Please test pre0: pugs.blogs.com/pre0/Perl6-Pugs-6.2.12.tar.gz | paste: sial.org/pbot/perl6 | pugs.blogs.com | pugscode.org | pugs.kwiki.org
Set by audreyt on 12 June 2006.
00:21 pfenwick joined
svnbot6 r10778 | audreyt++ | * JS Prelude: Fix an incorrect use of reduce. 00:24
meppl gute nacht 00:29
cognominal_ is ~{ whatever } synonymous of "{ whatever }" ? 00:35
audreyt ?eval ~{1}
evalbot_10746 "<SubBlock(<anon>)>"
audreyt no. :)
?eval ~{1}()
evalbot_10746 "1"
audreyt now that works
00:52 shachaf_ joined 00:55 hikozaemon joined 01:06 Khisanth joined 01:24 Ymmv joined 01:39 mjk joined 01:49 jiing_mIRC joined 01:51 jiing_ joined 02:29 ko1_away is now known as ko1 02:30 christopher joined 02:38 fglock joined
fglock christopher: in case you backlog - the Pugs::Compiler::Rule tree is at misc/pX/Common 02:41
svnbot6 r10779 | fglock++ | Pugs-Compiler-Perl6 - P6.pm doesn't need exec(); added some operators to the emitter
02:43 shachaf_ joined
christopher fglock: thanks. Is the one in perl5/ orphaned? 02:43
fglock there was a plan to move it there, but the development went on in pX 02:44
christopher fglock: Also, I got some regexes to work in the pugs shell (utilizing character classes) that don't seem supported yet in Pugs::Compiler::Rule. I'm confused about where this support is coming from. 02:45
02:45 shachaf joined
fglock so I removed the one in perl5 today 02:45
pugs uses the PGE implementation, which is written in Parrot 02:46
christopher Is the PGE implementation fuller in general? Are you writing P::C::R just to avoid the parrot dependency? 02:47
fglock the implementations are not in sync
there are missing features in both - some things are easier to implement in perl5 02:50
but parrot is more efficient
I hope P::C::R is easier for p5 people to contribute to 02:52
christopher I still haven't traced through the operation of P::C::R, but to implement character classes, it looks like maybe you just need a metasyntax handler for them. Are there any prototypes for metasyntax handlers yet, or would this be the first? 02:57
fglock see Pugs/Emitter/Rule/Perl5.pm - the metasyntax is handled by a big sub in the end of this file; there is a separate implementation for :ratchet in lib/Pugs/Emitter/Perl5/Ratchet.pm 03:00
christopher: do you have a commit bit?
christopher yes, a seldom-used one... 03:01
fglock re: avoid parrot - no, the goal is to have pure-perl6 implementation (the grammar is already written in perl6) 03:02
s/have/have a/
maybe using Pugs::Compiler::Perl6 03:03
P::C::R uses a source filter for $/<0> - the idea is to use P::C::P6 to compile these snippets 03:04
and then rewrite P::C::Perl6 to P6 using itself 03:05
03:15 stennie joined 03:21 _SamB_ joined 03:22 _SamB_ is now known as SamB 03:23 FurnaceBoy joined 03:28 stennie left
christopher fglock: Not sure why I thought character classes didn't work previously. 03:36
But, I see that if I replace "Pugs::Compiler::Regex" with "Pugs::Compiler::Rule" when constructing my rules, the from() and to() on match objects gets screwed up.
fglock christopher: Regex and Rule use different emitters - can you add a test? 03:47
christopher fglock: done. 03:48
fglock thanks!
christopher: this is a bit hard to fix - un-anchored match is a hack :( 03:53
Pugs::Compiler::Regex.pm line 126 03:55
from and to are not updated
03:56 mako132_ joined
svnbot6 r10780 | fglock++ | Pugs-Compiler-Perl6 - imported t/01-sanity; 01-sanity/01-tap.t runs. 03:56
r10781 | malon++ | Duplicate from() and to() tests from P::C::Regex to P::C::Rule
fglock sleep time here - I'll see it tomorrow & 03:58
04:00 fglock left 04:01 bsb joined
miyagawa nothingmuch: unknownplace.org/cpanrecent/rss/author/nuffin 04:12
bsb I just got the the "Ergo.. Everyone hates Ingy" slides... totally cracked me up 04:26
s/the/to/ 04:27
gaal bsb: moose! 04:29
bsb hi gaal 04:37
04:38 justatheory joined 04:47 justatheory joined 04:59 ko1 is now known as ko1_away 05:07 iblechbot joined 05:29 buetow joined 05:34 stennie_ joined
ingy everyone still hates me? 05:46
bsb not after about slide 125 :) 05:48
ingy: I trust you've seen that talk? 05:49
QtPlatypus bsb: Which talk is this? 06:04
ingy the anti-ingy talk!
bsb pugs.blogs.com/talks/npw06-deploying-perl6.pdf
although I've spoilt the punchline
ingy goes to sleep
hated...
bsb no! 06:05
ingy it's ok. be hated is the least of my worries
bsb that can't be good
wolverian who is your greatest worry?
ingy embraces the hatred 06:06
wolverian: the man
wolverian damn I hate that guy. 06:09
ingy all true mongers hate *that guy*
bsb and The Man hates us... 06:10
06:28 szbalint joined 06:40 stennie_ joined 06:59 SM_ax joined 07:37 kane_ joined 07:49 renormalist joined 08:12 marmic joined 08:27 mariuz joined 08:47 rgs joined, drrho joined 08:54 macroron joined 09:29 nine is now known as niner 09:30 niner is now known as niner_, niner_ is now known as niner, niner is now known as nine_r, nine_r is now known as niner 09:38 lisppaste3 joined 09:53 tup joined
renormalist ?eval say "hello" 10:27
damn :)
evalbot_10746 OUTPUT[hello ] Bool::True 10:28
renormalist ah, ok
?eval say 1; say 2; say 3;
evalbot_10746 OUTPUT[1 2 3 ] Bool::True 10:29
renormalist ?eval say for [1,2,3];
evalbot_10746 OUTPUT[ ] undef
renormalist What I'm trying to tell: My pugs doesn't print newlines after "say" if done in a "for" loop. Compare "say 1; say 2; say 3;" versus "say for [1,2,3];". What's the difference? 10:30
Stupid me. It has to be: say for (1,2,3); 10:33
Even "say for 1,2,3;" works. Ignore me. :-) 10:34
10:38 scw joined 11:00 chris2 joined 11:08 Aankhen`` joined
svnbot6 r10782 | audreyt++ | * add concurrency example directory 11:27
11:28 visq joined
svnbot6 r10783 | audreyt++ | * Pugs.Prim: Implement &retry and &Code::retry_with as specced in 11:28
r10783 | audreyt++ | liz's Concurrency Spec Draft.
r10784 | audreyt++ | * examples/concurrency/stm-retry-with.pl - demo from the STM
r10784 | audreyt++ | features "atomically", "retry" and "retry_with". Help needed
r10784 | audreyt++ | to convert that to a proper unit test.
11:30 cochrane joined 11:34 penk joined
svnbot6 r10785 | renormalist++ | - minifix in Overview.pod: elems gives number of elements, not last element 11:37
11:41 meppl joined 11:44 meppl joined 11:53 jsiracusa joined 11:55 bjoern_ joined 11:57 meppl joined
cognominal_ $text ~~ /<?ident>/ && print %hash{~$/}; # I suppose the necessity of a ~ means that a hash keys is not necessarily a string 12:13
audreyt ys
but otoh, in this case {$/} is prolly good enough 12:14
cognominal_ because the default value for a Match in scalar context is a string. 12:15
audreyt yeah.
cognominal_ note that Perl6 is overingeenired... just like is de facto a natural language after centuries of use. 12:16
overengineered...
audreyt well, perl is overengineered... 12:17
cognominal_ that will be a tough sell for people obsessed by regularity
theorbtwo People obsessed by regularity should feel free to put in redundant ~s. 12:18
audreyt for people obsessed with regularity, they know where to find Scheme
theorbtwo ...or use a more regular language.
cognominal_ well. I read a few mail in a recent thread. If the only argument we find for people not groking Perl is "I pity you". we are in a slippy slope. 12:20
QtPlatypus cognominal_: Why?
12:21 visq joined
theorbtwo I rather agree. 12:21
cognominal_ The happy few mentality is a dangerous one.
And it does not help to get a job doing what you like
12:22 elmex joined
QtPlatypus People should use a lanaguge that maps to there brainspace. Some people have unperlish brainspaces. Forcing them to use perl is just as bad as using forcing a perler to use something non perl. 12:22
cognominal_ I am not advocating everyone to jump on our boat either :)
we almost have the language, we probably have to invent the media... 12:23
audreyt The Medium is the Message... 12:24
</McLuhan>
rgs P5 is overengineered too, but almost noone notices
because it keeps dwimming
cognominal_ Like Woody Allen, /me pulls a Marshall McLuhan
p5 has been stretched over any reasonnable limit. 12:25
rgs we don't like being reasonable.
cognominal_ Considering the modesty of Perl1, it just proves that the initial concepts were right.
audreyt initial concepts, you mean like FORMATS? :) 12:26
theorbtwo We do, however, need to find a good niche on the slippery slope where we won't keep rolling downwards.
cognominal_ I don't know any language who started so modest and grew so big in compatible ways
audreyt er, uhm, BASIC
rgs FROMATS
cognominal_ apparently a good way to grow is to get infected by good functional memes? 12:28
audreyt well, anything far from the current gene pool is bound to be good 12:29
or shocking
most likely both
cognominal_ yea, that's my point, the Perl5 communauty was becoming shockingly inbred 12:30
s/Perl5/Core Perl5/
rgs well, evolving perl 5 is near to impossible 12:31
renormalist Sth completely different: Is docs/Perl6/Overview.pod meant to be document how it will work in future or should it match to current Pugs possibilities? (I'm thinking about a regex example in it, that would still need a rx:P5 currently, but not in future, probably)
cognominal_ rgs: tu l'as dit. bouffi 12:32
12:32 buetow joined 12:53 EarleMartin joined 13:01 paultcochrane joined 13:03 paultcochrane joined 13:14 dfan joined 13:19 Limbic_Region joined
Limbic_Region Juerd ping 13:20
audreyt renormalist: huh, rx:P5?
renormalist: you can use real regexes today in both Overview.pod and pugs code 13:21
it's meant to be how it ought to work
Limbic_Region audreyt - pugs uses pcre for p5 regexen correct? 13:22
audreyt yup 13:23
breakfast, brb
Limbic_Region so only those regexen features supported by pcre are available to pugs - do you know if that's @larry's intention for p6?
enjoy ;-)
audreyt eh pcre is a _superset_.
to p5 13:24
I havn't implemented the callback hook
but otherwise it should be a strict superset
13:24 bsb joined
Limbic_Region by callback hook, I assume you mean ? and ?? code constructs 13:25
or /e
audreyt yup
there is no /e
use
s:P5/.../{code}/ 13:26
which is far cleaner
callback hook is actually quite trivial to do
but havn't seen any need yet
Limbic_Region ok - so by superset - you mean that pcre shouldn't choke on anything p5 is able to handle and some things p5 can't - pcre can?
audreyt yeah 13:27
like variable length lookbehind
p5 can't do that
Limbic_Region oh, well features were what I was thinking of - but that's nice too
I was thinking more along the lines of hairy regexen which cause p5 to throw-up
audreyt nod. the bleadperl engine is much better 13:28
at not crashing the stack
Limbic_Region many of which were recently fixed (in bleed perl) when ...
yeah - just what I was about to say
audreyt yeah, but PCRE should handle those fine too
Limbic_Region ok - so do you know the plan for p6?
audreyt the :P5 is specced to support p5 regexen
I imagine PCRE is good enough
on environments where we can't reliably bundled a libperl runtime 13:29
Limbic_Region I know japhy (IIRC) was working on a p5 regex parser which would theoretically allow p5 regexen to be done in rules
audreyt that's a possibility too.
but as I said, it should handle what the p5 unit test suite
tests its regex
at least
Limbic_Region *shrug* - doesn't really matter to me. I am a regex neophyte anyway
audreyt and pcre I think does that.
Limbic_Region just wondering
audreyt k :) 13:30
brb
13:37 christopher joined
Juerd Limbic_Region: pong 13:40
Limbic_Region did you change the hardcoding in cgi:irc intentionally or did something happen I should make you aware of? 13:42
specifcally for #perl6 and #parrot
renormalist damn, now I missed the answers. Real regexes? I'm too slow for Pugs. Is there something special needed during compiling Pugs, eg. PUGS_EMBED="parrot"? 13:45
theorbtwo colabti.de/irclogger/irclogger_log/...2006-06-21 (start at the bottom). 13:46
Juerd Limbic_Region: That depends - what has changed? 13:47
timeout :) afk 13:48
13:49 ludan joined
audreyt renormalist: no 13:52
renormalist: just put parrot in your ENV path
that's all you need to use real p6 regexes
as well as rules and tokens
13:53 amnesiac joined 13:54 rzoz joined
renormalist during compiling pugs or at runtime? I had an old/defunct parrot during my last pugs-compile.In the meantime, parrot is ok. Do I have to Recompile Pugs? 13:54
My initial problem is, that the example regex from Perl6/Overview.pod doesn't work for me: $email ~~ /([^@]+)@(.+)/ while with rx:P5 it does match. 13:57
?eval [email@hidden.address] ~~ /([^@]+)@(.+)/
evalbot_10746 Match.new( ok => Bool::False, from => 0, to => 0, str => "", sub_pos => (), sub_named => {} )
renormalist ?eval [email@hidden.address] ~~ rx:P5 /([^@]+)@(.+)/
evalbot_10746 Match.new( ok => Bool::True, from => 0, to => 15, str => [email@hidden.address] sub_pos => (Match.new( ok => Bool::True, from => 0, to => 4, str => "affe", sub_pos => (), sub_named => {} ), Match.new( ok => Bool::True, from => 5, to => 15, str => "zomtec.org", sub_pos => (), sub_named => {} )), sub_named => {} )
Limbic_Region Juerd - sorry, got pulled away 14:00
audreyt at runtime
Limbic_Region Juerd - it is no longer hardcoded for #perl6 on freenode and #parrot on irc.perl.org
audreyt ?eval 'a@a' ~~ rx/([^@]+)\@(.+)/ 14:01
evalbot_10746 Match.new( ok => Bool::False, from => 0, to => 0, str => "", sub_pos => (), sub_named => {} )
Limbic_Region IOW - it is now possible to select #perl6 on irc.perl.org
audreyt it's just syntax. 14:02
renormalist:
?eval 'foo@bar' ~~ /(<-[@]>+)@(.+)/
Juerd Limbic_Region: That's weird.
evalbot_10746 Match.new( ok => Bool::True, from => 0, to => 7, str => "foo@bar", sub_pos => (Match.new( ok => Bool::True, from => 0, to => 3, str => "foo", sub_pos => (), sub_named => {} ), Match.new( ok => Bool::True, from => 4, to => 7, str => "bar", sub_pos => (), sub_named => {} )), sub_named => {} )
audreyt the charclass syntax changed is all
Juerd I think the dist upgrade changed this back :( 14:03
audreyt instead of [^abc] write <-[abc]>
cognominal_ WOW?
Juerd Limbic_Region: When I have time, I will fix it.
cognominal_ IRL?
oops
Limbic_Region Juerd - well, you told me to let you know if anything about cgi:irc changed
Juerd <[]> sucks so much :(
audreyt [] rocks so much
Limbic_Region no worries Juerd - I know which goes with which
audreyt tradeoffs, tradeoffs
Juerd Limbic_Region: Yep, and it's important that I know this :)
renormalist audreyt: Nize. Works. Should I update Overview.pod to use that? 14:04
Juerd audreyt: Well, the way I would like the regex syntax to be, <> would be available :) 14:05
audreyt renormalist: please do
14:06 fglock joined
svnbot6 r10786 | renormalist++ | - Overview.pod: corrected regex example 14:08
14:10 vel joined
audreyt renormalist++ 14:11
audreyt resumes chglogging and releng
fglock: update ChangeLog to reflect pX state? 14:12
bbiab
fglock audreyt: ok
any idea on how to make p5 'use v6' to call a module?
audreyt use v6-pugs; 14:13
note the dash immediately after
that's all
fglock didn't work the first time I tried - will try again
audreyt try at cmdline
perl -e 'use v9-pugs'
fglock it should emit an error, right? I got no message 14:14
audreyt eh 14:15
fglock perl 5.8.7 here
audreyt you prolly already have v6.pm installed
try change that to v9
or v10
pugs ships with a stub v6.pm at this moment
feel free to change lib/v6.pm to call into lrep
that's what it was designed to do
fglock cool - thanks! 14:16
audreyt so I'm glad pX is at a state where it can become useful
yay
fglock++
fglock did you see hello_world.pl ?
it runs already
audreyt yup :)
so please s/P6/v6/
14:17 rgs joined
audreyt and have fun :) 14:17
&
14:19 stennie_ left 14:26 pfenwick left 14:30 paultcochrane joined, buetow joined, bjoern_ joined, lisppaste3 joined, renormalist joined, kane_ joined, SM_ax joined, alch` joined, kanru joined, buubot joined, pasteling joined, kolibrie joined, evalbot_10746 joined, jabbot joined, locksy joined, revdiablo joined, dvorak joined, takesako joined, obra joined, perlbot joined, xerox__ joined
svnbot6 r10787 | fglock++ | Pugs-Compiler-Perl6 - s/use P6/use v6-pugs/ 14:53
14:56 miyagawa joined
svnbot6 r10790 | fglock++ | Pugs-Compiler-Perl6 - pass 01-sanity/02-counter.t 14:59
stevan audreyt: I could not agree more 15:03
when Perl 6 comes out,.. everyday *will* be like Christmas! 15:04
fglock audreyt: P::C::P6::v6.pm will conflict with Inline::Pugs::v6.pm if both are installed
we could have a main v6.pm interface that dispatches to an implementation 15:09
Limbic_Region what's the svn/svk command to report who made what change? 15:10
clkao blame 15:12
Limbic_Region thanks
svnbot6 r10791 | fglock++ | Pugs-Compiler-Perl6 - pass 01-sanity/03-equal.t 15:14
audreyt fglock: just kill the inline one 15:15
it's just stub.
"svk ann" is shorter than blame
and is more politically correct
"svk praise" if you don't mind typing
fglock: the inline one is a placeholder because PCP was not working
now that it is, simply take it off 15:16
use v6-pugs; # is just fine
to delegate to PCP, that is
fglock should I use Moose as the p5 target, or is this overkill?
audreyt I think Moose is the perfect p5 target 15:17
fglock use v6-moose
audreyt v6-pugs; should simply target moose
that should be implied
fglock ok
audreyt to take advantage of Moosee::Autobox and friends
stevan: glad you like my slides :) 15:18
90% of it is existing material... but seriously rearranged
obra audreyt: I was very amused to see the pieces of that talk
stevan audreyt: it was very cool,.. kind of a greatest hits collection 15:20
the Audrey Tang - Box Set if you will
289 pages of classic slides
audreyt and I maxed out the unicode use :)
stevan :)
rgs waits for the dvd deluxe edition 15:21
stevan will sell rgs one for 6 easy payments of $19.95 USD
plus shipping and handling
1-800-I-WANT-PERL6
operators are standing by
15:22 chris2 joined
stevan call now and we throw in a free Chicken Rotiseree 15:22
fglock do you have a web page?
mmm chicken - lunch time & 15:23
stevan fglock: re: Moose::Autobox - not that the API does not 100% match the Perl 6 API, and it is very experimental and prone to change (most likely to be closer to Perl 6) 15:29
s/- not/- note/
audreyt stevan: oh "subset" makes more sense than "subtype"
because subtype adds capability
and subsets takes away
stevan yeah true
well in Moose it kind of does both 15:31
the subtype is created (adding capability)
but it restricts (taking away)
audreyt uh, no 15:32
in normal OO usage
a subtype is more like adding new behaviour to an existing type's behaviour 15:33
but moose subtypes add no behaviour of its own...
or does it? 15:34
15:34 justatheory joined
stevan well,.. it creates a new constraint 15:34
its kind of odd,.. because there is no existing type system it can actually hook into 15:35
and it cannot be enforced outside of accessors
(well it can,... but its really verbose to do so) 15:36
but the subtype does actually create a meta object
15:38 vel joined
audreyt nod 15:40
stevan moose "subtypes" will never be correct 15:41
since we cannot hook into the existing type system
or actually create our own
(we could abuse tie(), but thats makes it less usable) 15:42
maybe Moose::Autobox will lead to a sane type system overlay 15:46
stevan really needs to start reading TaPL again
I think I have read pretty much every MOP book in existsence at this point
TimToady nm
beware, look what it did to Audrey... 15:47
stevan :)
obra TimToady++
chip good day
TimToady Scheme and Lisp, eh?
I think it's a good direction, under the covers at least.
15:48 visq joined
chip well, it's more just Scheme really. But I hate to dis the Lisp brethren 15:48
TimToady Actually, all you have to do now is port Haskell... :P
PerlJam and make it *fast* :) 15:49
TimToady details, details...
chip Having learned a new verb tense doesn't automatically qualify me to talk about time travel
TimToady chip++
chip a fellow on #scheme wrote a nice description of how unwind-protect could be reasonably defined for Scheme (where "I'm leaving the scope" is not a simple concept), and that clued me in 15:50
Taylor Campbell mumble.net/~campbell/blog.txt most of the content is silly, oddly enough 15:51
audreyt: can you use the equivalent of dynamic-wind productively, and do you agree that the third thunk (for "the scope has been left once and for all") is useful? 15:54
chip actually isn't too sure about that one
PerlJam chip: being an ignorant non-language-implementor-but-having-implemented-one-or-two-in-the-past, it seemed reasonable to me when I read your email. 15:55
chip PerlJam: struck me the same way :-) Thing is, the Scheme primitive only supports the "enter" and "leave" thunks; the third one, for "the last continuation has left the building" is actually kind of the odd one out. For example, if I've been holding a continuation, and through a GC cascade that last continuation is discarded, that event will trigger the "elvis" callback, at an obviously non-obvious time 15:57
it's the closest thing to "finally" you can get when you have all those continuations floating around 15:58
PerlJam I wonder if there are instances in the wild of where Schemers have implemented a work around to get those semantics. 16:01
chip it's easy enough with a closure and GC. Which maybe is my answer :-)
TimToady there's a sense in which that third thunk is the Monadic thunk, in the Haskell sense. It's what controls the side effects that aren't undoable. 16:02
as in a QM computer that's reversible until you try to observe it, the third thunk is the "observer". 16:03
all QM is just quantum foam plus persistent accounting errors. We're the accounting errors... :) 16:04
chip narrowly avoids quoting the Matrix architect's atrociously clunky monologue 16:05
fglock what is the perl 6 way to say: use v6; print 6; use v5; print 5; use v6; 16:06
audreyt ...exactly that? 16:07
chip fglock: system <</usr/bin/perl -e>>, "print 5" ?
chip kids
TimToady druther have something that looks well scoped, like we did with file-scoped packages vs embedded package blocks... 16:08
audreyt clkao: I havn't used the 3rd form of unwind
s/clkao/chip/ 16:09
chip audreyt: there is no third form in the real one 16:10
TimToady package perl5:foo { ... } or some such 16:11
chip audreyt: I introduced it since I figured people would want it
ingy morning
audreyt ingy: hey!
ingy hi audreyt 16:12
TimToady hey ingy, I don't care what everyone says, I think you're okay.
ingy are you going to ms today?
TimToady What's in Mississippi? :)
ingy TimToady: thanks 16:13
16:13 snowstalker joined
ingy a hillbilly monopoly 16:13
16:15 pmurias joined
pmurias hi 16:15
audreyt ingy: yes, probably going to hack CLR<->PIL today 16:16
ingy cool 16:17
pmurias fglock: can i run add-svn-props on PCP6?
16:18 iblechbot joined
audreyt ingy: so how's your sched looklike in the next 72hr? 16:18
ingy I leave for Chicago in 30 hours. free until then 16:19
audreyt hrm 16:22
I don't know... pickmeup and meet for dinner?
ingy at the hotel? 16:23
I remember how to get there... what time
snowstalker Hello all. If I know Perl 5, and know zilch about Perl 6, where can I start to contribute a little? :-)
ingy snowstalker: here 16:24
:)
TimToady the buck starts here, as it were...
snowstalker ingy, Great. At least I started on the right foot ;-)
TimToady, Greetings :) 16:25
fglock pmurias: sure
TimToady welcome!
audreyt ingy: well, actually we can meet the main campus building 35
TimToady biab &
audreyt snowstalker: you have a commit bit, right? ) 16:26
:)
snowstalker audreyt, yep :-D
audreyt ok. talk to fglock here as I'm about to run :)
snowstalker audreyt, although not sure how to use it yet
pmurias snowstalker: a perl5 implementaion can be found in misc/pX
snowstalker audreyt, okay 16:27
pmurias although it's (much) less advanced then the haskell one
fglock snowstalker: hi :)
audreyt ingy: do you know how to drive to main campus?
ingy audreyt: I don't know where that is.
snowstalker pmurias, All I need is perl 5 to run whatever is in misc/pX ? 16:28
ingy msg me an address
snowstalker fglock, Hello :)
audreyt ingy: "1 Microsoft Way"
ingy gah 16:29
fglock snowstalker: cd to Pugs-Compiler-Perl6, and perl -Ilib hello_world.pl (you'll need PadWalker and Parse::YAPP at least, I think)
snowstalker fglock, ok I need to install PadWalker, will start there :) 16:30
fglock you can install Pugs::Compiler::Regex from CPAN, it will get PadWalker
Pugs::Compiler::Rule, actually
snowstalker fglock, ok
ingy audreyt: what time?
audreyt ingy: does 5pm sound about right? 16:31
ingy ok. will you have irc there?
audreyt no... use cellphone 16:32
ingy ok, I'll ring you once and you call back
audreyt 'k
ingy for this job... 16:33
we should have shotguns 16:34
fglock oops - I lost my ChangeLog changes 16:36
svnbot6 r10792 | pmurias++ | fixed line endings 16:42
pmurias fglock: do you like Attribute::Handlers?
fglock: sorry i screw Emitter/Perl5.pm 16:49
svnbot6 r10793 | fglock++ | Pugs-Compiler-Rule swallows Pugs-Compiler-Precedence
pmurias fixing it...
pasteling "snowstalker" at 221.134.203.185 pasted "Pugs-Compiler-Perl6 doesn't compile :-)" (14 lines, 1.3K) at sial.org/pbot/17896
snowstalker fglock, I tried running perl -Ilib helloworld.pl but I get an error, is that normal? It's at sial.org/pbot/17896 16:50
pmurias run Makefile.pl from Pugs::Compiler::Rule
and make
or something similar
fglock please svn up again - I just changed Makefile.PL 16:51
snowstalker pmurias, When I try to install Pugs::Compiler::Rule using command line 'cpan', it says ftp://ftp.perl.org/pub/CPAN/authors/id/F/FG/FGLOCK/Pugs-Compiler-Rule-0.05.tar.gz 16:52
LWP failed with code[404] message[File 'Pugs-Compiler-Rule-0.05.tar.gz' not found]
fglock, oh okay
16:54 penk left
snowstalker fglock, I'm at rev 10794, but still see the same error... 16:54
kolibrie fglock: I'm getting the same on feather in my homedir
fglock checking... 16:55
kolibrie wonders if it is because perl5 is trying to use Pugs/Grammar/Rule before a .pmc of it has been made 16:56
fglock perl Makefile.PL / make should take care of the .pmc file 16:57
it works here
kolibrie fglock: so first go into Pugs-Compiler-Rule and perl Makefile.PL
and then perl -Ilib hello_world.pl ? 16:58
fglock yes, in PC-P6 - but pmurias needs to commit a fix :)
kolibrie twiddles his thumbs 16:59
snowstalker did what kolibrie said but sees new errors such as "use" not allowed in expression ... 17:00
kolibrie snowstalker: this is all bleeding edge, btw 17:01
17:01 penk joined
kolibrie fglock: is that what pmurias is fixing? 17:02
snowstalker kolibrie, I get that now. It's just that Day 1 in this p6 world is confusing to me :). Still wondering what a .pmc file is ... etc , Can you tell me which docs I should be reading for all of this?
fglock yes - a svn message slipped into the source file
kolibrie snowstalker: this particular hello_world.pl was only written yesterday 17:03
svnbot6 r10795 | fglock++ | Pugs-Compiler-Rule - TODOize 2 tests
fglock snowstalker: .pmc is a perl5 reserved file type for "compiled perl"
kolibrie snowstalker: hmm, the synopses have a lot about Perl 6 the language 17:04
fglock the perl6-to-perl5 compiler takes a .pl source and generates a .pmc, which contains perl5 code
kolibrie snowstalker: a lot of the pugs, particularly the pX part, are not documented so well, and change from week to week. 17:05
audreyt snowstalker: btw, add yourself to AUTHORS
kolibrie snowstalker: hanging out here helps, and reading pugs.blogs.com 17:06
snowstalker kolibrie, Yesterday? Damn! That *is* cutting-edge
fglock, what does the "mc" stand for ?
fglock fixed
kolibrie snowstalker: .pm (perl module) -> .pmc (perl module compiled) 17:07
svnbot6 r10796 | fglock++ | Pugs-Compiler-Perl6 - lib/Pugs/Emitter/Perl6/Perl5.pm fixed
fglock maybe "perl module compiled"?
snowstalker kolibrie, I'm looking at svn.openfoundry.org/pugs/docs/getting_started
fglock please retry hello_world.pl
snowstalker audreyt, but I haven't even done anything yet :)
kolibrie, aah ok
fglock, ok
pmurias flock: you were faster with the commit :) i got an merge message from svk 17:08
s/flock/fglock/
fglock pmurias: no problem :) 17:09
audreyt snowstalker: that's a good way to test commit
snowstalker fglock, yippee, It works :)
audreyt customarily the first checkin
snowstalker audreyt, oh okay :)
fglock snowstalker: look at the hello_world.pmc file
snowstalker fglock, btw it said unrecognized token '' 17:10
at pugs/misc/pX/Common/Pugs-Compiler-Perl6/lib/Pugs/Grammar/Perl6.pm line 14
pmurias fglock: what tests are supposed to pass
? 17:11
fglock that's just a warning - it had many more yesterday :)
snowstalker fglock, so it converted p6 to p5, if i understand it correctly ?
fglock, hehe
fglock the 3 first tests in 01-sanity should pass
snowstalker: yes!
snowstalker fglock, gr8
fglock try running hello_world.pl again - it will reuse the .pmc file 17:12
snowstalker fglock, yeah it looks like it did because it didn't give the warning again 17:14
audreyt fglock: want to carry your journal over or mirror it at pugs.blogs.com?
fglock audreyt: which you find better 17:15
17:15 bernhard joined
audreyt either is fine but I'd like you (and also any interested #perl6'ers0 to cowrite pugs.blogs.com 17:15
a journal bit is in your inbox now :) 17:16
fglock ok!
svnbot6 r10797 | snowstalker++ | Adding my name here because audreyt told me to ( colabti.de/irclogger/irclogger_log/...l=340#l550 ) :-) 17:19
fglock audreyt: is it ok to remove the shebang from .t files? it tells p5 to run pugs...
or is there a workaround?
audreyt killall #! lines i'd say 17:22
or change them to a neutral #!perl6
pmurias fglock: why does generating the .pmc take so much time? 17:25
fglock pmurias: it is using the old backtracking rules - :ratchet can't run all the syntax needed yet 17:26
audreyt: ok
pmurias what else is needed in :ratchet? 17:27
fglock hashes and arrays - maybe some extra details 17:28
currently ratchet and non-ratchet rules have problems calling each other, but that's not a problem if you convert everything to non-ratchet 17:29
17:29 frederico joined 17:30 bsb left
snowstalker Are the HTML versions of the synopses at dev.perl.org/perl6/doc/synopsis.html kept up to date with the svn versions ? 17:33
audreyt yup 17:35
17:37 amnesiac joined
ingy audreyt: gnip 17:42
ingy finally clues in on the meaning of Gnip Gnop (after 30 some years) 17:43
TimToady snowstalker: they should be at most 6 hours out of date, last I knew. 17:44
17:45 Southen joined
audreyt ingy: gnop 17:45
ingy: maybe 5:30p is better
(I'm just about to leave hotel now)
going to hack on Microsoft Phoenix it seems
that codename seems to be popular 17:46
snowstalker audreyt, TimToady, ok cool
audreyt hopefully it will not get changed to Microsoft Firebird ;)
ingy audreyt: maybe just call me when you get back to the hotel room
instead
pmurias fglock: my $fixity = $n->{fixity} doesn't seem to work when put in _emit
audreyt ok ok.
ingy we don't have to bind on dinner or else 17:47
but if you feel like hanging out, I'll gladly come over
ingy attempts to minimize stress 17:48
fglock pmurias: some nodes don't have a fixity
ingy audreyt: I wanted to ask you a q, I msged you over to the right
audreyt I didn't see it 17:49
try again
17:50 Ymmv joined
pmurias it is undefined for *all* nodes 17:51
i implemented a hash based dispath for those that have
but can't get it to work because of that 17:52
looked at P::E::R::P5, there's pretty quoting in there 17:55
svnbot6 r10798 | fglock++ | Pugs-Compiler-Perl6 - implemented 'if'; but parser error in 04-if.t
pmurias it could be cleaned up with ruby style interpolation 17:57
fglock pmurias: example?
pmurias eq. "#{$var}"
or 17:58
"#{expr}" more generally
fglock I'm ok with that 17:59
17:59 FurnaceBoy joined
pmurias will do it later 18:00
ingy fglock: ping 18:01
pmurias bbiab
fglock ingy: pong 18:02
ingy why is P-C-R/{Makefile|blib|pm_to_blib} checked in? 18:03
I'm getting conflicts on them
fglock probably by mistake - please remove them 18:04
thanks
ingy ok
nothingmuch ingy: møøø 18:05
merlyn complained about MC btw
he said that i should have written 'use tt' with Inline
ingy he's wrong
nothingmuch yeah, i told him so
but he does raise a small issue
ingy Inline will use MC
:)
nothingmuch =)
perlmonks.org/?node_id=556022
"which already has support for parameter passing" 18:06
ingy so eventually he'll be right
nothingmuch that thing
no he won't, 'use tt' is not introducing foreign functions in a nice way
but meh, details
anyway, can we iron out the way params are passed on the use line?
also, if i write a "touch_my_pmc" script, that takes a tarbsall and increments the mtime of the PMCs, would you roll it into the dist? 18:07
that really confused me for a short while
ingy sure 18:08
params can only be passed as a string...
nothingmuch due to the filter system's limitations?
ingy unless it is the *first* compiler module used
nothingmuch oh 18:09
can you explain why?
ingy yes
you see my boy...
the whole MC process is invoked by the first caller (which coincidentally gets passed a list of params)... 18:10
MC then plays god and parses the rest of the module and calls all the compilers and stitches up the result. so Perl mever even calls the other use lines. 18:11
nothingmuch ... because the moment 'use SomeFilter' is done it gets the rest of the file?
it can't say where to stop filtering? 18:12
ingy no not really
because it needs to filter everything to create a pmc 18:13
nothingmuch hmm
because otherwise the pmc thing won't know when it's done? 18:14
fglock bbiab &
ingy well the filter really throws away all the lines (up to the data section) 18:15
and rereads the file from disk
nothingmuch oh
ingy because it even needs the lines *before* it was invoked 18:16
to be part of the pmc
18:16 DaGo joined
nothingmuch hmm 18:17
conjecture:
you open the file and keep it in some closure var
ingy then after compilation it chops off the head number of lines from the result and passes the back to FUC
nothingmuch import/filter_add and filter_del are used for 'use' and 'no'
and another filter after that dups everything the the pmc file 18:18
filling the gaps?
i'm really scared about low DWIM factor for 'use some_filter $global'
so it might be worth it =/ 18:19
ingy well we could warn if we see $ in the params :P
nothingmuch okay, so use some_filter foo()' 18:20
it's endless
give me 5 mins
and i'll get back to you with an idea
18:20 ludan joined
ingy ok 18:20
18:21 nxu7 joined 18:29 qdlllbp joined 18:33 OuLouFu joined
fglock pmurias: re time to generate .pmc - it compiles the whole p6 grammar every time 18:40
there is no grammar cache currently
svnbot6 r10799 | fglock++ | Pugs-Compiler-Perl6 - pass 04-if.t 18:46
18:46 ruoso joined
gaal nothingmuch++ # random beingamooseness 18:47
fglock nothingmuch: does Moose aim to approximate the p6 grammar? 18:52
nothingmuch whores karma 18:57
fglock: not the grammar, but the feel
has $.foo is rw;
has foo => (
is => "rw",
);
ingy hi nothingmuch
did you come up with anything?
nothingmuch hi ingy
i think so
but phone confused me
either way it's too sick to do now
we'll have a chat at the conf
and then we'll implement it like the javascript hashcash idea 18:58
ingy I think for now, I should just document how it works
that it doesn't take a live list of args 18:59
mostly what I'd envisioned was that it would take no args.
use this; no that;
etc
that was the spirit. 19:00
19:00 miyagawa joined
ingy because these lines are compiler invocations 19:00
you could even argue that they shouldn't take live args
because those will change from invocation to invocation 19:01
but the compile should always be the same
nothingmuch heh
the womenbof thrtead is silly
ingy MenBOF!!! 19:02
nothingmuch well
BEGIN time is vague in Perl
but i see what you mean
that said
qw/..../ is nicer than ["...", "..."]
so i want some hack for TT
ingy either one you have to parse though... 19:04
we can talk more at yapc...
nothingmuch yeah
19:22 justatheory joined 19:48 ZuluOne joined 20:12 audreyt_ms joined
audreyt_ms Greetings from behind the restrictive firewall :) 20:12
ZuluOne evening 20:13
xerox The Great Wall? 20:14
audreyt_ms the MS corpnet firewall, as it is
Limbic_Region how much time did it figure out how to get here from there? 20:17
nevermind - you are cgi:irc ing
audreyt_ms yup :) 20:19
so, no effort at all
Limbic_Region oh, out of curiosity - was your entry into the US any less problematic this time? 20:20
audreyt_ms was actually very easy 20:21
though they still did a full baggage check
they are certainly much more polite :)
Limbic_Region glad to hear it - perhaps you might not be so apprehensive about visiting in the future 20:22
audreyt_ms we'll see :) 20:23
but yeah, perhaps I am less terrorist-looking now
Limbic_Region well - more likely, your reservations didn't involve any last minute changes but you didn't hear me say that 20:25
or type it for that matter
audreyt_ms hm, last time I didn't change it at last minute either... 20:27
20:28 Vutral joined, alch` left 20:36 dolmen joined 20:45 kanru joined
svnbot6 r10800 | audreyt++ | * remove our v6.pm to not clash with the one in misc/pX. 20:50
r10800 | audreyt++ | (eventually I'd like Pugs to install that one by default)
20:52 cmarcelo joined 20:55 fglock left
svnbot6 r10801 | fglock++ | Pugs-Compiler-Perl6 - 'sub' fix 20:55
20:55 fglock joined
obra Who's resposnsible for all the wonerfulness that is the pugs smoke server? 20:56
leo obra: $ svn blame util/smokeserv/smokeserv-server.pl 21:05
I think
obra sorry. "the whole set of tools"
leo e.g. search.cpan.org/~nuffin/Test-TAP-HT...pm#AUTHORS 21:07
search.cpan.org/~gaal/Bundle-Pugs-S...mokeKit.pm 21:08
see also smoke.parrotcode.org/smoke/ ;) 21:09
21:10 fglock joined
audreyt_ms fglock: I just checked in a new, easier-to-manage v6.pm 21:11
obra We've been doing some fun internal stuff to get ourselves a pugs-based smokeserver
pmurias g'night
svnbot6 r10802 | audreyt++ | * add a new v6.pm that uses Module::Compile :)
fglock audreyt_ms: thanks! did it work well for non-modules? 21:13
21:16 elmex joined
audreyt_ms fglock: yup 21:17
I wonder if instead of .pmc 21:18
having it as .pl.compiled
or .plc
because you could've have another same named .pm in the dir
sounds sane? I'll check it in 21:19
fglock sure
how about .t? .tc?
audreyt_ms see commit msg :) 21:20
(yes)
svnbot6 r10803 | audreyt++ | * instead of .pmc, use .plc for .pl files, .tc for .t files, etc 21:21
r10804 | audreyt++ | * v6.pm: localize $@ before entering p5eval context.
21:22 nxu7 left
nothingmuch www.scienceblog.com/cms/thirst-for-...10849.html 21:23
audreyt_ms Interestingly, the VB->PIL->Javascript plan will give Pugs a native code compiler backend as well 21:35
svnbot6 r10805 | Sage++ | First version of translation program
audreyt_ms the same one as used by MSVC
and it's even compatible with Rotor
TimToady Sage++
audreyt_ms (which means BSD license all the way)
ooh Sage 21:36
audreyt_ms checks it out
ZuluOne from nothingmuch's link, "mate selection correlates closely with perceived intelligence" 21:41
svnbot6 r10806 | audreyt++ | * translate.hs - add svn properties; fix the nested if/else
r10806 | audreyt++ | to use a proper case expression, as requested in comment
21:44 penk joined
svnbot6 r10807 | lwall++ | s/oeprator/operator/ 21:47
r10808 | fglock++ | v6.pm cleanup
r10809 | audreyt++ | * translate.hs - stylistic Haskell cleanup; use Parsec "try" instead 21:56
r10809 | audreyt++ | of fully qualified name; use case expressions instead of if/else
fglock does Moose has subroutine parameter binding or signatures?
spinclad is MSIL <-> PIL indeed going to go both ways? 21:57
audreyt_ms actually it's MSIL <-> CIL <-> PIL now 21:58
spinclad a crossover at which to trade front- and back-ends
audreyt_ms CIL being the MSVC intermediate language
it's far easier to work with
MSIL too baroque
and CIL is common denominator so to speak
as it's closer to GIMPLE
svnbot6 r10810 | fglock++ | PC-P6 - sub declarations; no parameter list yet
audreyt_ms also it has this wonderful Grand Unified Retargettable Language thing 21:59
which makes parsing and emitting object languages (like PIR, MSIL assembly, etc) trivial
so it's a nice midpoint
probably going to write Haskell bindings using cpphs
TimToady GURL power... 22:00
audreyt_ms indeed
going to do some tree comparison in the design meeting next hour
spinclad would PIL -> CIL -> PIR make sense in the next go around then?
audreyt_ms not sure yet
fglock audreyt_ms: there is a big community around Mono here - it would probably be a nice talk in CONISLI 22:01
audreyt_ms people here are not using CIL at all really, it being a MSVC technology instead of a CLR one
where "here" means "here in the SQL building with CLR/VB folks"
spinclad and BSD licensed? amazing, 22:02
audreyt_ms so trying to gather more info in the meantime
spinclad s/,$/./
audreyt_ms Rotor is BSD licensed; the CIL is apparently open as an IL
the actual tools -- not clear at the moment -- certainly the entire Phoenix chain is noncommercial shared-source -- which is not good enough -- but the Hs binding itself can be BSD 22:03
will report back in a more coherent fashion after meeting
fglock looks for a p5 module that implement named parameters 22:04
audreyt_ms Params::Validate? 22:05
22:05 audreyt_ms is now known as audreyt_ 22:09 Ymmv joined
fglock I mean, at the syntax level 22:09
there is Perl6::Parameters, but it is a source filter 22:11
22:13 Vutral joined 22:18 FurnaceBoy is now known as FB|afk
svnbot6 r10811 | fglock++ | PC-P6 - stub routine for param binding 22:32
Juerd There was one that hijacked prototypes iirc
Can't find it 22:35
svnbot6 r10812 | fglock++ | PC-P6 - fixed parameter list AST 22:44
22:44 fglock left 22:46 macroron joined 23:11 kattana_ joined 23:14 penk joined, cmarcelo joined, ruoso joined, OuLouFu joined, jsiracusa joined, scw joined, jiing joined, Zemyla joined, Kattana joined, TimToady joined, cognominal joined, qmole joined, viirya joined, knewt joined, Grrrr joined, amv joined, rw-rw-r-- joined 23:16 justatheory joined 23:28 cmarcelo_ joined 23:33 cmarcelo_ left 23:36 FB|afk is now known as FurnacePissy 23:43 cognominal_ joined 23:46 Vutral joined 23:52 shachaf_ joined 23:57 dvorak joined

Powered by Cro and the Raku® Programming Language.

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

Thank you!