pugscode.org <Overview Journal Logs> | smoke/win2k (occasional): xrl.us/fqum (25/4554, r2403), Mac OS X (62/4506, r2362), Linux (38/4508, 2381) | pugs.kwiki.org
Set by Corion on 27 April 2005.
jabbot pugs - 2454 - * split the Eval monad into two construc 00:11
nothingmuch the web is evil 02:32
obra yes 02:37
nothingmuch catalyst is being nice to me though 02:43
how is bamboo progressing?
ingy: ping 02:54
obra nothingmuch: been working on book and work 02:58
about to start to need bambooish things.
so maybe things will happen
nothingmuch what does it actually do? 02:59
obra right now, increment a counter
nothingmuch hehe 03:01
what are your plans for it then?
from what I gathered it is sort of a toolbox for webapps?
obra oh. we've got some spec. you seen the spec? 03:02
nothingmuch nope
obra hang on
nothingmuch is having a braindead moment:
obra svn.openfoundry.org/bamboo/trunk/li.../Guide.pod
nothingmuch shift->method(shift);
which shift will happen first?
the right one, right?
crysflame uh 03:03
my p5 brain says left first
nothingmuch (p5)
crysflame oh, totally left
i think
nothingmuch you're right 03:05
nothingmuch is happy
crysflame left
wow, it's still there 03:06
goto &{ splice( @_, ARG0, 1 ) } if ref( $_[ARG0] ) eq 'CODE';
that's what your question reminded me of
nothingmuch that's scary
what is it?
crysflame it transfers execution to the coderef in $_[0] 03:07
nothingmuch right
but why like that?
crysflame well, it was a POE state
i could have made it more wordy but that did what i intended
nothingmuch goto &{ shift } if ref $_[0] eq 'CODE'; ?
crysflame precedence with shift?
nothingmuch or maybe i have no clue what ARG0 is
crysflame also, um, ARG0 isn't $_[0]
nothingmuch okay
crysflame it's more like an ENUM that varies at runtime 03:08
nothingmuch what's the precedence issue?
crysflame (ref: POE)
if you shift during a goto, what happens
mugwump svn 100 software engineers shoot you 03:09
mugwump didn't remember typing svn
nothingmuch crysflame: i'm very sleepy, i don't see what's wrong
crysflame i was really wary, that was all 03:10
not like wary and goto splice ref CODE mix well
but, hey
nothingmuch ooh, fun 03:12
i think i will learn to use XML::Twig
someone tell me if i'm sane for doing this:
i have an XML, in an xml: 03:13
crysflame no
nothingmuch <root><elem><nested/></elem><foo/><verbatim> ... xml here ... <elem><bar/></elem> ... </verbatim></root>
i would like to parse the first <elem> with <nested> in it
and extract some values
but i would like to get the string that is the XML inside <verbatim> unparsed 03:14
or rather parsed
but as text, not as a tree
is there a better way than what I'm hoping XML::Twig can give me?
mugwump sounds like a job for XPath 03:15
nothingmuch: see the xql method in XML::Sablotron
nothingmuch xslt... scary 03:16
mugwump oh, ignore that. it just gives you a nice DOM
maybe libxml has an xpath lookup method too
nothingmuch i don't see an 'xml' method 03:17
oh, sql
xql
damnit
nope, not in the docs either 03:18
mugwump sec, I'll find it
nothingmuch oh, it's ::DOM
mugwump search.cpan.org/~pavelh/XML-Sablotr...DOM/DOM.pm
#xql
should be able to call toString on the fragments 03:20
nothingmuch sounds good 03:22
the only problem i see is that again i won't be learning to use XML::Twig
mugwump I recommend to learn the DOM first, it's transferrable knowledge... 03:23
OTOH XML::Twig has xpath and toString support too 03:24
nothingmuch so basically i make a new situation 03:25
give it to Sablotron 03:26
with some XML
and then construct a DOM::Document from the situation, after it has been moved around a bit?
mugwump The situation doesn't seem to do very much, I've always just treated it as something spurious the module needs to work properly :) 03:27
bsb TreyHarris: did you get an answer to the reduce qn? 05:10
gaal morning 05:42
bsb hi gaal, just the person I was after 05:47
remember the make ghci failure? 05:48
seems to be a class between:
/usr/lib/ghc-6.5/HSbase_cbits.o 05:49
src/UnicodeC.o:00000000 T blkcmp
gaal, could you please run: nm -A /usr/lib/ghc-6.*/*.o | g 'blkcmp' 05:51
05:54 Aankh|Clone is now known as Aankhen``
gaal grep you mean yes? :) 05:56
bsb yep, my alias, sorry 05:57
gaal anyway, not found on my linux box.
bsb You have ghc-6.4 from experimental, yes?
gaal sounds like we need to rename a symbol in our copy of the code eh?
yup
bsb I guess we do 05:58
I'm reading about ghc packages, may be an alternative there... 05:59
gaal it's not referenced anywher ein the project, btw, wo it's simple enough.
ooh, that might be a better alternative.
forum2.org/gaal/pugs/smoke-msys.html r2454 06:00
06:22 chady_ is now known as chady 06:33 Aankh|Clone is now known as Aankhen``
bsb gaal: no luck at the package level 06:35
the only symbol not in ghc's base package is u_iswdigit
Why would they leave only one out? 06:36
TreyHarris bsb: no, i never got an answer 06:41
bsb I think reduce is foldl1 in the prelude 06:42
could be wrong though, /fold[lr]1?'?/ all confuse me
TreyHarris my machine with ghc is rebooting... do you happen to know where the prelude is in hugs? 06:43
06:44 castaway_ is now known as castaway
bsb in hugs do :find fold1 06:44
":find foldl"
TreyHarris oh, cool. didn't know about :find
bsb It's nice 06:45
gaal TreyHarris: this may be useful too: haskell.org/ghc/docs/latest/html/li...index.html 06:46
bsb: try renaming the symbol in our copy of UnicodeC? 06:47
TreyHarris i see... so it's like i suspected. foldl1 calls foldl. since you can't make a single function both return a list and an element
gaal it's only mentioned one apart from the declaration
autrijus hi \camels 06:48
gaal the better solution would be to introduce symbol mangling but i don't know how to do that.
ehy autrijus
autrijus problem with latest ghc? 06:49
gaal permute<ehy>
obra hey autrijus
autrijus yo obra
in ghc 6.5 (cvs version tthat is) I suggest disabling our UnicodeC
bsb autrijus: yes, clash on UnicodeC.o
autrijus bsb: I'll try a fix. sec 06:50
bsb note that haskell lacks u_iswdigit
TreyHarris does !! end up getting optimized somehow, or is it really O(n) as the prelude specifies it? 06:53
autrijus TreyHarris: I think it's O(n) in the general case 06:54
obra hey Schwern
Schwern Hey.
Stevan asked me to have a look at Test.pm and the todo stuff. I don't understand what all the fuss is about, looks trivial to me. 06:55
TreyHarris autrijus: thanks. btw, i find pugs incredible amusing, because it's given me an excuse to do Haskell again after 15 years... it was my very first "real" language :-)
autrijus TreyHarris: that's good... 15 years ago you didn't have monads, right? :)
TreyHarris i don't think so.... i didn't even learn I/O, because it made no sense to me. either i've learned enough in the interim to understand monads now, or they were using something else then 06:56
autrijus I think they are using 06:57
main :: [Response] -> [Request]
back then
TreyHarris yeah, that looks familiar
i don't think Gofer even had that though 06:58
decay castaway: you didn't try to find us yesterday, did you?
gaal autrijus, time for clues on #line?
castaway Nope 06:59
autrijus gaal: give me some time to fix the Unicode.hs
gaal sure. i'd ask on #haskell but it looks like some of this is too pugs specific?
Schwern So the goal here is to make TODO tests fail visibly when developing Pugs?
How does one run an individual pugs test? 07:02
gaal Schwern: prove -v t/test.t 07:03
but you need to set up env first:
export HARNESS_PERL=$PWD/pugs # or ./pugs, whatever you prefer
export PERL6LIB=ext/Test/lib 07:04
Schwern make test TEST_FILES=foo.t seems to work
gaal that's cool too (does it do verbose?)
Schwern Probably. Its just makemaker
*** Error: Can't modify constant item: VStr "TODO: defined sub (symbol table)": NonTerm SourcePos "/Users/schwern/devel/pugs/blib6/lib/Test.pm" 195 9: NonTerm SourcePos "/Users/schwern/devel/pugs/blib6/lib/Test.pm" 67 5 07:06
Any idea what that means?
196 is $desc = "TODO: $desc";
Are function arguments read-only by default? 07:07
gaal yes, add 'is rw'
Schwern is rw works?
Str ?$desc is rw? 07:08
Maybe I want is copy
Yep
gaal i think they do, Schwern.
autrijus you want is copy. 07:09
Schwern Ok, force_todo is now unnecessary.
autrijus Schwern++ # mad perl6 skillz
bsb: please retry the build r2455 07:12
bsb: and let me know if ./t/syntax/unicode.t works 07:13
bsb ok 07:15
./t/syntax/unicode.t All tests successful 07:21
jabbot pugs - 2455 - * GHC 6.5 had the new unicode semantics
bsb make ghci only works if I remove UnicodeC.o 07:23
I didn't make clean though
gaal forum2.org/gaal/pugs/smoke-msys.html r2455 ghc6.4 msys 07:26
autrijus bsb: cool, then it works 07:32
gaal #line? :)
autrijus :r
err, ww
gaal: ok, url? 07:33
gaal sial.org/pbot/9738 and sial.org/pbot/9739 07:34
notes: 1. i know i'm doing something wrong aroung "ld <- try $ do" because fail "" from there leaves the function;
2. this breaks with e.g. t/statements/last.t - with very weird debug prints, e.g. reported line numbers that go up to line 72 - much more than the length of that file.
the second url is a test file that does work correctly. 07:35
Alias_ autrijus: #threatnet is finally alive :)
autrijus gaal: ok. I'll bring it to lunch :) 07:36
Alias_: what is that?
or rather, breakfast
gaal btw, you prefer the whole parser.hs file? 07:37
Alias_ autrijus: Spiffy idea 2004-D :)
autrijus: Globally networked greylisting system
gaal i can upload that too if you'd rather.
Alias_ many-many stateless over IRC
ali.as/threatnet/
could have sworn I mentioned it earlier
autrijus gaal: how about you just committing it? 07:38
gaal it breaks some tests
autrijus so?
gaal good point. :)
autrijus I thought the rule is it's okay to commit things as long as it doesn't break Test.pm :)
gaal okay, i'll merge with lastest r and clean up the trace prints. 07:39
castaway I thought it was "commit anyway, someone will fix it" :)
gaal castaway: i don't want someone to do my wokr for me 07:40
i want to figure out what i'm doing wrong, with help of someone who knows better than me :)
castaway then label it "ork in progress" ?
(preferably with a "w" :)
gaal orks in progress. 07:41
it breaks many tests where there's a comment starting at column 1. that scarier? :) 07:42
r2456 07:45
bsb Does assuming have an inverse, an unassuming method?
gaal in haskell that's called uncurry
bsb But that's a bit different, I think 07:46
gaal don't know about the p6 version
bsb And it doesn't sound as nice as "unassuming"
gaal agrees unassuming does sound nice.
bsb uncurry bundles the first to params in a tuple
gaal *curry* sounds nice to all indian food lovers though. 07:47
bsb Mmm. I think I'm getting one cooked for me tonight.
jabbot pugs - 2456 - #line directive, work in progress 07:52
autrijus woot. 07:55
autrijus goes to lunch &
tainted- hi 07:56
are newbies welcome here?
gaal of course!
tainted- #perl is incredibly intolerant of newbs
gaal tainted, this channel is primarily for development of perl6 though, so p5 Qs aren't very appropriate 07:57
however, if you have p6 Qs, fire away :)
castaway perlmonks.org is better for newbies.. :9
tainted- yea i need a perl newbie channel 07:58
gaal perlmonks.org is *excellent*.
castaway try the chatterbox on perlmonks
tainted- k
castaway (thats a website, not irc, btw :)
gaal there are some good p5 newcomer lists.
www.perl.org.il/mailman/listinfo/starter/ is one, hosted by Israel.pm. 07:59
tainted- wow great thanks guys
10 minutes in here > 2 hrs in #perl
castaway well we're probably less busy too, (the channel that is) 08:00
not that I've ever been in #perl
Alias_ Which #perl :) 08:02
castaway any :) 08:03
castaway waits to see when tainted turns up in the CB..
mornin osfameron 08:13
osfameron morning castaway
castaway how's fotango-land? :) 08:14
osfameron yeah, pretty good 08:15
castaway nifty
osfameron yeah, lots of exciting stuff going on 08:16
(though I'm pretty much doing stats automation ;-) )
castaway cool 08:17
castaway is pondering/attempting to apply for a job there ,)
osfameron good luck! 08:18
or break a leg, or something
castaway heh
actually a recruiter asked me if I was interested, and I think she's still waiting for feedback 08:19
osfameron Preferred?
castaway sorry? 08:20
osfameron I guess not then (name of the agency that I went through)
castaway oh the agency..
yup, them 08:21
jabbot pugs - 2457 - unbreak the build
pugs - 2458 - * The Eval monad is now split into (SIO/
autrijus omg. I'm seeing an unbelievable amount of speedup. 08:25
gaal autrijus, are the tests parsing for you? i think i've broken them all :)
autrijus I think at least 3x.
gaal: you've broken them all. 08:26
I'm fixing.
gaal at least it means they run a little bit faster. *duck*
autrijus uh no
if you would, update to r2455
(before your commit)
and apply r2458 by hand
I think it will result in a huge speedup.
mandel.p6 is blazing here. 08:27
gaal BTW: i've made it so pugs can build on cywgin 08:28
autrijus cool
gaal since wokring with msys is like wearing a hairshirt
08:29 Aankh|Clone is now known as Aankhen``
gaal that's a large patch, autrijus 08:31
autrijus it's in autrijus.org/tmp/r2458.diff 08:32
gaal oh, thanks, no i can manage with svn
i was commenting on the 08:33
hmm
insanity :)
wow, i hadn't realized cygwin was under so heavy development 08:34
i wanted to install patchutils which i embarassingly didn't have installed, now it's updating many mbs 08:35
prolly want a reboot sometime :( building on the other box in the meantime.
the trick to upgrading cygwin w/o a system reboot is to do it while all cygwin programs are closed. 08:43
sucks, but less than a complete reboot. 08:44
autrijus gaal: #line is in 08:47
now I really need to get food.
&
gaal thanks... i'd like to understand how it's done
oh
i'll ask you later then. bon apetite.
autrijus see Lexer :)
you can hack in filename support 08:48
&
gaal right.
jabbot pugs - 2459 - * fix #line directive. 09:01
gaal woot! autrijus, your patch provides something on the order of another 11% speedup on make smoke. 09:04
bsb S06/Pointy subs: does not require a preceding comma when included in a list 09:08
does that include: @a = ("one" -> $x { $x**2 }, "three")
or is it only to allow for @list -> $elem { ... } 09:09
jabbot pugs - 2460 - * Pugs can now build on cygwin 09:31
gaal forum2.org/gaal/pugs/smoke-cygwin.html r2460/cygwin 09:41
autrijus: ping 09:51
i'm afraid #line still doesn't work. the tests are very funny. 09:52
i.e. adding a subtest makes the ones before it pass where they previously had failed. 09:55
jabbot pugs - 2461 - basic tests for #line, see comment withi 10:01
nothingmuch good morning 11:06
sort of
castaway afternoon :) 11:08
theorbtwo Good sort-of-morning to you too, nm. 11:15
12:18 chady is now known as chady_
stevan hey all 12:37
Steve_p Good morning 12:38
theorbtwo G'morning, Stevan, Steve.
Erm, g'afternoon. 12:39
.oO(Time flies when you're cursing Make.)
stevan good morning guys :)
I see Schwern was in here helping on Test.pm, but I dont see any commits
any idea oh what he did? It looked as if he eliminated the need for force_todo()
which would be nice :) 12:40
jabbot pugs - 2462 - * rewrote position handling and statemen 12:41
pugs - 2463 - * bare block execution restored in the p
autrijus greetings. 12:42
I just rewrote everything :D
stevan autrijus++ # :)
autrijus (well, in this regard, it's everything in the parser part)
Steve_p heh
stevan autrijus: any idea what Schwern did with Test.pm? I am not seeing any commits
autrijus had no idea how many things will break
autrijus runs test
stevan: he added "is copy" to $desc or something
but I think he forgot to svk push
stevan autrijus: actually think he misunderstood what I was asking him about 12:43
Juerd autrijus: What keeps you motivated?
stevan oh well I will email him
theorbtwo Morning, autrijus.
autrijus Juerd: thank-you messages
yo theorbtwo.
we have a much cleaner Pos model 12:44
$ ./pugs -e 'die 123'
*** Error: 123 at -e line 1, column 1-8
theorbtwo I'm finally starting to get somewhere with the haddocking. Make sucks; the problem was in spaces-vs-tabs.
Now the problem seems to be in haddock's haskell parser. 12:45
Juerd Working on the same thing for 87 days, rewriting things over and over until they're perfect, communicating with almost everyone in the Perl universe... It's a form of commitment I haven't previously seen anywhere.
Juerd is very glad you're doing it, though :)
autrijus Juerd: see, seeing you're glad keeps me motivated
Juerd Admirable 12:46
autrijus Juerd: I don't really ask much... I just want a working Perl 6 :)
theorbtwo I'm also glad you're doing it -- and think you are mad as a hatter.
autrijus mmm hatter
stevan working perl6 ++ 12:47
theorbtwo (Heavy metal poisoning)--
Juerd autrijus: Right. That statement is in bizarre disagreement with itself.
stevan (((insanity driven-development)++)++)++
Juerd heh
theorbtwo wonders if the karma-tracking agents grok the parened form. 12:48
Juerd I don't think autrijus is insane. Mad, yes, but one has to be pretty sane and have good feeling for sanity to implement something as vague and futuristic as Perl 6 in current time.
stevan Juerd: IMO insanity is 1) not always a bad thing 2) fairly subjective :) 12:49
" normal " thought processes are not always appropriate when programming 12:50
Juerd In your opinion, it is subjective. Hm, nice.
stevan: insane is a form of unhealthy
stevan the thought complexity of programming requires a mind which can be heavy in the abstract, but still remain grounded
theorbtwo Hmm... haddoc doesn't like the ^# \d+ ".*" style of line directive (which ghc in .hs->.hpp mode outputs), nor does it like #ifdef.
theorbtwo wonders if he's using an old haddock, or if it really needs some perl lovin. 12:51
scook0 theorbtwo: I've been working on fixing stuff for haddock
stevan Juerd: its all diagnositic BS IMHO
jabbot pugs - 2464 - * the last block on a Stmts chain needs 12:52
theorbtwo Fixing haddock to work with pugs, or fixing pugs to work with haddock?
scook0 Fixing pugs
I found you can work around the #ifdef problem by enclosing each directive in {- -} comments 12:53
I'm looking at some other stuff as well
theorbtwo I'm doing it by preprocessing first, and am about to stick in a perl -i -pe after that.
scook0 How well does that work? 12:54
theorbtwo Donno yet. 12:55
Juerd autrijus++ # Just read the journal. I should make this my browser's home page.
theorbtwo RSS++
Juerd Yeah, but this is the first entire journal I'd like to RSS
theorbtwo (But I generally read it as soon as it's announced, before it hits my friends page.)
Juerd I haven't used RSS yet because use.perl has no all-journals feed. 12:56
autrijus heads up: fp.p6 infinite loops. fixing
Juerd Woohoo
car's ready
It now runs on LPG
theorbtwo LPG? 12:57
scook0: Have you hit the problem of haddock disliking TH yet? 12:59
Juerd theorbtwo: A cleaner fuel. 13:00
That happens to be less expensive too
theorbtwo Wow. Eval.hpp makes my xemacs' hilighter take lots of CPU. 13:01
Juerd In the netherlands, it's available at almost 50% of all service stations
theorbtwo Sounds great!
Juerd is now waiting for his brother 13:02
There he is
gone.
afk
theorbtwo Later!
scook0 theorbtwo: TH? (no...)
theorbtwo Template Haskell.
autrijus ...679 failed. 13:04
theorbtwo r2465 has a rule for making hpp files from hs files.
Try make src/Pugs/Compile/Haskell.hpp, then feed that to haddock.
13:07 Qiang_ is now known as Qiang
theorbtwo Ah, somebody on #haskell says that a new haddock is expected "soonish". 13:08
scook0 'Soonish'... 13:09
theorbtwo Yeah, I rather wonder when that is as well. 13:10
I don't see an accessable VCS on the haddock page, either. 13:11
scook0 I wonder if there is a way to tell Haddock to ignore chunks of code...
jabbot pugs - 2465 - Rule for making .hpp files for haddockin
gaal autrijus, ping 13:15
autrijus pong 13:16
kungfuftr *splat* it was a cherry bomb
gaal saw the amusing magicals/line_basic.t ?
autrijus yeah.
gaal that's on par with the kind of weirdness i've been moosing with... 13:17
but most of all i want to understand how i generalize try 13:18
so i can say something like { filename <- try ruleFileName ; setPosition if filename (....) else (...) } 13:21
autrijus sure
you have to test (null fulename)
gaal i can't seem to fail "" there
because that dose weird things to the rule
also there's still the weird bug that makes me think maybe the rule isn't terminating correctly 13:22
theorbtwo OK, have working haddock documentation, but only after a lot of manual work commenting out the bits of the .hpp that haddock doesn't grok. 13:23
(TH, GADTs, and some stuff that I don't know why it doesn't like.)
scook0 You haddocked the whole thing?
theorbtwo Yep.
autrijus ooh. 13:24
theorbtwo Now I have to either fix haddock, or figure out how to automate that manual bit. 13:25
gaal oh, and also: skipToLineEnd should probably try to consume the newline too, no? At least as an optimization.
theorbtwo The problem with just commenting out the bits it doesn't like is that then we loose documentation for them.
autrijus gaal: I'd rather you keep it... 13:27
not sure
gaal uh, oh, larry just posted about context :)
autrijus uh. 13:28
autrijus head explodes
gaal autrijus, okay, that's just a thought about optmization. what worries me much more is correctness, and i don't understand the bug where snipping lines from the end of a file causes passing subtests to fail.
autrijus "I think S02 is probably wrong. It should be unknown/list context. 13:30
in Soviet Russia, Synopsis writes Larry!
Draven_ yay 13:31
autrijus ok, there should be less test failures now 13:32
gaal smokes 13:33
autrijus gaal: example.t still hangs 13:38
Limbic_Region is having trouble coming up with a real world example of a list you might need to process that would be too big to fit in memory all at once
any suggestions? 13:39
autrijus Limbic_Region: a video stream?
integral do strings count? If so, genetic stuff
Limbic_Region integral - I was just thinking of genetic stuff
integral any area where you've got lots of experimental data I suppose, physics, medcine, etc 13:40
Limbic_Region real problem - typically the genetic stuff would be in a file and you would just read it one line at a time 13:41
gaal attention all smokers: if you have a script that uploads your smoke results automatically, put it in PUGS_SMOKE_UPLOAD and run-smoke.pl will call it for you.
Limbic_Region I am trying to have a real world example of iterators where the next item is calculated from the previous
bah - nevermind, I will come up with something
gaal pi?
integral random number generator?
gaal primes? 13:42
jabbot pugs - 2466 - * fold redundant Pos cells; this should
pugs - 2467 - * Allow user-defined upload command for
machack666 hola 13:44
Anyone want to interpret a portion of S04?
I'm trying to figure out if traits are introspective. 13:45
S04 seems to hint that, but does not go into detail about how we are doing this. 13:46
theorbtwo You mean if there's a way to ask "does this value have this trait"?
Er, s/value/container/
machack666 correct, but I'm dealing with cclosure traits
if &?BLOCK is the current block, I'm not sure of what the best way to specify syntax for (say) all of the ENTER blocks 13:47
but we also have the wrinkle where individual variables can be declared with traits
my $var will first { get_a_job() } 13:48
this statement should add an implicit FIRST{} block to the containing closure, with $var as the topicalizer. No problems there
but do we also need to have an introspection method for the $var variable itself --- $var.traits or something 13:49
theorbtwo I should think so.
machack666 so does &?BLOCK.traits make sense as well?
theorbtwo Yes. 13:50
machack666 I would think that we could just have traits be a rw hashref with the trait type as the key
theorbtwo The .traits method would be on Any, which includes Block.
So would I. 13:51
machack666 i.e., &?BLOCK.traits<ENTER>
would give us an array of the closures to be run on block entry
theorbtwo Makes sense to me.
jabbot pugs - 2468 - * add Pos back to structural folding.
theorbtwo Post to p6l about it, and get a larry to put in in an S? 13:52
machack666 but if we allow user-defined traits we need some sort of transofrmation rule for the control traits when they are defined.
theorbtwo Why?
machack666 my $var will later { run_away() } 13:53
if "later" is user-defined then does it get transformed into $var.traits<LATER>?
or do we just leave it as is? 13:54
we should be consistent about naming. Since closure-level traits are ALL CAPS, and variable-level traits (which may extrapolate to the block level) are lower, we need to resolve those for the triggers. 13:55
unless we want to try to execute all traits for both &?BLOCK.traits<ENTER enter>
i.e., either the variable traits become uppercase or the block traits become lower
well, gotta go 13:56
jabbot pugs - 2469 - * even more structural induction fixes o 14:01
pugs - 2470 - * even more folding
gaal gotta go too. see 'ya. & 14:07
autrijus testers, please don't write 14:23
is (...) 14:24
anymore
is(...)
is the only correct way
clkao you hate my space!
autrijus yes.
theorbtwo (Significant whitespace)--
autrijus uh. 14:25
print (4+3)*7
you think it should print 7?
I find that to be quite indefensible :)
theorbtwo It's a hard choice for me what the better of those two evils is. 14:26
jabbot pugs - 2471 - * abstract all induction into "Unwrap" c 14:32
pugs - 2472 - * whitespace fixes
theorbtwo My preference, I think, is that it prints 7, and warns of multipication in void context. 14:33
autrijus but it is a warnable offense in p5. 14:35
%s (...) interpreted as function
I don't know, really. 14:36
Juerd said it's been decided that bareword functions no longer globs space
theorbtwo It's only a warnable offense with print. 14:38
...and only with exactly one space. 14:39
autrijus which is weird too. :-/
I prefer consistent whitespace-sensitive rules over arbitary whitespace-insensitive rules :)
theorbtwo No, it makes sense; the foo-in-scalar-context warning covers it when it's a problem.
autrijus I think I'd rest my case. 14:41
I see now it's a defensible position, but my brain is not wired that way :)
jabbot pugs - 2473 - * unTODO a test 14:42
pugs - 2474 - * Noop is now consistently ignored durin
stevan hola Schwern 14:47
Schwern Hey 14:48
stevan can you commit what you did yesterday?
Schwern Yeah, sorry, I was working on it while waiting for a friend's flight to come in at the aiport... and she came in.
stevan ah 14:49
no big deal, I am just anxious to see what it was
Schwern So all it does is change all todo tests to be not todo unless %ENV<PUGS_RELEASE> is set. 14:50
stevan ok
Schwern And if the test failed.
stevan but there are many tests which should still TODO
as I mentioned in my email , there is kind of 2 layers of TODO
Schwern Oh.
stevan those which should work, but not yet 14:51
Schwern Then why not make two?
Wait... no, I really don't get it.
stevan and those which are not yet implemented and are far off
Schwern Ahh.
Make two levels of todo.
stevan take object for instance
they are way off (which in pugs land is a few weeks)
jabbot pugs - 2475 - * more whitespace.
stevan and then odd forms of for loop (for @array &sub) 14:52
which are just not on the radar, but "should" work
how would you suggest we go about 2 level-todos?
Schwern Gotcha.
stevan then of course, the whole policy of "when releasing failures are TODO"
which I know has its drawbacks/issues 14:53
Schwern That's simply wrong.
theorbtwo Perhaps :todo<6.2.8.3>, which is TODO until version 6.2.8.3, then becomes a hard failure.
autrijus Schwern: so you'd suggest releasing with todos?^Wfailures?
stevan theorbtwo: thats maybe a little too specific :)
Schwern theo: I doubt things are that stable.
stevan maybe :todo is regular TODO 14:54
Schwern autrijus: I'm against the idea that ALL failures should be todo when releasing.
autrijus Schwern: what would you do as an alternative?
stevan and :todo(2) is for "not even close stuff"
Schwern autrijus: Umm.... let them fail or don't release with failures.
autrijus Schwern: we are on times releases
and we did let them fail
Schwern autrijus: Then don't run the tests.
autrijus but then, lots of people download tarballs and mail me
saying they have failures.
Schwern did you *expect* them to fail? 14:55
autrijus yes.
Schwern Then mark them as todo!
autrijus we are on a timed release schedule.
sure, but luqui has this great idea
stevan Schwern, that is waht we do (with force_todo())
Schwern Like, *ding* its 12 noon... release!
autrijus of having two TODOness
one is for things that are should be fixed asap
(bugs)
one is for things we can do at leisure
Schwern autrijus: Right, that's what we're talking about right now.
autrijus (features)
stevan Schwern: more like, *ding* its 2am Tiawan time, ... release 14:56
autrijus s/2am/7am/
Schwern stevan: Is this release automatic?
stevan LOL
Schwern: not really
theorbtwo autrijus: what do you think about the idea of sprinkling #ifndef HADDOCK around?
Schwern Good. Then yes, take anything which is exepected to fail and mark it as TODO.
autrijus theorbtwo: I think it's fine.
stevan Schwern: that is what force_todo() is for :)
otherwise it is a tedious process 14:57
Schwern stevan: Except its horribly unmaintainable.
autrijus :todo<bug>
theorbtwo Perhaps :knownbug and :wishlist?
stevan (truet me I did it many times)
autrijus :todo<feature>
stevan Schwern: it isnt really maintained though
Schwern Exactl
y
stevan it is there for release, and removed afterwards
stop-gap measure only
:todo<some attribute>++ # I like that 14:58
autrijus and we can have some flag
Schwern PUGS_RELEASE is what I used.
autrijus that ignores certain flags
Schwern Its committed.
autrijus oh good.
you rc0k.
stevan Schwern: thanks
Schwern Well, Test.pm will honor PUGS_RELEASE but nothing sets it
I couldn't figure out the twisty maze of MakeMaker hacks. 14:59
autrijus Schwern: is it a ENV var?
Schwern Ayep
autrijus k.
masak help /unjoin 15:00
Schwern stevan: Ok, so you can handle the multi-layer todo? 15:01
jabbot pugs - 2476 - * repair hash unwrapping on rhs. 15:02
pugs - 2477 - Make all TODO tests fail visibly unless
stevan Schwern: yup!,... muchas gracias for the input
nothingmuch hola stevan 15:06
Schwern Excellent. My work here is done. 15:08
perlmonks-bot www.perlmonks.org/index.pl?node_id=450922
theorbtwo Who was that masked bot?
autrijus Juerd: all our S02 semantic went out from window in r2476 15:19
* * Larry invoked Rule 2. We deal.
stevan hey nothingmuch
nothingmuch is developing with Catalyst and loving every minute
stevan nice
I have been reading all that stuff, looks very cool 15:20
nothingmuch woobling.org:81/page/מי_אנחנו <-- currently somewhat broken
it is
but soon: two models (simple xhtml in xml files, and CDBI)
templating was very easy 15:21
control flow is logical
and it's very very flexible
jabbot pugs - 2478 - * Larry invoked Rule 2. We deal.
theorbtwo Hm, how many backslashes do I need for this insanity? 15:24
stevan '/' x Inf 15:25
theorbtwo I want to do a perl -pe 's/(...)/...$1.../' from a Makefile, which is generated from a qq() heredoc in a Makefile.PL.
Oh... but I need doublequotes on the outside for portability. 15:26
autrijus theorbtwo: you can invoke util/ instead 15:27
util/munge_haddock.pl
look at how util/src_to_blib.pl
was invoked
theorbtwo Hm, it feels silly to invoke a sepperate file for something as short as -i -pe "s/^(#.*)/{- $1 -}/"
autrijus trust me, it is for the best.
oh and -i is unusable on win32 15:28
have to roll you own
or use -i.bak
theorbtwo Ugh. Silly default mandatory locking. 15:29
jabbot pugs - 2479 - * fix tic-tac-toe in light of the contex 15:31
Juerd autrijus: Too bad... 15:42
Corion Did we remove most of the :todo(1) tests? I see lots of red in my smoke :) 15:44
Juerd They were made :todo tests if I read the journal correctly. 15:45
Not that :todo should result in anythign other than :todo(1), though
But if that's broken, everything using it breaks too :)
Corion BTW, the data on my machine (datenzoo.de) is still accessible via ssh, but it won't come up in non-recovery mode at the moment it seems. Even after I did a fsck already ... 15:46
theorbtwo Probably some important file gone missing.
/etc/fstab, at a guess. 15:47
Juerd Hehe
Corion theorbtwo: Hmmm. Interesting idea ...
Juerd fstab isn't important except at boot time :)
Corion BTW, I see both, failing and :todo(1) tests, so it's not :todo that is broken
Juerd Unless you boot with a writable root fs :)
Corion (dark green and red)
No, /boot is a separate partition and is likely mounted ro 15:48
Juerd /boot doesn't have to be mounted
It's accessed directly by the boot loader, before any fs is mounted. 15:49
And after that, it's no longer needed except for reconfiguration of the boot loader of course.
Corion ah, so it's merely mounted after that as a convenience
Juerd (Assuming Linux. I don't know how other systems do this)
Corion Juerd: Linux is correct :)
Juerd Corion: Well, not just that. It's usually mounted to indicate that it is indeed a separate partition
While otherwise installing packages might poison the otherwise empty mount point. 15:50
It's then better to get errors about not being able to write on a read-only filesystem than to think everything is alright, reboot and then discover you were putting the files in the wrong place.
theorbtwo Well, something is causing it not to boot normally, in my experince this normally means that it didn't fsck and mount -a OK. 15:51
(As other errors will normally get ignored.)
Corion Ah, likely the problem with the todo tests is either with the perl script or with harness - the test output still has TODO: in front of it :)
Juerd theorbtwo: Then boot manually instead of with init scripts, and find out what's wrong
Corion theorbtwo: Yeah, except that I just did a recovery boot+fsck
Juerd theorbtwo: Put "init=/bin/sh" in your kernel command line. 15:52
theorbtwo Exactly.
Juerd, I know.
Corion ... but maybe the box comes up soonish, or I'll have to do another reboot and go in in recovery mode again, to get a better hint :)
theorbtwo I'm currently trying for psychic debugging just for the hell of it.
Juerd theorbtwo: Though skipping fsck shouldn't mess things up that much.
Unless it is RH/FC :P
Corion Juerd: But skipping fsck makes things take a lot of time - and I don't know if fsck is waiting for input :)
theorbtwo The only way to fix a RH system is to install Debian over top of it. 15:53
Corion BTW, the Y operator now seems to work - I get 2 unexpected successes in all that red :)
Juerd If it's a headless box, you should configure fsck to not wait for input :)
theorbtwo++
Corion theorbtwo: I'm considering doing that with Debian over SuSE now, but as all the data is still available, I'm reluctant
Juerd: Even then, the reboot/fsck will take a lot of time without feedback
Juerd More time than not running it 15:54
But in my experience, never more than 5..10 minutes
Unless you forgot to enable DMA before fsck runs
In which case it can easily take an hour.
Corion ... I need a unicode keyboard, to properly enter Ā„
Juerd Corion: In vim, ^KYe 15:55
Corion Juerd: I'm impatient
theorbtwo Most modern kernels go to DMA mood straight out of boot.
Corion Juerd: I don't use vim :)
mauke compose = Y
Juerd Corion: Then I don't care anyway :P
mauke you need a compose key
theorbtwo Try altgr-shift-z on an xfree86 german keymap.
Corion theorbtwo: Neither altgr-shift-z nor altgr-shift-y work, but maybe I need newer keyboard drivers. Oh. You say xf86 - I use Win2k :) 15:56
theorbtwo I need to figure out how to get my 102nd key to work as compose.
(I use a German keyboard with an American layout.)
mauke theorbtwo: me too :-) 15:57
theorbtwo (American keyboards don't have the key to the right of the left shift, so it's not mapped to anything useful.)
Juerd theorbtwo: X? With xev, find out its key code, then run xmodmap -e 'keysym 123 = Multi_Key'
mauke theorbtwo: it's mapped to ōæ½xAB ōæ½xBB ōæ½xA6 here
U+00AB: LEFT-POINTING DOUBLE ANGLE QUOTATION MARK [ōæ½xAB]; U+00BB: RIGHT-POINTING DOUBLE ANGLE QUOTATION MARK [ōæ½xBB]; U+00A6: BROKEN BAR [ōæ½xA6]
Juerd Broken bar?
Are you living in 2003? Infix zip is Yen now ;) 15:58
What would you need a broken bar for?
mauke it looks almost right :-)
theorbtwo xmodmap: commandline:0: bad keysym name 'Multi_Key' in keysym list
mauke the normal usage of that key is < > | 15:59
theorbtwo It is on a german keyboard; on a british keyboard it's \ and broken bar.
ingy morning
Juerd theorbtwo: Then it's Multi_key or multi_Key 16:00
theorbtwo: I recall there being something funny with the caps
And if it's neither of those, then s/keysym/keycode/
mauke I use the "context menu" key as compose key
theorbtwo Hm, that had no apparent effect, but didn't give an error message either.
Juerd theorbtwo: Now try to use it.
theorbtwo mauke, I don't have one of those -- I have a pure 102 layout.
I did -- I get <, or > when shifted. 16:01
Juerd theorbtwo: <multi> <a> <e> should yield ōæ½xE6
theorbtwo I just get "<ae".
Juerd Did you use the right key code?
And try keycode, instead of keysym. I never know the difference
jabbot pugs - 2480 - The yen/Y/zip operator works now as yen
Juerd It works as itself? 16:02
theorbtwo Oooh, that worked.
Juerd theorbtwo: Show us.
theorbtwo xmodmap -e 'keycode 94 = Multi_key'
compose,a,e=Ʀ
(I'm UTF-8.)
Juerd pretty. Now switch to iso-8859-1 so I can read it.
:P
In a few years I'll be migrating to utf-8 16:03
When hopefully, screen will play that game too
theorbtwo No, you switch to utf8. Then you can read hebrew, chinese, and all sorts of other strange languages that you can't understand, but can go "oooh, pretty".
Also, it has useful characters like ā‚¬.
Juerd I don't care about hebrew and chinese 16:04
Cyrillic characters are often not written in utf-8, so that's not fixed by that either
And Esperanto people use ASCII nowadays
g-with-circumflex is gx.
theorbtwo Not for me it isn't.
machack666 theorbtwo: useful characters like a-hat not-equals carriage-return? ;D
autrijus well, in order to talk sensibly about perl 6
Juerd theorbtwo: It is. 16:05
autrijus this has to be a utf8 channel to reduce confusion for all parties involved :)
Juerd theorbtwo: You have to mentally transliterate it, which comes naturally after a few minutes of Esperanto chatting.
theorbtwo No, like the euro currency symbol, which is not in iso-8859-1.
kungfuftr theorbtwo: iso-8859-15 iirc
integral migrate to iso-8559-15 then :-)
Juerd Who cares? The official symbol is EUR anyway.
theorbtwo (It is in iso-8859-15, but if you're going to switch from 1 to 15, you might as well switch all the way to utf8 and be done with it.)
Juerd The E thing is a logo, not a symbol. It has strict specifications for the angle of serifs, for crying out loud. 16:06
Well, not serifs, but the line endings.
mauke ... which everyone ignores, it seems
Juerd It's not just a C with an = through it.
mauke: Most fonts do it right.
theorbtwo IIRC, it did only for a very short time.
Juerd theorbtwo: Switching to -15 is much easier. 16:07
theorbtwo I know.
Juerd If I want to switch to utf-8 I have to (HAVE TO) migrate all boxes with it.
And I have to give up being able to access things like my e-mail from any console I happen to find.
Which is too much to ask.
I use screen with alway-open programs in it. 16:08
I don't want to restart everything each time I connect.
integral hmm, pity screen can't reencode from utf-8 to iso-8859-15
autrijus sure can 16:11
^A:encoding utf8 iso-8859-15
^A:encoding iso-8859-15 utf8
for the other way
Juerd is going for another test drive
Hopefully, the roads are a bit less busy now
Stupid traffic jams.
machack666 Juerd: is your name pronounced like "Word"?
Juerd machack666: No. I'd write it like that then.
afk 16:12
mauke huhu, someone in EFnet/#perl was trying to read an @array with while ($line = <@array>) 16:13
it almost works
machack666 autrijus: is Eval.hs where the AST gets "executed"? 16:14
autrijus machack666: yes
machack666 so if I need to trigger block-level actions at runtime, I would need to add code there? 16:15
autrijus most possibly yes
machack666 and if something is done at compile-time, it would (could?) go into Parser.hs?
i.e., attqaching the traits to the appropriate block-object 16:16
autrijus yes. 16:17
theorbtwo Uh-oh. I can no longer use the key between 5 and 7 on the top row of the main section of my keyboard. 16:18
16:18 zuulvin_ is now known as mila
theorbtwo Restarting X, BBIAB. 16:19
jabbot pugs - 2481 - * limit PUGS_RELEASE scope for release t 16:21
machack666 bbl 16:25
Corion Grrr. I guess it was an unloaded network card module... Grrrreat. 16:27
(I wonder how the datacenter set up the machine :) ) 16:28
Maybe via serial console :)
jabbot pugs - 2482 - Small cleanup of Test::proclaim. 16:31
pugs - 2483 - * EOLs at EOFs
pugs - 2484 - * Beginning work of MonadEval class
pugs - 2485 - * Eliminated explicitly passing env arou
Corion Yippie... 16:32
ingy Juerd: ping 16:36
Juerd: ot. is there a way in irsii to jump to the window with the most recent activity 16:37
Corion Happyness is a machine reachable over the network.
ingy Corion: actually Happyness is merely a mispelled locution 16:38
as is mispelled ;)
Corion r2479 - datenzoo.de/pugs/win2k.html - 867/3781 16:44
pugscode.org <Overview Journal Logs> | smoke/win2k (occasional): xrl.us/fqum (867/3781, r2479), Mac OS X (62/4506, r2362), Linux (38/4508, 2381) | pugs.kwiki.org 16:45
obra autrijus? 17:02
What the hell changed in the last 12 hours with pugs?
autrijus hi. schwern made all our :todo fail.
or rather, all the force_todo fail
so you should see red. 17:03
don't panic :)
"env PUGS_RELEASE=1 make smoke"
will give the original green one.
Corion ah ;)
obra no. I mean it smoked 30% faster
autrijus obra: oh. yeah. I rewrote the monad
obra :)
autrijus++
Corion I found the all-green version better, so I knew when I broke something important :)
autrijus we are now in a six story monad 17:04
EvalT ContT ReaderT SIO (STM/IO) a
ingy I am using '1 * foo' to force foo() into scalar context in a numeric situation (in perl 5)
anyone think of other clever ways 17:05
mauke 0 + foo
Corion ingy: ~~ foo
ingy well that's about the same
autrijus yeah, ~~foo
obra pugscode.org <Overview Journal Logs> | smoke/win2k (occasional): xrl.us/fqum (867/3781, r2479), Mac OS X (62/4506, r2362), Linux (164/4642, 2484) | pugs.kwiki.org 17:05
Corion ~~ is the scalar context operator in Perl5 17:05
ingy ok nice
Corion (like its long-word equivalent, "scalar", but with different binding) 17:06
:-))
ingy it would be cool if +foo worked
Corion ... but that will maybe nummify your return value. Especially nasty with overloaded return values.
So "scalar" might be your best bet unless you go for obfuscation :-)
Oh - you're talking about "numeric situation" already. So ~~ it is. 17:07
mauke ~~foo doesn't numify
Corion mauke: Hmmm - true. Strings have xor too.
theorbtwo ~~ either nummifies or stringifies.
(It never returns a ref.)
jabbot pugs - 2486 - * $?CALLER::POSITION repaired 17:11
Odin-LAP mummifies! 17:12
:D
ingy mauke: I numify in foo, so it's ok by me
thanks folksies 17:13
Corion ... and I thought that this was not the channel for Perl5 questions ... 17:17
Ontolog does Pugs really need GHC 6.4? Won't it run with 6.2? (thats the latest version for Gentoo)
Corion Ontolog: Compile from source then. 17:18
Bwahahahahahahahahhahahaha.
Ontolog shove it!
Corion Ontolog: The reason for 6.4 is Template Haskell and lots of #ifdefs I think 17:19
mauke Corion: huhu, still compiling
Ontolog why does Haskell gotta be such a hassel?
Corion mauke: I thought Gentoo was really fast, because it's optimized exactly for your machine?
mauke I don't run gentoo
the binary release doesn't work because it requires glibc 2.3 17:20
I can't compile glibc 2.3 because of a parse error in pthread.h (wtf)
I think I need a new computer 17:21
Corion I luckily got around needing to reinstall a box.
Qiang looks at the P166 machine in the corner
jabbot pugs - 2487 - * #line tests largely passes now.
castaway Ontolog: gentoo has 6.4 too 17:42
you need to fc it some tho 17:43
Ontolog castaway: how do I unmask it?
castaway force
lemme look, sec
/etc/portage/package.unmask: >=dev-lang/ghc-6.3
Ontolog ahh 17:44
thanks
castaway it's probably easier if you install ghc-bin normally, then 6.4
Ontolog i have ghc-bin 6.2 installed now
jabbot pugs - 2488 - * move callCC back into MonadCont 17:52
paddyez hi 18:16
is there a good document on the history of Pugs? 18:17
obra yes
use.perl.org/~autrijus/journal
basically, daily updates from the very first day
paddyez mkay I was thinking on sothing more compact 18:18
obra well, it's only 3 months old
paddyez a summary would do
obra Autrijus Tang has been leading a team of hackers to implement perl6 in Haskell. 18:19
;)
paddyez " I caught some bad cold today, but still managed to start going through the failing tests, reducing them to a more manageable number." it starts with the weateher
obra use.perl.org/~autrijus/journal/22965
that's the first entry
autrijus paddyez: PA01 is the summary. 18:21
see pugscode.org, click on "Overview" :)
if you want concise history, read the ChangeLog file.
jabbot pugs - 2489 - * somehow sort VComplex.
paddyez thanx a lot all 18:22
autrijus np :)
paddyez: hm, maybe you can help improve the Wikipedia Pugs entry 18:23
while you're figuring the history out :)
paddyez yes that is what I am trying to do 18:24
but in the german language though
autrijus ah. I see the .de one is much better than .en
hm, how come the .de is not backlinked to .en?
paddyez well it is from today 18:25
I am doing a complete rewrite on www.wikiweise.de/ a german fork
autrijus a fork!
hm, the search form on wikiweise is borken 18:26
some tomcat error page
paddyez it is still buggy
the article does not exist yet
I am right at it creating the article to schow a friend of mine who is a perl freak how it is donde 18:28
done
autrijus btw, I've added interwiki tag 18:30
paddyez great
autrijus also [[Autrijus Tang]] :D 18:31
oh, duh, it's de.
never mind :) 18:32
paddyez *ggg* 18:35
autrijus journal up. see you tomorrow! 18:50
&
castaway night autrijus
jabbot pugs - 2490 - * remove "import List" from AST.hs 19:01
Corion Hmmm. Defaulting to $_ is broken... Or at least, it consistently doesn't work for abs.t and int.t (which I'm about to commit :) ) 19:19
... a non-pointy for loop doesn't seem to set $_ currently... 19:24
theorbtwo Add a test to t/statements/for ?
Juerd ingy: Yes, meta-a
Corion theorbtwo: I'm about to, resp. also looking into the problem ;)
Gah :( And I also committed my trace()-ridden test-build of Pugs - sorry ...
castaway roll it back? 19:25
Corion svn revert++
Juerd ARGH
IBM puts "quality seals" on hardware boxes
And I test hardware before going to clients 19:26
The most recent quality seals leave a pattern of words on the box
"VOID VOID VOID VOID"
theorbtwo Ugh.
Juerd That's not a nice thing to give your client :|
theorbtwo WD-40?
Juerd Now I have to re-box things.
I still love IBM though. 19:27
Don't get me wrong.
Corion Hmmm. Seems like I'm too dumb to revert my changes by checking out an older version. So I'll undo my changes locally and then recommit. 19:28
Juerd just built a new floor in his... what's the word for the big empty space behind the back seats of a car?
castaway depends which language ,)
boot
Juerd I made it out of polystyrene and hard board :)
theorbtwo Trunk.
Juerd castaway: I know that, and was hoping for a fight ;)
castaway (a trunk is a big chest to put clothers and things in ,) 19:29
castaway grins
Juerd Ah, right, those were the words.
theorbtwo Right. In order to transport them, for example.
Juerd I have an LPG tank instead of the spare wheel now
And it's a bit taller than the wheel
theorbtwo What do you do when you tire goes flat?
castaway tyre!
LPG?
Juerd castaway: Yep 19:30
castaway whats that?
Juerd theorbtwo: They gave me a can/tin of stuff that inflates and fixes/mends the tyre
castaway: Yep is to mean "yes"
castaway Grr.
What is "LPG" ?
Juerd You brought that up, I thought you'd know what it was :)
It's a fuel 19:31
castaway Thank you
Juerd Liquid gas
Liquid when under enough pressure
castaway sounds like a contradiction in terms ;)
theorbtwo en.wikipedia.org/wiki/Liquified_petroleum_gas
Juerd And new LPG injection systems never vapourize it, so it stays liquid all the time. So "gas" isn't exactly correct anymore.
jabbot pugs - 2491 - Added test for default argument of int() 19:32
pugs - 2492 - Move int.t below builtins/math, revert d
pugs - 2493 - Backed out tracing in Prim.hs
castaway 2ein gut'un
Juerd What is "2ein gut'un"?
castaway (oops)
a typo ;)
Juerd Is it valid intercal?
castaway ;)
masak :)
theorbtwo Juerd, "gas" never referred to the third state of matter; it's short for "gasoline". 19:33
Juerd theorbtwo: I know, but the current most common word for it is gas.
Regardless of how the words became 'gas', both LPG and gasoline are called gas, and they're both liquid :) 19:34
castaway (if you like, think away the 2, and translate from german)
Juerd castaway: What is 'un?
castaway silly ammis ;)
Juerd What is ammis?
castaway its a written rendition of a spoken word/sound
americans 19:35
Juerd finds it funny by the way that people outside western europe all ask me what LPG is.
This way you'll never satisfy kyoto :)
castaway do they?
Juerd Many, many do.
castaway shrugs
Juerd Anyway, I now drive cleaner and cheaper 19:37
And I always wonder why not many more people do this 19:38
castaway grats
Juerd I'm sorry I keep asking you, but what is a grat?
(You speak some kind of non-technical english that I don't understand! :)) 19:39
castaway its short for "congratulations"
Juerd Oh - I knew the "congrats" abbrev.
Thanks anyway :)
castaway :)
learning++
Corion Hmmm - I think I see the problem ... the "implicit topic" $_ is passed in as a subroutine parameter to the for-loop-code-block, but there isn't seen as the global $_ ...
Juerd Corion: Big mistake there! 19:40
Corion: $_ is lexical in Perl 6!
Corion Juerd: Yes, but I think the lookup is done wrong
Juerd Not dynamically global as in Perl 5
Corion Hmmm - I could write a test against that, actually. 19:41
Set $_ to something, and check what value it has in $_ = "Hello World"; for "Goodbye World" { say $_ }; # should be "Goodbye World"
jabbot pugs - 2494 - Test for implicit loop topic (re)used as
pugs - 2495 - Made Pugs.Compile.Parrot mostly work aga
Corion Err - { say } # of course - I need to check defaulting. 19:42
Juerd If you do this, put in a test for say for "Goodbye World" too
Corion Hah. And it outputs "Hello World"
Juerd As that's even more likely to go wrong
Corion Juerd: I don't get you? 19:43
Juerd say for "g w"; should also say "g w"
And after that, $_ should be back to what it was before the for-statement 19:44
Corion Ah - yes
Juerd This is scoping without { } and thus a special case
Corion Juerd: Thanks for the hints - they hopefully made the test clearer... Now, if I could find out where that stuff is actually implemented ... :) 19:49
jabbot pugs - 2496 - Changed test for checking for lexical $_ 19:51
gaal aloo! 19:54
theorbtwo Allo, gaal. 19:56
Corion gaal! 19:57
gaal how're things? :) 19:58
gaal backlogs
theorbtwo Things are going slowly, but progressing, for me.
gaal still haddocking, or the other stuff? 19:59
theorbtwo still haddocking.
It doesn't like TH or GADTs.
gaal if i were in its place, i wonder how i'd react to them! 20:00
castaway pokes HTML+CSS
theorbtwo Eh. 20:01
It should make use of ghc, rather then being a copy-and-paste-then-edit fork of it. 20:02
gaal is curious, taking a look at this haddock thing now 20:03
masak is haddock really sufficient for newbies?
Corion Hmmm. Where do the magical default variables (*_) come into play? In Parser.hs, I don't see them fudged in ...
masak feels like some external documentation would be good, too
theorbtwo www.haskell.org/ghc/docs/latest/htm...index.html is straight out of haddock.
castaway bro! 20:05
theorbtwo masak: Certianly would be.
But starting somewhere is good.
gaal Corion: see AST.hs:1184
masak theorbtwo: true
gaal (i think) 20:06
Corion gaal: That's right at the EOF for me ... Do you mean Typeable1 ?
gaal sorry, 667 - defaultScalarPAram
no, scratch that 20:07
it's not used anywhere. no idea then :)
Limbic_Region pokes his head in
Limbic_Region looks up to see that between work for $ and an article he hopes to get published @perl.com, he hasn't even peeked in to see what is new and exciting in pugs/p6 20:08
theorbtwo Oooh, nifty, Limbic. 20:09
gaal ah, I think it's called 'topic' in Parser|Eval|Monads
Corion gaal: I'll poke the stuff there then ;) I exposed a nice global $_ vs. local $_ bug (resp. implicit lexicalisation of $_)
What I wonder though is, that "for" does not mention any topic :) 20:10
castaway it should
Limbic_Region stevan ping 20:11
Corion ... yes - I guess that's part of the problem ;-)
theorbtwo What's the article on?
Corion ... I just don't know how to tell it that, resp. where a bare block without params specifies that it has one parameter...
gaal ah, a favorite p5 gotcha area :)
you know 5.10 has 'my $_', yes?
forum2.org/gaal/pugs/smoke-cygwin.html r2496
Limbic_Region theorbtwo - www.perlmonks.org/index.pl?node_id=451278 20:12
dragonchild said I should talk to chromatic about publishing it
not something I had in mind when I wrote it, but I contacted chromatic anyway 20:13
Corion gaal: Yes, I'm aware of that. There isn't much difference between my $_ and local *_ though, except that "my $_" works in closures :)
Limbic_Region chromatic said that the code snippets had to be both working and practical before he would consider it
Corion btw, my smoke test for r2496 is also up 20:14
gaal there was a very long thread and interesting about it on clpm once
castaway practical snippets++
Corion Feh. And I was stupid. I inserted fresh tests into for.t, and didn't push the force_todo numbers ...
So now we get unexpected successes there :)
castaway *g* 20:15
stevan Limbic_Region: pong
Limbic_Region did you see the tests I added for last? 20:16
stevan yes, very nice :)
Limbic_Region found a bug - which was really cool
I also modified next to be in align with last
gaal wow, i see that in the 6 hours i was gone there was *another* speedup (after the 11% of earlier, the 30% of yesterday, the ~50%(?) of the day before that).
Limbic_Region but I am not sure I did it right
stevan Limbic_Region: they are hard tests to wrap your head around 20:17
Limbic_Region before the force todo patch - I was under the impression that all tests marked as :todo wouldn't report as failures
but when I run pugs t/statements/next.t - it told me there was 1 failure
stevan it takes time and patience (and I had neither, hence i commited the minimal tests)
Limbic_Region I am still not sure next.t is right
but my pugs debugging fu is non-existant 20:18
so.... if you get a chance you might want to look at it ;-)
theorbtwo Oh, a nit that isn't mentioned in the replies -- under "what is an iterator", a filehandle isn't an iterator, but the readline operator is -- <$handle> in scalar context, not $handle.
Limbic_Region thanks theorbtwo - will change right away
stevan Limbic_Region: I will try
Juerd Hmm?
theorbtwo: That would mean that Higher Order Perl uses the wrong terminology throughout the book 20:19
stevan Limbic_Region: I think the force_todo patch Schwern did may need some re-thinking
Juerd theorbtwo: Something I find hard to accept, because MJD is a hero/god.
theorbtwo Juerd, how do you pronounce your name?
stevan it is too strict
Juerd theorbtwo: As Juerd
stevan LOL
theorbtwo Um, thanks.
Juerd theorbtwo: See juerd.nl/name or perlmonks.org/?node=Juerd
Or both
I'm not into sound samples though
qmole that's a popular question today :)
Juerd So you'll have to know what the german u-umlaut sounds like (or the french u in une) 20:20
Limbic_Region stevan - I don't expect to be around much this weekend, but if you need anything next week let me know
I should also check to see if ingy applied the Perl6::Bible patch I sent him 20:21
Juerd theorbtwo: Is this satisfactory?
stevan Limbic_Region: my pugs-time is pretty limited lately, my $work is getting in the way :) 20:22
theorbtwo I think so, yes.
Limbic_Region stevan - me too since my free time is non-existant (closed on the new house this past Wednesday) 20:23
ok, so ingy didn't apply the patch I sent him but he didn't not apply it either 20:24
I could have swore he said he would be releasing last night
castaway heh?
Limbic_Region oh well
theorbtwo It's some sort of strange heisenburgianpatch?
Limbic_Region theorbtwo - yep, until ingy uploads the next revision for everyone to see (opens the box) - the patch is neither applied nor not applied 20:25
castaway *g*
Limbic_Region ok, so talked to stevan and looked at Perl6::Bible - next on the todo list is Jonathan
theorbtwo You could check the SVN repo.
Limbic_Region who doesn't appear to be around.... 20:26
theorbtwo - huh? Wouldn't it have gone to CPAN first?
castaway you have a todo list full of people?
Limbic_Region people I needed to talk to - yes
theorbtwo His wife might get angry if he went around doing people.
...and very confused if he went around doing men.
stevan Its the list of people to "take care of once and for all" I am worried about 20:27
Limbic_Region chuckles and is reminded of Ovid use.perl.org/~delegatrix/journal/24450
theorbtwo Hm? Ovid has lots of names, or what? 20:29
Limbic_Region you aren't aware that Ovid recently had his CC # stolen and tracked down the criminals himself? 20:30
theorbtwo Er, right. 20:31
Nevermind.
I thought stevan's mention of people to take care of once and for all reminded you of ovid, and you meant to link to one of ovid's several entries about it. 20:32
(I read his LJ.)
Limbic_Region ahhh - well in any case, it is weekend so I am off
TTFN all
theorbtwo Later, Limbic!
Have fun packing. ;)
stevan add theorbtwo to his "list" ;)
theorbtwo Oh, you have a LJ?
stevan no just a "list" (muwhahahahahah) 20:33
theorbtwo Yeah, that was a joke.
stevan :P
Corion BTW, www.kill-them-all.de/ has lots of German punk mp3s (only by Die Ƅrzte) but it's good (live) punk :) 20:37
theorbtwo (Oooh, I can type Ƅ without switching keyboard layouts now.) 20:39
Can you suggest a good one for an introduction to the band? 20:40
(compose key)++ 20:41
Corion theorbtwo: "Elke" 20:42
Wait - I'll look for some URLs :)
mauke "es fing an, als sie mich anrief ..." 20:43
Corion www.kill-them-all-space2.de/bootleg...salles.mp3 www.kill-them-all-space.de/bootlegs...attivo.mp3 www.kill-them-all-space.de/bootlegs...am1tag.mp3 20:44
mauke: Yeah, I'm just looking for a speed-punk version instead of the one live-funk version I've got here :) 20:45
Wie am 1. Tag isn't bad either
_metaperl Corion, do you live in Germany? 20:46
castaway ;)
good question that
theorbtwo wgets.
castaway Is the Pope German?
stevan Corion is the Pope? 20:47
mauke no, bavarian (huhu)
Corion stevan: No, I just pull his strings
castaway good point mauke ,)
stevan Corion: I knew there was more to you than Haskell and Perl6 :P
Corion stevan: I want to streamline the process of string-pulling, that's why I need hyper-operators and junctions! 20:48
stevan Corion: ahh, I see, so Pugs is one part of your plan to take over the world 20:50
Corion stevan: Yes. Just in case my flying monkeys fail
stevan I thought that was what the >>+<< operator was 20:51
a flying monkey
just very far away
Corion stevan++ # >>+<< is the flying monkey operator 20:52
castaway HTML is weird! *sigh* 20:55
Juerd No, >>@<< is the flying monkey operator! 20:56
All we need now is an infix @. Suggestions?
(What, we have a \W ASCII character that has no meaning yet? MADNESS!) 20:57
stevan Juerd: it only looks like that when the monkey gets closer
castaway (hmm, or I'm silly.. )
this song is kinda boring
Juerd castaway: HTML is weird AND you're silly. What do you think of this compromise?
castaway sounds about right
"Ist das alles" is the entire text? 20:59
Corion Oooo - my favourite song - Madonna's Dickdarm - www.kill-them-all-space.de/bootlegs...ckdarm.mp3
mauke no, it's the refrain
Corion castaway: That's the refrain, yes :)
stevan has no german speed punk, so he just queues up California Uber Alles 21:00
Corion stevan: See "Madonnas Dickdarm" - there is some talking though, before the guitar hits :) 21:01
Hah. I've found why $_ doesn't get propagated into for blocks. for 1 { say } # doesn't mention $_ within the block, and thus the name never gets propagated to the block-names ... Now, how to fudge this :) 21:06
( retVerbatimBlock is the offending part, together with ruleBlockLiteral, both in Parser.hs)
I guess a special Block just for ruleBlockLiteral is the first fix ...
theorbtwo Woo! 21:10
Grr.
Well, it was working, then I did an svn up, and now it isn't.
Juerd Corion: Then the rule has to change, because you can't at compile time know that a sub will or will not access $CALLER::_
Corion Juerd: Hmm - no, it's not about $CALLER::_, which is fully specified resp. needs padwalking anyway. 21:11
theorbtwo: Maybe I broke some line endings at the end of Prim.hs again?
Juerd Corion: for 1 { $say() }
Corion: You can't know whether $say will use $_
Corion Juerd: Yes, but $say won't see the topical $_, as $_ is lexical 21:12
Juerd Corion: It will if it uses the $_ in CALLER::
Corion The implicit topicals ( say() ) are now an exception to the rule ...
Juerd Corion: Which is exactly how defaulting works.
Corion Juerd: Hmmmmmmmmmm - that's true...
theorbtwo Nah, up conflicts. I'll fix them fairly quickly.
Corion Juerd: That's another thing that needs to be fixed, true... Care to write a test "for" that case too? 21:13
Juerd No, sorry
Well, in fact: yes, but I won't
I have to minimize coding as much as I can. It's too stressful, time consuming and addictive. 21:14
(Although not-coding is also frustrating)
Corion Juerd: Ah ;)
Oh :(
Juerd As an example, I MSIE-proofed the design for convolution.nl/nieuw/site.plp
As you can see in its source, it took quite some javascript to kick MSIE in order to do simple PNG things
I also had to change the entire CSS based layout for this crappy "browser" 21:15
castaway cheats by devving in htmlkit, which previews using ie ;)
Juerd It worked in firefox/mozilla/camino/omniweb, konqueror/safari, opera.
nothingmuch www.ccc.de/biometrie/fingerabdruck_...anguage=en
Juerd And then I spent over 5 hours making it work with MSIE.
castaway: That doesn't fix any of MSIE's problems. I needed the alpha channel of PNGs, which MSIE does not support in <img> tags. 21:16
nothingmuch Juerd: there's a js hack
Juerd castaway: You have to trick it into using the right library, by using a proprietary css attribute "filter" with the right thing
castaway oh.. didnt nm just solve domething similar?
nothingmuch yup
although it's not quite solved 21:17
Juerd nothingmuch: I didn't know that, but I wrote some JS myself. This was pretty easy.
nothingmuch it's hacked
Juerd nothingmuch: It also doesn't quite work on backgrounds
Corion Huh? Where is "retTopicalizedBlock" defined? I only find it mentioned in Parser.hs ...
nothingmuch woobling.org:81/ <-- view source
Juerd nothingmuch: So I needed more JS, to handle horizontal repetition...
nothingmuch oh, i dunno about backgrounds
castaway hands Corion a grep -r
nothingmuch IE sucks so bad
Juerd But the end result works: convolution.nl/nieuw/site.plp
nothingmuch: The font is all weird and the page is mirrored! DEFACE! :)) 21:18
Corion castaway: I use Win32 find-in-files, which doesn't find it below src/ - which is why I'm asking :)
theorbtwo Corion: These guys talk a lot, but sound like they'd be good not-live.
nothingmuch yeah, some idiot made it all hebrew ;-)
holy crap! your site is also gibberish
Juerd nothingmuch: woobling.org:81/pngfix.js
nothingmuch htdocs/pngfix.js 21:19
sorry
Juerd nothingmuch: That's some weird display, I expected plain text. Nice!
nothingmuch thank catalyst
castaway doesnt find it anywhere at all ;)
nothingmuch that's what 'die' looks like
Juerd Thank you, catalyst
nothingmuch woobling.org:81/htdocs/pngfix.js
Juerd nothingmuch: Your JS does more or less the same as mine, but in a more complex way
nothingmuch updates the template
it's not my js
i'm just using it 21:20
=)
Juerd nothingmuch: I just abuse a 1x1 transparent pixel GIF
nothingmuch: So I don't have to create a new object with equal size
nothingmuch: See my code
nothingmuch ah
i won't know what it does
i don't really know what mine does either 21:21
just that it loads the png in a different way
Juerd Mine just adds the filter and changes the image to something transparent
That's all
And it does so for all images, not only PNGs. I'm currently not using any other images anyway.
The repetition code sucks though
nothingmuch browsers suck 21:22
Juerd And when I was done, I realised I wrote a lot of code for just three images :)
var xtotal = 785;
var xwidth = 273;
That's 3 tiles only :)
Just writing the HTML manually would have been easier probably
But I didn't think of that at the time.
nothingmuch would normally wait for ie7 to come out 21:23
but this site is not mine
=)
Juerd When it's out, it'll take at least a year before most people actually use it.
It's easier to force them to use a better browser 21:24
theorbtwo IIRC, MS has stated that there won't be an upgrade option other then buying a new OS that has it.
nothingmuch be
Juerd However, this is a commercial site and I wouldn't want to lose potential customers at the very first stage.
nothingmuch it's easier to not develop web content
theorbtwo Or not use non-binary PNG transparency, at least. 21:25
Juerd One thing I hate about MSIE is that it requests my transparent 1x1 pixel GIF file *six times* per page.
Because it is used 6 times.
MORONS
theorbtwo Though it should be noted that maps.google.com uses it heavily.
nothingmuch haha
theorbtwo Juerd: Get your webserver to send an expires header with it. 21:26
Juerd theorbtwo: Google managed to forget most of our planet.
theorbtwo I don't think they forgot about it so much as haven't gotten around to it yet.
Juerd theorbtwo: It *does* 21:27
castaway germany is next
Juerd theorbtwo: It's a static file...
theorbtwo Oh. OK, that's kind of scary then.
castaway supposing they can find an advertising partber ;)
Juerd I'm thinking about including it in the JS source
theorbtwo Corion, you still here? 21:28
Does the PM CSS caching hack work with IE?
Corion yep (I close Chatzilla when I'm away)
uh ... dunno, should, but I didn't test it I think
Hmmm. Seems like I'm getting nowhere tonight with my hacking on Parser.hs ... Have a good night all! 21:41
theorbtwo To play with haddockability, try find src -name \*.hs|perl -pe 's/hs$/hpp/'|xargs make && haddock `find src -name \*.hpp` (after getting r2497). 21:44
jabbot pugs - 2497 - We have haddockability! (There's not ye 21:51
theorbtwo Oh, fuck, I broke the build I bet. 21:53
Guest41009 anyone got mac os i can ftp
? 21:54
theorbtwo Guest41009: Do we know you?
Guest41009 lol
i dont think so
pjcj src/Pugs/Help.hs:52:13: Not in scope: data constructor `PUGS_VERSION'
known problem?
castaway apple have one
Guest41009 anyone have a mac os x i can ftp or p2p fomr them 21:55
castaway download.info.apple.com/Apple_Suppo...ion_7.5.3/
theorbtwo This isn't the place to be asking, Guest41009.
osfa obPerl6: $macosxwarez.(any('ftp', 'p2p')) 22:00
gah, dunno the right syntax
Guest41009 e
lol 22:01
theorbtwo given $macosxwarez {any($_.ftp, $_.p2p) }
Guest41009 anyone
got mac os i can ftp?
jabbot pugs - 2498 - Whoops, broke the build!
pugs - 2499 - More fixes for my stupidity -- this shou
Guest41009 anyone 22:02
got mac os x i can ftp
osfa noone has ops here?
pjcj Guest41009: do you know this channel is logged?
Guest41009 no
whats logged
pjcj now you do
Guest41009 whats logged
hattmoward coughs 22:03
pjcj who needs ops? ;-)
castaway laughs
theorbtwo Well done, pjcj.
osfa theorbtwo: wouldn't that call both methods and return the values?
You'd want it to call either method whichever was available?
theorbtwo Hm, g'point. 22:04
Limbic_Region this is what happens when you drive on the wrong side of the road news.bbc.co.uk/1/hi/england/hampshi...493713.stm 22:07
castaway pokes Limbic 22:08
Limbic_Region hopes castaway doesn't poke too hard as he just finished eating dinner 22:09
theorbtwo hopes he didn't break the build. 22:16
src/Pugs/Parser.hs:940:4: The last statement in a 'do' construct must be an expression
(I changed a line there, but only the formatting of a comment.
Anyway, goodnight, all. 22:17
pjcj does anyone fancy applying pjcj.sytes.net/tmp/pugs.patch? 22:44
or just making the change
And here's another, if any committers are around: 22:51
pjcj.sytes.net/tmp/pugs2.patch
those two get the build going again
Limbic_Region pjcj - do they fix the build on Win32?
pjcj no idea, but I doubt they harm it 22:52
Limbic_Region I was just doing a fresh checkout after theorbtwo b0rk stuff and went to sleep
well, I would be happy to check them in for you but I am afraid to with my local build being b0rk
Limbic_Region blames theorbtwo 22:53
pjcj no problem - they're both trivial anyway
ingy Limbic_Region: I forget, did you email me the patch? 22:55
Limbic_Region: maybe resend to [email@hidden.address] 22:56
Limbic_Region you are kidding right?
you told me to nopaste it
so after I did it went into the ether
ingy right sorry
revdiablo pjcj: I can check them in if nobody else does. 22:57
Limbic_Region just kidding
Limbic_Region will regen
ingy very funny
revdiablo tries building
ingy you scared me
Limbic_Region well - the patch itself did go into the either
pjcj revdiablo: thanks
Limbic_Region but I still have the modified file
ingy ok, I'm stepping out for a sec
I'll apply it as soon as i get back online 22:58
Limbic_Region emails it and then nopastes it
ingy my friend rking wrote WebService::NoPaste
revdiablo pjcj: ok, it built, I'll commit 22:59
ingy diff -u foo bar | nopaste url
returns pasted url 23:00
anyway I must eat now...
laters
revdiablo pjcj: should be good now
Limbic_Region perlbot nopaste
perlbot Paste your code here and #<channel> will be able to view it: sial.org/pbot/<channel>
Limbic_Region ingy - I forgot the -u option on the email 23:01
but I will do it correctly in the nopaste
jabbot pugs - 2500 - apply patches from pjcj to fix build
pasteling "Limbic_Region" at 69.250.48.119 pasted "Perl6::Bible patch for ingy" (53 lines, 1.6K) at sial.org/pbot/9779 23:02
pjcj revdiablo: thanks again
Limbic_Region kicks theorbtwo for breaking the build
Limbic_Region then heads off to watch jeopardy
revdiablo pjcj: btw, they hand out commit access very liberally around here. you'd just need to find an admin to send you the invite...
theorb Eeep. 23:03
I'm an admin.
Email
Er, Email?
pjcj if I get a bit of time to hack on things I'll take the comitter bit, for now it doesn't seem worthwhile 23:04
I'm hoping autrijus' latest Eval stuff will let me add coverage support 23:05
but I know autrijus had a few ideas about that a while back
theorb pjcj, thanks for fixing the build. 23:09
pjcj no problem - I wanted to see how much faster things were ;-)
Limbic_Region theorb - the email comment was meant for ingy not you 23:12
Limbic_Region wanders back off for the last hal
theorb No, L~R, I was asking for pjcj's email addr to +&=COMMITTER him. 23:13
Limbic_Region oh 23:28
pjcj no problem - I wanted to see how much faster things were ;-) 23:32
qpalzm
hmmm - that went to the wrong window ;-) 23:33
Juerd theorb++ # +&=COMMITTER 23:57
perl6++ # making irc much more efficient :)
theorb Don't ++ me too easy; I meant +|= 23:58
Juerd Oh, you weren't going to make him committer-only?
Shillo Hullo, all. :)

Powered by Cro and the RakuĀ® Programming Language.

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

Thank you!