2.11?? 02:20
dduncan rather, if a 6.2.11 went out, I wanted to get some of the smaller but significant changes in, like file renames and directory reorganizations
stevan dduncan: well that is up to you then
dduncan eg, all the SQL/Routine is being renamed into Rosetta/*
stevan I am not sure when that will be
dduncan yes
stevan renames are quick things :)
dduncan yes 02:21
but have a big affect on interversion diffs
stevan dduncan: not as much in svn as in CVS,.. or are you refereing to the ones on CPAN? 02:23
search.cpan.org that is
dduncan yes, cpan
stevan ah, well it's up do you,.. I think it is probaly not something you shoudl worry that much over 02:24
dduncan just my perfectionism
if I know when "feature freeze" is, I can time things for before or after
stevan I hear yah 02:25
02:36 gaal joined
rep www.imdb.com/title/tt0427340/ 02:44
dduncan rep, so are you going to see that? 03:03
rep heh, no way
dduncan should work well as a double-bill with imdb.com/title/tt0418279/ 03:04
those were both based on toy lines that sold to the same crowd when I was young, in the '80s
stevan no way,.. transformers would so kick He-Man's ass
dduncan and had cartoons on around the same time
I saw them both 03:05
still, besides nostalgia for twentysomethings, I'm not sure these have anything going for them
stevan Optimus Prime would squash Castle Greyskull
dduncan but John Glover seems to be a good actor 03:06
or at least memorable
stevan and there is no comparing the Decpticons to that looser with the skull face
stevan can't recall who he is :P
dduncan memorable from Back to the Future and Smallville
in BTTF, he was George McFly
stevan remembers now 03:07
dduncan in Smallville, big daddy Luthor
at least thats what my memory says
BTTF was my favorite film series in the '80s
I, at least, hope the Transformers movie doesn't call the leads "Autobots" 03:08
they should use the names from other versions
stevan what else would they call them?
dduncan like Cybertronians
or maybe Maximals 03:09
Cybertronians is better though
stevan does not recall those versions
dduncan as I recall, the or an original Transformers from Japan had the two main factions from different planets 03:10
the leads were from Cybertron, and the adversaries from Destron
so they were Cybertronians and Destrons
by contrast, Autobots and Decepticons sounds rather dumbed down
and the first also doesn't fit since a lot of the characters weren't vehicles ... and the vehicles thing is more of an earth adaption than how they were at home 03:11
a lot of good ideas came out of the Mainframe Entertainment revival of the franchise in the late '90s 03:12
in that case, it showed that the transformers, upon landing on a planet, scanned the area for indigineous things and based their transformations on something similar
which is a reasonable explanation for their looking like earth things, though they weren't from a place anything like earth 03:13
fyi, the Mainframe revival was called Beast Wars
all CGI and stuff
dduncan despite all that Transformers trivia, is not actually that much a fan of the franchise 03:15
but its better than He-Man
dduncan I meant to say it that way
dduncan this way
stevan :) 03:16
03:19 hcart1 joined 03:20 hcart1 left
stevan so I have a question for the lazy-programmers in the house 03:41
If I have a lazy sequence,.. I define it as an initial start value, and then a function which given the initial value computes the next item in the list 03:42
then say I want to C<map> over this list 03:43
oh wait,.. lemme make one other point before we get to map 03:44
this is an immutable structure,.. so head() returns the start value and tail() returns another lazy sequence with the computed next item as the "initial value" in it 03:45
now to map over this
I can keep it lazy by just wrapping my "compute next value" function with the function I wish to map over my list with 03:46
but this all falls apart if my "compute next value" function has a stoping clause in it
aka- it is not an infinite list 03:47
s/aka/i.e./
the reason this breaks is because the stop-clause is now acting against the value which already has the map function applied 03:48
so for instance
(1 .. 5) would look like this:
LazySeq.new(start_value => 1, next_value => sub $current { if $current == 5 { bool::false } else { $current + 1 } } ); 03:49
but if I were to do this: map { $_ *2 } (1 .. 5) the stoping clause would break cause it would never be reached 03:51
because $current would never equal 5
svnbot6 r8612 | Darren_Duncan++ | r1842@Darren-Duncans-Computer: darrenduncan | 2006-01-08 19:30:30 -0800 03:58
r8612 | Darren_Duncan++ | /ext/Rosetta-Incubator : rename of SQL::Routine[|::*] out of 'SQL', part 1
r8613 | Darren_Duncan++ | r1843@Darren-Duncans-Computer: darrenduncan | 2006-01-08 19:36:45 -0800
r8613 | Darren_Duncan++ | /ext/Rosetta-Incubator : rename of SQL::Routine[|::*] out of 'SQL', part 2
04:21 Cryptic_K joined 04:39 Amnesiac joined 04:45 mjl69 joined 04:57 joepurl joined
audreyt yo 05:08
dduncan: there's no feature freeze in pugs development
dduncan I wasn't being literal 05:09
I just prefer that a cpan tarball isn't made in the middle of a multi-commit update
audreyt and yes, I'll look at 6.2.11, most notably new parrot targetting and PIL cleanup, today
nodnod... you have 48+ hours
should be quite sufficient for multicommit updates
dduncan right now I'm making the main changes I want in prior to that 05:10
audreyt *nod*
in any case... pugs won't release until you are happy with the releas
dduncan thanks for the time frame
geoffb Just read Audrey's journal about the syck-athon ... y'all are doing some damn fine work, it seems.
audreyt (actually, pugs won't release when anyone here is unhappy :))
geoffb: thanks :) I just released another version. 05:11
and clkao did do a YAML-RPC
dduncan fyi, unless it was fixed recently, one of the main issues I'm aware of regards private methods and/or other subs ... calling them fails in a lot of ext/ modules ... or did last week
geoffb go Audrey!
audreyt but today seems more PIL and JavaScript and PIR oriented so far
geoffb and go clkao
audreyt dduncan: sure, I'll look at it after PIR/JS/PIL
dduncan this leads to a lot of the deaths in smoking ext/
audreyt clkao mentioned that he did not really do YAML-RPC -- they were just mocking my journal. heh. ;) 05:12
geoffb What are the states of the backends at this point? 05:13
heh
svnbot6 r8614 | Darren_Duncan++ | r1848@Darren-Duncans-Computer: darrenduncan | 2006-01-08 21:15:46 -0800 05:17
r8614 | Darren_Duncan++ | /ext/Rosetta-Incubator : rename of SQL::Routine[|::*] out of 'SQL', near end of main work
audreyt JS is still functioning, modulo an array.method bug that clkao wants to see fixed.
PIR is thoroughly broken due to the lexpad.
P5 is in the process of migrating to PILN.
stevan audreyt: morning 05:18
audreyt I don't think anyone is maintaining PIL1-based P5 backend
stevan: hey!
stevan: we need to settle some terminology
stevan audreyt: ok
geoffb With the little time I've had to look, I understood PILN == the bottom layer that the backend needs to support, and then stuff will Just Work (tm). Is that correct? 05:19
audreyt geoffb: yes
stevan: Tuple is fully evaluated fixed size immutable number of things
geoffb Cool
stevan yes maam
geoffb Morning, stevan
stevan morning geoffb
although IIRC, you have enough 3 hours to go until morning :) 05:20
geoffb Heh. I tend to greet people in their timezone, if I can manage it. :-)
stevan geoffb: I am only 19 minutes into my monday morning here 05:21
geoffb East coast NA, right?
stevan yes sir
audreyt stevan: Range is an iterator with a starting, iterating, and stopping condition
dduncan it's 9:21pm here
geoffb dduncan: western Canada, as I recall, yes?
audreyt stevan: it's concrete range, not continuous range
dduncan correct 05:22
stevan audreyt: yes (but I have some questions related to that, but we can do those later)
concrete?
geoffb still retains a few brain cells, shockingly. :-)
audreyt 1..10 doesn't contain pi
stevan yes
audreyt er, I think I mean discrete
geoffb audreyt, nod
stevan "range which is easily generated with simple tools" 05:23
like addition
audreyt or multiplication.
stevan yes
audreyt I'm not sure whether we want to fix on addition
seems a bit pointless
stevan what is a continuous range then
audreyt a continuous range is something we don't have to worry about because it's not part of spec. 05:24
:D
stevan :)
stevan sets the ignore bit
audreyt now, Seq is I think a Tuple of Tuples/Ranges 05:25
stevan still fixed size and immutable?
audreyt and maybe /Generators (not sure if we want to separate Ranges that rely on side effects and those that does not probably KISS for now) 05:26
stevan reads audreyt
's
description and realizes he is asking dumb questions
audreyt Seq is immutable. its size e not yet known until you force it
stevan yes
audreyt once it's completely forced it's essentially a tuple
stevan yes
audreyt so it's not mutable (assignable/pushable) from the outside 05:27
stevan as it is then fully evaluated
audreyt but its internal repr can change.
yes.
stevan I have a question re: laziness and ranges
map { $_ * 2 } (0 .. Inf) 05:28
will that return a lazy list?
audreyt sure, it builds the Range by binding the original Range obj
stevan ok
audreyt and have the iterating condition set to iterating the original one
and _then_ run *2 05:29
stevan hmm
audreyt and exhaust whenever the original one is exhausted.
that's why I'm not sure we can have just one Range
stevan so we may need a MapRange, GrepRange, etc
audreyt actually they can be represented using the same Range structure
using closure
but I'm not sure if it will be easier or not 05:30
certainly its outside API stays a Range API.
stevan is not finding that approach to be easier ATM, but it might be my current Lazy list impl
still does Range,.. but internally is diff
audreyt although, we want a .reverse operation on Range API
05:30 apple-gunkies joined
stevan yes, thats not as bad though 05:31
audreyt and different internals can implement it either very easily (preserving laziness) or very difficultly
stevan TIMTOWTDI :P
audreyt yeah, I think we want different concrete classes
and a Range API to unify them
stevan I have spent the past 2 hours trying to find a way to not use diff classes,..and I am pretty sure your right on that one :) 05:32
audreyt the enumerable API we were talking about may include:
Enumerable(all?, any?, collect, detect, each_with_index, entries,
find, find_all, grep, include?, inject, map, max, member?, min,
partition, reject, select, sort, sort_by, to_a, to_set, zip)
and Range object itself may include:
==, ===, begin, each, end, eql?, exclude_end?, first, hash,
include?, inspect, last, member?, step, to_s
stevan anything which has a stopping clause needs a special case class to handle this stuff
audreyt (this is from "ri Range" ;))
yes.
stevan yes I noticed the ruby method? stuff 05:33
anru is now known as kanru
geoffb audreyt, ri?
audreyt geoffb: ruby's perldoc is spelled as ri
stevan its Ruby for Perldoc
geoffb "Ruby Info"?
nod 05:34
audreyt stevan: oh. re repr types. I'm not so sure ignoring invocant is a great idea
stevan: what do you think about "opaque"`create ? 05:35
stevan what do you mean?
audreyt instead of ::whatever`create_opaque
this may be a silly idea.
but does allow passing repr type as a string easily
instead of doing a (potentially nasty) switch/case on CREATE time
stevan I am fine with create_opaque() (function, not method)
oh ,.. wait I see
hmm, I like that 05:36
audreyt excellent :)
audreyt is happy that her silliness is appreciated
ok, I'll make it happen Now then
stevan switch/case in PIL^N is nasty too :)
geoffb Not just now, Now
audreyt yeah, it's a boxed now
but I need to move away from this silly japanese television stuff the hackathoners are watching 05:37
and go back to my room and haxx0r (which means no wifi for a while)
miyagawa lol
stevan audreyt: maybe not the string,.. maybe we use some new special notation? 05:38
<opaque>`create()
actually that could get ugly
nevermind
audreyt :) 05:39
dduncan any specific shows the hackathoners are watching? 05:43
audreyt miyagawa may know the proper japanese name for that
I have no idea except I'm going to hide a bit :) & 05:44
miyagawa it's King-Chang's disguise show
www.ntv.co.jp/kasoh/ 05:46
05:47 mtve joined 05:48 qwacky joined 05:51 j0sephi joined, tewk joined, broquaint joined 05:52 jp-autark joined
svnbot6 r8615 | stevan++ | PIL/Native/Bootstrap/Classes/LazySequence.pil 05:56
r8615 | stevan++ | - fixed this to be a plain old generator driven lazy list rather than
r8615 | stevan++ | the pseudo range thing I was building. Anyway, this may all be merely
r8615 | stevan++ | an acedemic exercise anyway :) (learning new things)++
stevan hmm, are ranges always lazy? 05:59
stevan thinks the answer is yes, and so proceeds (for the moment) under that assumption 06:00
svnbot6 r8616 | Darren_Duncan++ | r1854@Darren-Duncans-Computer: darrenduncan | 2006-01-08 22:10:20 -0800 06:11
r8616 | Darren_Duncan++ | /ext/Rosetta-Incubator : some SQL::Routine->Rosetta::Model cleanup; updated Copying and TODO, synced versions of Rosetta.pm and Validator.pm (were 0.490.0) upwards to match Model.pm (0.710.0)
PJF encourages anyone who's interested to edit the freshly created Pugs entry on PerlNet at perl.net.au/wiki/Pugs in whatever way seems fitting.
stevan commits notes on Range objects and goes to sleep 06:20
be back in ~8 hours
gaal $morning!
stevan :)
morning gaal
gaal night stevan :)
stevan goodnight gaal :) 06:21
stevan &
06:21 rantanplan_ joined
gaal ?eval 'my $x = eval "--- !pugs/object:IDoNotExist\nbaz: 42" :lang<yaml>; print $x.yaml 06:21
valbot_860 is now known as evalbot_8616
evalbot_8616 Error: unexpected "m" expecting "\\", "$/", "$", "$!", "'", word character, "::" or "q"
svnbot6 r8617 | stevan++ | docs/notes/range_implementation_notes.pod
r8617 | stevan++ | - adding the notes from #perl6 converstation about ranges
r8617 | stevan++ | and their internal implementation's
gaal ?eval my $x = eval "--- !pugs/object:IDoNotExist\nbaz: 42" :lang<yaml>; $x.yaml
evalbot_8616 "--- !pugs/object:IDoNotExist \nbaz: 42\n"
gaal class autovivification :) 06:22
?eval my $x = eval "--- !pugs/object:IDoNotExist\nbaz: 42" :lang<yaml>; $x.baz = 17; $x.perl
evalbot_8616 Error: No compatible subrountine found: "&baz"
gaal oops :)
what's the correct behavior when loading an object from YAML for which a class doesn't yet exist? error? default class? default class with default accessors? 06:24
06:30 Aankhen`` joined
dduncan okay, I would like some feedback ... 06:41
I need to have a good, short name for the language that Rosetta uses at its core to represent database instructions and schemas and stuff
my first thought ... 06:42
or one of them ...
was to piggyback off some of Pugs' naming conventions ...
and call it "RIL", or "Rosetta Intermediate Language" 06:43
its purpose is very similar to PIL
also, if one wants a more amusing variant, "RILE"
(of course, by the same token, PIL can become PILE) 06:44
This may also be used as a module name, but if not, at least the language name
wolverian dduncan, what does the E stand for, then? 06:46
dduncan the E is just meant to make it easier to pronounce ... but it can come from the last letter of "language"
on the other hand "RIL" could be pronounced without spelling the letters too 06:47
names can look like acronyms, without being acronyms 06:48
...
in fact "RIL" could also be short for "Relational Intermediate Language" ... a double meaning 06:49
wolverian so what does RIL do?
only call it that if it truly is relational, unlike SQL :)
dduncan RIL is sort of like desugared SQL
but that it is better than SQL too
wolverian sounds good to me
do you have the syntax designed yet? 06:50
dduncan the syntax is loosely designed
I hope to have a draft of that committed in the next few days
wolverian (domain specific languages)++ 06:51
gaal GRILL?
dduncan standing for what?
06:51 xinming1983 joined
gaal Great RIL Language? 06:52
wolverian heh
dduncan I'd rather not be recursive in its definition
wolverian it's not recursive: Great Rosetta Intermediate Language Language
gaal hey, at least it isn't infinitely recursive.
for that it needs to be GGRILL
Or GGRIL 06:53
I must away to $work.... have a nice one :)
PJF Bye Gaal. Happy working. 06:54
dduncan or if I wanted to be an ass, I could call it GIRL, for "Great Intermediate Relational Language"
wolverian only if it's relational :)
gaal PJF: argh, I started editing the pugs page but got distracted, sorry :( 06:55
dduncan here's a question, wolverian ...
can something be called relational if it is broad enough to define both relational and non-relational alternative concepts? 06:56
06:56 GeJ joined
dduncan assuming that's okay, plain IRL could actually work well 06:57
not so similar to PIL
but mainly it doesn't reference Rosetta
wolverian not when it comes to SQL, since it is not relational by design
dduncan so if I rename Rosetta to something else later, IRL will still work as a language name
the "Intermediate" in "Intermediate Relational Language" could also double in meaning for something that covers relational definitions, but non-relational ones too 06:58
that is, while it can handle truly relational database definitions, it also has to be able to model the behaviour of existing non-relational databases 06:59
07:15 iblechbot joined 07:30 lisppaste3 joined 07:32 azuroth joined
gaal ?eval (1 .. 3) >>**<< 2 # ok 07:36
valbot_861 is now known as evalbot_8617
evalbot_8617 (1/1, 4/1, 9/1)
gaal ?eval 1 .. 3 >>**<< 2 # legit precedence issue, or a bug?
evalbot_8617 Error: Hyper OP only works on lists
wolverian do we have a precedence table? 07:37
besides the one pugs uses :)
gaal yes, in S02 I think. 07:41
aaahhhahah!! closure doesn't work!? 07:42
?eval sub make_tag ($tag) { -> $text { "<$tag>{$text}</$tag>" } } my $bold = make_tag("b"); my $ital = make_tag("i"); say $ital("lightweight syntax for closures is { $ital('fun') }!");
evalbot_8617 OUTPUT[<i>lightweight syntax for closures is <i>fun</i>!</i> ] bool::true
gaal note all the tags are <i>
buu Oh dear god!
wolverian ?eval my &closure = do { my $x; { say $x++ } }; closure for 1..3 07:43
evalbot_8617 OUTPUT[0 ] Error: cannot cast from VBool True to Pugs.AST.Internals.VCode (VCode)
wolverian really.
gaal, I prefer my &ital :) 07:44
gaal wolverian: that's a stylistic issue... that closure is broken pains me more :( 07:45
wolverian your example looks fine to me, as it only uses $ital
gaal er.
um.
gaal looks for a nearby wall
wolverian :)
gaal++ # so cute 07:46
gaal cute or no, walls are in short supply. good thing we have larry coming in on February (though I doubt he'll appreciate it if I bang my head against him) 07:47
wolverian :) 07:50
make make_tag take a block instead of a string, so that it can construct a DOM tree instead of a flat stirng 07:51
s,stir,stri,
gaal wolverian: your ball: perl.net.au/wiki/Pugs 07:52
07:53 r0nny joined
PJF gaal: Thanks for the edits! Very much appreciated. 07:53
gaal Very much in need of further hackage :) 07:54
PJF gaal: Indeed, but it's a good start. :) 07:57
07:58 Alias_ joined
Alias_ seen audreyt? 07:58
jabbot Alias_: audreyt was seen 2 hours 14 minutes 48 seconds ago
08:09 elmex joined
svnbot6 r8618 | audreyt++ | * Representation types layout. 08:16
audreyt rehi
stevan: repr types layout is in.
dduncan hey hey hey 08:17
08:18 xinming joined
audreyt hey dduncan 08:21
svnbot6 r8619 | audreyt++ | * updated prim table for the default reprs.
dduncan last commit of today should be in any minute
audreyt woot 08:22
I'll go finish the repr types in a couple hours and check them all in 08:23
gaal audreyt: when the syck gets a tick, please see if I'm doing anything obviously elliptic in Rule deserialization 08:24
audreyt okay 08:25
bbiab...
dduncan pushing now... 08:27
done
svnbot6 r8620 | Darren_Duncan++ | r1864@Darren-Duncans-Computer: darrenduncan | 2006-01-09 00:26:12 -0800 08:30
r8620 | Darren_Duncan++ | /ext/Rosetta-Incubator : finished known tasks to merge old SQL::Routine -> Rosetta, renamed 'the SQL::Routine language' to 'IRL' (Intermediate Relational Language), changed NAME pod of Rosetta::Model and Rosetta::Language
08:33 xinming1983 joined
dduncan good night, whomever 08:44
clkao audreyt: pil-array-p? 08:57
09:11 kane_ joined 09:24 PJF joined 09:42 joepurl joined 10:02 nnunley joined
drbean what's a good way of remembering the name of the perl.org page where perl6 people's blogs are posted 10:11
integral planetsix.perl.org you mean?
drbean yeh. I thought it was rainbow or something. why planetsix? 10:12
integral I just remember that these things are called planets :-/
drbean What things are called planets? 10:13
integral these blog aggregators
drbean OK. I guess it is the accumulation of matter at one point. 10:14
The 6th planet is mercury,venus,earth,mars,jupiter,saturn
GeJ it's a reference to planet (planetplanet.org) which is a feed aggregator. Since there's already a planet.perl.org, planetsix.perl.org was created for perl6 only 10:19
10:36 ingy joined
audreyt planet.pugscode.org works too 10:37
bbiab... 10:38
svnbot6 r8621 | audreyt++ | * Finished transcribing PIL.Native.Objects methods into the new 10:40
r8621 | audreyt++ | PIL.Repr* methods -- still some missing from the spec.
10:50 cognominal joined 10:59 G2 joined 11:14 SamB joined 11:18 nooga joined
nooga hem.. hello 11:19
r0nny re 11:29
11:34 nooga left 12:03 nothingmuch joined
obra hellow from RT training 12:30
12:37 iblechbot joined
nothingmuch hola obra 12:41
nothingmuch really needs to scan more slides/negs
all the photos on my page are aging
12:46 chihchun joined 13:02 chris2 joined 13:21 Limbic_Region joined, kolibrie joined 13:30 dada joined 13:45 orafu joined
svnbot6 r8622 | audreyt++ | * PIL^N: Switch to representation type is now complete. 13:47
r8622 | audreyt++ | * See the updated docs/notes/piln_object_repr_types for the `create
r8622 | audreyt++ | syntax. Autoboxing is not supported yet but shoudl resume soon.
stevan audreyt: :) 14:25
svnbot6 r8623 | stevan++ | piln_object_repr_types.pod - change sigils to be more self-documenting in audreyt++ code examples 14:35
14:36 loftcity joined
stevan audreyt: I think you forgot to commit PIL.Repr.Internals :) 14:41
14:41 theorbtwo joined
svnbot6 r8624 | stevan++ | PIL/Native/Bootstrap/* - removing the :: stuff 14:44
14:55 bsb joined 15:06 vel__ joined 15:14 eric256 joined
eric256 is there a list of currently supported pugs features somewhere? just curious since i've been out a while and there is never an easy way to get caught back up ;) 15:17
gaal eric256: smoke.pugscode.org ? 15:25
pugs.blogs.com also 15:26
15:26 Limbic_Region joined, hexmode joined
gaal standard answers, sorry; there's no other functional inventory that I know of. 15:26
I doubt there's a pristine version of that inventory too ("the" p6 feature list) 15:27
eric256 true. just hoping some magic had taken place ;) 15:29
15:30 lisppaste3 joined
gaal at some level, "the smoke report *is* the feature list" is a great answer. maybe it's a little too detailed, but the graphical matrix at least gives you a quick summary with links to further info 15:34
eric256 that true 15:36
just a list of , here is where the object system is, here is the type system, etc, would be nice 15:37
Limbic_Region that's a problem I tried to address on the list a while back 15:46
it is hard to know what is left to be done if there isn't a framework and list of things done and to what degree
this was more WRT the language specifications then the implementation, but the same principals apply 15:47
gaal Limbic_Region: can you characterize such a list? everyone wants a different level of detail
I'll give you an example:
Limbic_Region gaal - for design sure, for implementation no
gaal take a look at whytheluckystiff.net/syck/ and scroll a page and a half down 15:48
at first glance this is great - a bounded list of features, with reasonable level of detail
Limbic_Region right, that is a high level list
with a short summary of detail 15:49
gaal but pugs is about 40 times bigger than that
Limbic_Region but again, I am not as concerned with pugs
because implementation follows design
gaal so problem #1 is that you'll have a big list
problem #2 is that if you're interested in a particular feature, and it isn't listed as 100% complete, then the list doesn't help you 15:50
and you can't "link to the tests"
Limbic_Region well, that is a matter of opinion and perspective
gaal because there is no one test that encapsulates all there is to know about the status of this feature
many (most?) of the tests in t/ are functional tests, "does this feature work?" 15:51
if you see split.t with 214 tests, all passing, you are fairly confident it is complete 15:52
but some tests are more subtle than that. -- some features are more subtle than that!
eric256 a high level list with links to test files would be nice though 15:53
Limbic_Region is on the phone
gaal and I'm not talking of bizarre coner cases like "does that florgh burglue when zorlicity levels achieve comecitude"
eric256 it coudl even specify what files need to pass in order to consider that feature "done"
gaal yeah, but then it'd be hard to maintain and nobody will maintain it. :) 15:54
eric256 i'm not sure how hard maintaining would be.
what about adding a description to test files? the compiling a summary based on % pass and the summary. ? 15:55
gaal you have descriptions: most .ts have a =kwid description
the .yaml doesn't carry that information but it easily could\ 15:56
eric256 although there are LOTS of tests. so we might sub divide it by directory. hmmm. i could tie that description into my Test/Documentation link thingy (which i've now lost and need to refind)
getting sick sucks, getting sick on the holidays realy sucks, i've spent about a month with the flu now! what fun! ;)
gaal :(
I suppose someone with plenty of HTML-fu can make a "folding" report 15:57
top level: 8000/10000 pass 15:58
then [ 01-sanity: 12/12 ; builtins: 240/300 ; data_types: 100/109 ; ... ] 15:59
and so on for each level
Limbic_Region gaal - here is a link to my argument groups.google.com/group/perl.perl6....a75de53310
in a nutshell - it is impossible for anyone outside of @larry to help flesh out the design because there is no framework saying what is and is not done or what, in total, needs to be done 16:00
eric256 loves his boss. migrate to oracle, oh and while your doing that can we proceed with these 15 other projects. blah
Limbic_Region originally, the AES were supposed to follow the chapters of the Camel with all of the RFCs following into 1 or more chapters
that has gone out the window
so there is no foreseeable end to the constant hashing out of design on the list 16:01
because no one knows what consitutes the entire framework
of course there will always be things we didn't think of, but come on - those should be far and few between
*shrug* 16:02
gaal Limbic_Region: well, maybe no one does really know. That doesn't make p6 impossible.
hard to get into, sure
Limbic_Region no - it makes it frustrating
gaal well, $larry's quote there is honest about it. 16:03
stevan Limbic_Region: I feel your frustration, and then some, as we are trying to actually *implement* all this fuzziness 16:04
stevan thinks if ever there was a language with lazily evaluated hand-waving function it is Haskell :) 16:05
audreyt undefined :: forall a. a
svnbot6 r8625 | audreyt++ | 15:41 < stevan> audreyt: I think you forgot to commit PIL.Repr.Internals :)
stevan audreyt++ :)
gaal that isn't to say I don't think things can be improved; but I don't know how it might be grand-adminned, and I can sure see how overadminning will cause underparticipation :/ 16:06
stevan audreyt: how do you say "thank you" in chinese?
gaal xie xie
eric256 okay i missed something... where did audrey come from? did you change your name or just feel like confusing all us part timers? ;)
gaal evil :: a -> b
stevan eric256: see pugsblog
eric256 i did. i didn't see any mention of why though 16:07
audreyt eric256: pugs.blogs.com/audrey/2005/12/runti...pecas.html
stevan although the true reason was to mess you part timers,.. but shhhh you didn't hear that from me
audreyt lol
stevan thinks it might be a long way to do for a joke, but then audreyt never really seeems to do anything without diving into it completely ;) 16:09
eric256 swears he read the journal and never saw that post.
audreyt eric256: there are two journals :)
eric256 that would explain it.
theorbtwo pugs.blogs.com/audrey/2005/12/runti...pecas.html
audreyt pugs.blogs.com and pugs.blogs.com/audrey
stevan swallows his new ./pil :)
audreyt :D 16:10
stevan: I think it's sane.
it also brings our architecture much closer to parrot.
stevan yes
audreyt (or, at least, parrot according to leo's new design docs)
stevan leo's docs have been bouncing around in my head lately
audreyt it's so close that I'm tempted to call Repr PMCs ;)
Limbic_Region leo's design docs or chip's design docs?
stevan LOL
audreyt Limbic_Region: leo's, based on pugs quickrefs 16:11
s/based/inspired/
gaal btw ruby has a nice quickref doc too (was it the inspiration for docs/quickref?) # www.zenspider.com/Languages/Ruby/QuickRef.html 16:12
16:12 rantanplan_ joined
audreyt gaal: ingy figured out ways to roundtrip between pugs and YAML.pm via YAML::Marshall 16:12
gaal woot! 16:13
audreyt gaal: I think objspace serialization format can easily be YAML at this point
there's really no reason not to use that
ingy audreyt: no I didn't
well actually I did
gaal yes, we even have clean separation beween userland objects and magic stuff
ingy but not before you said that
gaal (!pugs:object/Foo vs. !pugs/Role) 16:14
audreyt SerTH _may_ be slightly faster (like Storable.xs), but we completely lost interoperability; dumping/loading xml in HsXml comsumes far more memory than syck
gaal: yeah
ingy audreyt: let's make the 'perl/' be $YAML::TagPrefix for now
audreyt ingy: sure
ingy :)
gaal audreyt: which reminds me! when would be the right time to drift the prelude and make it unslow? 16:15
ingy adjusts
audreyt gaal: after 6.2.11 16:16
and before 6.28.0, I think
I'm done with PILN for today
I'll now move to PIL/PIR and releng
gaal whee! what can I do?
audreyt gaal: fix "make smoke" ;)
gaal it's broken!? eep 16:17
audreyt like, figure out why -Mblib isn't working anymore.
we need some PIR interop fix -- I'll do that first, now parrot 041 is out
16:17 mncharity joined
ncharit is now known as putter 16:17
gaal hey putter :)
putter hi gaal :)
gaal svn ups parrot and prepares a clean pugs build 16:18
putter audreyt, stevan: ping?
stevan hey putter
audreyt putter: pong
putter hi stevan, question/comment re piln "cond"
hi audreyt 16:19
gaal audreyt: where is -Mblib borkage encountered? or will i see it quickly? :)
audreyt "make smoke" should fail
gaal quick enough for me. :)
16:19 masf joined
stevan putter: go ahead,.. although audreyt might be better equipped to answer 16:19
audreyt (and if you want a sugar form for "cond", or a sugar form for multiple binding ($a, $b) := (1,2) in PILN surface syntax, the answer is a definite "go ahead") 16:20
gaal during svn up:
L examples/cookbook/08file-contents
and it aborts. 16:21
s/up/st/
audreyt mumbles something about svn and svk.
putter Each time I look at piln, the "cond" method jumps out at me. cond of course has a very old, specific, consistent, and universal meaning in lisp world... and that doesn't seem to be it. I don't know of any countervailing traditions. Could I interest you in postscripts "ifelse"?
gaal :)
putter ;) 16:22
stevan putter: why not just rename cond to something else? 16:23
audreyt probably rename it to if_else
putter: please go ahead and do a s///g.
stevan search and replace is much easier than changing all the code to a new concept/approach
putter stevan: err, yes, renaming is what I had in mind. Sorry. Back to using an irc client that has a record of truncating what some folks see. Brevity->ambiguity. Oops. 16:25
nifty. will rename now. thanks :) 16:26
audreyt :)
putter: btw, do you still remember the @array.shift PIL failure?
i.e. do you remember what was the issue, now that I have some cycles for a quickfix?
putter hmm... I think it was 16:28
my @a=([]); push(shift @a, 3, 4); shift gets called with three args. that one?
gaal has a cycle home now. see you in a few... & 16:29
audreyt hm
probably not
putter audreyt has cycles, gaal cycles... hmm... maybe I should...
audreyt $a = @shift.shift();
is the thing that triggered clkao's smoke failure I think 16:30
gaal putter: you in boston, no? I've a friend there who cycles to work even in winter
but really & now
audreyt decides to journal first, then look at PIR/PIL
putter err, rushed pun. should have been audreyt has cycles, gaal has a cycle, mumble. gaal.& 16:31
putter looks at irc log
colabti.de/irclogger/irclogger_log/...sel=30#l55 16:33
16:33 Eimi joined
audreyt putter++ 16:34
hm, I should also write DrIFT Yaml 16:35
putter It looked like shift @a,$morestuff would eat morestuff. thats the test which was fialijng
err, failing
audreyt --- !haskell/Maybe/Just 3 16:36
audreyt thinks it's an attractive thought
audreyt cuts that thought and goes back to pugs
putter: ok. 16:37
putter 's wish-list: rules, pil transition, oo. apropos the general life advice "If you want something, ask for it." ;) 16:42
16:42 konobi left
audreyt what do you want from rules? :) 16:43
hm, the lack of "proto" support is the real cause behind the PIL parsefail. 16:44
putter ah 16:46
re rules... 16:49
gaal rehi 16:53
putter to whatever extend rules are still a flat global namespace, that's a problem. eg, I know while <b> works, <A::b> doesnt (PGE parse error). grammar inheritance would be convenient, but isnt a showstopper. at that point I can actually try running large grammars. not sure what will happen then.
gaal lives near work? or cybercycle? ;) 16:54
gaal trades many things for convenience 16:55
16:55 justatheory joined
gaal and also cycles like an idiot :) 16:55
putter err, not PGE parse error. my fuzzy recollection is <A::b> behaves like <undefinedrulemumble>.
:)
lisppaste3 putter pasted "if_else or ifelse - which is nicer?" at paste.lisp.org/display/15550 16:58
eric256 considers useing his bosses email on the internet as a spam trap ;)
putter notes channel is logged. :) 16:59
audreyt :)
gaal urgh, random ghc errors again. i hope my cpu isn't melting. 17:00
eric256 hehe
audreyt putter: if_else -- or ifte
but ifte is obscure 17:01
if_else then
putter gaal: as long as the smoke doesn't escape, you're fine.
audreyt: ok
17:01 bsb left
gaal putter: that's unfortunate considering I'm trying to get it to smoke :p 17:01
eric256 gall, you no that fan thingy, you should leave it on ;)
gaal lol
I tried underclocking but my bios is for l33t g4merz, it only shifts speeds one way 17:02
audreyt h'thon photo, fro miyagawa: www.flickr.com/photos/tags/chupei/
gaal whee
17:03 elmex joined
putter gaal: oooh. but computer's run on smoke - when it escapes they stop working. my I suggest following microwave instructions for frozen pasta cheese things, while skipping the add sause step. *lots* of smoke. 17:04
gaal ingy looks very concerned about that soup
putter: yeah, i know :)
putter lol
gaal eeepepp!! tell me that doesn't say "zhu9" there in pinyin 17:06
17:07 orafu joined, elmex_ joined
eric256 someone likes to take photos of everything they eat ?? ;) 17:07
gaal err here I mean: www.flickr.com/photos/bulknews/8441...otostream/
(9 would be a very large n) 17:08
eric256 stairs blankly at gaal. okay it doesn't say "zhu9" 17:12
gaal eric256: in mandarin chinese at least, there are four tones each syllable can take; they are part of the meaning of the syllable. 17:13
audreyt xin1 zhu2
gaal chinese speakers differentiate between the tones easily. others often find it tricky to pick up.
audreyt: my sanity thanks you.
eric256: (to continue the explanation) "9" would have implied rather more tones :) 17:14
miyagawa eric256: I always take photos when I eat something, especially I'm in foreign countries 17:15
gaal audreyt: my latest pugs (against parrot HEAD) looks for libparrot only in system locations.
audreyt gaal: even if you set the PARROT_PATH env?
gaal yeah, fully qualified path even. 17:16
strace++
where's that stuff configured? I'll take a shot at fixing it. (If you aren't checking in a fix now :) 17:17
17:17 avinash240 joined
audreyt I'm not 17:17
gaal never moose, grepping for it 17:18
audreyt src/Pugs/Embed/Parrot.hsc: dir <- getEnv "PARROT_PATH"
gaal thankee
audreyt np
gaal oh wait, it doesn't even get there - it's ld.so that giving the error. 17:21
(I compiled with embedded parrot)
ldd pugs | grep parrot -> libparrot.so.0.4.1 => not found 17:23
leo looks like an installed parrot - what was the --prefix? 17:24
gaal of parrot? i didn't make install
leo ah - which arch? 17:25
gaal linux (well, colinux) / debian
leo then you should have rpath_blib set
gaal leo: what's that? set it in pugs build or in parrot build? 17:26
note that it's pugs not finding parrot's so
(which does exist in blib/lib/libparrot.so.0.4.1 in my parrot tree)
svnbot6 r8626 | putter++ | Renamed PIL "cond" to "if_else". 17:27
leo ldd parrot
libparrot.so.0.4.1 => /home/lt/svn/parrot/leo/blib/lib/libparrot.so.0.4.1
gaal: $ grep rpath lib/Parrot/Config.pm 17:28
in parrot root
gaal leo: 'rpath_blib' => '-Wl,-rpath=/home/gaal/parrot/blib/lib',
leo yeah - and that should be visible in ldd too
gaal ldd parrot works OK
ldd pugs doesn't resolve parrot 17:29
leo ah
putter re cond->if_else: I can't test, so may have missed something. mainly any "cond" (not `cond) in a .hs file outside of src/PIL (there was lots of grep noise). docs/notes/recursive_polymorphism_and_type_inference was intentionally not updated.
leo then you need above rpath_blib for linking pugs w libparrot
or link statically 17:30
gaal leo: thanks, I'll look for the right place to meld it together with pugs' PARROT_PATH setting
can pugs link statically with parrot today?
putter gaal: geee, you get a ld.so error! I'm envious. I just segfault. ;) maybe rpath...
gaal putter: strace is your friend :) 17:31
putter :) 17:32
leo LD_LIBRARY_PATH is your friend :)
gaal leo: LD_PRELOAD :)
17:33 avinash240 left
svnbot6 r8627 | audreyt++ | * Ingy has renamed his Chinese name from ?\229?\147?\135?\229?\147?\136?\232?\140?\182 to ?\230?\135?\137?\229?\144?\137?\229?\164?\167?\232?\129?\182, 17:33
r8627 | audreyt++ | and English name from "Brian Ingerson" to "Ingy d?\195?\182t Net",
r8627 | audreyt++ | so now he has the unique status of having _two_ Unicode names.
r8627 | audreyt++ | (and you thought I was weird.)
gaal actually, if the user bothers to set PARROT_PATH, should we not use that to update rpath (if that does what I think it means?)
17:33 hcarty joined
leo gaal: or you just install it 17:34
gaal I'll set LD_LIBRARY_PATH. :-p
putter re names/weird, lol 17:37
gaal I think maybe export broke, because "plan 9" gives funny error messages. 17:39
# *** No such method in class Int: "&plan"
checking.
hmm. Export didn't *universally* break. But tests don't see plan nevertheless. 17:40
ahhh! I think someone is being naughty and oeverwriting PERL6LIB :) 17:41
17:43 Odin- joined
gaal aaaaahhh no :) the tests say "require", so the exporter never exports! the bug is in the tests. 17:45
audreyt gaal++ gaal++ 17:46
audreyt experiences one of those "oh duh" moments.
putter I'm off. cheers & gaal++ ;)
gaal bye putter :) 17:47
uh, oh; use_ok is quite tricky to implement now. 17:48
its job is to get exports into somebody else's lexical environment.
audreyt I think this can trigger a p6l. 17:49
I raised that problem in toronto
but lwall didn't get around to this about that.
gaal grep -r use_ok t ext | grep -v svn {- no jokes about svn please -} | wc -l => 103 17:50
p6lling.
audreyt (for now, change them to "use" ?) 17:51
gaal use + pass "kludge" 17:52
to keep plan
stevan Ingy looks to have aged in this picture, and is shorter than I recall :P www.flickr.com/photos/bulknews/83962913/
gaal yeah, i'll do that.
I think I'll get some dinner first
audreyt: are we relenging for, like, tonight?
audreyt gaal: no, this weekend 17:53
gaal cool
brb
stevan putter: re: testing PIL
a simple sanity test is to call some methods on ^Class in the REPL loop 17:54
putter: ^Class.get_method_list() is a good one since it returns a large amount of stuff
audreyt: everything seems to be working well BTW 17:55
audreyt stevan: :D
good, then my afternoon wasn't in vain
stevan I need to rethink the new->bless->CREATE->BUILDALL-*BUILD thing though 17:56
audreyt we can wrap IO stuff in a IO PMC^WRepr too
nodnod
stevan yes, I was thinking IO would be nice
stevan wants to make PIL self-testing :)
audreyt parrot people does a OS PMC
putter stevan: tnx! :) note to self - find pil repl loop...
&
stevan ^OS`OUT`print() would work :)
audreyt with "chdir" "rm" "stat" "umask" "chroot"
but that I think is a weird idea 17:57
stevan I dont need that really
^OS`OUT and ^OS`ERR and I can port Test::Builder :P
audreyt let me sleep on it 17:58
and I'll get back to you :)
stevan ok
audreyt (mostly because it will no longer be in STM)
stevan tried to hack it with `trace already,.. but it didnt work
audreyt and my "atomic" plan needs some thinking
stevan ok
stevan really should get some $work done today anyway 17:59
stevan & too
Limbic_Region no meet on the pizza? 18:00
Limbic_Region is too lazy to be a vegetarian 18:01
stevan Limbic_Region: meat and cheese is very un-chinese IIRC
Limbic_Region very un-kosher stevan, not sure about the chinese though 18:03
stevan Limbic_Region: when was the last time you saw Chedder Lo-Mein? :P 18:04
Limbic_Region well, I live in America and here we don't have chinese food for the most part. For the most part, we have an americanized version of chinese food 18:05
gaal doesn't remember seeing cheese in the little of china he glimpsed
Limbic_Region I did live in S. Korea for a year and a half though
and it was very common to have a slice of cheese in your raman noodles
Limbic_Region forgets the hangul name for the dish though
stevan milk cheese or yogurt cheese?
gaal there's a famous (in some circles) paper about an American writing teacher in China that starts by telling about how businessmen visiting China gave cheese as a gift and it offended their hosts. 18:06
oh yogurt I saw a lot of. yummy :)
stevan Limbic_Region: seems that maybe things are changing -> www.globalpolicy.org/globaliz/cultu...cheese.htm 18:07
Limbic_Region I have been wondering about S.E. Asian countries and what kinds of fish are common for eating (especially WRT pregnant women) 18:08
18:08 avar joined
PerlJam gaal: cheese is offensice? 18:08
er, offensive
stevan Limbic_Region: fish is best consumed minimally when pregnant, mostly because of the high amount of toxicity they cary 18:09
some fish are worse than others for this, but I dont recall which ones
gaal PerlJam: apparently for some it is 18:10
Limbic_Region stevan - that's not exactly accurate
it is the mercury that is a problem
and it is actually recommended to get 2 servings per week of low mercury fish
stevan Limbic_Region: which is very toxic to pregnant women
Limbic_Region no, it is toxic to the babies nervous system - but yeah
Limbic_Region already knew that 18:11
stevan Limbic_Region: some doctors (ours in particular) said to get the nutrional aspects from elsewhere
but TIMTOWTDI
PerlJam mercury is toxic to *everybody's* nervous system :)
Limbic_Region 's question had to do with what fish were common in S.E. Asian countries especially WRT pregnant women
PerlJam - yes, but a developed nervous system is less prone to be influenced and the body is capable of purging itself of heavy metals over prolonged periods of time 18:12
stevan stops drinking his mercury & Mountain Dew
Limbic_Region a developing fetus doesn't have such a luxury
stevan - well, I assume that advice all depends on your practicioner as our's was all for eating fish in small quantities and of the safe variety 18:13
18:13 nothingmuch joined
stevan Limbic_Region: our last obgyn was a vegan, which probably explains it 18:13
Limbic_Region heh 18:14
Limbic_Region was a vegetarian for a few years but definately is too lazy to be a vegan
stevan everything in moderation I say
PerlJam Limbic_Region: shrimp and tuna are "low mercury" sea food. I don't know about SE Asia, but tuna is fairly popular in American "asian cuisine"
stevan resumes drinking his mercury & Mountain Dew
eric256 is sure there is some pun in there about bad poetry (yes i know its vogan, buts its gotta be in there somewhere anyway)
Limbic_Region ok - bringing this moderately back on topic
stevan eric256: I am sure there are many bad vegan poets 18:15
Limbic_Region what exactly does audreyt and/or other asians use for quick fuel when prolonged hacking?
doritos/pizza/mt. dew probably aren't it
eric256 hehe we have a family friend who doesn't eaty anything cooked, and no meat, so he eats raw fruit and vegetables and uncooked bread, makes vegans look pretty normal. ;)
PerlJam "uncooked bread"? 18:16
We call that dough around here.
18:17 hcarty joined
stevan Limbic_Region: "Chinese cuisine was therefore unusually well suited to Jewish tastes because, unlike virtually any other cuisine available in America, traditional Chinese cooking does not use any milk products whatsoever" maybe not meat and cheese, but cheese itself (from this article www.soc.qc.edu/Staff/levine/NYJews-...inese.htm, interresting read too) 18:17
PerlJam: uncooked bread is actually good, but it is not dough 18:18
Limbic_Region except as I said, chinese food in America (from my experience anyway) isn't very authentic
PerlJam stevan: bread isn't bread until cooked.
Limbic_Region though it has also been my experience that other than very orthodox jews, maintaining kosher is more about believing everything is ok then actually being ok 18:19
stevan PerlJam: one sec,.. lemme find a link,.. I eat this stuff all the time,.. it's very good (and I am pretty sure there is no cooking involved)
PerlJam Limbic_Region: *nothing* in america is "authentic". It's all "american version of <insert other thing here>"
Limbic_Region: it may not even be about believing any more. There are kosher products in local mega-marts. I can't see *anyone* believing those items to be "okay" as far as kashrut law goes 18:21
Limbic_Region PerlJam - actually, almost every product in America has a Kosher or Unkosher designation - just look for a little K or U which may or may not be in a circle 18:22
gaal "chinese food" (a blanket term for scores of cuisines) uses pork a lot, which kinda gets in the way of kosher laws :) 18:28
theorbtwo gaal: You simply need to adopt a few american jewish kosher laws: (1) if they wanted you to know what was in it, they would have put it on the menu. 18:30
(2) Chinese food is always kosher.
gaal ha! I googled for the author of that essay I mentioned earlier. the picture I came up with is on topic: www.cas.sc.edu/engl/faculty/emeritu...alene.html
clkao gaal: have you fixed the test.pm require problem
18:31 eric256_ joined
gaal theorbtwo: personally I have no problem with kosher laws. I eat kosher too. 18:31
clkao: no, I just wrote p6-l about it
clkao k
gaal clkao: there are two problems really
1. s/require Test/use Test/ # easy
clkao hate software
theorbtwo gaal: I would think maintaining kosher is pretty easy when you live in Israel. 18:32
gaal 2. s/use_ok/something else/ # hard
Limbic_Region heh - considering kosher loosely means separate
I would tend to agree theorbtwo
gaal you misparsed :)
18:33 bernhard joined
theorbtwo misparsed? 18:35
clkao so to fix smoke.. ? 18:36
18:37 stevan joined
gaal clkao: working on it 18:37
18:38 Corion joined
gaal theorbtwo: I won't refrain from eating something because it's kosher. 18:38
theorbtwo Oh. 18:39
gaal hmm, only tests in ext/ did "require Test". I wonder why...?
eric256_ the answer.com definition of kosher is ...weird. does kosher realy mean you can have meat and cheese prepared in the same kitchen? /me is ignorant as to why that would be 18:40
theorbtwo I thought you meant that you only eat things that are.
18:40 orafu joined
gaal theorbtwo: "too"'s arity is underdefined. 18:40
eric256_: kosher is weird. 18:41
18:41 Corion joined, Corion left
gaal not my department though. 18:41
theorbtwo Essensially, you can't eat milk and meat together. Just how not together they have to be depends on who you ask. 18:42
eric256_ but why? i mean is there some fundemental reason i'm missing? (yes i know asking why of relegious matters is often pointless, but i'm realy curious)
Limbic_Region eric256_ you should read the wikipedia entry 18:43
theorbtwo Yes, you should.
Also, L_R, you should read [18:42] [theorbtwo(+ix)] [6:#perlmonks(+npt)] [Act: 3,5]
Whoops.
en.wikipedia.org/wiki/Hechsher
Limbic_Region but the one that makes the most sense to me is the separate one
Limbic_Region has some time and does so
theorbtwo The K, without the circle, is meaningless because it's untrademarkable. 18:44
Limbic_Region learned an awful lot about Judaism and Kosher when in the Philippines
theorbtwo There's a lot of philipino jews?
Limbic_Region nope 18:45
Catholic converts
they actually have a daily program on TV educating catholics on their old testament heritage
svnbot6 r8628 | gaal++ | ext/*/t -
r8628 | gaal++ | * if a test "require"s Test.pm, it doesn't get its exports. fix: s/require/use/.
theorbtwo Ah.
Limbic_Region which always ends with why they should remain catholic
theorbtwo Heh. 18:46
gaal oh, crap. the exporter is broken on multi subs. fixing. 18:47
svnbot6 r8629 | audreyt++ | * "pop" and "push" should parse as unary; use an evil hack for that.
r8629 | audreyt++ | Thanks clkao++ for keep pestering me about this. :)
gaal (should we not have noticed that the whole test suite was failing for a few weeks now?)
theorbtwo wonders if there isn't a good way of priortizing tests to make a test run shorter. 18:48
audreyt gaal: I noticed it last week but as it's in midst of lexical export hack, I thought it's fine to break a bit :)
18:49 GeJ joined
gaal audreyt: means "skip" is broken 18:49
audreyt celebrates chip's return-to-life in #parrotsketch
gaal: I noticed :)
gaal i think it's fixable, looking into it. 18:50
audreyt clkao has been ranting about this "can't run tests" thing for a while now
cool, gaal++
18:54 khaladan joined
gaal yeah, the thing is, a sub's multiness has to propagate to the exporter somehow 18:56
oh, but there's another problem! its name can't be the key in the %*INC<module><exports> hash 18:57
because since this is a multisub, there's more than one (duh!)
PerlJam gaal: The long names are unique though
gaal every so often I wish perl could promote a scaalr to an anonylist 18:58
PerlJam: long names, including the sig you mean?
PerlJam aye
gaal I don't think I have an easy way to access the sig text inside the parser. I wonder if "show styp" works though. (nasty hack!) 19:00
ugh, nonono, the sub must be still keyable by its name. a list is what I need. 19:01
okay, so for regular subs let it be a singleton list.
Sanity test: you can't import only some definitions of a multi sub, right? 19:02
19:03 Amnesiac joined 19:12 r0nny joined 19:15 elmex joined, planrantan joined 19:17 mjl69 joined
gaal we would be very nice indeed to catch errors like 19:19
multi foo(A) is export ; multi foo(B) # no export
but that's for another day.
audreyt nod.
stevan gaal: (multi foo(A) is export ; multi foo(B) #no export)++ 19:34
its a very common erlang (and IIRC lisp too) to export based on arity
s/very common/very common in/ 19:35
19:35 marmic joined, chris2 joined
stevan I have always liked that technique :) 19:35
audreyt of course, maybe what we really want is "proto" support :) 19:37
stevan yes, but that is nice too
so the public API takes the single param,.. then the internal API does the pattern match on the args 19:38
great for accumulators and such
stevan wants all the cool functional language toys :)
audreyt :)
stevan: re the iterator thing in our dispatcher 19:39
stevan yes
gaal no sanity;
theorbtwo Meh, what's wrong with naming the public api foo and the internal api foo_?
stevan theorbtwo: nothing :)
audreyt do you think we want a new repr type for it?
stevan audreyt: I will ponder it
audreyt www.parrotcode.org/docs/pmc/iterator.html 19:40
stevan maybe a code repr type? ;)
audreyt it's there to provide multiple .each visitors on a hash
and/or an array
gaal argh. In RuleParser, how do I iterate over elements of a VList?
audreyt Code will be there for sure
gaal: a VList is just a List.
so, not sure what you mean.
gaal hand a closure to uneafeEvalExp?
hmm I think it's an Eval VList 19:41
audreyt you need to unsafeEval then.
gaal but actually maybe not!
we'll see in a sec :)
stevan audreyt: re: iterators,.. it makese sense to be like Parrot, so I think it might be a good idea
audreyt mmm the sense of adventure
gaal I'm no longer afraid of the inferencer :) 19:42
audreyt stevan: maybe, though it violates encapsulation badly
gaal++ # so nice to hear that
stevan: it may make more sense for the aggregates to return iterators
instead of creating iterators and passing aggregates as args
gaal audreyt: www.livejournal.com/users/gaal/181614.html
audreyt or maybe I was not understanding the problem corrcetly
in any case it's nearly 4am so I'll journal and sleep :) 19:43
gaal :)
audreyt gaal: niice :)
stevan audreyt: I will ponder,.. i think iterators are a good thing, not sure if they need to be their own repr though
audreyt k
stevan is currently doing manual PDF layout so his mind is a little mushy
audreyt: how much value is there in converting some of the PIL^N stuff to PIR? 19:47
gaal grins at finding out what forM was all along
audreyt gaal: I was tired to flip mapM ;)
gaal and that he'd reinvented it in the yaml stuff
19:47 avar joined
gaal audreyt: I can see why 19:48
audreyt stevan: I asked allison that question... seems codegen to PIR instead of POST is still the way to go
stevan: so if you'd like to write up some compilation rules based on sample transformatinos, that'd help
stevan so compiling PIL^N to PIR??
audreyt stevan: though I'd like to see the S12 metamodel running happily on the new reprs first :) 19:49
stevan yes, that would be first priority of course :)
audreyt stevan: we need to bring our bootstrap code to PIR to recreate MM on it
stevan I guess I am wondering where all this PIL^N stuff will fit in when things move more towards parror
audreyt stevan: so the codegen is not avoidable 19:50
the question is whether higher level PIR2 code compiles to PIR or thru PILN first
and the latter is more simple, so we do that first
but the former is equally valid, and we can tackle it later
stevan ok
so Perl 6 to PIL2 to PIR, first via PIL^N, but then straight to PIR once the MM is ported 19:51
audreyt hm, all metamodel tests pass, but simple.t fails. 19:52
stevan: yes, I think that is a sound plan.
stevan is happy he won't have to go back to PIR for a little while :)
audreyt actually, exactly the same thing as the ->P5 plan.
stevan I rather like my lambda friends in PIL^N
audreyt the ->JS plan is special because PIL2->JS is easier.
heh. I'd happy for PIL^N be the PIR that happens to nest ;) 19:53
hm, array.t and pair.t also failing.
stevan audreyt: yes, they should be
the class stuff should be the only thing which works actually
audreyt k 19:54
stevan I will fix all that other stuff this evening (~ 6 hours from now) 19:57
audreyt woot 19:59
once I wake up I'll carry over the other repr types
from old runtime, as detailed in my journal
stevan sounds good to me
audreyt do note that this is only for sane minimal OO p6 to happen 20:00
because after that we can write more "native" classes in p6 :)
stevan :)
gaal
.oO( the sun machine is coming down and we're gonna have a party )
 20:01
20:05 dduncan joined
dduncan greetings 20:07
audreyt yo
dduncan fyi, there is now another set of commits I'll do before the 6.2.11 release
audreyt that's fine; as you saw, I got some $job assignments 20:08
so release will have to wait till weekend :/
dduncan okay
20:11 nnunley joined 20:18 vel__ joined
gaal yes! it compiles. ship it etc. 20:20
audreyt woot, ship it!
gaal darn, runtime error. those aren't supposed to happen :) 20:23
audreyt heh.
stevan gaal: clearly a case of bad/malformed input :P 20:24
gaal ah, a VRef <Array>, not a VList. hmpf. 20:30
which, come to think of it, should have been entirely obvious to me because this is a hash value 20:32
what's p6 for @{ } ? 20:33
Ann Cxt CxtSlurpy? 20:35
audreyt Syn "@{}" 20:37
gaal thanks 20:39
audreyt np :) 20:40
gaal beh, my fix didn't work :( 20:41
tracing some more. 20:42
it should have worked :( 20:48
audreyt: still have some wakies to spare? 20:49
I'll commit; please uncomment line 860 and note that reinstallation of symbols is supposedly taking place like it should 20:50
assuming the name need not be fully qualified.
ext/Algorithm-TokenBucket/t/test.t is an example failing test.
20:51 macli joined
svnbot6 r8630 | gaal++ | lexical exports - stab at fix for multi subs not being exported properly 20:57
21:05 nnunley joined
gaal sleeps & 21:21
21:25 eric256_ left
va is now known as avar\away 21:35
21:52 peacock joined 22:35 hcarty left 22:53 rep joined
var\awa is now known as avar 23:03
23:30 Odin- joined, lisppaste3 joined 23:55 feng joined
rafl Any shedule for a new release? 23:56

Powered by Cro and the Raku® Programming Language.

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

Thank you!