Summer of Perl | 6.2.12 released! | paste: sial.org/pbot/perl6 | pugs.blogs.com | pugscode.org | pugs.kwiki.org | www.treehugger.com/files/th_images/paradigm.jpg
Set by nothingmuch on 12 July 2006.
svnbot6 r11520 | kolibrie++ | test for overriding <ws> in a grammar 00:01
00:06 trym joined
kolibrie audreyt: I guessed at the correct output for t/rules/grammar.t (r11520), because it does not work here 00:10
clkao 456 tests pass today 00:11
kolibrie clkao: awesome!
clkao now i really want a diff tool for ttm. nothingmuch: you said you have one? 00:12
kolibrie audreyt: can you test it on your laptop and elsewhere?
nothingmuch clkao: yes
look in the TTH docs
all you need is to initialize it with several models
and it'll create Test::TAP::Model::Consolidated for you
but note that there are a few bugs
i can iron them out if you get errors in the template
clkao are you gogin to fix them?
nothingmuch because I know what they are
but i just haven't cared so far since nobody used that feature
nothingmuch bounces 00:15
exit row seat exit row seat exit row seat!
svnbot6 r11521 | fglock++ | v6 and PCR - various fixes in the precedence table 00:17
kolibrie pmichaud: okay, your nopaste yapc_schedule.pl worked, though is not so elegant 00:20
pmichaud: however, I had to change all the backslashed angles to the <'literal'> form, otherwise I got a PGE error 00:21
pmichaud: *** Cannot parse PGE: <schedule> 00:22
clkao summon fglock 00:25
nothingmuch FGLOCK FGLOCK FGLOCK RISE FROM THE PITS OF DOOM AND WALK AMONGST US! 00:26
pmichaud kolibrie: if you're getting "Cannot parse PGE", do you have PARROT_PATH set? 00:27
I don't think it's a PGE bug
nnunley gets bleading MAD.
pmichaud it worked fine for me with the backslashed angles
nothingmuch gives nnunley some valium
nnunley The source was not strong with me. 00:28
pmichaud also, just because the message says "PGE" in it doesn't mean that the error is in PGE :-) 00:29
clkao fglock, common, all we need is [1..4] to get 02-test-pm all pass 00:30
kolibrie pmichaud: the only change I made was converting all the \<small\> to <'<small>'>, etc. 00:36
pmichaud: with the backslashes it gets the error
TimToady j 00:37
kolibrie pmichaud: it could be something outside of PGE, maybe. I don't know parrot or the pugs bridge really well
clkao fglock, you broke && 00:38
gaal nothingmuch: ping 00:39
nothingmuch gaal: you should be spleeping!
anywho
gaal true dat. 00:40
nothingmuch i planned on that actually
gaal but I missed my sleepy time
nothingmuch so i'm kinda busy
oi =(
well, at least no worky tomorrow
use orz
clkao do more hiking
gaal use HebrewWeek; use Friday;
nothingmuch short story: i'm not sure the slot calculation stuff is 100% correct in the MOP.hs file 00:41
so i wanna go over and be sure
since I waws very very picky over this to stevan
till he got it into moose
gaal clkao: made a small urban trip today with a friend who just bought a bike
nothingmuch because I designed it to be really for perl 6
gaal nothingmuch: excellent, please drop everything else and let's go over it now. :)
nothingmuch ug ug ug
i am breaking moose right now 00:42
slaughtering it, in fact
gaal don't do that
nothingmuch on a branch
it's a branch, it's OK
gaal yes, moose have no business on trees anyway
nothingmuch but i'm on ritalin withdrawl and I can't be distracted
gaal if thy venture there, I guess it's fair game.
nothingmuch or i will definitely not get it done ;-)
damn you audreyt !
gaal blah blah
withdrawl alek 00:43
nothingmuch ;-)
gaal two days and you're a junkie
nothingmuch actually i was kidding, i'm actually on a fairly good day
00:43 weinig is now known as weinig|foood
nothingmuch despite various distractions i managed to get a bit of stuff done 00:43
but either way I'm out of the ritalin, at least till i go see a doctor or meet ingy/audrey ;-) 00:44
gaal so, is stevan about? I can badger him
nothingmuch no, he's watching star trek with the kids
also, i'm the evil guy behind that part 00:45
i am sorry
i just really promised to finish this
gaal that's okay. have fun 00:46
when's your flight?
nothingmuch tomorrow morning
i leave at around 10ish
and fly out at 16:30
gaal oh.
nothingmuch (arriving an hour too early at the airport, but at least it's a direct train)
gaal proposes a short TLV Hackathon
nothingmuch i land on saturday morning
next week?
or maybe i can even come in the middle of the week 00:47
gaal I assume it'll take you a week to recover from jet lag
nothingmuch now that i work on my own terms
gaal I don't, not quite :/
nothingmuch usually 2 days or so
no wait, it's the bad direction
gaal yeah
nothingmuch i'd say 3-4 days
gaal anymoose, we can talk when you get here 00:48
nothingmuch yes
we shall
gaal what row are you on the plane?
nothingmuch we shall!
oops
i was typing "exit row woohoo yeah!" and accidentally hit the up arrow while pressing shift 00:49
gaal what row *number*? :)
nothingmuch gaal++; # suggesting online check in
32
what does it matter? i don't even know what kind of plane it is
gaal awrk. rows <= 20 have 'lectricity
nothingmuch i don't work well on planes anyway
my brain shuts off after an hour and a half 00:50
i never finished the battery even now that i don't have 2 of em
so i'm not worried about wasted time
gaal just enough time to svk ci notes into MOP.hs :)
nothingmuch i'll try
but i'm really stupid and burnt out on flights ;-)
gaal anyway, I'm not going to sit here and be your procrastinatory excuse! go do those things you have to do :-P
and: 777, likely. 00:51
nothingmuch thank you sir
gaal for whipping you to work or for the pickle jar? 00:52
(wow, that was an obscure joke) 00:53
nothingmuch yes it was
nothingmuch chuckles in a fake manner
gaal eyes the whip
nothingmuch ponders getting an aeron chair
stevan has an old one 00:54
i've been sitting on it all week
and it's very nice
gaal go work already :)
nothingmuch i am working
gaal ha!
lumi gaal & nothingmuch: Could I stare at that stuff with you two when you hack it? 00:58
gaal lumi: please do join us!
lumi Eggcelent. :P 00:59
btw is there a trick to using rules? I'm getting "Broken pipe" when I try anything at all
A trick to building rulesfully
clkao nothingmuch: i think we need some aggregated ttm - for a test in rX, environment Y, backend Z
so we can do easy query on X-1 same environment, or X other environment 01:00
etc
gaal lumi: is your parrot embedded?
i'm guessing not
01:00 justatheory joined
lumi Yeah 01:00
Should be anyway, I set the path, and put it in the env var
gaal strace?
lumi Eek 01:01
I'd need nuffin's help, I haven't used the Mac trace facilities
gaal Eek indeed. Scary Macs.
I think it's aix like. strart trace daemon, do stuff, stop trace daemon and analzse dump. 01:02
lumi Eurgh. I have a trace. What next? 01:04
gaal what happened before the signal?
and: is thre an exec there?
(if there is, you don't have an embedded parrot)
lumi AFAICT yes, it's spawning a parrot 01:05
Which is weird. Maybe I needed to make clean at some point
gaal looks it. I'm betting you're running against a different version than the one you compiled against. 01:06
also check carefully the perl Makefile.PL output to see if it gave up on embedding
nothingmuch clkao: that's what my smoke server does
ask zev to integrate the code
i showed him how easy it is =)
01:07 khisanth_ joined
nothingmuch lumi: ktrace 01:07
and then kdump
gaal one mouse button, two trace commands.
01:07 khisanth_ is now known as Khisanth
lumi I had a vague recollection that it was ktrace, and man did the rest :P 01:08
I may have one mouse button, but I have four modifier keys
nothingmuch gaal: it's called "decoupling" ;-)
5
lumi Oh um
Yes, shift is a modifier key indeed
nothingmuch also, esc does mod when alt isn't meta
in e.g. terminal stuff
lumi A real space cadet kb
nothingmuch clkao: so how do i get your neighbors' tea? 01:09
svnbot6 r11522 | fglock++ | v6 - 'use perl5:CGI' 01:11
lumi Parrot.hsc:346:5: error: token ""0.4.5-devel"" is not valid in preprocessor expressions
But everything keeps running
Building (sorry)
clkao nothingmuch: uhm, send me a hiveminder task about how much you want etc 01:13
obra clkao++ # corporate shill
nothingmuch clkao@bp ?
clkao nah. it's actually useful
nothingmuch: yes please
otherwise i lost track
nothingmuch how does it know to complete your email? 01:14
isn't that bad for privacy?
or is it because we share a group or you've mailed me before?
gaal aren't you in the same group or something?
01:14 fglock_ joined
gaal hmm though the inviter isn't necessarily trusted 01:15
point.
fglock_ walks in
lumi Do I need a particular parrot? I built latest from svn 01:16
gaal lumi: probably good choice
fglock_ clkao: && works for me 01:17
clkao fglock_: you broke 03
lumi gaal: Any idea what that error I had is?
gaal lumi: nope, but let me look...
(hsc is preprocessed into hs, for ffi stuff) 01:18
ooof svk pull is so slow :(
clkao of course that's svk's fault 01:19
gaal it's pugs' fault for being so active
and light's fault really for being so slow
but clkao go ahead, fix that :) 01:20
is there a lot of round tripping in a pull that basically gets n remote updates? or is the data to be xferred computed once and transmitted as a bulk? 01:21
because roundtripping to taiwan hurts
audreyt raises the "set up svn on feather and migrate to it" janitor task again 01:22
gaal audreyt: is that a covert means of encouraging you to erdos a lot? 01:23
audreyt yes. 01:24
fglock_ clkao: fixed
vsoni audreyt: Are there any thread safe scanner/lexers
audreyt vsoni: Lemon
gaal ah, I get it. the version string "0.4.5-devel" is not comparable with the ">" that's in the hsc 01:25
audreyt gaal: oh?
aw.
svnbot6 r11523 | fglock++ | v6 - 'say' emits a function
audreyt vsoni: flex?
gaal I suppose, lumi, that hacking your parrot working copy to discard the -devel from the announced version will "fix" that...
lumi Yeah, I guess it ought not grab the "devel" maybe
kolibrie audreyt: can you check that test I added: t/rules/grammar.t
lumi Changing whatever re's grabbing it that way might be better 01:26
gaal probably :)
lumi I'll do it when build ends maybe
audreyt kolibrie: checking
lumi Right now my poor g4's cycles are being eaten by ghc
vsoni audreyt: is not reenterant?
gaal has no pity for g4 cycles 01:27
audreyt vsoni: flex with -R
lumi Heh
gaal lumi: I don'
t know where PARROT_VERSION is schlepped from
it may be a parrot include. audreyt?
kolibrie audreyt: I had to change the backslashes to <'literal'> form, otherwise I was getting PGE errors. Patrick said it's not PGE. Any ideas?
audreyt vsoni: oh wait, only in C++ mode 01:28
vsoni audreyt: yep...
lumi gaal: Yeah, iswim, problem
gaal well we can always autogen the hsc file :) 01:29
lumi I don't know cpp well enough
gaal no, cpp is too late
lumi Hmm
01:30 fglock_ left
gaal we need smth -> hsc -> hs -> c -> nirvana 01:30
long is the path of the righteous etc.
lumi hsc can #include "parrot_ver"
Shorter
gaal heh, indeed.
faster too.
lumi Pugs AST has too many internals 01:31
gaal no kidding!
audreyt vsoni: re2c
vsoni: re2c.org/ 01:32
lambdabot Title: "re2c Home"
lumi If only you could #include `parrot -v | perl -e ...`
cpp is insufficiently advanced
vsoni audreyt: thanks audreyt
gaal yes, instead of magick you get only the ick.
audreyt vsoni: also c2lib may be interesting, but unrelated 01:33
vsoni: re2c+lemon looks like a winner.
gaal lumi: src/Pugs/pugs_config.h is the place to put this probably. use something like CLEAN_PARROT_VERSION 01:34
commit and see if it works? (or the other way around)
I think I'll try to sleep a bit now :)
good night :) &
kolibrie gaal: isn't it morning there? 01:35
gaal soon
kolibrie sleep well
gaal thanks :) 01:36
lumi Nite gaal
gaal z&
audreyt kolibrie: <''> doesn't appear to play well 01:37
kolibrie audreyt: I had just the opposite, see if backslashes work
lumi What gens this pugs_config.h? 01:38
audreyt oh wait
token speaker { [\w|\s]+ }
you had it
token speaker { \w[\w|\s]+ }
the first worked for me
kolibrie oh yes, that is different, sorry 01:39
audreyt and that alone fixed it.
I changed it to \w+
r11524
see if it worksforyou
brb 01:40
svnbot6 r11524 | audreyt++ | * grammar.t -- make it pass here by adjusting the rules a bit.
audreyt and maybe <''> would work too
if so, change the tests back
kolibrie pulls
nothingmuch wants his prodigy album returned 01:43
pasteling "kolibrie" at 66.160.90.8 pasted "grammar.t output when there are backslashes instead of <''>" (7 lines, 283B) at sial.org/pbot/18385 01:46
"kolibrie" at 66.160.90.8 pasted "grammar.t output with <''>" (12 lines, 343B) at sial.org/pbot/18386 01:47
kolibrie audreyt: so the first error is kind of weird, and seems to mess up the PGE bridge? completely 01:48
audreyt: and when PGE runs, the final test matches nothing here 01:49
ingy hi nothingmuch 01:52
nothingmuch ingy: hi
what's up? 01:53
svnbot6 r11525 | kolibrie++ | grammar.t - switched back to <''> form
ingy 8^B
nuffin hunny
nothingmuch look, i have an ascii moose on my webpage: nothingmuch.woobling.org/ 01:54
lambdabot Title: "Yuval Kogman's Link Farm"
nothingmuch ain't it preeeetty
ingy yes darlin
audreyt kolibrie: all grammar.t still pass for me
kolibrie: head parrot
nothingmuch yay! all tests pass
(all failing tests removed ;-) 01:55
audreyt kolibrie: it is possible that you are running old parrot, I think
but hm.
nothingmuch clkao: star merge sanity check:
can i branch a branch in a trunk/branches topology
and then push back up wards to the previous branch
and then back up from that
in two steps?
clkao is so happy to be able to run: ./prove6 t/rules/grammar.t
nothingmuch or will smerge go all the way 01:56
clkao nothingmuch: yes
nothingmuch hugs clkao
clkao nothingmuch: so, you want tea. in different batch?
nothingmuch i want the green tea we had
and maybe something else too
kolibrie audreyt: I have tried on three computers, the grammar fails to match on any of them
nothingmuch if you can think of a nice idea
clkao you want them sent in different batch? 01:57
audreyt kolibrie: indeed. do we have a new parrot built on feather?
kolibrie: I can try on feather
nothingmuch clkao: one package is good
clkao ok
nothingmuch it doesn't need to be partitioned
clkao i will process that
lumi So tell me of this tea, please
clkao uhm, more tea requests?
kolibrie kolibrie: I built on in ~kolibre/parrot
nothingmuch lumi: clkao lived (lives?) by a plantation in .tw
that makes very nice tea
clkao i think best practical should start selling hackers tea 01:58
nothingmuch that would be a nice service
kolibrie audreyt: but now I'm getting: Cannot find the parrot executable in PATH
nothingmuch will it be in webapp form?
audreyt kolibrie: add it to PATH
export PATH=$PATH:/home/path/to/parrot
clkao we need to build a jifty ecommerce site
nothingmuch lumi: anyway, i'm asking him to send me some from .tw
because i really liked what I had during the obrathon
and while hiking
clkao but you need the water from the pond 01:59
nothingmuch we may still have some ;-)
lumi And some evil eggs?
nothingmuch actually
did we make the tea from the cola bottle?
oh wait
the last tea was definitely from garfield pond
kolibrie audreyt: PATH worked, grammar.t still fails last test on feather 02:00
audreyt kolibrie: what's the PATH to parrot on feather?
nothingmuch clkao: i hate svk because if i have vim .swp files it won't let me branch
err
not hate
svk sucks becauase
(Sorry, i got carried away ;-) 02:01
kolibrie audreyt: ~kolibrie/parrot
audreyt kolibrie: it's all my fault 02:02
sorry for wasting you one day of time
parrot r13788 :) 02:03
r 13288
(I had noncommitted local changes.)
kolibrie audreyt: ood
ooh, rather
audreyt: now it works?
audreyt yup
sync up to latest parrot 02:04
and it'd work
kolibrie ups
does parrot need realclean? 02:05
audreyt no
just "make"
not even "make" is needed in this case
kolibrie that easy?
audreyt yeah
kolibrie audreyt: looky there!
audreyt anticlimatic, I know 02:06
kolibrie audreyt: will pmichaud be upset? he keeps claiming it shouldn't work?
audreyt I don't know
maybe he'll fix PGE so that it stops working?
j/k :)
kolibrie that would be just my nightmare 02:07
kolibrie envisions presenting to programming staff at $work and suddenly having it stop working 02:08
for production I'll really rather do it under v6, though 02:09
audreyt: you may have noticed the namespace test it grammar.t, that was to test namespaces, but now that is gone, and untested 02:10
kolibrie isn't quite sure how to test importing a grammar anyway, so the test may not have even worked before 02:11
audreyt: anyway, thank you much and & 02:12
02:27 fglock_ joined
svnbot6 r11526 | fglock++ | v6 - parser: fixed circumfix:<[ ]>, postcircumfix:<[ ]>, $exp.[], .[] 02:29
r11526 | fglock++ | - fixed an infinite loop in parser (retry null match)
r11526 | fglock++ | - no semicolons inside [] yet
meppl gute nacht 02:30
fglock_ Schlaf gut 02:33
meppl ;)
03:57 ilogger2 joined, leptonix joined 03:58 ajs_home joined, thepler joined 04:06 szbalint joined 04:25 orafu joined 04:27 pasteling0 joined 04:37 Yappo joined 04:43 xinming joined 05:13 lambdabot joined 05:15 xinming joined 05:20 miyagawa joined 05:50 xinming joined 05:59 merlyn joined 06:08 lambdabot joined 06:13 lambdabot joined 06:20 xinming joined 06:47 mauke_ joined, xinming joined 06:50 cmarcelo left 07:02 mauke_ is now known as mauke 07:04 Aankhen`` joined 07:10 iblechbot joined 07:17 drrho joined 07:20 xinming joined 07:29 masak joined 08:09 nothingmuch joined
gaal nothingmuch: are those just random hikers you met on the trail? 08:29
08:29 wilx joined 09:04 xerox is now known as xerox_ 09:05 xerox joined, revision17_ joined, xerox joined 09:07 jabbot joined 09:27 beppu joined 10:22 andara joined 10:58 beppu joined 11:12 chris2 joined 11:22 ludan joined 11:42 wesle1 joined 12:14 bpphillips joined
nothingmuch gaal: yes 12:36
12:44 froh-doh joined 12:46 scook0 joined
nothingmuch see you guys in ~18 hours 12:53
(worst case scenario that does not include a crash ;-) 12:54
13:13 nova-alpha joined
nova-alpha hi 13:13
broquaint Hola. 13:17
13:19 Jedai joined 13:20 xinming joined 13:27 wesle1 joined 13:28 lisppaste3 joined, lumi joined 13:36 mj41_ joined 13:37 stevan joined 13:53 buetow joined 13:59 buetow joined 14:01 vel joined 14:07 fglock joined, nova joined 14:08 Qiang joined 14:20 lisppaste3 joined 14:21 kanru joined 14:26 cjeris joined 14:32 andara joined 14:43 kane_ joined 14:46 f0rth joined 14:51 f0rth_ joined 14:52 hexmode joined 14:54 chris2 joined 14:58 mj41_ joined 14:59 orafu joined 15:02 lollan joined 15:03 mako132 joined 15:10 elmex joined
kolibrie I tried running examples/rules/yapc_schedule.pl against the actual schedule (yapcchicago.org/the-schedule/monday) and on feather it took forever (20 minutes) so I killed it, and locally I get *** Cannot parse PGE: <schedule>\n*** Error: no ICU lib loaded 15:14
lambdabot Title: YAPC::NA 2006 » Monday
15:15 ivanfrey joined
ajs Garrrr! I missed the Boston.pm talk! 15:34
<--- ijut
cjeris I wrote a few tests in t/operators/hyper.t for hyperoperators distributing into subarrays like -<<(1, [2, 3]). Since I'm new to p6, could someone spare a moment to idiot-check them before I commit? 15:36
pasteling0 "cjeris" at 140.247.124.140 pasted "new hyperop distribution tests" (63 lines, 2.1K) at sial.org/pbot/18394
kolibrie cjeris: generally the rule is commit and then people can update 15:37
cjeris kolibrie: ok. I didn't want to break something if the rule was in the other direction 15:38
kolibrie cjeris: thanks for asking, and welcome aboard 15:39
cjeris kolibrie: thanks!
kolibrie apt-gets libicu34-dev 15:43
15:44 justatheory joined 15:47 DaGo joined
masak cjeris: the tests look good by me 16:01
cjeris masak: cool, thank you! 16:02
16:09 ruoso joined 16:17 iblechbot joined
clkao appearantly the installable devel::caller is already on cpan 16:17
16:35 amnesiac joined
stevan the Connecticut hackathon is now over, nothingmuch is on his train (to catch his plane) 16:38
much m??sey goodness has come out of it though :) 16:39
xerox You do lots of hackatons, meetings, and conferences, men!
kolibrie (moose)++
stevan We got some good m??se pics of the kids and the dog as well 16:40
xerox There seem to be always some conference going on here, unlike other programming language channels.
stevan xerox: I think it is really a seasonal thing
16:40 dbrunton joined
xerox Since you are lambdacamels it may be of interest to you <haskell.org/haskellwiki/History_of_Haskell>. 16:41
stevan things calm down in the winter (nothern hemisphere winter to be exact)
xerox: do you know where I can find the SPJ paper on the history of Haskell? I was told it is an excellent read
stevan reads the first paragraph and withdraws the question 16:42
xerox Check that link.
It is also from Hudak, Hughes, and Wadler. All the sages :-)
stevan I should read first,.. ask questions later :)
PerlJam stevan: be sure to read section 12.4.4 16:47
16:47 Daveman joined
fglock clkao: ping 16:50
audreyt g'morning 17:07
stevan: so nothingmuch isn't going back to Boston?
stevan audreyt: no, he is flying back to .il
audreyt awww. 17:08
too bad
guess I'll meet him in OSDC.il then
stevan yeha we wanted to try and come back up, but $work is too crazy for me right now 17:09
audreyt: how much longer will you be in Boston? 17:12
audreyt stevan: 8 days
stevan then to OSCON?
audreyt yup
stevan nothingmuch and I are going to work up 3 talks for YAPC & OSCON next year 17:13
and maybe try to submit a paper to OOPSLA on Moose/MOP/Perl6MM
audreyt I'll be in OOPSLA
stevan cool
audreyt or rather, DLS 17:14
stevan DLS?
audreyt www.oopsla.org/2006/submission/trac...osium.html
lambdabot Title: OOPSLA 2006
stevan ah yes
I think I read about this on LtU
audreyt <- this year's invited speaker
stevan very nice :) 17:15
audreyt maybe I get to present your paper :D
stevan we are thinking next year
audreyt ah ok.
stevan by then Moose/MOP stuff should be worked out
audreyt nodnod
that'd b wonderful
s/b/be/
stevan we actually discussed a nice refactor to the method protocol which will allow us to layer a really nice query interface over the MOP 17:16
which will make MOP actions really really simple
audreyt query as in structural query?
or as in relational query?
stevan uhm,... neither, and kind of both :)
it is still a rough idea
audreyt mooseish query then 17:17
stevan yes :)
we basically realized that several features would be greatly simplified by a more unified and consitent iteration interface
audreyt decides to fix Win32 build breakage first
ah. yup
stevan and what good are iterators if you cant filter then nicely :)
I will let the ADD do the rest of the connecting of the dots :) 17:18
it also simplifies the Moose::Visualized module which I have been pondering for a while now
17:20 SamB joined
audreyt yup 17:21
a vision for moose 17:22
stevan m??se g??gles
17:30 kudra joined
kudra hi 17:30
kolibrie hi kudra 17:31
audreyt kudra: hi 17:41
kudra that was a double hello :)
ajs Question about enum: Would "enum x <<a b c>>; sub foo(x :$n,$m); foo :n<a>, 2;" pass :n<'a'> or :n<x::a> to foo? 17:43
In other words, does the a get auto-quoted when the named parameter is expecting an enum?
PerlJam ajs: X<a> is equivalent to X{'a'} always 17:46
(or so I believe :)
for the latter, you'd have to do :n(a) (assuming that really works)
audreyt yup and yup 17:48
ajs Well, gah! That's just unfun.... 17:49
Thanks for the info, though. 17:50
So, the example for the round function will be: my $n_rounded = round :dir(RoundEven), $n;
PerlJam Why is it called :dir? 17:55
round :semantics(RoundEven), $n 17:56
round :semantics(IEEE), $n
(or something)
dir implies "direction" to me and that doesn't make sense. 17:57
pmichaud_ actually, I first read it as "directory"
"What do directories have to do with rounding?" <-- my first reaction
PerlJam round :semantics(Floor), $n; round :semantics(Ceiling), $n; # :-)
ajs PerlJam: ... because it's a direction. There are seven directions associated with rounding numbers: up, down, positive infinity, negative infinity, even, odd, random.
Don't confuse this with floor/ceiling. That's not a parameter to round! 17:58
pmichaud_ almost makes me wish for :up, :down, :even, :odd, ...
ajs round only does the closest whole number. The only question is what you do with .5
PerlJam pmichaud_: but then what happens when they ask for round :up :down, $n ? 17:59
ajs I really dislike adding :foo and :bar when the they are mutually exclusive
pmichaud_ but it *reads* better
ajs round :up, :down, $n does not read better
PerlJam ajs: who actually would use "random" ?
maybe for testing. 18:00
but not "for real"
ajs PerlJam: random rounding directionality is used very often in statistical circles, though even is more popular in those circles
pmichaud_ so, we're going to increase the huffmanization for the correct case to avoid the potential for an incorrect one?
ajs It's VERY common. In fact, I know of at least two production applications that affect your wallet on a regular basis.
kudra bye
ajs later kudra
PerlJam can't we always add a constraint on the sub so that (at compile time), there's an error if mutual exclusivity is violated? 18:01
pmichaud_ we don't always know at compile time
ajs Yep, you can always call round([,] $capture) 18:02
A run-time check could catch that
pmichaud_ also, why only round to integers?
ajs pmichaud_: We round to complex as well 18:03
for complex input
er, actually, Complex
PerlJam ajs: But I want to round to the nearest tenth
pmichaud_ what about rounding to the nearest tenth, nearest half, etc.?
fglock kolibrie: re mail - that looks like spam
pmichaud_ nearest three, nearest ten, nearest 100, etc. 18:04
kolibrie fglock: thought so, just wanted to make sure your computer wasn't infected or anything
pmichaud_ fwiw, php uses "precision" for round: $n = round :precision(0.5) $x
ajs pmichaud_: so, you want: our Int multi Math::Basic::round ( Num $x, :$up, :$down, :$infinity, :$neginfinity, :$even, :$odd, :$random, :$precision, :$base)?
pmichaud_ maybe just precision: $n = round :precision(RoundEven) 18:06
PerlJam eh?
ajs That's not a precision.
That's a direction for tie-breaking
pmichaud_ ohhhhh
I get it
PerlJam ajs: round :tie(RoundUp) :precision(0.001) 18:07
pmichaud_ personally, I like $n = round $x :up to $n = round $x :dir(RoundUp)
but I don't have a strong opinion on it
ajs Why is :tie preferable to :dir? More to the point, what are we solving for, here?
PerlJam ajs: dir still doesn't make any sense.
( to me)
why not "direction"? 18:08
ajs Simply length. I try not to have named parameters which exceed the length of the function name.
18:08 chris2 joined
pmichaud_ notes, fwiw, that unix command-line options have no difficulty with mutual-exclusivity 18:09
PerlJam ajs: do you do any C programming under unix? Have you ever used creat(2)?
ajs Quite a bit
pmichaud_ bows out for a while
PerlJam I think dir is the same kind of mistake as creat 18:10
ajs creat is a mistake because it sacrifices readability for one character. "dir" is a commonly used abbreviation (esp. in programming) for "direction", and more than halves the length of the token. I don't think this is a reasonable comparison. 18:12
If there were any possible confusion with "directory", then I'd say it was a bad choice
kolibrie prefers verbosity over ambiguity, unless it is something used so often that it is no longer ambiguous 18:13
ajs But the problem is that "tie" isn't what we're doing here. We're selecting a direction in the case of a tie... I'm not sure that works adverbially...
kolibrie: And you don't think rounding direction is used very often? Read: www.pldesignline.com/howto/showArti...=175801189 18:14
18:14 GeJ joined
kolibrie ajs: I think it should have a nice default so I don't often have to specify my direction 18:15
PerlJam ajs: you'll note in that article that they mention "round-floor" and "round-ceiling" :-) 18:16
kolibrie: what's a nice default?
ajs Well, it does have a nice default, and I don't expect most people to specify it, but those who need it, use it A LOT.
PerlJam: That's not the section I was refering to. Perl does and always has made a distinction betwen "rounding" and "flooring" as different operations. 18:17
But if the concern is that signatures are too short, and should allow for all possible combinatorics of uses of the abstract sematic represented by the function, then S29 is going to beat out War and Peace
(and I'll not be doing all of it) 18:18
PerlJam Sure, but are the names given there representative of the names used in common parlance? If so, shouldn't we adopt them?
kolibrie PerlJam: I don't round enough to know, that's why I'll use the default
ajs PerlJam: We have adopted them. That's why there's a floor and ceiling function. These are well used names.
PerlJam So this is a case where you think TIMTOWTDI shouldn't apply? 18:19
ajs But "round" presents an ambiguity, and to resolve that ambiguity, you have to pick a direction to go in case of tie.
kolibrie we could use unicode arrows :) 18:20
ajs WHAT ... ARE ... WE ... SOLVING ... FOR?!
bpphillips notices the bikeshed, starts thinking about what color it REALLY should be ;-)
kolibrie I was just saying that for people who round very infrequently, it would be nice to have adverbs that are verbose enough for me to understand them 18:21
PerlJam ajs: 1) dir is a poor name (at least IMHO). We need a better one. And 2) I'm lobbying for RoundFloor/RoundCeiling because I think that there's some largish subset of the population who think of floor and ceiling as rounding operations and would naturally look for them there.
(Certianly not in POSIX without some help) 18:22
kolibrie the regex adverbs have one-letter and verbose forms 18:23
theorbtwo PerlJam: That's what documentation is for, not making the name longer.
kolibrie could have :d and :direction
cjeris $n = round $x :toward(Even) ?
bpphillips cjeris++ 18:24
theorbtwo I don't think we need to put every concieveable variation of rounding in the core.
PerlJam theorbtwo: I can accept that. I think.
theorbtwo ceil(), floor(), and a round() with unspecified tiebraking are plenty.
ajs Decide amongst yourselves. I'll move on to something else for now. I don't have time to have a 20 minute conversation about every function. There are just too many. If you want a run-on signature, though, you'll have to tackle the formattign question. Right now S29 dodges that by having, in general, short signatures.
PerlJam kolibrie++, and I can accept that too. (In fact, earlier when we were talking about precision, I was going to suggest :p rather than :precision, but why not both? ) 18:25
ajs: sorry to distract. keep up the good work! ajs++ :-)
kolibrie yes, ajs++ 18:26
ajs Thanks guys. Sorry to be cranky, it's just that this work is SO slow... I have to slog through the synopses multiple times for every single function, as they always end up bringing up their own odd questions. 18:31
elmex elmex++
perlbot What kind of idiot karmas himself? Your kind of idiot!
ajs And, then there's the 3 years of mailing list archives... 18:32
perlbot--
PerlJam ajs: sounds like you need to delegate. Now that S29 is broken up into pieces, recruit volunteers for each piece
ajs: though you'll have to be responsible for quality control
(since you're the function pumpking :-) 18:33
tewk hears PerlJam volunteering :)
ajs Well, that will happen. I need the framework first. Ultimately at least half of the API documentation needs to wait until AFTER the first implementation of the modules that it documents. It will be iterative, and I hope the poeple who implement will be willing to document what they do 18:34
PerlJam tewk: indeed. If I remembered how S29 was divided, I'd grab a piece that looked interesting enough to keep me working on it.
ajs: oh! Then what needs implementing still? Maybe make RT tickets for each function/module so that people can be encouraged to get to work :) 18:35
ajs PerlJam: Every module that S29 documents (e.g. Math::Basic, Scalar, OS, etc) will have a corresponding API doc.
And a corresponding module that people can actually "use"
PerlJam needs to get a parrot commit bit or something ... I keep getting more and more interested in it as perl6 gets nearer 18:36
ajs If you want to start writing modules, based on Functions.pod, feel free. Those who write code, typically get commit bits around the first time they say, "I have code, who wants to commit it?" ;-)
tewk PerlJam: do you have a pugs commit bit? 18:39
PerlJam tewk: I do. 18:40
(who doesn't? ;-)
stevan matz
ajs Oh, I was misunderstanding your statement...
stevan guido never picked his up
tewk My thoughts exactly after asking, stupid question.
ajs stevan: heh
tewk but I'm sure it's waiting for him.
stevan tewk: yes :) 18:41
maybe this OSCON audrey can get them to sign up
tewk matz seems like the type who would definitely sign up.
is matz going to be at OSCON?
18:51 evalbot_11535 joined
audreyt Win32 build repaired. 18:51
kolibrie audreyt++ 18:52
fglock svnbot is down
audreyt bringing it up now
cjeris audreyt++ 18:53
ntfs-- : svn: Can't open file 'docs\Perl6\API\.svn\tmp\text-base\Math::Basic.pod.svn-base': The parameter is incorrect.
ajs Doh 19:04
What should I call that
Math-Basic.pod? Or do we tell NTFS to go touch itself? ;)
audreyt Math/Basic.pod 19:05
I think
makes CPAN's life easier
ajs Really? Ok. I thought subdirs would be annoying.
Will do
xerox How does one see the list of people's karma? :D
fglock audreyt: svnbot still doesn't work 19:06
19:06 svnbot6 joined
audreyt fglock: need a ?join 19:06
fglock thanks!
svnbot6 r11540 | ajs++ | Moved Math::Basic so that systems that cannot do :: in filenames will cope
ajs done 19:07
Ok, that's it for me until Sunday or Monday. Have a great weekend, y'all 19:10
kolibrie ajs: happy weekend
audreyt have fun :) 19:16
svnbot6 r11541 | fglock++ | v6 - TODO update 19:25
r11542 | audreyt++ | * take away another "ar" call on Win32 to reduce confusing error msg. 19:31
audreyt bbiab... 19:36
kolibrie audreyt: I got yapc_schedule.pl to work on the real schedule (yapcchicago.org/the-schedule/monday) if I removed the '—' (\x{2014}) characters, which were not matching . 20:01
lambdabot Title: YAPC::NA 2006 » Monday
nothingmuch SO!
guess what
did my flight checkin go smoothly? 20:02
kolibrie hey nothingmuch
nothingmuch of course not!
nothingmuch wants to sleep
so much
kolibrie are you stranded?
gaal nothingmuch: what happened?
nothingmuch my train was like 20 mins late 20:03
but that was OK
it was still enough time to actuallly make it
only the eticket machines kept crashing with my info 20:04
gaal eek
nothingmuch i finally found someone who listened
so she tried it
and it turns ouit checkins to israel are restricted
and i need to go to some other place
from that place i got sent to yet another place
where I discovered that somehow my booking was canceled
gaal !!!
nothingmuch (this is about 30 minutes before departure)
gaal oh
nothingmuch (i was running around the airport for over an hour by then) 20:05
gaal hmmmm what now then?
nothingmuch so they put me on a different flight
departing at 22:00
kolibrie ugh
nothingmuch so that was fixed, kinda
gaal what time is it now? 16:05
kolibrie yes
gaal where are you?
nothingmuch and to compensate me, i discovered that I "was selected by Continental airlines"
gaal you were selected for what?
nothingmuch what this means is that some guy with gloves gave me a full body massage 20:06
gaal heh
nothingmuch if i'm really lucky they will tell me that i'm on standby
gaal an upgrade to business on a transatlantic flight might be a nicer selection...
kolibrie nothingmuch: you at JFK?
gaal are you waiting in the airport? 20:07
nothingmuch kolibrie: EWR
i'm now on the gate next to mine
(Israel's gate is blocked off)
gaal I wonder if you can check yout luggage somewhere and hop off to manhattan
newark is alittle clumsy to reach manhattan by with cheap routes
nothingmuch gaal: i'm already checked in 20:08
i don't want any more mess
gaal you need to take the bus to Newark the city, then a PATH to exchange place, switch to the other line and bam, you're in 33rd St
under five bucks
but that's a few switches and potential confusion.
obra will be in nyc tomorrow 20:09
kolibrie and how many hours?
gaal kolibrie: not very long, under 1.5 hr
nothingmuch gaal: that's ~1.5 hrs in and ~1.5 hrs out
gaal yes
nothingmuch while I can have fun it's going to be risky
this doesn't include recheckin time
and leaving the gate
gaal or take a taxi, about fifty bucks evrey way :(
nothingmuch naaaaaaaah
gaal well you do have six hours
obra nothingmuch: you've got internet, right? 20:10
gaal oh, less than that really
kolibrie or helicopter, they have those too
gaal well, look on the bright side :)
we get to moose :)
nothingmuch obra: it appears so
gaal: yes
gaal where are you pulling internet from btw? I was wondering that at the gate 20:11
before my flight that is
gaal 's crummy ol' NIC doesn't have scan
nothingmuch concourse
gaal let me guess you're at 99? 20:12
98 maybe
stevan nothingmuch!!
gaal stevan: moose
nothingmuch hoi esteban
gaal: 96 right now
or 94
gaal so, let's antler a bit!
nothingmuch but it's a few meters from 98
gaal: give me 15 mins to finish some Class::MOP silliness 20:13
gaal my first question is about the method tables
sure
stevan gaal: moose
gaal hey, maybe we can gobby this? 20:14
nothingmuch they are not tables ;-)
gaal doesn't have SEEE
nothingmuch they're iterators
gaal huh? over what?
gaal looks at Class::MOP for reference
kolibrie audreyt: do you know where \x{2014} not matching . in a grammar is caused, and how easy to fix? 20:16
nothingmuch gaal: i was kidding
i will explain soon
gaal k 20:17
fglock which perl5 module implements 'reduce'? 20:18
gaal List::Util probably
brb 20:19
nothingmuch: foo says hi 20:22
well, in Class::MOP it's simply the Perl 5 symbol table. so I'm guessing a Map Ident Code simply? 20:23
nothingmuch gaal: pet foo for me 20:24
gaal++; # helping me out last year when my flights also went screwey
gaal will do
nothingmuch gaal: anonymous classes might do whatever
it should ofcourse, be opaque
but Map Ident Code is good enough
if Code has the meta data
gaal nothingmuch: oh! need me to call anyone or are you good on that front?
nothingmuch (i don't know Code, i thought VCode...) 20:25
what time is it in israel?
gaal nothingmuch: the new Code data is pretty rich
nothingmuch i mailed my 'rents
gaal 23:25
nothingmuch but crap
they won't check email
please wake them up
08-651-2264
gaal I'd be happy to give them a call
nothingmuch <3 gaal
gaal what's your TLV ETA?
nothingmuch checking 20:26
gaal + flight no. 20:27
nothingmuch 90
gaal k, give me 5 min
(phone on emergency charge...) 20:28
nothingmuch gaal: flight 90, arrives at 4:15 pm
oh shit
svnbot6 r11543 | kolibrie++ | yapc_schedule.pl - updated to work with yapcchicago.org/the-schedule/monday
r11543 | kolibrie++ | * still fails to read past the \x{2014} characters,
r11543 | kolibrie++ | so delete those from the html until that gets fixed
lambdabot Title: YAPC::NA 2006 » Monday 20:29
nothingmuch they have a scheduling issue
gaal in the meanwhile, if you're bored, take a look at src/Pugs/AST/CapInternals.hs
nothingmuch *GUGHHHHGGGGGGG*
and no fucking trains
nothingmuch dies
gaal nothingmuch: crash here
you'll have to pay 120nis for a cab to my place (sorry no car to pick you up with)
but you're welcome to stay here till you're unstranded
nothingmuch <3 once more 20:30
oh wait, when it's mozash?
fglock gaal: reduce works :)
gaal fglock: yay :) 20:31
svnbot6 r11544 | fglock++ | v6 - say [*] 1..5
gaal whee! fglock++
fglock: remember to also do [\*]
nothingmuch: tomorrow no trains Lod -> Beer Sheba due to maintenance. 20:32
fglock gaal: submit a patch to List::Util? :) 20:33
nothingmuch gaal: *sigh*
gaal: tell them that you can host me if they can't pick me up 20:34
they have some sort of engagement related to my sister
gaal fglock: actually part of the point of scans (productions) is that they're lazy, so in perl 5 they're less efficient
nothingmuch but i don't know what it is
luckily i have cloths and stuff
pmichaud kolibrie: (\x{2014} not matching) that's a parrot implementation issue -- it doesn't deal well with unicode characters if ICU isn't loaded 20:35
gaal nothingmuch: the train site is kinda confusing, could be only the Lod->bs service is closed but there's a train from tlv leaving 22:11
kolibrie pmichaud: ICU is loaded (before I loaded it I was getting ICU not found errors)
clkao fglock: pong 20:36
pmichaud then it ought to match
kolibrie hmm. Any particular version of ICU I need?
nothingmuch gaal: =/
gaal: i have you as fallback
so i will see what I can do
pmichaud not that I'm aware of -- anything that causes parrot to build cleanly ought to be okay
kolibrie pmichaud: ok 20:37
fglock clkao: I fixed some syntax problems
clkao cool
gaal nothingmuch: your folks will pick you up tomorrow :)
just spoke with them 20:38
fglock clkao: t/data_types/hash.t parses now (emits with errors)
pmichaud kolibrie: you might run prove t/compilers/pge/p6regex/metachars.t in your parrot instance, it has some unicode checks in it
gaal oh, even if there weren't a train you could have taken the bus, it's not too bad.
but.. they know the flight number etc.
pmichaud I'm not sure how good the bridge is when dealing w/unicode. I suspect it's okay, but...
svnbot6 r11545 | fglock++ | v6 - updated ChangeLog 20:39
nothingmuch gaal: thanks
gaal nothingmuch: sure thing
nothingmuch <3 gaal
gaal++
kolibrie pmichaud: t/compilers/pge/p6regex/metachars....ok
clkao fglock: WHOOT 20:40
kolibrie pmichaud: could be a bridge thing, maybe.
clkao fglock: with hash working we can start rewrite the emitter in perl6
gaal oh, I didn't know List::Util was XS
fglock clkao: it's time to split the emitter in submodules - it is starting to get big 20:42
nothingmuch clkao: you'll mail me again, right? becasue you marked the task as finished
20:42 rodi joined
clkao nothingmuch: yes, i created a task for my friend to report the postage 20:42
nothingmuch: once i got it i will and it to your total balance and request payment by paypal ;) 20:43
nothingmuch =) 20:44
gaal: join moose? 20:46
audreyt: shit, i forgot to tell you that stevan couldn't visit boston 20:47
clkao moose
fglock clkao: the whole compiler can be written in v6 :)
clkao yes, very soon ;) 20:48
gaal nothingmuch: well
?
20:52 cmarcelo joined
nothingmuch 6 20:53
svnbot6 r11546 | fglock++ | v6-alpha-0.009 in cpan
clkao fglock: are you aiming to roll releases every two days? 20:55
21:03 knewt joined
kolibrie nothingmuch: stevan mentioned it here this morning 21:03
clkao fglock: need [1..3] 21:05
fglock clkao: say [1..3].perl works here 21:07
clkao oh ok
02-test-pm/1basic still fails to parse 21:08
nothingmuch kolibrie: the not being able to go to boston? i saw audreyt asking about it in my away log but not the remainder of the conversation 21:11
kolibrie nothingmuch: yes
I don't know how it concluded, either, but I think she was aware
nothingmuch okay =) 21:12
fglock clkao: re 1-basic.t - subs must be moved from 'statement' to 'term'
lumi nothingmuch: Maybe I can pick you up?
nothingmuch lumi: no, they are picking me up 21:13
lumi Oh, k
nothingmuch but thanks!
clkao fglock: i though it is already. otherwise my $foo = sub {} wouldn't work?
anyway, off for a bit
fglock clkao: 'my $foo = sub {}' works because the emitter is smart about it 21:17
svnbot6 r11547 | gaal++ | * Pugs.MOP: both Module and Class only Maybe have Package and Module, respectively
fglock clkao: "Release Early, Release Often" 21:18
gaal openfoundry is kinda slow tonight :( 21:20
cjeris I'm trying to fix Pugs.Prim.List to distribute hyperops over nested arrays, but I think what I just coded is wrong. 21:22
pasteling0 "cjeris" at 140.247.124.140 pasted "trying to add hyperop distribution" (22 lines, 533B) at sial.org/pbot/18402
svnbot6 r11548 | fglock++ | v6 - removed empty file 'moose-1.pl'
cjeris It needs to return a reference in the second slot, so it looks like (1, [2, 3]), right?
21:42 DaGo joined 21:53 kanru2 joined 22:02 cjeris left, thepler joined 22:07 kanru joined 22:09 froh-doh joined
svnbot6 r11549 | cmarcelo++ | * ByteString support for Map and MapSL. 22:16
r11549 | cmarcelo++ | * Naive HsJudy-based entry for K-Nucleotide Shootout Entry.
r11549 | cmarcelo++ | Three times better than GHC #1 entry. Couldn't check times for
r11549 | cmarcelo++ | GHC #2 and GHC #3. But worst than Ketil's entry in Hawiki
r11549 | cmarcelo++ | (Haskell old wiki).
nothingmuch pulls an audreyt 22:17
22:19 Jedai joined 22:25 rindolf joined
rindolf Hi all. 22:25
xerox cmarcelo: How's your SoC doing? :D 22:26
audreyt apparently pretty well :) 22:32
<3 the ByteString support 22:33
rindolf Is the Google SoC open for people who already graduated from university? 22:34
obra hey audreyt 22:35
rindolf Hi auder
Hi audreyt
audreyt obra: hey
rindolf audreyt: have you given a talk in Boston?
audreyt yes.
rindolf Or Cambridge, MA?
audreyt: yes, there's a discussion about it in the Boston.pm mailing list. 22:36
audreyt I don't know where I've gave the talk
rindolf Whose archive is only for subscribers... <sigh />.
audreyt: Oh.
audreyt MIT campus, I think.
rindolf audreyt: you didn't drive?
audreyt not sure
no, obra drove me there
rindolf audreyt: OK.
audreyt pays very little attention to Reality in general
obra MIT Campus. E51 22:37
rindolf obra: does MIT has other buildings spread in Cambridge? 22:39
obra: also, does Harvard teach Engineering?
A distant relative (from marriage) is from Brookline, MA originally. He started studying in Harvard. First Law, then Philosophy. He quit with one semester left. 22:40
Or so I heard.
He also sold his startup to an Internet sports news company. 22:41
www.couchsurfing.com/founders-dan.html 22:43
lambdabot Title: CouchSurfing - Daniel M. Hoffer
rindolf Seems like he moved to Silicon Valley. 22:44
No, wait - www.couchsurfing.com/profile.html?id=15AU he's still in Boston. 22:46
lambdabot Title: CouchSurfing - HOFFER
22:46 fglock left 22:47 rashakil joined 22:48 putter joined 22:52 DaGo joined 22:56 putter left
nothingmuch couchsurfing is back? 23:08
wolverian yes 23:12
couchsurfing++ # great for summer festivals
vsoni seen audreyt 23:43
jabbot vsoni: audreyt was seen 1 hours 7 minutes 15 seconds ago
audreyt vsoni: yes? 23:46
vsoni audreyt do we use .pcc_sub directive in PIR
aureyt: for subs
23:46 diotalevi joined
audreyt no. 23:46
it's all .sub now
.pcc_sub + label 23:47
is pasm only iirc.
vsoni so .pcc_sub is deprecated
audreyt well, only in PIR
vsoni ok.
thanks audreyt 23:48
clkao zz
audreyt vsoni: np :) 23:50

Powered by Cro and the Raku® Programming Language.

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

Thank you!