geoffb's column: www.oreillynet.com/pub/au/2333 | Pugs 6.2.10 coming soon | pugscode.org | pugs.kwiki.org | paste: sial.org/pbot/perl6 | www.geeksunite.net
Set by autrijus on 5 October 2005.
autrijus ok, "make test" should yield 100% now. 00:03
consider this a preliminary preflight
still awaiting rafl
and I can't stay up any longer
clkao i am doing smoke-js 00:04
autrijus cool, clkao++
clkao go for another sauna?
autrijus we'll release 6.2.10 tomorrow.
autrijus waves &
nah too tired.
scook0 autrijus: oh, I see my patch broke Prelude--I had a feeling something would go wrong 00:15
I'll look into it
autrijus cool. 00:18
journal up -- g'nite!
scook0 bye
dduncan I'm back ... but about to leave again ... I'll svn up in a few hours and smoke with precompiled Test.pm and lib.pm 00:24
svnbot6 r7454 | stevan++ | Perl6-MetaModel - 02:02
r7454 | stevan++ | * adding 2 pictures :
r7454 | stevan++ | 1) normal method disptaching
r7454 | stevan++ | 2) method dispatching with eigenclasses
02:43 Cryptic_K is now known as K
scook0 autrijus: I have a revised version of my Parser patch, but I might wait until after 6.2.10 in case it breaks other stuff 05:10
svnbot6 r7455 | qtplatypus++ | Adding a simple DFA. This is in preperation to build a simple regex 05:40
r7455 | qtplatypus++ | implementation as a work around for "Noting having PGE" installed problems.
QtPlatypus ?eval my %hash = {foo => "bar",baz => "buz"};%hash{"foo"|"baz"} 05:47
evalbot_7455 [undef]
QtPlatypus ?eval my %hash = {foo => "bar",baz => "buz"};$temp="foo"|"baz";say %hash{$temp} 05:48
evalbot_7455 Error: Undeclared variable: "$temp"
QtPlatypus ?eval my %hash = {foo => "bar",baz => "buz"};my $temp="foo"|"baz";say perl %hash{$temp}
evalbot_7455 \undef bool::true
QtPlatypus Damn. 05:49
gaal morning 05:59
QtPlatypus gaal: Good morning.
gaal we've moved to winter time :( 06:00
QtPlatypus Summer time for me in a few weeks. 06:01
gaal and the livin' is easy
dduncan I'm back 06:09
gaal wb 06:17
QtPlatypus ?eval Bool::True 06:35
evalbot_7455 Error: No compatible subroutine found: "&Bool::True"
QtPlatypus ?eval bool::true 06:36
evalbot_7455 bool::true
QtPlatypus Is there somewhay to tell perl to recurse deeper? 06:43
scook0 QtPlatypus: recurse deeper? 06:44
QtPlatypus There seems to be a limmit to the depth of structures that .perl will give you 06:45
?eval ([=>] (1,2,3,4,5,6,7,8,9,10,11,12,13,14,15)).perl 06:46
evalbot_7455 '(1 => (2 => (3 => (4 => (5 => (6 => (7 => (8 => (9 => (10 => (11 => (12 => (13 => (14 => 15))))))))))))))'
QtPlatypus mmmsss.
Its spitting out (... => ...) for me in some situations 06:50
gaal QtPlatypus: backlog a few days, this was discussed here 06:55
i think autrijus changed the limit to 1024 or something
svnbot6 r7456 | gaal++ | use parametric $(MAKE) instead of assuming `make'. Should fix build for some systems. 06:56
gaal here's a musing. now that we have auto smoke uploads, i want to tag each smoke matrix with "clean source" or "has local mods" (and, for bonus points, a list of the locally changed files). 06:59
not sure where to get that information, though. the easiest way is to check mtimes on source files
but that's obviously not correct. 07:00
also, we probably don't want to maintain a list of interesting files manually
and finally, i don't want to couple this too tightly with any source control system :) 07:01
QtPlatypus makes a request for the lambda camels to build some debugging tools.
gaal there's alread a debugger in perl6, QtPlatypus
only it's not for perl6 :) 07:02
examples/obfu/l33t.p6 :-p
QtPlatypus looks puzzled.
gaal i'm just being silly :)
jp-autark Hm, I'm still getting build errors with a fresh checkout of 7455 07:04
*** Wasn't able to find 'Syck_stub.o', aborting... 07:05
I have Syck_stub.o in ./dist/build/src/Data/Yaml/Syck_stub.o
nothingmuch morning 07:06
gaal hola and aloha nm 07:07
nothingmuch discovered, to his great shock, that he is actually neat and tidy
scook0 howdy nothingmuch
nothingmuch my family went away for holiday
gaal alek
nothingmuch and I stayed due to lack of vacation days from work
the house is clean
the kitchen is spotless
i ironed my shirt
s
gaal really? you should come visit more often. 07:08
nothingmuch there is no hairbrush hair or toothpaste around the sink in the bathroom
no mess has been made of the living room
and the real issue is that I'm usually blamed for this stuff when they're home1
nothingmuch has shift key sync problems
mmm... someone brought nice chocolate to work 07:11
gaal you're already at work? doing good with the clock change i see. 07:13
nothingmuch didn't I gain an hour? 07:15
24h ago it was 10:15, right??
nothingmuch considers porting Path::Class instead of File::Spec 07:24
and assimilating File::Spec into Path::Class
gaal nothingmuch: yes.
off to work &
nothingmuch ciao 07:25
it should be delightful with $path.=subdir("foo"); 07:26
autrijus wakes up to drink something and sleep again 08:10
scook0: please don't hold off patches... you can do a "make test" :)
if you see parsefails, then you know that your Parser.hs refactoring breaks things 08:11
so you'd know where to look; otherwise it's as safe as it can be and please commit ahead.
bbiab& 08:12
scook0 autrijus: "make test" would take my box a few weeks :)
autrijus you don't have a feather account? :) 08:13
nothingmuch scook0: if you'd like you can test on my box
or probably on feather
autrijus ah, there you go
autrijus goes back to sleep
scook0 well, I'm currently hunting down a subtle bug in it 08:14
there's a temporary workaround, but I'd rather do a proper fix
autrijus *nod* take your time :) scook0++
nothingmuch scook0: /msg me if I can help CPU power wise... irssi is in the BG
scook0 ooh, I think I found the bug! 08:15
apparently `foo($bar, $baz,)` is legal 08:16
building... 08:17
meppl guten morgen 08:21
scook0 ooh, it works! 08:22
nothingmuch hates the fact that in order to write multitasking code in perl 5 you have to think too much up front 08:24
i wish we had lightweight threads that behaved like closures
rep well, concurrency is not easy 08:25
(in any language)
nothingmuch right
but this stuff is really not complex concurrency
it's separate parts
that must run at the same time
and report to the same parent
threads introduce too much of a headache (all data must serialize properly, can't rely on e.g. Tie::RefHash for proper identity keying, etc) 08:26
the issue is that you not only have to design your app
but you also haave to figure out how you are going to implement everything at the lowlevel
like what tradeoffs you're willing to make to work with forks 08:27
or whether you can fake concurrency by using an event loop
if your code has good enough latency
scook0 nothingmuch: care to test my patch for me? 08:28
nothingmuch scook0: sure... nopaste a unidiff
pasteling "scook0" at 220.253.96.162 pasted "Parser patch" (109 lines, 4.5K) at sial.org/pbot/13573 08:30
nothingmuch full test suite?
scook0 I don't know 08:31
whatever you're happy with
nothingmuch okay 08:32
scook0 it shouldn't be changing any behaviour 08:33
nothingmuch just see that it compiles?
scook0 (at least any behaviour that was already working)
oh, it compiles 08:34
but last time I committed it, it caused parsefails in Prelude
it runs Prelude fine now
I'm just not sure if it will break anything else
nothingmuch fsck 08:39
nothingmuch 's puter made a weird sort of freeze-ish thing
scook0 eep
nothingmuch i wish osx had virtual consoles
the player kept making noise
and the UI sort of responded
i think the UI server crashed or something
you could move the mouse, but not click anything 08:40
and the cmd+tab app switcher thingy was frozen in place
and now i forgot if I applied your patch already or not ;-)
seems like I did 08:41
scook0 's computer semi-freezes due to thrashing every time he builds pugs 08:42
nothingmuch ah, that I don't get... =)
scook0 which sometimes crashes nautilus :(
nothingmuch that's not nice
nothingmuch bought plenty of ram 08:43
src/Pugs/Parser.hs:1734:0: Warning: Defined but not used: `parseParamList'
scook0 the fact that I'm running GHC + Gnome + Firefox with 256m probably doesn't help
nothingmuch is that supposed to be there?
hehe
scook0 oh, I'll take it out when I commit
the code that used to call that now calls my new code 08:44
nothingmuch ah
scook0 thanks for telling me though--I'd forgotten to take it out 08:45
nothingmuch dum dee dum... still compiling
nothingmuch likes Schwern even more after the recent Test::Builder::Tester issues =) 08:53
use.perl.org/~schwern/journal/27075
scook0 so he made an update to a test module 08:56
which then broke other test modules
nothingmuch right
scook0 which broke large parts of CPAN
wow
nothingmuch Mark Fowler was sick so he took over T::B::T
now it's being fixed
overall it was handled quite well
dum deee dum 09:00
i think it's been compiling Pugs.Run for the past few minutes now =)
perhaps I shouldn't have built Test.pm into it ;-)
scook0 I remember once my compiles were taking heaps longer then usual 09:02
so I looked in Makefile.PL 09:03
and found that somebody had accidentally made `make unoptimised` and alias for `make optimized`
nothingmuch i'm making optimized =)
scook0 & 09:05
nothingmuch ciao
pdcawley Does pugs still need haskell plugins installed? 09:10
nothingmuch pdcawley: no, that's only for eval_haskell('...') last I heard 09:11
pdcawley Something I'm unlikely to do in the immediate future. 09:12
By gum, it's been a while since I brought pugs up on this box.
nothingmuch did you get the compilation stuff fixed?
pdcawley I hope so; got the ghc binary package installed
I'm about to find out :) 09:13
nothingmuch =)
pdcawley It's even been a while since I last did an update of parrot. 09:15
nothingmuch how can you live with yourself? 09:16
pdcawley Easily enough.
nothingmuch okay =)
pdcawley I've been doing Ruby On Rails stuff in my programming life.
nothingmuch if I had more time I would probably learn that 09:17
pdcawley Ruby's not as nice as Perl 6, but it has the massive advantage of being here now.
And, as I don't know either C or Haskell...
Well, I *read* C. But I don't write it.
nothingmuch as you don't know them... ?
pdcawley I can't really go about implementing much of either Parrot or Pugs. 09:18
nothingmuch pugs can use lots of non c knowlege
what about a PIL backend on ruby to help both pugs and your ruby learning? 09:19
i think that would be uber cool
pdcawley Yeah. I might well come back to that; hence the getting pugs built again.
nothingmuch write with rails in perl 6
pdcawley :)
Oh. Bollocks. Build failed.
RTLD_NEXT undeclared.
nothingmuch it's been a while since I've seen that... let me see if I can remember 09:20
can you nopaste more of the braindump?
pdcawley Ah... I think I may have gcc_select 4.0 in power.
nothingmuch ghc 6.4 or 6.4.1? 09:21
pdcawley 6.4.1
nothingmuch i thought that was supposed to be fixed =(
pdcawley And I didn't do a make clean, so starting again.
nothingmuch scook0: tests seem to be passing
scook0 nothingmuch: cool 09:22
I just have to finish cleaning the kitchen &
nothingmuch t/builtins/caller..................................***
unexpected "c" 09:23
expecting ";", statements or "}"
at t/builtins/caller.t line 12, column 19
t/builtins/caller..................................dubious
is that your fault?
scook0: ?> 09:25
scook0 hmm
maybe
I'll look into it
does it work in normal pugs?
nothingmuch let me see
scook0 ?eval sub b_sub { try { caller.subname } } 09:26
evalbot_7456 undef
scook0 must be the patch
pdcawley Oh bollocks, bollocks and bollocks some more. Still throwing the RTLD_NEXT error.
nothingmuch scook0: yes, on the other machine it works
scook0 &
nothingmuch pdcawley: did you relclean and regenerate the gcc
pdcawley Did a make clean. Isn't that enough? 09:27
nothingmuch regenerate the makefile
pdcawley Yup, did that.
nothingmuch got confused (i was talking to someone else while typing)
realclean or distclean ought to fix it
pdcawley K
Shame there's not a 'make realclean_but_keep_makefiles' 09:28
nothingmuch that wouldn't be really clean, would it? =)
pdcawley Well, it depends what you're trying to clean.
make clean_generated_c_files ? 09:29
nothingmuch i think what you really want is 'make optimized_but_really_quickly_without_making_a_fuss'
pdcawley Bollocks!
nothingmuch buys some suspiciously cheap tea (cgi.ebay.com/ws/eBayISAPI.dll?ViewI...400323366) 09:30
scook0: more parse barfs, please wait
t/oo/attributes/instance.t parses on regular pugs, but not with your patch 09:31
pdcawley distcleans this time.
pasteling "nothingmuch" at 212.143.92.226 pasted "./pugs -Iext/Test/lib t/oo/attributes/instance.t output" (4 lines, 411B) at sial.org/pbot/13574 09:32
pdcawley And more bollocks. 09:36
nothingmuch pdcawley: =(
do you have an account on feather?
pasteling "nothingmuch" at 212.143.92.226 pasted "more parse fails for scook0: t/operators/operator_overloading.t" (4 lines, 200B) at sial.org/pbot/13575 09:38
scook0 I see a common theme 09:42
it doesn't like `foo.bar`, `Class.bar` or `Class.bar()` 09:43
nothingmuch i'm glad testing is helping =)
scook0 nothingmuch: thanks 09:44
I can see the general shape of the problem
but I'll need to do some more investigating
pdcawley Yes. But I'm going to try a build on the laptop first. 09:45
Which seems to have a clean ghc installation.
No, I don't know why.
nothingmuch pdcawley: okay... if you need an OSX that works I can give you ssh to $home_box
scook0: it still hasn't finished the test suite
pdcawley It's a local box that works I want more than anything.
But thanks for the offer. 09:46
nothingmuch scook0: out to lunch - will report more later 09:49
scook0 ok
?eval class Foo {}; Foo.new 09:52
evalbot_7456 Foo.new();
scook0 ?eval foo.bar 09:53
evalbot_7456 Error: No compatible subroutine found: "&foo"
scook0 ?eval sub foo {}; foo.bar
evalbot_7456 Error: No compatible subroutine found: "&bar"
scook0 ?eval sub foo {}; sub bar {}; foo.bar
evalbot_7456 undef
scook0 ?eval sub foo {}; sub bar { say $_ }; foo.bar
evalbot_7456 Error: unexpected "b" expecting block construct, ":", term, term postfix, operator, postfix conditional, postfix loop, postfix iteration, ";" or end of input
scook0 ?eval sub foo {}; sub bar { say @_ }; foo.bar 09:54
evalbot_7456 undef
scook0 ?eval sub foo {3}; sub bar { say @_ }; foo.bar
evalbot_7456 3
scook0 ?eval sub foo {3}; sub bar { "A: @_[]" }; foo.bar
evalbot_7456 3
pdcawley How is foo.bar legal in those contexts? 10:04
rep win 13 10:06
oops
svnbot6 r7457 | autark++ | * Don't abort when Syck_stub.o is in the correct directory. 10:12
jp-autark is linking with parrot-0.3.0 supposed to work at this time? 10:13
scook0 pdcawley: I don't think any of those is *supposed* to work 10:18
`foo.bar` should mean `foo().bar()` 10:19
and `MyClass.new` should mean something like `(::MyClass).new()` 10:20
but I'm not yet sure how pugs currently makes those work
nothingmuch: when the test suite finishes, could you email me the results as a file? (@gmail) 10:22
QtPlatypus ?eval multi sub foo(bool $test) { say "Match" };foo(bool::true); 10:33
evalbot_7457 Match bool::true
QtPlatypus ?eval multi sub foo() {say "MissMatch"};multi sub foo(bool $test) { say "Match" };foo(bool::true); 10:34
evalbot_7457 Match bool::true
svnbot6 r7458 | qtplatypus++ | Can now mark some state as "Final". 10:42
QtPlatypus gaal: Can you check if Mine::Base64 still hangs? I think I've resolved the issue, but since my copy never hanged I don't know if that is the case. 10:47
svnbot6 r7459 | qtplatypus++ | Resolving conflicts with the ruleless version of Base64. Test still 10:48
r7459 | qtplatypus++ | diked out untill I can confirm there non hanging state.
rafl_ autrijus: What are gmakeisms? :-) 10:51
autrijus rafl_: for foo in perl5 :) 10:54
rafl_ Oooch.. why? :-) 10:56
10:56 Lopo_ is now known as Lopo
autrijus rafl_: because nmake does not understand that. 10:56
rafl_: so, you going to check in the privlib changes? 10:57
rafl_ Yes, but I'm going to write them first.
autrijus cool. I'll adjust the hs side now
rafl_ How will the hs side know where possible locations are? Do we set something in PugsConfig? 10:58
Whops: 11:01
Interface file inconsistency:
home-package module `Pugs.AST.SIO' is mentioned,
but does not appear in the dependencies of the interface
ghc-6.4: panic! (the `impossible' happened, GHC version 6.4):
forkM Declaration for retError{v}
wolverian I just love errors with "panic" and "the impossible happened" in them 11:02
rafl_ I don't for some reason.
autrijus rafl_: we already have access to privlib and sitelib so I'll check both 11:03
rafl_ autrijus: OK.
autrijus rafl_: hm, did you restart the build?
incremental nonclean builds in ghc 6.4 is known to be troublesome
rafl_ Hrm, I hate waiting for a full build.. 11:04
OK, I'll be offline until the build finished now.
autrijus k. 11:07
meppl *g 11:19
rafl_ autrijus: You said you reverted my install vs. copy changes. But I can't find a $setup install anywhere..
gaal QtPlatypus: at $work now for at least five more hours, better get someone else on win32 to check it 11:20
rafl_ It get's install though. Dumb rafl.. 11:21
nothingmuch beer++
autrijus rafl_: right... see util/ghc_setup.pl 11:22
rafl_ nothingmuch: Indeed!
nothingmuch scook0: no more errors in the test run
autrijus (r7460)
nothingmuch beer at lunch++ for 1 .. 5
nothingmuch is underslept (also works when oversleeping)
and it really helps me calm down
big calorie boost
the cultural aspect is also significant
rafl_ autrijus: Ah, works well.
nothingmuch and ofcourse a small amount of alcohol
gaal beer and peanuts. you're hitchhiking? 11:23
nothingmuch gaal: nope, just went out to lunch
svnbot6 r7460 | autrijus++ | * do install instead of copy when there's no DESTDIR... 11:24
gaal nothingmuch: you just reminded me of ford prefect, is all. 11:26
nothingmuch ah
=)
nothingmuch is rereading the guide right now
with the 5th part
it's going a bit slow due to too many nights spent doing other stuff, but i'm really having fun 11:27
svnbot6 r7461 | rafl++ | * Install perl5 into perl6 {sitelib,privlib}/auto/pugs instead of pugslibdir. 11:30
r7461 | rafl++ | * Removed pugslibdir from util/PugsConfig.pm.
r7461 | rafl++ | * Make clean and realclean for perl5/ not fail if there were errors.
scook0 erp 11:33
I got `foo.bar` to parse again...as `foo ($_.bar)` :(
it might be time to cheat a little... 11:34
rafl_ OK, the HS, PIL-Run and PIL2JS site needs to be fixed WRT r7461 11:35
autrijus already fixed. testing 11:36
nothingmuch 3okmij.org/ftp/papers/zfs-talk.pdf
cool!
i wonder how it performs
autrijus nothingmuch: not terribly amazingly; the memory is never freed. 11:37
and also GHC did not have native support for the thousands of delimited continuations oleg uses.
hence, his request to me that GHC should get native support for resume.
nothingmuch starts backing up 30GB of music 11:38
integral does it even have exposed support for any kind of continuation?
autrijus integral: Control.Exception.throw/catch is escape continuations 11:39
nothingmuch what exactly are delimited continuations?
autrijus nothingmuch: see ccshan's "shift to control" slides
integral oooh. 11:40
autrijus www.eecs.harvard.edu/~ccshan/recur/talk.ps.gz
nothingmuch what about this? www.eecs.harvard.edu/~ccshan/recur/recur.pdf
autrijus same, but in paper form 11:41
svnbot6 r7462 | autrijus++ | * adjust the helper launchers to find auto/pugs/ in either
r7462 | autrijus++ | privlib or sitelib, now pugslibdir is gone.
scook0 nothingmuch: there's also a good explanation of delimited continuations way back in the #perl6 logs
from when autrijus explained them to me
nothingmuch scook0: how way back?
nothingmuch gimbles over to the chatlogsa 11:42
scook0 nothingmuch: a few months
nothingmuch oh my, that won't be easy
scook0 wishes he'd started keeping his pugs notebook earlier 11:43
then I'd have the date
nothingmuch scook0: could you share that notebook? 11:44
scook0 nothingmuch: it's made of dead trees 11:45
nothingmuch ah
scook0 so sharing it wouldn't really be feasible
autrijus nothingmuch: 11:46
"reset introduces a new scope, in which you can call shift to get a partial continuation. Calling this partial continuation will result in starting at reset again. At the end of the shift, shift will return from the reset."
rafl_ autrijus: Does it work now? Because we installed into perl6 privlib/sitelib, not the perl5 equivalents.
nothingmuch what does reset do? 11:47
postpones execution and returns a continuation instead of a result?
autrijus nothingmuch: no, reset on itself does nothing.
it's jsut a mark.
rafl_: oh, I totally misunderstood you.
rafl_: I thought we were installing into perl5 sitelib/privlib.
nothingmuch so shift does what I thought reset does up to the parent reset? 11:48
autrijus rafl_: fixing
nothingmuch: er no, shift is just like call/cc 11:49
it takes a block, then runs that block with the cc as the argument.
the only difference is that the cc is not until end of program
but only until the reset
nothingmuch oh, i got it backwards
scook0 nothingmuch: there's a conversation starting at colabti.de/irclogger/irclogger_log/...4,Sat#l283
nothingmuch scook0++
autrijus nothingmuch: so unlike normal call/cc which never returns
nothingmuch ... it returns via it's caller
integral I think it looks really confusing in scheme where the future is to the *left* of the reset/shift not the right :-/
autrijus shift can return just fine
scook0 nothingmuch: also, look at the haddocks for shiftT/resetT in pugs
integral: shift/reset is SO much easier to understand in Haskell :) 11:50
nothingmuch callCC is essentially a clean goto
right?
integral scook0: I agree :-)
scook0 eep, it's late 11:51
hopefully I can finish my patch tommorow evening 11:52
otherwise it'll be at least a week :(
bye all
nothingmuch ciao!
autrijus nothingmuch: it's goto that passes values, but is
otherwise comparable 11:53
nothingmuch and how does shift/reset compare to goto?
autrijus the good thing about delimited continuations is that they act more like normal functions
so they are like... gosub!
nothingmuch gosub?
autrijus (i.e. they can be reasoned about like normal functions.)
never mind, crude BASIC joke. 11:54
nothingmuch ah
which are closer to coroutines?
integral they also label more than a goto label does; you typically can't goto out of the current function for example
autrijus true
nothingmuch integral: assume goto is sub goto { $*INSTRUCTION_POINTER = $_[0] } 11:55
and the rest is just compile time mishmash
integral nothingmuch: yep. a continuation OTOH sets the IP, and fiddles the return stack
nothingmuch is very proud that in his p5orth that's how jumps are done
EXECUTE keeps it's stack frame but overwrites the instruction pointer such that the passed in address looks like it was really the caller 11:56
and it's all pure forth
shift reset is basically call with caller continuation, but also give it the current continuation on the side as if we called it regularly? 11:57
autrijus nothingmuch: reset marks the boundary of shift's return 11:59
nothingmuch gives up on the distinction 12:00
rafl_ autrijus: Seems to work now. Thanks. 12:01
Anyone who want's to fix PIL-Run and PIL2JS?
autrijus rafl_: they are implicitly fixed. 12:02
integral nothingmuch: reset puts a label in the return stack at that point. shift as opposed to call/CC only saves the stack upto the last reset marked into the continuation it makes
autrijus rafl_: try "pugs -BPerl5 -e 'say 123'" after moving the srcdir aside.
nothingmuch integral: the problem is that I'm reading too fast to really grok now... @work. I'll read carefully later
but that is a good explanation
autrijus iblechbot: PIL2JS will die with "permission denied" if run with nonroot for precomp prelude/Test
svnbot6 r7463 | autrijus++ | * look into perl6 sitelib/privlib, not perl5 one...
autrijus iblechbot: you think it's wise to pregen the Prelude at build time for JS? 12:03
nothingmuch phew! all installed CPAN modules were updated
integral++; # you made it real clear as I was formulating my "thanks but no thanks" reply =) 12:04
integral np ;)
autrijus nothingmuch: I think you'll enjoy this too 12:05
www.eecs.harvard.edu/~ccshan/nepls8/talk.pdf
it's ccshan's dissertation -- using delimited continuations to solve linguistic problems
nothingmuch the harvard server is responding very very slow from here
rafl_ Hrm.. how would you explain this: /usr/bin/make install DESTDIR=/home/rafl/projects/perl6/pugs_local/debian/tmp 12:12
make[1]: Entering directory `/home/rafl/projects/perl6/pugs_local'
make[1]: *** No rule to make target `tmp-Prelude.pm', needed by `pm_to_blib'. Stop.
autrijus rafl_: I think you ran Makefile.PL during a strange time that tmp-Prelude.pm is present. 12:13
rafl_ autrijus: OK. I'll just clean it. 12:15
autrijus rafl_: I fixed that.
rafl_ autrijus: btw, ./Setup copy wants --copy-prefix, not --prefix. 12:16
autrijus is apalled that gobby is not in freebsd ports. 12:20
svnbot6 r7464 | autrijus++ | * s/Prelude-temp.pm/Prelude/
autrijus rafl_: noted
rafl_ autrijus: What also would be much easier is to do ./Setup copy --copy-prefix=/\$(DESTDIR). Then the default copy prefix is / which is what ./Setup install does, AFAIK. 12:21
autrijus rafl_: but register takes prefix as DISTDIR?
rafl_: er no, it's not / it's /usr/local
hence my problem
rafl_ Oh, of course. 12:22
autrijus rafl_: try r7465.
rafl_ autrijus: I'm not sure about the register stuff. I don't use that target. It's only important for me that it is not executet implicitly with install.
autrijus: I take .installed-pkg-config and register manually with ghc-pkg in a script that is executed when I install the Debian package (postinst). 12:23
autrijus ok, I'll keep it that way.
rafl_ autrijus++ # Works 12:25
svnbot6 r7465 | autrijus++ | * copy wants --copy-prefix according to rafl.
autrijus eggcellent.
so... do we please the Great Hierarchy sufficiently for this release now? :)
rafl_ I'm not sure your sentence makes sense, but from what I can guess from it: Yes, I think the build system is perfekt for this release. :-) 12:28
autrijus excellent! 12:29
autrijus goes back to tidy up changelog during the lunch.
rafl_ Only the JS Prelude should be generated during the build and also get installed. 12:32
autrijus I was afraid you'd say that. 12:33
if you or iblech can add that in, sure 12:34
rafl_ It tries to write to /usr/share/perl6/5.8/auto/pugs/perl5/lib/Prelude.js which it can't otherwise.
autrijus but I really need to find food :)
nod.
rafl_ OK, I'll try to do that.
iblech rafl_: ./runjs.pl --compile-only -we 1 >/dev/null will generate Prelude.js and Test.js 12:48
rafl_: (This command does not need an installed /bin/js) 12:49
svnbot6 r7466 | iblech++ | * Usual svn props. 12:52
r7466 | iblech++ | * PIL2JS: Makefile.PL: Clean Prelude.js, Test.js, output.html.
rafl_ iblech: Oh, I just hacked it in in another way. But it doesn't find the prelude yet. Maybe you can hack that in? 12:53
iblech rafl_: ./runjs --p6preludepc=/path/to/where/the/Prelude.js/should/be/created --testpc=/path/to/where/Test.js/should/be/created --compile-only -we 1 >/dev/null 12:55
rafl_ iblech: Could you please hack that into Pugs.hs? The prelude will be in sitelib/auto/pugs/js/lib/PIL2JS/Prelude.js or in the same path in privlib. 12:58
iblech rafl_: Hm? Pugs.hs? Not into some Makefile.PL? 12:59
rafl_ iblech: Pugs.hs calls runjs in doHelperRun
iblech rafl_: Yep. And runjs will precompile the Prelude and Test.pm if not already done. But I thought we wanted to force precompilation at build-time? 13:00
rafl_: (pugs -CJS --p6preludepc=... --testpc=... -we 1 >/dev/null works too) 13:01
rafl_ iblech: I think the options should go there.
iblech: The PIL2JS Makefile already generates the prelude and get it installed.
Hey, why aren't my changes pushed to the repository? Stupid me..
Ah, hasn't been merged because you changed perl5/PIL2JS before.
svnbot6 r7467 | iblech++ | PIL2JS: jspugs.pl, runjs.pl: Document the --testpc option and added --metamodel-base. 13:02
r7468 | rafl++ | r16904@ata: rafl | 2005-10-09 14:48:46 +0200 13:03
r7468 | rafl++ | * Generate precompiled JS Prelude and install it.
rafl_ iblech: Yeah, it's precompiled at build time, but runjs needs to know where it is. Therefor it needs to be called with the right --p6preludepc args.
iblech: Does runjs really need a precompiled Test.js?
iblech rafl_: Yep, Test.js is unconditionally linked in. (This is because there's no real reliable way for runjs.pl to read the sourcecode and search for /use Test/ (think -e, -M, etc.)) 13:05
rafl_ Ah, OK.
iblech rafl_: Is pugs already installed when PIL2JS's Makefile.PL:pure_all runs? 13:07
rafl_ iblech: No. 13:08
iblech rafl_: (Because it'll either need a "pugs" in $PATH or we've to give it a --pugs=... option)
rafl_: Hm.
rafl_ iblech: But there are only two paths where the precompiled stuff can be.
p6{sitelib,privlib}/auto/pugs/js/lib/PIL2JS/Prelude.js
iblech: pugs is already built at this time. ../../pugs should work. 13:10
iblech rafl_: Hm. Currently, PIL2JS.pm expects Prelude.js and Test.js in ./, where . is the dir runjs.pl/jspugs.pl is in. 13:12
rafl_: But if Prelude.js should be in js/lib/PIL2JS/, then we'd have to update PIL2JS.pm accordingly, resulting in runjs.pl/jspugs.pl not working when PIL2JS is not installed... 13:13
svnbot6 r7469 | rafl++ | r16910@ata: rafl | 2005-10-09 15:08:44 +0200
r7469 | rafl++ | * Also generate Test.js as build time.
rafl_ iblech: Yes, therefor Pugs should pass a proper --p6preludepc to runjs.
iblech: Shouldn't be too hard to let it look into two places. 13:14
iblech rafl_: Right, but the definition of "proper" varies: if PIL2JS is installed, --p6preludepc should point to p6{sitelib,privlib}/auto/pugs/js/lib/PIL2JS/Prelude.js, but when developing, it should be $srcdir/perl5/PIL2JS 13:15
rafl_ Oh, right.
iblech: Breaking it when it's not installed it maybe acceptable for this relase only if it's too much work to get both working..
iblech rafl_: Actually, it is, because we can't simply check for the existance of Prelude.js/Test.js (as we can do with other files, as PIL2JS.js or the metamodel) 13:16
rafl_: Hm. What about if install Prelude.js and Test.js unconditionally in the dir jspugs.pl/runjs.pl get installed into?
rafl_: This would work, I think.
We'd only have to cp Prelude.js Test.js blib/lib/ 13:17
rafl_ And not that hard to do. commited.
svnbot6 r7470 | rafl++ | r16912@ata: rafl | 2005-10-09 15:18:35 +0200 13:18
r7470 | rafl++ | * Install precompiled js stuff next to runjs.pl.
iblech rafl_++ # great! :)
Only thing left to do is setting $PIL2JS::cfg{pugs} appropriately 13:19
rafl_ -v please.
iblech Never mind, autrijus had already hacked that in into PIL2JS.pm 13:20
So we're set, I think :)
rafl_ Yay!
Generating some .debs to test it.
* Error: Your precompiled Prelude is outdated. 13:21
iblech This means that Prelude.js is older than Prelude/JS.pm.
rafl_ Possible. The order in which they get installed isn't predictable.
(in the debian package) 13:22
iblech Hm, Prelude.js/Test.js get re-compiled each time one issues make... Fixing that.
rafl_ OK. 13:23
Could you also disable that mtime check?
iblech Is this a one-time error only? If so, I would s/die/warn/ 13:24
rafl_ No, it happens every time when I run pugs -BJS. 13:25
For the installed pugs only.
iblech Ok, removing the check. 13:26
rafl_ Thanks.
svnbot6 r7471 | iblech++ | PIL2JS: 13:32
r7471 | iblech++ | * PIL2JS.pm: Removed the mtime check for the Prelude, as it breaks Debian.
r7471 | iblech++ | * Makefile.PL: Only precompile the Prelude and Test.pm if not already done.
r7471 | iblech++ | Also make use of autrijus's $PIL2JS::cfg{pugs}, so Makefile.PL will find
r7471 | iblech++ | ../../pugs or ..\..\pugs.exe.
rafl_ Generating new packages.. 13:33
Works. Gereat! 13:36
iblech++
iblech: pugs -BJS -e'"Foo".say'
iblech: Why is '1..0' printed at the end?
iblech FYI, just make cleaned, rm -rf /usr/lib/perl6/site_perl/auto/pugs/, make install'ed, mved $srcdir, typed pugs -BJS, and it worked! :) 13:37
Because Test.pm is unconditionally linked in, and Test.pm always installs an END {...} block
(Again, runjs.pl can't detect reliably whether Test.pm is really needed.) 13:38
QtPlatypus needs someone to test his MIME::Base64 in Win32 13:39
iblech Look at runjs.pl:71 -- I tried to use some magic which'd make pugs-smokejs unnecessary, but failed because pugs wasn't able to parse the combination of -CPerl5, -M and -e I used... 13:40
14:04 rafl_ is now known as rafl
svnbot6 r7472 | rafl++ | r16916@ata: rafl | 2005-10-09 15:56:13 +0200 14:04
r7472 | rafl++ | Debian:
r7472 | rafl++ | * Let the pugs package suggest perl and spidermonkey-bin.
r7472 | rafl++ | * Removed usr/share/pugs from pugs-modules.install. Those stuff is now
r7472 | rafl++ | installed in usr/share/perl6.
r7473 | rafl++ | r16917@ata: rafl | 2005-10-09 15:59:39 +0200
r7473 | rafl++ | * Added perl5/*/t/*.t to test files.
r7474 | rafl++ | r16918@ata: rafl | 2005-10-09 16:02:14 +0200
r7474 | rafl++ | * Reverted accidently changed util/yaml_harness.pl. I forgot to remove the 14:05
r7474 | rafl++ | patches applied for the Debian package before commiting.
rafl iblech: Why is Test.pm handled specially at all? 14:06
iblech rafl: Performance of the testsuite
rafl :-( 14:07
iblech: You told me a cool way do disable smoking/sending smoke results once. I forgot it. Can you tell me again how it worked.
s/.$/?/
iblech rafl: Sure, PUGS_BUILD_OPTS="smoke_upload=0" (but untested) 14:08
rafl Great, thanks.
iblech rafl: I quite like seeing JS smoke results after a hour (instead of at least three) :) 14:09
rafl iblech: Buy better hardware!
iblech In the future, we'll probably need /var/cache/pugs/js/Prelude.js, with appropriate measures that no one can write precompile an evil modul and save it as Prelude.js 14:10
rafl: Do you want so sponsor me? :)
s/so/to/
rafl iblech: Yes, nice feature. 14:11
iblech: Well, actually... no.
iblech: It would be nice to include the name of the smoker in the overview of smoke results on smokeserv. 14:12
And maybe also display the architecture. 14:13
pugs doesn't work on mipsel, arm, mips, ia64, etc. yet. 14:15
It kind of works on hppa, but lots of tests seem to fail. 14:16
iblech Should both be relatively simple to implement. I'll have a look at it, probably tomorrow 14:17
rafl On s390 as well.
Thanks.
iblech IIRC putter runs Pugs on ia64, not sure
rafl OK, I'm off till this evening. I hope a have a new release to put in Debian then. :-) 14:18
nothingmuch wonders how much effort it'll be to make smokeserv.pl into a catalyst based app
with slightly more abstraction WRT the naming
and have it submit storable dumps instead of htmls
rafl iblech: Well, it doesn't work with Debian's ghc6 on ia64. Haven't tested other compilers.
iblech rafl: Ah, ok.
rafl nothingmuch: Hm. I'd volunteer to do that if iblech agrees.
nothingmuch it's opensource, iblech doesn't have to agree ;-) 14:19
rafl Well, I don't want to crack the machine it runs on to install it. :-)
nothingmuch if that becomes a problem i'll host it
rafl Beside iblech agrees ;-)
iblech I don't have any knowledge of Catalyst, but if rafl wants to do it, I'm fine :) 14:20
nothingmuch rafl: what I would really like is that once it gets storables and slightly more structure, we use Test::TAP::Model::Consolidated to display diffs
rafl OK, I'll start on that this evening if the build of the Debian packages I currently do is finished and fine.
nothingmuch rafl++
iblech Installing Catalyst on the smokeserver is absolutely no problem :)
rafl++ indeed
nothingmuch iblech: what about mod_perl?
rafl OK, bye. (really now) :-) 14:21
nothingmuch ciao!
iblech nothingmuch: I'm root on the server, so that'd be no problem either :)
integral why storable rather than something more inter-language like XML, YAML, or JSON?
iblech bye rafl
nothingmuch yummy
integral: performance
YAML takes 40 seconds to load
XML is probably not as bad
but Storable is optimized for speed 14:22
and the structure is very big, at 10000 unit tests
integral that's irritiating about YAML
nothingmuch very =(
rafl: if you backlog, use Data::Serilizer so that the serialization backend is pluggable ;-)
gaal autrijus: re r7464, sorry for the bikeshed paintjob suggestion, but imo "Prelude.pm" is better than having no extension at all... 15:09
autrijus gaal: problem was if it failed it left a .pm around for MakeMaker to pick up 15:12
gaal: maybe a END{} block to clean that away?
geoffb See, now *there's* a good reason to be offline 15:18
nothingmuch darnit 15:20
svnbot6 r7475 | autrijus++ | * "make haddock" now complies with Cabal convention and 15:25
r7475 | autrijus++ | put the HTML docs into dist/doc/html/. This should keep
r7475 | autrijus++ | the src/ tree cleaner.
r7476 | autrijus++ | * Expose all Pugs.*, Emit.* and PIL.* modules for Haskell programs. 15:31
r7477 | autrijus++ | * Inline::Pugs: revert to the previous, non-PIL-Run UI
r7477 | autrijus++ | as the API has changed and I don't quite feel like to
r7477 | autrijus++ | chase it right now...
gaal autrijus: or put in a tmp/ directory MM won't touch? 15:35
(end block where, build_pugs.pl?)
autrijus gaal: end on genprelude
svnbot6 r7478 | autrijus++ | * Pugs.Run.Perl5: remove commented-out debug statements. 15:36
r7479 | autrijus++ | * Pugs.hs: refactor out noEnvDebug so external programs
r7479 | autrijus++ | can elect to run without debug traces.
autrijus gaal: done. 15:38
gaal too fast for me, as usual :)
autrijus :)
svnbot6 r7480 | autrijus++ | * adopt gaal's bikesheding correction to paint the tmp Prelude back 15:42
r7480 | autrijus++ | as Prelude.pm, and unlink it at END of gen_prelude.pl.
r7481 | gaal++ | add --keep option for Prelude debugging
gaal current music, incidentally: Paint It Black
autrijus chuckles 15:43
gaal coincidence, I swear :)
gaal imagines a PHB giving a speech: "... now this, our bikeshed feature for which we had worked so hard in design meetings..." 15:45
(corporate malapropisms)++
autrijus civilization is wonderful: it replaces bloodshedding with bikeshedding 15:47
gaal cats are perfection of this process, just shedding. 15:48
autrijus lol
I think I'll be in OSDC. 15:49
(preferably if damian will also be there)
# www.osdc.com.au/
gaal monash! 15:50
sounds close to home for him.
autrijus bleh, ::Pair also participated in magical pairs 15:51
I mean the ::Pair class object itself
simply because it .isa Pair
autrijus feels like hlaghing
gaal Pair: Pairs are indeed recursive 15:53
autrijus down to two failures. 15:58
svnbot6 r7482 | autrijus++ | * even more giving-up on magical pairs: stringify ::Pair in is()
r7482 | autrijus++ | calls as otherwise it would participate in Pair magical binding
r7482 | autrijus++ | because it .isa Pair. Special casing for that feel wrong, so
r7482 | autrijus++ | let's also write it off and wait for the sane-magic-pairs revolution...
r7483 | autrijus++ | * ditto for Perl6-Value-List/t/.
autrijus relooks arcfour
gaal $works :( 16:02
autrijus enjoy :)
gaal :)
I like your attitude! 16:03
stevan morning autrijus, gaal
gaal yoyoyo
stevan thinks that p6 will be the new LISP
we just need to get nothingmuch's exceptuations accepted
leo__ autrijus: are you planning to write some official messages about Pair's insaneness? 16:04
autrijus leo__: it's been extensively documented on p6l by luqui et al
with a working proposal even
s/proposal/counterproposal/
so I'm not sure what I can add to this, other than outlining what I'd like to support instead in pugs... 16:05
leo__ ok - rather warnocked iirc
autrijus nod. 16:07
hm, arcfour seems to have actual problems 16:13
as the ciphertext did not match and the garbled chars killed my terminal
gaal unicode problems probably
our io assumes utf8 16:14
getc etc.
autrijus oh hm. yeah.
though we have binmode
gaal but that doesn't help if the prim reads more than one byte and assumes an encoding :) 16:15
autrijus true.
gaal binmode is Haskell's binmode, i think it just does eol/eof?
autrijus you are correct 16:16
okay, we need to integrate the Data.FastPackedString based IO.
(it's already in src/Data/)
gaal we can add some metadata to VHandle
autrijus *nod* 16:17
gaal there's a dummy $layer arg to open :)
autrijus I saw that :)
gaal P.hGet assumes latin-1 16:18
also, i don't see a character based read 16:19
but it's very nice that they have gz io.
can Haskell strings contain NULs? 16:20
svnbot6 r7484 | autrijus++ | * todo-fail arcfour.t again as its test failures garbles my
r7484 | autrijus++ | terminal and there is no obvious piece at fault; gaal
r7484 | autrijus++ | suggests utf8ness, which is plausible.
autrijus gaal: of course... 16:21
gaal: CStringLen too
gaal (i mean other than at the last byte ;-)
cool
@t Char
er, not here
pasteling "iblech" at 62.216.196.182 pasted "luqui's pair proposal" (10 lines, 295B) at sial.org/pbot/13579 16:28
autrijus iblech: that, and also more importantly 16:31
$pair = :foo(1)
blah($pair) # positional
blah(*$pair) # named
iblech ok. I like the proposal very much :) 16:32
How will we implement the normal pair/magical pair distinction? Two classes (Pair and Pugs::Internals::PairUsedForNamedParameters)? 16:33
autrijus no, the named for parser is a parser form 16:34
it's not really a pair in any sense
it's syntactic.
iblech Ah, makes sense. But then how'll the PIL for foo(a => 42) look like? Currently, it's simply serialized as foo(infix:Ā«=>Ā»("a", 42)) 16:35
autrijus yeah, in vanilla core we use Syn=> to distinguish
I think the App form needs extension to support all annotations
so instead of using substructural forms, it's all represented at App level 16:36
iblech example please?
PApp { pFun = ..., pNamedParams = {...}, pPositionalParams = [...] }? 16:37
autrijus yeah, have a structure for each zones sounds prudent 16:40
iblech ok. And *$pair will simply add $pair to pNamedParams then. Makes sense and is easy to implement in PIL2JS :) 16:41
autrijus which is always good to hear :) 16:42
iblech goes to write tests for it 16:44
autrijus :todo<revolution> 16:45
iblech :)
autrijus seriously, <unspecced> would do 16:47
would be good to have it in the release and then posted to p6l.
svnbot6 r7485 | iblech++ | * util/gen_prelude.pl: Removed now unnesessary unlink() call.
r7485 | iblech++ | * PIL2JS:
r7485 | iblech++ | * jspugs.pl, runjs.pl: Better handling/ignoring of -BJS.
r7485 | iblech++ | This means make test-js works again.
r7485 | iblech++ | * runjs.pl: New option --precompile-only (only precompile the Prelude and
r7485 | iblech++ | Test.pm, do not run or compile any user code).
r7485 | iblech++ | * Makefile.PL: s/compile-only/precompile-only/
16:51 reffie is now known as rep
leo__ autrijus: what do you think of ~lt/named-args.txt 16:58
(at feather)
pasteling "autrijus" at 194.145.200.126 pasted "lt's proposal" (20 lines, 647B) at sial.org/pbot/13580 16:59
autrijus leo__: sure, that is the same with luqui/iblech/me/etc 17:00
the only addition is a *$pair form
akin to *%hash but only for one pair
nothingmuch dvd 7 is now burning.... *sigh* 17:01
leo__ not quite, luqui still had some magic in his proposal I think
nothingmuch wonders what we're talking about
leo__ e.g. difference between a=>'b' and :a(b) or :a<b> - whatever
*$pair is *%hash with excatly one kv elem 17:02
autrijus leo__: yeah, I think that :a<b> and a=>'b' should at all times be interchangeable 17:03
makes no sense otherwise
iblech agrees
leo__ yep 17:04
autrijus good we are in vehement agreement.
leo__ great
autrijus so the entire "unexpected pairs jump into named zone" foo can go away. 17:05
leo__ for me - yes ;-) 17:06
autrijus good good.
leo__ a Pair arg is passed like a Foo arg - no special case at all
autrijus leo__: parrot of course has support for nonslurpy but still nameless arguments?
will it have named-only support?
nothingmuch you guys are discussing touples vs parrot calls? 17:07
autrijus nothingmuch: we guys are talking about demagicalizing the insane magic pair semantic :)
nothingmuch ah
has there been any attempt besides lexical sensitivity? 17:08
leo__ named support just depends on the 'saneness' of implementation possiblities - currently named args aren't supported
autrijus leo__: right, but iirc chip is working on a pythonesque support?
leo__ yep
autrijus nothingmuch: *$pair acts like *%hash with one pair
svnbot6 r7486 | stevan++ | Perl6::MetaModel -
r7486 | stevan++ | * new version of the Mini-MetaModel which has explicit EigenClasses for
r7486 | stevan++ | all classes created. This is an implementation of the diagrams I did
r7486 | stevan++ | yesterday. I still need to resolve how to deal with the class pointer
r7486 | stevan++ | since it now points to the eigenclass instead of the real class. But
r7486 | stevan++ | this is a somewhat minor issue really.
nothingmuch versus?
dduncan after back-logging, I agree with the idea of *$pair being like *%hash, making it bind to a named arg, whereas no * means it is positional 17:09
leo__ ther is really no need for a special *$pair
autrijus nothingmuch: versus no such syntax.
nothingmuch reads for more context
as for :a<b> and a => b interchangiable - me too 17:10
autrijus leo__: because it may be written as *{$pair} ?
dduncan likewise, I infer that one can also say foo( *$pair1, *$pair2, ... ) etc if they had several pairs to bind in the same method call
autrijus dduncan: aye
nothingmuch when there is one pair tuple with one named and hash are indistinguishable? 17:11
leo__ may point is: if you want named args: do it in the call or pass a *% name hash - nothing else is needed IMHO
autrijus leo__: say I have
foo($name => $value)
and then I want to refactor it out: 17:12
my $pair = (get_name() => get_value()); foo(*$pair)
then
sub get_pair { ... } ; my $pair = get_pair(); foo(*$pair)
surely that is sensible?
leo__ my proposal would need my %pairs - ... foo(*%pairs) 17:13
ther is no special *$pair at all
nothingmuch thinks that 'sub foo (*$a) { other(*$a) }' alone is worth the complexity of the tuple type
svnbot6 r7487 | autrijus++ | * fix a genuine Test::Builder bug uncovered by correct class-method 17:14
r7487 | autrijus++ | handling; Test::Builder.create should create a new builder object
r7487 | autrijus++ | without falling back to .new for singletons.
autrijus leo__: what nothingmuch said
dduncan back in about 3 hours ...
leo__ ther is aready *@ and *% support - why just another level?
nothingmuch because *@ and *% aren't all encompassing
*$a is much more transparent 17:15
so it's suitable for delegation
while the others are suitable for more "normal" slurpiness - accepting a variadic number of args
leo__ for me this looks like a premature 'pessimization' against 1 elem hashes 17:16
nothingmuch leo__: the tuple type can hold names and positionals
svn.openfoundry.org/pugs/docs/notes/theory.pod 17:17
see =head2 Tuples
my $tuple = ("foo",);
foo(*$tuple);
that passes in a positional
my $a = (1,2,3); # a tuple with positionals 1,2,3 17:18
my $b = (0,*$a,4); # a tuple with positionals 0,1,2,3,4
leo__ this is just a mix of zones imho and not needed at all
# when used in a call of course 17:19
nothingmuch then how do you do 'sub foo ($my_param, *$their_params) { $my_param + other(*$their_params) }' or something like that?
it's always been hard in perl
leo__ if *$their_params isn't modified at all (partially consumed e.g.) just pass an array or list 17:22
else use a named hash
iblech leo__: But then you lose type information etc.
nothingmuch no, because that doesn't capture the "magicness" of names, and no because you can't losslessly represent positionals in named hash without name mangling
also, type inferrencing will benefit from opaqueness
that way foo's signature can statically be other's signature plus one positional if static typing is enabled 17:23
leo__ I don't see how *$ can help here, or is better then *% or *@
nothingmuch because it makes no implication (even psychological) about whether the parameters are named or positional
autrijus *$ in this view is a zone-free device 17:24
it is indeed beyond the original spec of perl6 zones
(pugs had never really implemented zone separation either)
preflight. 17:26
leo__ ok, when I get that right Tuple ... ~lt/named-args.txt 17:27
nothingmuch are these zones in the metaphorical, syntactic, or implementation sense?
leo__ oops again
autrijus pugscode.org/tmp/Perl6-Pugs-6.2.10-pre0.tar.gz
nothingmuch where is named-args.txt?
leo__ a Tuple ~lt/named-args.txt
shit
autrijus nothingmuch: read the spec... it's syntactic.
nothingmuch autrijus: that is the distinction between a => b and (a => b) in a call? 17:28
leo__ yeah, wanted to paste ;-)
anyway the tuple is the param list, it seems
what I'm talking about is the implementation of that thing 17:29
nothingmuch as I see it in it's most naive form it's structure with an array and a list, where most of the time it's structure is known in advance
array and hash, sorry
autrijus nothingmuch: er no, no
nothingmuch autrijus: what spec then? theory.pod?
autrijus nothingmuch: A06 17:30
nothingmuch ah
gaal oooh preflight! /me wants to go home now 17:32
nothingmuch gaal: how was your pm talk?
autrijus: can preflight precompile Test.pm etc? i see no config.yml
gaal it's generated the first time 17:33
nothingmuch ah
gaal the talk was nice, the turnout was low 17:34
home is where the pugs is &
autrijus gaal++ # home is where the pugs is
nothingmuch autrijus: make smoke started on preflight 17:36
autrijus excellent
nothingmuch goes to make dinner
hah, airtunes is choking when I'm scping a big file =) 17:38
svnbot6 r7488 | autrijus++ | * we cannot add perl5/*/t/*.t to the test files because those are harnessed 17:46
r7488 | autrijus++ | with pugs as the main executable, and we had not yet mastered the art of
r7488 | autrijus++ | re-execing perl5 on non-perl6 files.
r7489 | autrijus++ | * add Setup* into MANIFEST.SKIP.
putter fyi, a recent (hmm, how do I check version when pugs doesnt run...) (within a few hours) pugs on FC3 x86_64 w embedded perl, dies with 17:49
./pugs: error while loading shared libraries: libperl.so: cannot open shared object file: No such file or directory
anyone creating /lib/ paths and forget to include /lib64/ ? 17:50
autrijus putter: hm, libperl.so is not in LD_LIBRARY_PATH?
putter checking...
my LD_LIBRARY_PATH is undefined. 17:51
autrijus where is libperl.so then
putter /usr/lib64/perl5/5.8.5/x86_64-linux-thread-multi/CORE/libperl.so 17:53
and perl -V shows /usr/lib64/perl5/5.8.5/x86_64-linux-thread-multi at the top of @INC
autrijus hm, so we'd need to chase @INC and add it to LD_LIBRARY_PATH
autrijus has flashback from the nightmarish PAR myldr days 17:54
autrijus screams
hm.
so, what if you ln it to /usr/local/lib etc and rerun?
putter I try to avoid a global LD_LIBRARY_PATH, so this is a new problem. But I've also upgraded ghc recently... but this seems more likely to be pugs.
autrijus this is pugs 17:55
Preflight #0: xrl.us/hxhj | geoffb's column: xrl.us/hxhk | pugscode.org | pugs.kwiki.org | paste: sial.org/pbot/perl6 | www.geeksunite.net 17:56
svnbot6 r7490 | iblech++ | New t/syntax/pairs.t: Tests for luqui's (modified) pair proposal, :todo<unspecced>. 17:58
putter well, ln'ing into /usr/local/lib didnt help. and I mangaged to rm my libperl.so, so my ability to test just went away. :/ 17:59
autrijus :/ 18:00
"yow".
iblech: hmm, I wonder if t/unspecced/ is a better place. 18:01
iblech: if t/syntax/, that probably means a p6l thread immediately :)
iblech autrijus: ok, p6ling :)
autrijus iblech++ 18:02
iblech: try to number or tag each orthogonal features :) 18:05
iblech autrijus: I'll nopaste it for review here :) 18:06
autrijus cool 18:07
autrijus praises the #perl6 hivemind
obra 'morning. 18:09
how goes the release cycle?
autrijus obra: see topic 18:11
in pre0
obra goes to look :) 18:12
do you already have an ubuntu build?
s/build/test/ 18:13
autrijus no
obra ok. building
autrijus "make test" would be appreciated
pasteling "iblech" at 62.216.196.182 pasted "p6l post: "Sane pair semantics"" (42 lines, 1.4K) at sial.org/pbot/13582
autrijus mm there is no sanity on p6l... only the less and least insane 18:14
autrijus again thinks of doing a "the least insane" comic strip in the style of Why's "the least surprised"
obra autrijus: should I worry about no smokekit, no parrot, no p5embed no ihaskell?
autrijus obra: nope. 18:15
gaal pf0 smoke started here w32
autrijus cool 18:17
gaal uh nothingmuch if you're precompiling test remember to increase heap size 18:19
sorry for not remembering to remind you earlier :)
autrijus iblech: I disagree
arrays containing pairs
my @foo = (1, a=>2, 3); 18:20
blah(*@foo); # surely all positional
iblech autrijus: Right, of course.
autrijus so just "Same for hashes"
iblech yep
autrijus otherwise, it's least insane. 18:21
iblech ok, sending... 18:22
autrijus iblech++
iblech (Less insane pair semantics)++ :) 18:23
sent.
putter ok, works with libperl.so ln'ed in /usr/lib64/. so what broken? @INC used to be used and that went away? 18:26
autrijus putter: letmethinkaboutit.
putter k :)
autrijus putter: try r7492 18:28
not sure if it works
svnbot6 r7491 | iblech++ | t/syntax/pairs.t: *@array passes its elements all positionally, of course.
r7492 | autrijus++ | * shared libraries are to be linked. thanks putter for noticing.
r7493 | iblech++ | * pugs::hack: Added ext/FA-DFA to the directory tree. 18:32
r7493 | iblech++ | * ext/FA-DFA/Makefile.PL: Typo fix.
putter trying...
pasteling "gaal" at 192.115.25.249 pasted "build fail on windows" (10 lines, 498B) at sial.org/pbot/13583 18:33
gaal iblech: looks like your turf
oops about never having attempted to build js on win32... 18:34
iblech gaal: r7494 18:35
svnbot6 r7494 | iblech++ | PIL2JS: PIL2JS.pm: Fix for Win32.
gaal looks like plenty more is built now by default than 24h ago! 18:36
iblech: ACK
proceeding with r builds now.
there's an annoying problem with processes not terminating correctly in windows. this break things because you get write errors on make targets (Setup.exe, pugs.exe). no idea why this happens and how to solve it programmatically though (i just kill the process manually when i see it happening) 18:41
this isn't a new problem
obra paste.husk.org/4334 18:45
paste.husk.org/4334 <- build failure
iblech obra: You need Class::Rebless from CPAN. But I'm unsure whether we want to force all users (even users who don't want to -BJS) to install it. autrijus? 18:46
obra iblech: it should have bloody-well warned me when I ran the makefile.pl, no?
putter autrijus: re r7492, nope. :( 18:48
iblech gaal: Hm, the fix for Win32 breaks PIL2JS on Linux...
gaal oy.
cm meow
iblech obra: Probably. Waiting for autrijus to decide 18:49
gaal: specialcasing for $^O eq "MSWin32"...
obra iblech: is it an optional feature?
autrijus putter: hm, hrmph, hrmph. 18:50
putter: letmethinksomemore.
iblech obra: Yep, it's only if you want to compile to JS
obra iblech: how does one specify that? 18:51
sure sounds like an optional feature.
And Module::Install supports those
iblech obra: pugs -BJS -e 'say "Hello World from JS"'
obra iblech: no. at build time
iblech obra: Ah, dunno...
gaal oughta be from config.yml i suppose.
svnbot6 r7495 | iblech++ | PIL2JS: PIL2JS.pm: Unbreak PIL2JS for non-Win32. 18:53
iblech Alternatively, we just make the precompiling fail gracefully and add a warning to Makefile.PL ("If you want to use the JavaScript compiler, you'll have to install following additionall CPAN modules: [...]") 18:54
gaal the user might not notice it - we have plenty of messages already; and then they'll wonder why things don't work when they try them (unless the message can be said again later?) 18:55
iblech warning later again is not a problem 18:56
obra I'd love to see the messages go away in favor of feature configuration. 18:57
gaal then i like your suggestion. :)
obra but a graceful dependency missing message seems a win
svnbot6 r7496 | autrijus++ | * try again to find libperl.so for putter. 18:58
iblech Ok. I'll make runjs.pl fail gracefully 18:59
gaal iblech++ # grace
autrijus putter: still slive? 19:00
s/slive/alive/
svnbot6 r7497 | iblech++ | * Added missing shebang lines in ext/*/t/*.t and t/*/*.t.
r7497 | iblech++ | * t/builtins/control_flow/try.t: "The following warning is expected".
gaal iblech: your fix wasn't enough. :( 19:01
*** Reading input from "-"...
...and hang.
iblech oh :(
unsure how to fix 19:02
gaal temp file (not sure what's going on there really) 19:04
?
iblech Problem is that system "... > ..." will eat shell metachars 19:05
Does reopening STDOUT etc. work on Win32?
svnbot6 r7498 | iblech++ | PIL2JS: runjs.pl fails now gracefully when Class::Rebless (needed by
r7498 | iblech++ | PIL::Parser) is not installed.
gaal not really sure. checking. 19:06
iblech Idea: Instead of using open "-|", which doesn't seem to work under Win32, I use IPC::Open2. 19:07
gaal is ipc::open2 core in 5.6.1? 19:09
yes. 19:10
svnbot6 r7499 | iblech++ | PIL2JS: PIL2JS.pm: Fix for Win32, for real. (IPC::Open2 instead of open "-|").
iblech gaal: r7499
gaal (since perl 5, wow.)
thanks, testing.
btw, do you get this too in your precomiplation? 19:12
*** Useless use of a constant in void context:
VInt 1
oh, it's reading from "-" again? 19:13
iblech [warnings] yep
[-] yep, but this time STDIN got mangled by IPC::Open2, I think 19:14
gaal maybe i'm just impatient? how long after the `*** Reading input from "-"...' message should I wait? 19:16
autrijus check your cpu level
gaal on par with the -CPugs Prelude.pm?
low cpu. :( 19:17
iblech :(
autrijus :((
should I boot to win32? 19:18
gaal is there an effective way i can give you control of my box?
autrijus first, you install this BO2K thing... 19:19
*smile*
gaal heh
wait, vnc should work, no?
autrijus nah, I'll just reboot. bbiab
putter: I believe the issue of not finding libperl should be fixed now. 19:20
wolverian autrijus, sounds like you need vmware :) 19:21
autrijus I like flyakite :)
cm autrijus: cute! :) 19:24
didn't know that 19:25
autrijus it and cygwin are the two things that makes win32 bearable. 19:28
cm visual studio does that job for me ;) 19:29
autrijus any smokes up for #0? 19:38
iblech still smoking here
autrijus k
gaal can't smoke on win32... 19:39
actualy
i be stupid. i can smoke manually. thud. 19:40
autrijus ingy_: yo. I noticed that pugs.kwiki.org has as its title "The Perl 6 Interpreter"
ingy_: can we change it to "A Perl 6 Implementation" please?
obra autrijus++ 19:41
vladtz smoke fail on linux with PIL2JS needs the Class::Rebless here
autrijus wow, is this for real? perlonpoles.blogspot.com/ 19:42
obra autrijus: why isn't pugs updating to the perl6 repo?
gaal hahahaa 19:44
autrijus obra: because uhm I think svn.perl.org disallows checkin
I'll switch to https:// now.
integral either that or it's using the word "synergistic" seriously :)
obra autrijus: that _must_ be a joke. 19:45
autrijus yeah, not a bad one at that
I certainly laughed.
obra It's not quite surreal enough
It filled me with dread
autrijus mm black humour
obra :) 19:46
obra went to check pugs on cia.navi.cx
and it's not there ;)
autrijus nodnod, wait 10 mins 19:47
btw, irc.pugscode.org should now propagate far enough.
obra as a freenode synonym? 19:48
autrijus no, as a irclog redirection
same idea as smoke.pugscode.org
obra *nod*
Random perl5-aside. Does tying the symbol table work this year? 19:49
autrijus I don't think that ever works.
or will ever work.
obra :)
That's really for the best. But is mildly inconvenient for Jifty::DBI
gaal (r7499 win32 smoke entered t/ here)
cm eigenpatterns :o) 19:50
19:50 revdiabl1 is now known as revdiablo
autrijus "A Short Conspiracy Theory: Ruby is actually Perl 6. / Pugs is a decoy. / Nobody?s ever seen Matz and Larry in the same room together ..." 19:51
# www.exactlyoneturtle.com/wordpress/?p=45
cm :D
obra ...I have.
cm But then who is this autrijus guy?
autrijus ...me too.
cm: autrijus is an asian committee.
cm :o 19:52
autrijus wow, this is amazingly funny. if only dankogai wrote in English... 19:58
# blog.livedoor.jp/dankogai/archives/50077921.html
gaal you know japanese? 19:59
autrijus I can understand almost all the kanji 20:00
gaal and that's enough to get the jokes!?
teach me the kanji, man! :)
autrijus yeah, because almost all the proper nouns are in kanji.
wolverian is ruby's unicode support still broken? 20:03
eh, I'll check out myself, I'm too lazy. sorry :)
gaal win32 smoke up. 20:05
autrijus I'm fixing pil2js/win32.
gaal WTemplate is a PGE thing. 20:06
autrijus ext\Tree\t\visitor.t ext\WTemplate\t\basic.t
gaal (i think)
autrijus ok, please put the neccessary guard.
(you can slurp one from t/rules)
also please look atthe dying visitor.t
gaal right.
fixed Tree. 20:08
svnbot6 r7500 | gaal++ | add missing plan
iblech x86 Gentoo Linux smoke up as well. The failing from_perl6_rules tests are probably because I use Parrot HEAD instead of 0.3.0. 20:09
gaal *sigh* WTemplate claims to pass (skip) when run manually 20:10
autrijus *sigh*
gaal skipping completely?
autrijus yeah.
iblech: actually, there's no PGE changes afaik between 0.3.0 and head... 20:11
svnbot6 r7501 | gaal++ | skip this test completely; it confuses the harness (two subtests were skipped but something thought they failed) 20:13
gaal useless figure of the evening: by our next release, we may have more svn revisions than parrot. 20:15
autrijus ooh :) 20:16
gaal i've a friend who managed to build a newish perl on msys once. i must recruit him; if i am successful, we may get perl5 embedding on win32. 20:19
autrijus that'd be nice. 20:20
autrijus is still in open2 hell.
gaal is it a hot hell or a cold one? 20:21
autrijus it is freezing over. 20:22
gaal that's a good sign.
autrijus lol.
if you got spare tuits, proofcheck changelog for sanity? 20:23
gaal sure
(I intend to eradicate any trace of sanity.)
well, the pf0 version didn't have an r number. dunno if that was intentional. 20:25
autrijus nope.
nothing is intentional. please fix away
gaal n+2? :) 20:26
autrijus +5
gaal does the test count traditionally include ext/? 20:30
autrijus only for vanilla 20:31
gaal i mean in the changelog. the last release had "8100+" 20:32
we're now 10300+ by that count.
also: did we fix no bugs? :)
autrijus we did but I despair at looking at nearly 2000 commits. 20:34
gaal: yes, 10300+ sounds correct.
iblech Hm, need to sleep now. autrijus, I'd be entirely fine with not precompiling Prelude.js,Test.js on Win32 for this release
autrijus iblech: too late, I fixed it.
gaal I'm consolidating the tab expansion changelogs.
autrijus good night.
gaal heh.
night iblech!
iblech ooooh! autrijus++ :) 20:35
Night all :)
svnbot6 r7502 | autrijus++ | * Give up on Open2 pipe sanity on win32; use temp file instead.
autrijus gaal: please confirm my fix in a bit 20:37
then I can reboot
gaal r '02?
autrijus yeah 20:38
autrijus ponders whether we should do something about the useless warnings.
gaal then why in a bit? :)
autrijus I thought you are reviewing sanity :) 20:39
gaal am, but builds are bg
szabgab good evening
gaal heya szabgab-in-Hungary
autrijus greetings. 20:40
szabgab just an unrelated question: is there anyone in here who is in a country bordering Hungary?
hi gaal and autrijus 20:41
pasteling "gaal" at 192.115.25.249 pasted "close but no reboot" (22 lines, 976B) at sial.org/pbot/13588
autrijus :(
the crux is Test.pm can't be located. 20:42
which is not surprising
gaal is "Can't bless non-reference value" a trailing bug? 20:43
szabgab anyway, if you know anyone around please direct him to www.perl.org.hu/english/ he might be interested in HPW where $Larry will give the keynote
autrijus yeah
gaal szabgab: hurrah - folk from nearby nonbordering countries may also be interested in that ;-) 20:44
szabgab yeah, probably 20:45
autrijus gaal: do we have blib6\lib\test.pm ?
i.e. in your computer at this moment
gaal the changelog is disproportionate. there's nothing about tests or bugfixes and a whopping 32 JS backend bullets
szabgab we have one guy from Poland already which is not border
autrijus gaal: right, see if you con consolidate some of the backend ones. 20:46
obra Sadly, I'll be at the NPW that wekeend
autrijus ditto with obra.
szabgab We will try to setup the usual ?? video link
gaal autrijus: yes, it exists.
Test.pm
szabgab we just don't have internet yet :)
autrijus time to train pigeons 20:47
and implement rfc 1149
gaal hee
szabgab VOP? 20:48
gaal which MM does -BJS do now? 1.0? (to make consistent with the analogous line in Perl5) 20:49
autrijus JS follows its own implementation
rafl autrijus: What's the reason to not execute the perl5 tests? I don't fully get it from the commit message.
autrijus but the MOP is MM1ish
rafl: our make test sets ./pugs as the HARNESS_PERL. 20:50
gaal *nod*
autrijus rafl: a "perl5_test" target may work
but I'm not very enthusiastic to implement it myself
rafl autrijus: OK, when is the release planed?
autrijus rafl: ideally before I sleep.
which, since I've taken some antiflu medicine, may be soon. 20:51
if you'd like to Do It Right we can always delay until tomorrow :)
putter back. trying...
rafl autrijus: That's cool at all, but to soon for me to implement that. Isn't that important anyway.
No, always doing the right thing is boring. :-)
autrijus rafl: what if you have another 12hrs? :)
...but boring is the essence of kwalitee :) 20:52
gaal ?eval_sanity "JavaScript backend / Switched to much faster `-CPerl5`"
autrijus gaal: need some explanation
gaal: the JS codegen was parsing -CPIL by hand
now it's just taking the PIL AST as -CPerl5 and rebless. 20:53
rafl autrijus: Well, no. I'm busy getting the Debian packages ready. I don't think it harms not to run them.
autrijus rafl: okay then.
cool!
gaal autrijus: k
autrijus gaal: btw, precomp fixed (I think.) 20:54
gaal testing
autrijus r'03
I think this no-freeze policy is turning out well. 20:55
at each release I was tempted to reevaluate, but each time it turns out it's better :)
pasteling "gaal" at 192.115.25.249 pasted "works but may look like it doesn't" (9 lines, 437B) at sial.org/pbot/13590
autrijus yeah, those useless useless useless warnings. 20:56
gaal i'm unhappy with the changelog but am wary of doing something autrijus was wary of doing.
svnbot6 r7503 | autrijus++ | * Make sure that Test.pm is found for pil2js precomp.
autrijus gaal: maybe you can focus on the modules part.
I hear that is easier.
svk log -l 1000 modules 20:57
gaal "wanna hear about my modules?" echoes 20:59
autrijus lol
gaal what's modules, anyway
ext?
putter sigh. google's translation of 'Whole New Programming Republic of Perl6' is... very beta.
translate.google.com/translate?u=ht...uage_tools
autrijus gaal: aye 21:00
putter: yeah, I tried, can't trust that
putter: also you need to know something about current japanese politics to get the joke.
rafl want's to rewrite smokeserv with catalyst. Is the p6 port already available?
autrijus sri__ started working on one but didn't put it to svn iirc 21:01
I might be wrong.
gaal grrrr, how do i get upstream r numbers in svk log? i don't know if my r1215 is in this release...
forget it, i'll go by date
svnbot6 r7504 | autrijus++ | * put large DON'T PANIC buttons on PIL2JS messages
autrijus gaal: the log has (orig)
rafl gaal: For me it's displayed in brackets. 21:02
autrijus you can trace that.
gaal not here... r1216 (orig r1215): Stevan | 2005-03-27 03:14:41 +0200
autrijus the 1215 is upstream.
oh wait, local branch?
svk log -l 1000 //mirror/pugs/ext 21:03
then
or rather, 2000.
1000 is too few. (gasp)
rafl autrijus: OK, using p5 then.
gaal thanks 21:04
what's the svk option to -l that gives files modified in the r? 21:06
autrijus svk desc
gaal found it
autrijus oh.
-v
gaal i mean -v, yeah
Perl6-Container-Array is the new Array::Lazy? 21:07
autrijus I think so.
gaal k, i'll investigate...
autrijus cool 21:08
dduncan I'm back
autrijus greetings dduncan.
gaal yo
dduncan now svn upping and reviewing the p6l message, then will backlog
autrijus I'm waiting for this win32 test run to finish (soon)
then reboot
gaal is :)
dduncan I had some ideas on the pairs thing, but will see what's on p6l first
sri__ rafl: not enough time to work on Catalyst6 :( 21:11
dduncan finished the p6kl 21:13
s/k//
Ingo's basic presentation looks good 21:14
I have to think on how it would look to define a nested structure though 21:15
gaal in less(1), anyone know how to get a few lines of context above a search match?>
autrijus sri__: any code or sketch?
gaal like automatically hitting 'kkkkk' after 'n'.
putter autrijus: r7502 finds libperl.so. yay. some js warnings/errors, and some .a warnings of multiply defined .o's. otherwise looks usual (ie, just warnings for syck missing prototypes, and p5 redefs). 21:16
running test...
autrijus putter: ok, I think that's sane. I don't have tuits to squash those warnings, and I don't think shem as showstopping anyway 21:17
putter k
wolverian gaal, dunno, but '5k' works as well, if that helps. 21:18
gaal d'oh, Cabalization isn't mentioned in the ChangeLog :)
putter wonders if backend renaming should be revisited. calling the haskell backend "pugs" is a bit confusing... "pugshs"?;)
s/renaming/naming
gaal wolverian: actually i was hoping for something better, since if i do that the nexr time i hit 'n', i have to hit it twice :) 21:19
sri__ autrijus: nothing useful
autrijus sri__: k.
putter err, "a" ~~ /a/ Segmentation fault seems a bit unfortunate
autrijus putter: interp, or hs1
putter: er uhm, with both embed? 21:20
putter parrot 030 embed
autrijus very unfortunate. will fix after reboot
try external meanwhile?
with p5embed
gaal autrijus: are you waiting for me for reboot? 21:21
putter I'll let the current test run (the first) finish, then I'll build an external... unless anyone prefers otherwise. 21:22
autrijus gaal: no, still running tests
JAPJuggler hello everyone 21:23
gaal do Pugs.cabal.in and Pugs.sln disappear for you too sometimes?
e.g. after an aborted build 21:24
autrijus yes and I know why 21:25
fixing.
good catch!
JAPJuggler what do you guys think about something like mouseHole for perl6? 21:26
sri__ autrijus: maybe i'll have more time for fun projects next year :/
JAPJuggler I'm torn; I like the fun that Why and the ruby guys are having, but I want to stay true to perl. 21:27
we've copied their eigenclasses, so why stop there? 21:28
obra what _is_ mouseHole?
JAPJuggler I'm also thinking about creating some videos staring some camels to talk about current issues
svnbot6 r7505 | autrijus++ | * chdir now may fail, so try{}
autrijus it's a web proxy/ 21:29
mousehole.rubyforge.org/wiki/wiki.pl?MouseHole
JAPJuggler obra: mouseHole is a web proxy (in ruby) that allows you to run special ruby scripts that can fool around with a web page before your browser sees it
obra oh. is it hoodwink.d, but not in js
JAPJuggler kinda like greasemonkey
yeah
autrijus JAPJuggler: writing it in p6 would be feasible, but a bit challenging, today. 21:30
JAPJuggler my thinking was that if it was in perl, then you'd have access to all of cpan
dduncan oi
JAPJuggler autrijus, i see
autrijus JAPJuggler: check out how WWW::Recorder works?
I think there may be something like that in p5 already 21:31
JAPJuggler ok
obra HTTP::Proxy
JAPJuggler I've started to think of a lot of ideas lately, but not many people to talk about them with
cm you should blog about them, maybe in your use.perl journal 21:33
or do it the shapr way, implement them all :-)
svnbot6 r7506 | gaal++ | ChangeLog updates, todo bugs section
r7507 | autrijus++ | * do not accidentally nuke pugs.cabal and pugs.sln on case-insensitive, case-uncaring, case-brutal and case-apathetic systems.
autrijus mmm "code is speech" ;)
JAPJuggler heh, yeah, jfdi
autrijus gaal: there, fixed 21:34
gaal cool :)
autrijus I'll proceed to reboot now :)
gaal i'll proceed to hax0r svk log (ooh, fast!) to give me only bugstuffs. 21:35
JAPJuggler HTTP::Proxy seems like a good start
The slides O'reilly showed at OSCON about how perl has less hype have got me thinking how we can turn that around 21:36
i think pugs and all of the mad hacking that is going on will start changing this
autrijus JAPJuggler: the perl hype engine has gained some boost after publication of several Very Good Books.
JAPJuggler but I'd like to create more buzz and excitement by copying Why's techniques 21:37
autrijus I think we met at yapc::na
autrijus are you one of the jugglers? 21:38
did we meet at hotel lobby?
wait, there is just one juggler.
JAPJuggler yes, i was that juggler
i met you at the checkout desk 21:39
autrijus yeah.
excellent juggling, btw.
JAPJuggler thanks!
autrijus any video survived?
JAPJuggler that was my first performance ever
fulko said he'd send me a dvd, but i still don't have it. i was thinking of shooting him an email
I want to work on some more jokes and tricks for future performances.... :) 21:40
autrijus the perl6 one was brutally true and I love it :) 21:41
gaal how to get svk logs in reverse order?
autrijus I wish pugs can live up to your performance.
JAPJuggler i almost didn't do the perl6 one b/c I can't really juggle 6 balls... :) I kinda got lucky! :)
gaal 6! that's a lot of balls
autrijus gaal: something about reverse and alternate linebreaks
gaal heh 21:42
autrijus (meaning, I don't know, and I think a patch to SVK::Command::Log is desirable)
JAPJuggler I think I got 8 catches of 6 ball
gaal i can flash three, more or less.
:)
oh and i can drop seven!
putter re rules.t moving to parrot land... err... parrot getting a copy is nice... but as for it disappearing from pugs... um. while I've been known to enjoy coding without testing, perhaps all the current and future pugshs,pugsjs,pugsp5 rules code still needs to be tested. ;) i'm not sure what was envisioned here.
JAPJuggler heh
autrijus putter: er no, it's moving to Test::Base land 21:43
putter: the idea is to make them sharable.
putter: I have no problem with having a p6 harness for that spec of t::b
putter ah. got it.
autrijus I really think T::B is amazingly useful :) 21:44
JAPJuggler So I started a blog about a semi-fictional project: Perl on Poles
It's mostly a joke, but maybe some useful code will come out of it:
putter re "skipped tests - this file will be moved to parrot land"... so why are the tests being skipped?
autrijus ah it is _your_ doing.
JAPJuggler The official announcement is here: perlonpoles.blogspot.com/2005/09/im...ng-of.html 21:45
autrijus putter: because they were failing randomly depends on the PGE level
putter ah, so its "skipped tests - currently having problems with PGE" 21:46
autrijus yeah.
JAPJuggler: very amusing.
autrijus ponders "Perl on Pails" 21:47
JAPJuggler Perl on Pipes would probably be more correct, as we live in Vancouver....
gaal Perl on Praries? # for Mongolian hackers 21:48
autrijus mmmm Perl on Pugs 21:49
JAPJuggler heh
autrijus Perl on Pugs on Parrot on Ponie
JAPJuggler there is no pugs or parrot for fink, eh? 21:50
autrijus the amazing PoPoPoP technology
JAPJuggler heh
autrijus I don't know fink
probably not
JAPJuggler apt-get install pugs would be cool
:)
gaal ..and is available already :)
autrijus rafl did that for debian :)
JAPJuggler Hmm.... I guess it isn't in the osx tree
So if I want to get up to speed on perl6 and become more involved in pugs, is the synopses the best place to start? 21:51
gaal there are too few bug fixes in this release. we'll have to wait. 21:53
* Fixed infinite loop on sub { 1 }.pairs
* "sub foo" no longer silently means "multi foo"
... AND!?
putter r7502 test report - js complains again about not finding Test; is "TODO baz tests" a new category?;); dubious undef.t; 3 dubious pugsbugs/; pervasively dubious rules/; dubious eval_yaml pugs: toEnum{SyckKind}: tag (214748364802) is outside of enumeration's range (0,2) 21:54
gaal (yes, i did grep the log)
wolverian argh, debian doesn't have kwiki in apt? :(
gaal wolverian: apt-get install dh-make-perl ; dh-make-perl --cpan Kwiki :-) 21:55
putter and ext/ tests not run. end report.
wolverian gaal, true enough :)
21:55 typester is now known as sleepster
putter ;) 21:55
wolverian gaal, oh, thanks, also s:) 21:56
s,s:,,
er. s,\),:),
gaal uh, autrijus, do you remember what the adverbial block bug was all about? r7317 etc
heh
hmmm i also wanted to `nmake install`... lesseee 21:57
autrijus gaal: foo {1}.blah was parsing foo({1})
.blah
gaal this isn't the same bug as r6351 right? (sub { 1 }.pairs) 21:59
autrijus affirmative. 22:00
(no, it isn't.)
gaal on windows (vc++2003 built perl5), pugs.exe installs to c:\perl\bin\pugs.exe . not that i can think of a better place for it to go, but that's weird. 22:01
i mean it's not surprising, but it's weird.
wolverian gaal, oh, it didn't work. :(
autrijus yeah, been doing that for ages, now
gaal wolverian: :( it doesn't work on every package, apparently it (grep INGY META.yml)s 22:02
wolverian haha
gaal autrijus: ACK
wolverian oh, haha, this again 22:03
what is junoscript?!
I think I'm going to install something else than kwiki
stevan autrijus: you write the comic strip.. and I will draw it :)
autrijus ooooh attractive!
stevan autrijus: remember I wasted many thousands of dollars on art school :) 22:04
putter re PoPoPoP tech, there is always the dreaded Ruby On Perl6 On JavaScript On Perl6 On Perl5 On Rails with RbP6Js client scripting. ;)
autrijus I think cute foxes, while nice, are rather passōæ½xE9 now
JAPJuggler ōæ½xB5heh
autrijus (pass'e)
JAPJuggler how about cute camels?
gaal JAPJuggler: obviously you never smelled one
stevan long haired crazy asian camels ?
JAPJuggler :)
autrijus mmm dialogue between an lambda, a camel and a pugs
stevan talking lambdas are good 22:05
and pugs are funny no matter what :)
ok,.. dinner now,.. but I will backlog &
JAPJuggler have you guys noticed that the rails dudes are super paranoid about java? they seem so pre-occupied with it
dduncan regarding the discontent about the changelog ...
JAPJuggler i never noticed this with per lpeople
dduncan I suggest that Pugs Internals can have sub-sections for shared components and for each back -end 22:06
does kwid support ==== ?
likewise, Bug Fixes can be sub-divided similarly
JAPJuggler i went to Van ruby meeting at my office, and everything was referenced to java
no wonder they think rails and ruby is so cool, if you come from the cruel java world 22:07
(I'm installing ghc, this is why I'm so chatty... :) 22:08
autrijus dduncan: I think it does
there are binaries for ghc :)
dduncan if no one has any objections, I will modify the changelog likewise ... though it may need further massaging 22:09
autrijus and if you are getting 6.4.0, abort now :)
dduncan: go ahead.
JAPJuggler I found How to install Pugs on Mac OS X 10.4 in 22 easy steps
gaal dduncan: can ye hold on 1.5 min? i'll ci my changes first
JAPJuggler it suggests using DarwinPorts to install ghc
dduncan yes 22:10
JAPJuggler nice talking to you all, I hope to see you guys again soon... :) I know where to find you. ttyl
dduncan I also suggest another renaming/reorg of sorts ... 22:11
"Pugs Internals" should be more specific to say something like "Feature Changes"
because "Bug Fixes" also applies to Internals
gaal go ahead, Darren. 22:12
svnbot6 r7508 | gaal++ | more ChangeLog updates
dduncan re-upped
putter JAPJuggler: language communities sometimes define themselves by contrast with the competition. ruby's main competition in it's now largest nitche, business webbing, is likely java. python community developed as a "not perl". 22:13
JAPJuggler i see... interesting 22:14
gaal JAPJuggler: bye, and thanks for "Po" x Inf 22:16
autrijus ttyl, JAPJuggler 22:17
JAPJuggler heh, np... maybe I can commit some code next time... :)
gaal hey, /msg us your email and you can commit code now :) 22:18
JAPJuggler [email@hidden.address] should work 22:19
dduncan okay, I committed
JAPJuggler gtg, ttyl
gaal I'm done changelogging, myself
dduncan feedback is desired
gaal, I don't THINK I overwrote what you did
but check anyway
gaal :)
dduncan I didn't add any sections ... the other backends may want a mention
fyi, I have to go now for family meal, but will be back 22:20
gaal where's Beyond Compare when you need it...
dduncan I remember that 22:21
fyi, I moved bugs up to go above modules and tests 22:22
and backends mostly above that
otherwise, I added and changed nothing
svnbot6 r7509 | Darren_Duncan++ | Reorganized newer ChangeLog to rename Internals to Features and split it into main and backends - this is rough and should be massaged 22:24
gaal dduncan: looks good 22:27
autrijus, still awake?
no outstanding issues for win32 afaik. 22:28
autrijus good. 22:29
I'm awake; has been cooking
(literally)
putter x86_64 external parrot working... running make test... 22:31
autrijus I'm investigating in parrotembed and PGE. 22:33
bradfitz (hi. gaal twisted my arm into stopping in....) 22:35
leo__ autrijus: pmi and I exchanged some mail wrt status of PGE, maybe I can fix some issues tomorrow
autrijus leo__: good... but this is wrt 0.3.0 still
leo__ ah ok - then this doesn't help much 22:36
gaal whheeee! hello :)
autrijus hello gaal :) 22:42
gaal hee, that was actually aimed at brad 22:43
but hello chef!
autrijus is this the one you were referring about getting us a ruby backend? 22:44
gaal ah, no, the ruby backend was someone else
autrijus ah.
bradfitz: hi. :)
gaal twice over - it wasn't me who mentioned him :)
i mentioned someone who can get us perl5 on win32
bradfitz would love to write his baby in perl6, because he loves both. 22:45
s/write/rewrite/
autrijus oh. ah.
actually, I'm not sure that perl5 embedding would not work on win32 now, given a mingw perl 22:46
gaal autrijus: yeah, but we can get an msys perl5.8
autrijus care to try then? 22:50
gaal well, i'll have to whet my lasso skillz. 22:52
autrijus geoffb: we have been /.ed.
putter "frivolous"!?! humph. grumble. 22:55
autrijus blame chromatic.
revdiablo "from the haskell-has-never-been-or-will-ever-be-fun dept"
Gotta love the clever slashdot editors 22:56
autrijus note the "or"
gaal ... /. is not even in my url history
putter note! its a fresh /. article. only 3 comments. opportunity to shape the discussion. quick, what do we want to say!? ;)
gaal the 'or' is clever though. 22:57
"moose"?
autrijus putter: testify how fun it is?
sell merchandise?
autrijus has no idea
announce the imminent release?
gaal speaking of release, what's on the checklist? 22:58
autrijus final pass of chglog 22:59
Jooon Score:2, Insightful - I'm sure many more people would have a cleaner house if cleaning their house was fun. 23:02
putter well, could mention that -Ofun was also a matz objective for ruby 23:03
svnbot6 r7510 | autrijus++ | * edits of ChangeLog.
r7511 | gaal++ | fix typo
autrijus yeah and you can sneak an eigenclass joke into it.
putter have a message with links to pugs resources, irc, etc, so "maybe interested"s don't have to dig... 23:04
autrijus putter: sure, and then I can integ that to the homepage
gaal "Run Perl 6 Today! apt-get install pugs" etc.
rafl++ # did we say it enough?
putter pugs.wiki is currently non-vandalized... 23:06
elmex autrijus: why blae chromatic? 23:07
i liked his thoughts about the types...
dduncan this is a minor thing, but if anyone's still editing the changelog, the Perl 5 backend title needs a blank line beneath it 23:08
putter re /., currently up to 15 comments. I won't have a chance to post one. geoffb++ bbl &
autrijus elmex: he submitted that /. story so the "frivolous" is his doing
dduncan: did so
elmex ? 23:09
i don't know
what story?
autrijus developers.slashdot.org/article.pl?...mp;tid=156
putter 1 not having any serious purpose or value. 2 (of a person) carefree and superficial. 23:10
elmex 'a frivolous project (a toy Perl 6 interpreter) in an uncommon language (Haskell). ' -lol
wt?
wtf?
autrijus I guess chromatic was trying to be cute
elmex pugs is the first perl6 interpreter, and it's as toy as the first perl interpreter was
(not a toy) 23:11
autrijus not really cute, but hey, it's fine, I'm still grateful :)
elmex ;)
pugs is great, and haskell was a good choice for language research
autrijus elmex: instead of saying it here you could've been poisting to /. :)
*wink* 23:12
putter: PGE does not segfault here. :-/
putter test report, FC3, x86_64, external parrot, one failure, eval_yaml. ext/ tests didnt run.
autrijus mm, eval_yaml.
error log?
putter t/unspecced/eval_yaml..............................pugs: toEnum{SyckKind}: tag (214748364802) is outside of enumeration's range (0,2) dubious Test returned status 1 (wstat 256, 0x100) DIED. FAILED test 2 Failed 1/2 tests, 50.00% okay 23:13
Jooon autrijus: you won't by any chance go to stockholm after euro oscon? conferences.yapceurope.org/npw2005/....html#when
autrijus Jooon: yes.
talking about Pugs and Haskell.
the organizer was kind enough so I won't have to prepare new slides.
svnbot6 r7512 | autrijus++ | * evne more edits. this be preflight #1.
bradfitz can I make myself useful and run a make/make test on a bunch of different machines? if so, where's the release candidate?
Jooon autrijus: yay. I can't go to euro oscon, but I will be there 23:14
autrijus bradfitz: uploading.
gaal bradfitz: apt-get install ghc-cvs and cpan Bundle::Pugs::SmokeKit 23:15
actually maybe you have a goodenough ghc already (6.4.1 is great) 23:16
putter re segfault, "a" ~~ /a/ yields 0x0000000000c0364a in real_exception (interpreter=0x10bb080, ret_addr=0x0, exitcode=30, format=0xe3d91e "no get_params in sub")
at src/exceptions.c:702
702 the_exception->severity = EXCEPT_error;
autrijus Preflight #1: xrl.us/hxkk | geoffb's column (/. ed): xrl.us/hxhk | pugscode.org | pugs.kwiki.org | paste: sial.org/pbot/perl6 | www.geeksunite.net 23:16
autrijus bradfitz: xrl.us/hxkk 23:16
aka pugscode.org/tmp/Perl6-Pugs-6.2.10-pre1.tar.gz 23:17
bradfitz got it.
gaal 'hxkk' is a nice name.
autrijus gaal: indeed
elmex autrijus: i don't post @ /. :)
autrijus ok :)
dduncan I just noticed the '=' stripping ... good idea since a single main file heading appears not to be coming
getting preflight ... 23:18
autrijus ooh the cuteness backfired. 23:19
# developers.slashdot.org/comments.pl...d=13753083
dduncan the pugs distro has grown a lot btwn the last 2 releases ... a little under doubling in size, vs around 10% increments between previous releases 23:21
pasteling "putter" at 66.30.119.55 pasted "segfault bt" (66 lines, 2.9K) at sial.org/pbot/13591 23:22
gaal freepan's in a lull
dduncan that /. poster's another example of people not knowing about what they speak 23:23
putter bbl &
dduncan the first sentence particularly
mdmkolbe eval? perl 1+2 23:24
?eval perl 1 + 2
evalbot_7512 '3'
mdmkolbe ?eval help
evalbot_7512 Error: No compatible subroutine found: "&help"
mdmkolbe ?eval 23:25
?eval ?
evalbot_7512 Error: unexpected end of input expecting term
gaal ?eval "it does work"
evalbot_7512 'it does work'
gaal ?eval 6*9
evalbot_7512 54
autrijus putter: still looking into your bt.
gaal # base 13
mdmkolbe ?eval perl ((1=>2)&(3=>4)).value
evalbot_7512 '({ref:<Scalar::Proxy>})'
mdmkolbe How up to date is the evalbot usuallly? Is it built nightly? 23:26
gaal r7512 is the latest.
it's built all-the-timely.
autrijus: bump the version number :) 23:28
autrijus gaal: where? :) 23:30
gaal pugs -V
pugs_version: Perl6 User's Golfing System, version 6.2.9, August 3, 2005
svnbot6 r7513 | autrijus++ | * more changelog edits.
autrijus k done 23:33
leo__: ping 23:34
gaal smoke started, ETA 25 23:35
svnbot6 r7514 | autrijus++ | * adding throw_away to MANIFEST.SKIP.
r7514 | autrijus++ | * bump version number.
gaal ?eval $?PUGS_VERSION 23:37
evalbot_7514 \'Perl6 User\'s Golfing System, version 6.2.10, December 10, 2005 (r7514)'
gaal hey, that's funny 23:38
why is there a leading \ there?
?eval $?PUGS_VERSION.perl
evalbot_7514 '\\\'Perl6 User\\\'s Golfing System, version 6.2.10, December 10, 2005 (r7514)\''
revdiablo That is kind of bizarre 23:41
autrijus ?eval say $?PUGS_VERSION
evalbot_7514 Perl6 User's Golfing System, version 6.2.10, December 10, 2005 (r7514) bool::true
revdiablo Is evalbot running in the future?
December 10, 2005?
autrijus revdiablo: see my commit. 23:42
putter: reproduced; was triggered by unclean build.
I suspect Parrot_call_sub is not happy.
revdiablo Oh, so autrijus is from the future. That explains a lot!
svnbot6 r7515 | autrijus++ | * It being the 10th month this year does not make it December.
r7515 | autrijus++ | I blame the Romans. s/December 10/October 10/
revdiablo ;) 23:43
autrijus :)
gaal That doesn't explain the leading '
leading \
leading lambda! 23:44
may be *that* explains it.
autrijus gaal: that signifies a rw container.
since it is dumping the container
gaal ah!
has there been any change to t\rules\rules.t in this pf? 23:55
it takes a while before it's skipped here.
autrijus gaal: the rules are passed to macros. 23:57
I fixed that.
I mean, I made those not parse at all.
gaal will there be a pf2? i think i can't stay. 23:58
autrijus probably not, this will be release if nothing wrong happens
gaal smoke up. 23:59

Powered by Cro and the RakuĀ® Programming Language.

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

Thank you!