6.2.2 is airborne! | pugscode.org <Overview Journal Logs> | pugs.kwiki.org
Set by autrijus on 1 May 2005.
Odin-LAP But those may have been a factor -- however, there's also indication that it was for natural reasons. 00:00
Activity increased slightly after an earthquake in 2000, I recall.
theorb calls it time to get off the computer.
ninereasons the author of that page lists as "destroyed" any geyser that's been tapped for geo-thermal power.
Odin-LAP WTF? 00:01
Is he using geyser synonymously with 'hot spring'?
ninereasons perhaps a partial explanation there, for his belief that Geysir was damaged by people
Odin-LAP wonders how he feels about the blue lagoon, then. 00:02
ninereasons anything that was tapped at Rykir, he calls "destroyed" 00:03
www.wyojones.com/destorye.htm
Odin-LAP My intarweb isn't working. :/
ninereasons that's the gist, anyway.
Juerd How the hell can one get a literal [ in Kwiki? 00:31
As in @foo[5]
[=@foo[5]] 00:34
ninereasons isn' it with =verbatim or something like that?
Juerd pugs.kwiki.org/?Perl6Nomenclature 00:35
Feel free to add!-
Oh, wow, for something using bracketed delimiters, [=...] is pretty STUPID 00:37
[=@foo[5]] gets rendered as <tt>@foo[5</tt>]
Ah well, working around this is too much work. 00:38
wolverian spelling fixes 00:43
Juerd wikis++
kwiki-- # no diffs 00:45
wolverian yes.
s,lamba,lambda,g
now, sleep calls. bye. :)
Juerd Bye 00:46
mugwump jabbot: who made you, and where is your source? 01:17
jabbot mugwump: We were discussing you -- not me.
mugwump slaps jabbot
mugwump finds autrijus' SVN::Notify::Config 01:21
bah, I can't put a hook script in my OpenFoundry repo 01:24
obra that would be a nice security hole
(and one that sf.net supported unitl recently)
mugwump heh
obra mugwump: jabbot was created by gugod)
mugwump ok, so ideally, I'd just like to go to a page on rt.openfoundry.org and set up an address to receive commit messages 01:25
but without either of these options, I'll need to write something to periodically check for new versions then run the appropriate diff commands 01:26
actually, will svk run post-commit hooks etc?
maybe that "something" could just be "svk sync"
Is there an irc channel / rt queue for OpenFoundry? 01:28
mugwump finds rt.openfoundry.org/Foundry/Project/...?Queue=125 01:29
pasteling "revdiablo" at 66.218.36.163 pasted "more extensive tests of weird flattening/subscripting behavior" (55 lines, 1.8K) at sial.org/pbot/9904 04:06
revdiablo any thoughts on these? 04:07
dunno if I'm duplicating anything that already exists, but I just wanted to get down all the behavior we were seeing earlier 04:08
maybe I should check this into t/pugsbugs/flattening.t? or somewhere else? 04:09
ok, I'll take the silence as a yes. :) 04:17
04:25 simcop2387-vnc is now known as simcop2387
jabbot pugs - 2667 - test some broken flattening behavior 04:25
machack666 fg 04:47
gaal_ forum2.org/gaal/pugs/smoke-cygwin.html # eval_yaml causes segfault 04:59
jabbot pugs - 2668 - a few tests illustrating a .kv bug 05:05
06:34 castaway_ is now known as castaway 06:40 chady_ is now known as chady
nothingmuch oh my 06:49
i just saw a banner: "write 10,000 lines of code in 10 minutes"
generating code is evil 06:50
why isn't the runtime smart enough to do this stuff?
castaway ugh 06:51
mornin nm
nothingmuch morning
worst of all, to actually get your app to do what you want you have to write something
reading 10,000 lines of code does not take 10 minutes 06:52
and without reading what you have, how do you know what to add?
*sigh*
castaway depends.. 06:55
stuff like GUI designers do a fair bit of behin the scenes coding, that isnt all bad
"behind"
(at least IMO) 06:59
nothingmuch i'm not going to trust a system i don't know to generate N thousands of lines for me 07:07
ever
=/
even if i know what it's supposed to generate
unless i'm forced into it, in which case i will learn what it generates (probably not going to save me time) 07:08
or write one myself
(not bloody likely, i'd rather do this stuff at runtime)
castaway hmm, but you trust compilers to translate to machine code, and lots of other things 07:11
castaway shrugs
10,000 seems several orders of magnitude too many, but still, i like code generation for quick stuff.. VB works pretty well for its field, for example 07:12
(If you dont trust things you dont know, how do you ever do anything? :) 07:14
mugwump the lambda calculus of true knowledge can never be reduced to a normal form 07:23
nothingmuch castaway: compilers take something that's mine and makes it into something else 07:42
code generation of an entirely new web app 07:43
with DB connectivity
and other goodness
at 10,000 slocs, in 10 minutes
is not very transparent
i have no real place to specify input... You just can't sum 10,000 slocs in 10 minutes of any form of conveyance (english, perl, whatever)
there has to be some large void defaultness that just doesn't make sense to be that big 07:44
jabbot pugs - 2669 - Remove second instance of self. 07:45
castaway yeah, I said that number was overdoing it mightly.. 07:46
nothingmuch so we are in agreement =) 07:47
castaway GUI designers do the same tho, take what you pick as design elements, and turn it into code
nothingmuch but that's known input
castaway right 07:48
it seems so, yes :)
nothingmuch i say "make me some code that impelements the interface to respond to that widget"
and you get a bajillion tubs
castaway what are these 10k-lines-of-code things proposing to do?
nothingmuch s/tubs/stubs/
castaway lines of code are pointless unless they have some directed purpose :) 07:49
nothingmuch a "web application which features: db connectivity, err, something else, high perfomance, no bugs"
oh, 'stunning graphics'
castaway ah :)
misc_ nothingmuch: "b2b capability 07:50
castaway sounds like one to toss right in the bin
arcady why not just have a metaclass that does the code generation for you?
castaway or, just very bad advertising
arcady buzzword bingo!
castaway right
nothingmuch arcady: for this purpose see also (Catalyst|Maypole|RubyOnRails)
castaway If it implemented that stuff as widgets, with a nice interface to allow use, it would make sense, but generating the code seems crazy
(I should get around to using zApp someday) 07:51
nothingmuch misc_: i don't understand
arcady if you have to generate code, it's generally because you need a better language
nothingmuch agrees
misc_ nothingmuch: just a buzzword to add :)
castaway right 07:52
nothingmuch i even have the same view about snippets
ah =)
misc_ nothingmuch: for "something else"
nothingmuch xrl.us/fysk <-- content type screwey. it's a gif 07:53
that's the banner 07:54
this is where the banner links to: www.ironspeed.com/products/DownloadNow.aspx
07:55 rob__ is now known as rob_
castaway $1995 !? 07:56
arcady hey, it's a good deal! 07:57
castaway sale, enterprise edition for just $3145 !
no thanks :)
arcady look how many lines of code you can write!
perl and ruby programmers suck, they write so many fewer lines of code
castaway of gobbledegook? lots :)
arcady that's so much less productive : )
castaway *g* 07:59
lightstep omg, i can't wait for the video to load 08:01
castaway video? 08:02
mugwump lists.utsl.gen.nz/mailman/listinfo....gs-commits # for those interested, svn commit messages from pugs repo 08:03
lightstep yes, where they show how to build an app in 5 minutes (actually, the video is 6:19)
castaway ah
lightstep ironspeed.com/products/ProductVideo.aspx
examples/shuffle.p6 is broken 08:11
castaway then fix it, dear lisa, dear lisa.. :)
lightstep i don't understand the problem
that's sumething with the `sub' syntax 08:12
oh, it's not that 08:13
08:22 metaperl_ is now known as metaperl
jabbot pugs - 2670 - fix examples/shuffle.p6 08:25
bsb nothingmuch around? 08:55
jabbot pugs - 2671 - * Implement (rvalue only) `%hash.pick`,
09:20 integral1 is now known as integral
bsb Does this cause anyone else trouble: my $s = 'a'; sub eg { $s .= 'b' } 09:51
luqui tries
geez 09:52
bsb Internal error?
luqui every time I want to try something I have to recompile everything
because there's so much progress on pugs
bsb Ok, let me know in 20 minutes
luqui haha
bsb I get: Internal error while running expression: unexpected "s" expecting ";" or end of input 09:53
luqui I always find it disconcerting when I install build a new piece of software and see "no rule to make target test"
I think the best thing Perl has done to software development is to make test suites culturally mandatory
bsb What's missing test? 09:54
luqui rosegarden 09:55
castaway umm, isnt it somethihng other than .= these days? 10:03
luqui ahh that would be the problem
~=
castaway ~= .. right
arcady .= would be an odd operator indeed
luqui it exists 10:04
nothingmuch bsb: pong
castaway (unless you're writing some odd looking method stuff
bsb ah!
no, just p5 thinking
nothingmuch try a \n between the semi colon and the sub decl
it looks to me like parser smell 10:05
luqui I think he wants to call the 'b' method (quotes and all) on the string 'a' and assign it back into $s
nothingmuch i think the error happens before that
luqui there really is an error? 10:06
nothingmuch bsb: I get: Internal error while running expression: unexpected "s" expecting ";" or end of input
bsb nothingmuch: in t/subroutines/splat_rw.t there's a test for *@list is rw , what's the source?
nothingmuch bsb: that's l~r's test
luqui no error here
nothingmuch he was making sure that splatting and is rw don't make sense when used together
luqui oh 10:07
bsb S06/Perl5ish has sub swap (*@_ is rw)
is it a special case?
nothingmuch oh well =P
it's either a bug in the test or the design ;- 10:08
)
luqui design
nothingmuch so how do you do assign to an arbitrary number of args
i think it's bad sstyle 10:09
but it might be useful
luqui there's something going on there like sub swap (*@_ is context(rw)), for a 'context' ne 'context'
bsb I'm not getting the 'context' 10:13
castaway wonders if someone needs to go thru Pugs and try out "Perl5 syntax" stuff, and make it give sensible errors
luqui that would be very wise for "real" perl 6
so it might be a good exercise 10:14
nothingmuch in "real" perl 6 wont you have to say 'use v6' to make it p6?
bsb Aren't we trying to say @_ contents are rw?
luqui yeah, but people are still going to make p5 brainfarts
all the time
bsb I can vouch for that
luqui no, *@_ contents shouldn't be rw 10:15
well, unless it's declared without a signatue at all
then they are
and the implicit declaration is the one that's tripping us up
jabbot pugs - 2672 - $1 now $! in eval test
castaway nothing, yes.. but my point is that you might forget halfway thru programming, and you'll get those lovely "internal errors" instead of "you idiot, use '~=' instead of '.=" :) 10:16
luqui right 10:17
call them the YIEs
castaway YIE?
broquaint Or PIEs.
luqui or LIEs
"Legacy Idiocy Errors"
castaway ,)
luqui redhat needs to be overthrown, and gentoo needs to get easy to install so that it can overthrow it 10:21
castaway gentoo is hard to install?
luqui last time I checked
castaway just followed the instructions
luqui pretend you'd never seen a shell command before 10:22
castaway ah, well
It could do with a frontend, yes
luqui plus, the net drivers don't always work (which is why I'm on fedora)
castaway and something better than those fugly python (?) scripts ,) 10:23
luqui haha
yeah, dood, get with the times, OSX uses perl for everything!
for various definitions for everything
castaway really all that annoys me there is that ctrl-cing out of most of them gets a whole page of python errors
its a little shocking at first
luqui I think larry wants that for perl to some extent 10:24
castaway the errors?
luqui don't know what to think
yeah
castaway eh?
luqui eh?
castaway really, its much nicer to have to apply a commandline arg to get more error text, than the other way around
luqui the question is, what if you have trouble reproducing your error 10:25
and we'd rather not leave dump files lying around 10:26
castaway shrugs 10:27
I'd like for the user to have a clear way to do things..
preferably with no errors when genuinely just trying to close an app :)
luqui maybe ^Z; kill -SEGV % 10:28
then you can have your pretty ^C
castaway icky :)
luqui maybe perl can intercept SIGQUIT to do an icky exit
so ^C: clean exit; ^\ icky exit
castaway for example 10:29
I dont really care how.. I just hate those error screens.. :)
nothingmuch C+ +, oh how i loath thee 10:38
well, MFC really
*sigh*
bsb "control exceptions" in a block are redo, last, next and the like, right? 10:47
broquaint Yep. 10:48
bsb Thanks 10:50
jabbot pugs - 2673 - test pointy subs 10:55
nothingmuch hola theorbtw1 11:10
~~
gaal|work luqui, there-p? 11:14
jabbot pugs - 2674 - * `$thread.kill`, `$thread.detach`, `$th 11:15
pugs - 2675 - * minor tweaks on cashiers.p6 syntax
pugs - 2676 - :todo<feature> added
castaway ooh, threads? 11:16
theorbtwo Allo, nothingmuch. 11:17
nothingmuch castaway: yeah, we've had them for a while =)
castaway hmm, irssi shouldnt care on /msg, what server someone is using 11:18
nothingmuch don't messages get sent to your server, and from there to someone's server, and from there to someone? 11:21
luqui gaal|work: #t
castaway right, but Im on slashnet and freenode at the same time
11:22 chady is now known as chady_
nothingmuch oh 11:22
so it doesn't know what network, not what server? 11:23
but the user name is unambiguous? yeah, that is silly... ={
tr/{/(/;
castaway umm, yeah
castaway isnt into IRC jargon
nothingmuch it's not jargon, it's structure ;-) 11:24
in p2p jargon: an IRC server is an Ć¼ber-peer in a p2p network
only that it's Ć¼berpeer friends are known in advance
and, err, that's it =) 11:25
theorbtwo We've had threads for a while now, castaway. 11:26
castaway I guess I missed them.. are they specced?
nothingmuch i think not 11:32
grep -r 'async' t examples
lightstep revision 2676 doesn't compile
some ugly code makes a mess 11:33
is autrijus on it, or should i look?
nothingmuch hasn't seen him in a while 11:34
gaal|work luqui - just thought it's worth mentioning how java dumps a stack trace for the thread list when you enter ctrl-break on the console. 11:38
nothingmuch death should not include a stack trace though 11:39
gaal|work this does not require death, nothingmuch 11:40
nothingmuch it really hinders user friendlyness when trying to first run a program, the error message for "component foo is missing" is part of a page long dump
luqui we'll probably do something like that by default
theorbtwo IRC isn't very p2p, even if you ignore the clients.
luqui overridable, of course
nothingmuch gaal|work: yeah, i reckoned =)
theorbtwo It requires centralized planning; otherwise you get loops in the network structure.
nothingmuch theorbtwo: well, that's true... but the message passing itself is a bit like p2p 11:41
but in that sense SMTP is also p2p
gaal|work then again the developer is always glad if errors include a stack trace :)
^error reports
luqui loves firefox
nothingmuch gaal|work: that should be easy to add
lightstep funny thing (hebrew): www.tapuz.co.il/tapuzforum/main/Vie...d=52154687 11:42
theorbtwo SMTP is very much p2p.
luqui now all we need is a way to embed pugs in it
lightstep the nazis used fluorine, how can the state of israel put it in water?
luqui <script language="perl">...</script>
gaal|work luqui, activestate perl did that 11:43
luqui really?
cool
probably didn't catch on...
castaway sure, perlscript++
gaal|work called activescript, i think
castaway well it only works on win
gaal|work perlscript, thatis' it
luqui: on another note, see XUL::Node on CPAN
castaway AND with activestate installed.. Id assume thats not really a popn worth targettin
luqui woah 11:44
WOAH
castaway hmm?
luqui I've been wanting this forEVER
castaway kept meaning to try it and never did 11:45
it works in anything that vbscript and jscript etc works in, on win.. irrc (WSH)
luqui I expect this is how gmail does its thang
castaway yet you never looked?
nothingmuch the XUL/rails demo was very impressive
luqui XUL isn't exactly a descriptive name
castaway well thats also browser specific 11:49
luqui mornin' folks, good night 11:56
jabbot pugs - 2677 - fix type error 12:05
pugs - 2678 - Cosmetic fix to t/builtins/arrays_and_ha 12:35
pugs - 2679 - Added svn:ignores to src/. 12:46
pugs - 2680 - Added tests for state(). 12:55
pugs - 2681 - Added usual svn properties to t/.
pugs - 2682 - Cosmetic fix to t/var/state.t.
Limbic_Region greetings and salutations all 13:00
jabbot pugs - 2683 - * svn:ignored "tests.yml" in /. 13:05
Limbic_Region www.perlmonks.org/index.pl?node_id=453821 # Adventures in multilingualism (p5, python, p6) 13:07
out of curiosity - do coroutines work in pugs yet ( yield ) ? 13:11
nothingmuch the problem with pugs is that if you have taken an 8 hour mental break 13:14
the quickest way to check for these things is grep -r t/ examples/
and then figure out if it's todo or not
like, say, you went to sleep
then you're no longer up to date
puetzk nothingmuch: svn log -rBASE:HEAD is a really nice feature :-)
Limbic_Region nothinmuch - I just about fell out when hardburn told me the reason why he wasn't using Pugs was because it didn't have a feature he wanted 13:15
I asked him when was the last time he checked
he replied - "a couple of weeks ago"
nothingmuch hah
what was the said feature?
jabbot pugs - 2684 - Added tests for temp().
Limbic_Region I don't remember exactly but it was something to do with OO 13:16
something that could be done now even without full OO implementation though
IIRC
nothingmuch ooh well 13:17
Limbic_Region I keep poking and prodding him
he is another very smart young man
nothingmuch later he might blame you for addicting him
Limbic_Region hmmm - he probably would take offense to that 13:18
nothingmuch - that's probably true
so since I am using a binary distribution that doesn't include the t/ directory - you mind grepping for "yield" for me? 13:20
nothing in the examples directory that I see 13:22
autrijus Limbic_Region: where is said that perl6 will have yield()?
we have yield() for threads 13:23
(yielding control)
Limbic_Region autrijus - I am not sure I saw "yield" specifically, but coroutines for sure
autrijus where exactly?
Limbic_Region will look for references
autrijus I see nothing about coroutines in synopsis
except in a passing reference in lazy gather/take
which is not quite the same. 13:24
puetzk has seen some coroutine discussion for parrot, though that doesn't mean perl6 will expose it
autrijus sure, parrot has coro 13:25
Limbic_Region puetzk - that's probably my problem
autrijus but I need concrete syntax and semantics to implement this sort fo thing :)
Limbic_Region spent so long not paying attention to p6.l and only paying to p6.i that I think my boundaries are getting blurred
jabbot pugs - 2685 - Added OO aspects of temp() to t/var/temp
autrijus I think we can trivially do coroutines with lightweight share-everything threads. 13:26
in the interpreter core, that is
but not sure at all how to expose it at the language layer. 13:27
Limbic_Region is still looking
lightstep or as a perl6 library with reified continuations? 13:28
like the examples in scheme 13:29
autrijus lightstep: sure, that works too
Limbic_Region autrijus - found it 13:30
but it hasn't been written yet
www.perlmonks.org/index.pl?node_id=332117 # Perl6 Timeline By Apocalypse 13:31
chapter 17
autrijus well, write it!
docs/s17.pod
Limbic_Region that means I would have to read all the RFCs and see what Larry has said about them right? 13:32
17 talks about threads, asyncronous I/O - theres a lot more there than just coroutines
lightstep where is the perl6 timeline picture that was inspired by "wearing the hair shirt"? 13:33
autrijus lightstep: haskell.org/hawiki/AutrijusTang has one
Limbic_Region autrijus - how about I post to p6.l with a specific coroutine question 13:34
I don't think I am qualified to write S17
autrijus Limbic_Region: sure.. 13:35
13:37 sbkhh is now known as Odin-
puetzk somehow that made me think of www.frozenreality.co.uk/comic/bunny....php?id=91 13:43
but hopefully not :-)
perhaps given the pace of recent development, something more along the lines of www.frozenreality.co.uk/comic/bunny....php?id=44 13:44
Odin- Bunny is crazy.
puetzk indeed
Limbic_Region autrijus - message sent 13:48
clkao Writing Makefile for PAR 13:52
The 'Autrijus' method does not exist in the 'inc' path!
Please remove the 'inc' directory and run Makefile.PL again to load it.
Autrijus method!
lightstep Makefile.PL worked here a few revisions ago, and hasn't changed since then 13:53
Limbic_Region clkao - the Autrijus method is well defined "trade test for feature"
gaal|work Juerd++; # two-dimensional zip example in p6-l - lol
autrijus faints 14:01
www.nntp.perl.org/group/perl.perl6....uage/20827
Limbic_Region autrijus - does that make your life easier or harder? 14:06
lightstep compiling pugs really is no fun 14:08
theorbtwo I think Larry is going insane. 14:12
Er, insaner.
autrijus Limbic_Region: I think I'll have a LarryWhim.hs 14:21
that defines the "knobs"
theorbtwo Just make it change behavior on rand > .5?
autrijus MkLarryWhim { unlimitedDereference = False, ... } 14:22
nnunley_ 15:01 < autrijus> www.nntp.perl.org/group/perl.perl6....uage/20827
Doh.
autrijus lightstep: "make unoptimised"
lightstep autrijus, but then it's no fun
autrijus eh? :) 14:23
fwiw, I mostly just use "make ghci"
lightstep if the wow, cool
s/ if the//
Limbic_Region autrijus - so far I haven't gotten any answer expect WRT coroutines 14:24
autrijus there's two answers
but not from @larry
Limbic_Region thought luke was in @larry ?
autrijus oh, he is
had not see his mail yet 14:25
Limbic_Region he was the one with the gather/take answer
which isn't at all what I was asking
autrijus actually gather/take isn't a _bad_ answer
because lazy list is a form of coroutine
but problem is the interleaving semantics may or may not be the same as a traditional coroutine is 14:26
esp. when there are other lazy lists involved
in the computation
haskell.org/hawiki/PythonGenerator maybe related. 14:27
Limbic_Region and now Luke and I are accidently having a discussion off-list 14:30
scook0 I just realised why my 'make optimised' was taking so long
Somebody had put "unoptimised :: optimized" in the Makefile 14:31
Fixed now
autrijus danke
scook0 s/make optimised/make unoptimised/
Limbic_Region heh - that's one way to skew the numbers *grin* 14:38
off-list to on-list discussion has been restored - I see Luke's point and don't disagree with it - I am asking questions not making suggestions
autrijus Limbic_Region: fwiw, I see Luke's point too. 14:39
it's just he is simplifying stuff a bit.
because when there are multiple Lazy lists involved
in each other's production
and at least one of it involved external IO
then a python generator's behaviour does not neccessarily agree with a lazy list based generator. 14:40
but maybe we can explicitly define the lazy list generator's semantics as a coroutine. 14:41
that may be the way to go.
Limbic_Region did my last reply come through?
autrijus i.e. a gather{} always saves control stack
Limbic_Region I hope so
basically I said I am not indicating behavior I would like to see only asking if a decision has been made so I can write tests 14:42
and then I say, ok - I lied - I do have an opinion (not on behavior but syntax) - it should at least resemble coroutines in other languages
autrijus I believe iblech offered a reformulatin. 14:43
Limbic_Region is catching up 14:45
I spent a fair amount of time with Luke getting our entire conversation back on-list
so gather/take is the only ruled on stuff at this point 14:48
*shrug*
gaal|work nothingmuch, ping 15:02
nothingmuch pong
gaal|work hey. remember sial.org/pbot/9831 ? that bare 'svn' in the doc stanza, it's not supposed to be there is it? 15:03
nothingmuch svn up 15:04
gaal|work but you're already after an svn up.
Juerd Hey, freenode's back 15:29
Neat.
gaal|work hey Juerd.
Limbic_Region didn't know freenode went missing 15:30
wolveria1 irssi-- # what do you mean can't call my ->new()? 15:42
jhorwitz autrijus: ping
autrijus jhorwitz: pong
perlbot: nopaste 15:43
perlbot Paste your code here and #<channel> will be able to view it: sial.org/pbot/<channel>
jhorwitz if you can eval an IMC buffer, will you still need to be able to read .imc files?
autrijus no 15:44
evalling a buffer is enough
jhorwitz that's the answer i was looking for. ;-)
pasteling "autrijus" at 220.132.132.105 pasted "for jhorwitz" (24 lines, 1.1K) at sial.org/pbot/9919
autrijus ok. I need to run for dinner
that nopaste may or may not be useful to you -- I think you may already have it figured out 15:45
autrijus eagerly waits for eval_parrot()
be back in a bit... &
16:03 nnunley_ is now known as nnunley
Limbic_Region ok - hash versus junctions question - if you want to only add items to an array if they are not already in the array, would you prefer push @foo, $value if $value eq none(@foo) or the standard if ! exists $seen{$value} ? 16:40
Juerd Limbic_Region: Hash, for great justice, ehm, efficiency. 16:49
Limbic_Region: I'd probably create a type that does both Array and Hash, where the hash uses reverse logic: give the value, get back the index
Limbic_Region Juerd - I thought about it right after I wrote it and smacked forehead 16:51
Juerd Limbic_Region: This can be a repeated linear search as with $value eq none(@foo), but when needed optimized to use a hash internally
Limbic_Region if it was a one shot deal it is up in the air for me
Juerd Limbic_Region: Abstraction's good.
Limbic_Region if you are going to keep doing it is silly to use a junction
lightstep /quit
Limbic_Region Juerd - will share with you why I was asking
Juerd Limbic_Region: Parse error
18:54 < Limbic_Region> if you are going to keep doing it is silly to use a junction 16:52
I don't know what that means
Limbic_Region perlmonks.org/index.pl?node_id=453774
meaning if the push @array, $value was only going to be executed one time then performance really isn't a concern
if you are going to test each time you add an element and you expect that to be a number of times it is silly to use the junction 16:53
Limbic_Region realized that right after he wrote it
Juerd I find your perl style funny, with the whitespace within ().
Limbic_Region Juerd - I started doing that with single argument/parameters but not with multiple - haven't really decided if I like it yet 16:54
Juerd If it helps you decide, know that I don't :)
I use whitespace in brackets that construct 16:55
Not in brackets that group
Or in subscripting brackets
(1, 2)
[ 1, 2 ]
foo(1, 2)
Limbic_Region I know you are very particular about your style
Juerd @foo[1, 2]
Limbic_Region I am still quite maleable 16:56
given the way I used to code in BASIC - "I've come a long way baby"
Juerd Now comes the funny example: foo(( 1, 2 ), 3)
But that should have been written as foo([ 1, 2 ], 3) anyway.
Limbic_Region: I come from BASIC too and have learned to not use that as an excuse :) 16:57
Limbic_Region I'm not using it as an excuse - just as a reference point for relativity
Juerd Same thing
:) 16:58
Limbic_Region not at all
if you are going to evaluate where you believe someone is on a path to where you feel they should be - to judge their progress you need to know where they started 16:59
Juerd In where people start, I choose to ignore the BASIC part entirely 17:00
As that's too different from Perl.
Limbic_Region brb - helping diotalevi with a Pugs/GHC issue at the Monastery
Juerd - I am still not sure you are seeing my point though I am not sure it is worth trying to make 17:03
nothingmuch_ ooh! diotalevi is coming on board?! HURAAH!
Limbic_Region trying very hard nothingmuch
nothingmuch_ (you can lure him by saying he will already have positive karma, before even showing up! 17:04
Limbic_Region++
diotalevi++
Limbic_Region right now his problem is knowing if he built GHC right (wants to know if there is a test suite) since he can't find or doesn't believe there is a binary build for Sparc/NetBSD
nothingmuch_ pugs is a good enough test suite, methinks =) 17:05
Limbic_Region nothinmuch - would you mind joining the CB and helping him 17:06
I need to run to a meeting
nothingmuch_ sure
should I pretend to be naive?
ah, nevermind
17:07 nothingmuch_ is now known as nothingmuch
Limbic_Region Juerd - I think anyone who uses what they used to do as an excuse for not moving forward is just that - an excuse - someone that is willing/trying to move forward but explaining past influences to help others appreciate the relative progress isn't making excuses 17:07
and on that note - I am off
nothingmuch stevan: bonk 17:09
(which is more violent than ping)
kill CONT => pidof(stevan); 17:13
17:18 wolveria1 is now known as wolverian
Juerd Today's a very busy day on p6l 17:37
Over 60 messages already
afk 17:55
Corion r2697 - datenzoo.de/pugs/win2k.html - 4741 ok, 51 failed 18:16
ninereasons What are the svn repository targets for perl6_doc and Perl6-Bible ? 18:19
(alluded to in sial.org/pbot/9831)
nothingmuch see the top of util/catalog_tests.pl 18:20
and delete the line with just 'svn'
it's bullshit
ninereasons thank you nothingmuch
nothingmuch =) 18:21
please note that my machine is SMP
ninereasons heh, I mean, "thank you much, nothingmuch"
nothingmuch you might not want to do parallel testing, etc
nothingmuch thinks of smoking his weekly cigarette 18:22
ninereasons is the perl6_doc the "official" set of AE&S , nothingmuch ? 18:26
nothingmuch yep 18:27
from svn.perl.org
Corion Heh. Seems like I got $*PID working on Win32 now ... 18:43
... so I can now hunt down some other stuff that needs Win32 API interaction. I'm not sure if my code is good, because I couldn't get it to compile with an "unsafe" sprinkled in. 18:44
Whoops - the build broke, because I left a syntax error in ... :( 18:50
lightstep here is breaks at Pugs/Compat.hs 18:53
Corion lightstep: Yep - just committing the fixed version, sorry
r2700
lightstep funny, it's not clear what the error is 18:55
Corion lightstep: It is the "unsafe". Once that gets removed, it compiles. 18:56
... not that I'm not sure if the "unsafe" would be significant. But as the pid doesn't change during the course of the program, I don't care much.
lightstep unsafe is the default 18:57
but i get a warning on line 47, and then make exits with "Error 1" 18:58
Corion lightstep: Weird - with r2700 ? 18:59
lightstep yes
Corion Aiieeee - and now I get a segfault from Pugs.
lightstep: Then maybe my #ifdef magic is broken ...
you're not using Win32, are you?
lightstep it probably is
oh, wait, not 19:00
but pugs_config.h is strange
Corion lightstep: I haven't touched that ... I hope ...
lightstep i took a look at it, and it still says PUGS_VERSION "6.2.0" PUGS_DATE "April 13, 2005" 19:01
Corion lightstep: That's weird. Kill it and start fresh? 19:03
lightstep tried it 19:04
apparently, you should remove the import of System.Posix.Internals
that fixed it here
or, alternatively, move it to the non-posix part (if it needs it) 19:05
Corion lightstep: Ah - that's a leftover of my wild goose chase... Sorry! 19:06
(fixed in r2701) 19:12
lightstep yeah, i saw
Corion lightstep: I mentioned it for the convenience of the backloggers :) 19:13
lightstep it's really bothering there're no regular smoke tests 19:14
Corion lightstep: Well, it would bother me more to have my Win32 box run all the time. But yes, an automated message onto irc with the smoke results would be nice 19:15
jhorwitz autrijus: ping 19:21
ninereasons wiping out and then replacing the src/ directory sometimes makes weird problems go away. 19:38
lightstep probably when Makefile changes too much that make clean isn't effectivy any more? 19:39
ninereasons or perhaps a few Haskell artifacts from previous builds do not really go away with make clean; 19:40
Corion I think some .o aren't getting cleaned out properly, but I always wipe src/ and check out fresh whenever that happens instead of investigating 19:41
mj41 r2700, WinXP, -todo 933/4790 subtests failed, 80.52% okay. 19:47
80.52 percent Perl6. Nice work. 19:50
Any volunteers for t/statements/for.t ( 30 3 10.00% 28-30 ) ?
'for @list.values -> $num is rw { $num++ }' and such 19:52
lightstep the `for' implementation is a little long 19:53
lines 315-339
nothingmuch autrijus++ 19:55
i just realized pugs has made me a more effective programmer, due to exposure to haskell
even though I don't know haskell yet
ninereasons competence by osmosis ?
nothingmuch yes, you could say that =) 19:56
althought his is not so much competence
Juerd nothingmuch: Can you give a real world example of what you did now and what you would otherwise have done?
nothingmuch i could do this stuff before
it's just nowI'm doing it nicer
Juerd: plentiful recursion to clean structure up is coming out much more naturally
i didn't think of using the more functional way that often before
being an OOP head
Juerd I see 19:57
OO is harmful if it's the only tool you have
nothingmuch but a bit of hacking in haskell forced me to get used to it
it's not the only tool I had, it's just the one i used most naturally, but I agree
i don't feel harmed by my preference
Juerd You said it was less effective
nothingmuch my previous preference 19:58
you could s/effective/elegant/
maybe not maintainable
but actually I agree, and I shouldn't argue
lightstep what is cxtSlurpyAny ?
Limbic_Region nothingmuch - how did things with diotalevi go?
nothingmuch although I don't agree with OOP being the only tool I had ;-)
Limbic_Region: he is bold, and I have no time
he's bootstrapping by the docs
and I have to get real work done 19:59
Limbic_Region nothingmuch - his claim is that there is no binary build of GHC for Sparc/FreeBSD 20:01
nothingmuch i have no idea if that's true or not
but he's bootstrapping
Limbic_Region yeah - I told him I didn't know how to help him either because of the bootstrapping problem 20:03
he said what problem (being bold) 20:04
that's about the time I had to go to a meeting
Corion Pugs/Compat.hs is mostly evolving into Pugs/Compat/Win32.hs - should this be split up one day, or should this be kept in Compat.hs for the time being? 20:44
Juerd I HAVE A NEW DOMAIN NAME 20:49
AND I LIKE IT VERY MUCH
I HAVE NO IDEA WHAT TO DO WITH IT THOUGH
It is perl6.nl
Corion Juerd: You can alias it to pugscode.org for the time being :-))) Or set up a nopaste handler for #perl6 :) 20:50
Juerd The problem with aliases is that they create expectations
I want it to be its own site 20:51
That can be a single page with a link
You see, I want to keep the possibility of using it for commercial purposes when perl6 is released
webmind like what ? 20:52
Juerd Like a listing of dutch perl professionals who can help learn and/or migrate to perl6
webmind give it some what generic perl6 related content that'll fit with a purpose you might have later for it ? 20:53
Juerd Duh :)
The content is the hard part.
I could make it a wiki 20:54
But that'd require finding one that doesn't suck first
Corion Juerd: CGI::Wiki is a decent wiki toolkit IMO. There is also mojomojo, which has fancy features. 20:56
(like "live preview")
Arathorn likes moinmoin, if you can swallow the .py medicine.. 20:57
integral write a perl6 wiki...
Juerd integral: Hm, nice idea
A pugs based web server with a tiny wiki
I quite like the sound of that 20:58
Do we have HTTP::Daemon yet? :)
Arathorn wasn't one of the first perl wikis only about a 40 line CGI?
Juerd I don't want it to be CGI
Only self-hosting is enough showing off
obra Juerd: irc there is an http::server::simple in there somewhere 20:59
Juerd Let's see if I can find it 21:00
There's only HTTP::Headers so far
obra maybe I'll port it tonoght
tonight
Juerd And a server example, non-modular
obra http::server::simple is easy ;)
Juerd Good enough for what I want though :) 21:01
obra s/easy/trivial
Juerd obra: It'd be great if you ported this
Hm, that HTTP server is VERY simplistic :)
It ignores the request and has hard coded output :) 21:02
obra Juerd: noted.
obra has customer stuff. but after that will try
Corion discovers haskell.org/hawiki/FfiCookbook 21:04
Hmmm. we don't have t/builtins/system.t ... :) 21:21
I'll hack my test into t/magicals/env.t for the rw support of %*ENV 21:22
theorbtwo Corion, what can system() portably to test it?
Corion theorbtwo: system('pugs','-e','say %*ENV<FOO>'), for example 21:23
theorbtwo OK, works for me.
Don't your pugsrun tests cover that?
Corion theorbtwo: No, they don't. (and it's from the wrong side, there) 21:24
theorbtwo: But in a sense, yes they cover that too, by using it.
But I prefer explicit tests over implicit assumptions of working features
theorbtwo /metoo
cwest url 21:51
meppl guten abend 22:00
meta_perl guten abend 22:09
22:09 meta_perl is now known as _metaperl
meppl good evening meta_perl 22:13
_metaperl good evening 22:21
Corion mpeppl! 22:27
Hmmm. does undef %ENV{PATH} # enter an endless loop for anybody or did I break my build? 22:28
Err - delete %*ENV{PATH} # typoe
Gah. delete %*ENV<PATH> # is what I really meant 22:29
ninereasons r2696 hangs on that, Corion 22:30
Corion ninereasons: Thanks. So it's a bug and not caused by me. That's somewhat good.
ninereasons in fact, all the way back since r1876, pugs has exhibited this behavior 22:32
an old bug, Corion 22:33
Corion ninereasons: I just stumbled on it now, as I'm trying to make %*ENV work on Win32 :)
ninereasons i'm looking at Linux and FreeBSD here, fwiw 22:34
Jonathan_ Hmmm....Pugs segfaults (on WinXP) while running t/unspeced/eval_yaml 22:59
Don't have time to look into it right now. 23:00
Corion Jonathan_: Segfaults here as well (Win32/Win2k) - I think it's a bug in the C code. 23:07
Jonathan_ Yeah, makes sense. Should I post it on the list, or will someone notice it here? 23:10
Corion Jonathan_: Autrijus also reads the logs here, so he (or whoever put the YAML C code in) will notice here 23:12
Jonathan_ OK, cool. :) 23:15