Pugs 6.2.6 uploaded by cpan-upload.p6! | pugscode.org/ <Overview Journal Logs> | pugs.kwiki.org | paste at sial.org/pbot/perl6
Set by integral on 2 June 2005.
Limbic_Region in any account - she isn't finished mowing the lawn yet - so it looks like I get to keep "playing" 00:00
and yes - she is the one that does the yard work, takes out the trash, etc (100% maintenance of the house) 00:01
00:01 batdog is now known as batdog|gone
theorbtwo Yeah, I'm horrible. 00:02
Jess works /and/ maintains the place. 00:03
(Mostly.)
Limbic_Region I suspect that we will have hords of lamdaheads after hw2005 00:06
I don't understand any of the haskell code, but I can fully comprehend the impressiveness in the text that follows 00:07
geoffb It's mildly amazing that he has said so much, and yet has so much left to write 00:09
Limbic_Region well - he is in a bit of a catch-22
the speed of pugs development necessitates rewriting while writing 00:10
geoffb Any xchat users here? 00:11
kelan irssi for life!
crysflame has defeated irssi for the first time 00:12
kelan beat it into submission?
crysflame persuaded it to do as i request
it's /save didn't put chatnet = ""; on the servers
theorbtwo chatnet? 00:13
crysflame code word for "the object that has my network-specific nickname"
theorbtwo Any lambdaheads around? 00:16
instance Value [Word8] where fromval = ... is kicking my ass.
Limbic_Region geoffb - in reading it, it looks like the places that need to still be written are there as stubs
theorbtwo I've got all the others converted to fromVal from doCast, but I can't figure out [Word8], mostly because I can't figure out how to get the individual bits out.
geoffb Limbic_Region, nodnod . . . that's how I knew how much was left to write (approx.) 00:17
ninereasons geoffb, I'm using xchat, but I don't really know what I'm doing :-)
geoffb ninereasons, heh 00:18
theorbtwo is using irssi, but doesn't really know what he's doing either. 00:19
kelan yeah, i use irssi but i mostly just type in stuff and it sends them 00:21
geoffb kelan, I used irssi once or twice, and that's about the level I was at. discoverable--
kelan i don't have anything complicated i want it to do, so it works out 00:22
gaim can do irc. i actually liked the interface quite a bit, but i *hated* not being able to just hit the up arrow to get the last line i'd typed 00:23
geoffb really hates undiscoverable interfaces with poor docs (or at least they were at the time)
theorbtwo kelan: control-up.
kelan oh
theorbtwo discoverable++ discoverable++
kelan but!
theorbtwo Yeah, undocumented interfaces suck. 00:24
kelan when you tab-complete a nick, it doesn't put a colon at the end
theorbtwo I discovered that by accident.
I prefer it not putting anything after it, because sometimes I want a comma, sometimes a colon, sometimes a space, sometimes a period...
kelan i mostly just want a colon if the nick is the first word or a space if not, which is what irssi does 00:25
theorbtwo Hm, so it does.
I mostly want a comma.
If I'm going to use the direct address, I'm going to use it correctly. ;)
kelan i think you can change it. but don't ask me how:)
awwaiid use the source, luke
er 00:26
kelan or ask in #debian!
awwaiid hah
kelan thats what i usually do for linux questions
i do actually use debian, so i feel justified
geoffb kelan, asking in #debian is a true roll of the dice. 00:27
theorbtwo awwaiid!
#debian has been pretty crazy the last few days, for rather obvious reasons. 00:28
geoffb I rarely get a good answer from there, but it may be just the type of questions I ask. The most useful responses I get are often "um, try this other channel"
kelan apparently the answer is /set completion_char
geoffb theorbtwo, I believe it!
kelan theorbtwo, test
it works!
theorbtwo Hmmpf, setting it to "" just results in... odd, it changes it per-network... anyway, in theorbtwo"", which was not the effect I was going for. 00:31
geoffb Why does pugs in interactive shell mode fail on things that work when it is run on a file instead?
kelan geoffb: its more fun that way 00:32
geoffb I know that interactive mode not only runs the code, but also displays the return value -- that's not what I'm talking about. I'm talking about erroring on code that runs correctly in non-interactive mode 00:33
geoffb rolls his eyes at kelan
theorbtwo geoffb"", can I get an example.
geoffb Something deep must be different between them. 00:34
kelan theorbtwo: you can set it like this: `/set completion_char ` (note two spaces after) and it will set it to space
but then you get this `nick ` on completion
geoffb theorbtwo, give me a sec to minimize the test case
theorbtwo kelan, I don't want a space, generally... ah, well, /set completion_char , is closer to what I want, at least.
Limbic_Region ok - lawn has been mowed 00:35
guess it is time for me to go
TTFN all
geoffb theorbtwo:
#!/usr/bin/pugs
use File::Basename--perl5;
our &basename := File::Basename.can('basename');
say basename('/this/is/a.test');
theorbtwo geoffb, I can't run that example at all, as perl5 embedding is broken for me. 00:36
geoffb It prints a.test when run as a file; pasted into pugs interactive mode, it errors, twice
theorbtwo Blasted include files.
geoffb theorbtwo, BLEAH
theorbtwo :r 00:37
Whoops.
geoffb Anyone else have p5 embedding working?
theorbtwo I should get a feather account. 00:38
geoffb theorbtwo, definitely
Juerd, is feather's pugs built with embedding?
right, middle of the night in Europe, forgot 00:40
theorbtwo 2:50 AM. 00:47
I should probably get to bed soon.
My work is so close to the point where I can convivebly check it in without breaking the entire world, though.
geoffb Grrrr, Alt-4 != Alt-F4 00:48
theorbtwo
.oO(Windows users.)
00:49
geoffb theorbtwo, what are you working on?
theorbtwo, nope. Gnome defaults.
theorbtwo Getting rid of doCast in favor of fromVal.
geoffb theorbtwo, . . . which would gain us locality for cast failures? 00:50
theorbtwo Yep. 00:53
geoffb THANK YOU
00:59 fibonaci is now known as luqui
meppl gute nacht - good night 00:59
theorbtwo G'night, meppl.
You're up quite late. 01:00
meppl ;)
theorbtwo (Assuming you are in germany.)
Allo, Luke.
luqui hello
meppl i am
*zZzZ
macGuy what is the proper way to end a subroutine in a foreach loop 01:16
luqui what do you mean? 01:17
vcv- return; ?
macGuy ok
im in a foreach loop and i dont want it to continue if ive already done all that i need to do
so i want it to just break out
luqui yeah, you can return out of the sub
or you can 'last' out of the loop 01:18
macGuy thanks
im going to use last
theorbtw1 last exits the for loop, return exits the subroutine. 01:24
If the for loop is the last thing in the sub, then they are equiv, but in the general case, they mean quite different things.
geoffb macGuy, 'next if condition;' and 'last if condition;' are very valuable idioms . . . worth using often, because they are so strongly self-documenting 01:33
hmmm, I wonder if pugs has working labels . . . ?
?eval my @foo = <1 2 3>; my @bar = <a b c>; my $foo = '4c'; OUTER: for @foo -> $outer { for @bar -> $inner { $foo = $outer ~ $inner; last OUTER if $outer eq '2' } } $foo; 01:36
evalbot6 Error: unexpected "O" expecting ";", statements or end of input reserved word or Only one invocant allowed
Chewie[] !seen Aankhen``
geoffb seen Aankhen``
jabbot geoffb: Aankhen`` was seen 5 hours 2 minutes 30 seconds ago 01:37
geoffb Hmmm, judging by the evalbot6 output, I'd guess not . . .
01:38 theorbtw1 is now known as theorbtwo
theorbtwo I think it has labels, but only for goto. 01:42
?eval my $x; goto FOO; $x='a'; FOO: $x~='b'; $x; 01:43
evalbot6 Error: unexpected "F" expecting ";", statements or end of input Only one invocant allowed
theorbtwo ?eval my $x; goto 'FOO'; $x='a'; FOO: $x~='b'; $x;
evalbot6 Error: unexpected "F" expecting ";", statements or end of input Only one invocant allowed
theorbtwo Or not.
QtPlatypus I know that goto isn't implemented. And I expect it to be very low on piroties. 01:45
kelan ?eval chars( "Ā»" ); 01:46
evalbot6 *** unexpected ( or "(" expecting term postfix, operator, postfix conditional, postfix loop, postfix iteration, ";" or end of input at -e line 4, column 4
theorbtwo ?eval "Ā»".chars 01:48
evalbot6 *** unexpected ( or "(" expecting term postfix, operator, postfix conditional, postfix loop, postfix iteration, ";" or end of input at -e line 4, column 4
theorbtwo I suspect somewhere in the p6->p5->p6 cycle, there are encoding issues. 01:49
kelan why is it going through p5? 01:50
theorbtwo I'm not quite clear on that myself. 01:51
Something about getting the safe mode, and maintaining persistance at the same time.
02:43 batdog|gone is now known as batdog
pasteling "qtplatypus" at 202.7.69.25 pasted "This makes makes my copy of pugs die with a bus error. What about you?" (11 lines, 223B) at sial.org/pbot/10867 03:20
QtPlatypus Can someone look at that code and see if it causes your copy of pugs to die. 03:23
Hi tall_man
03:33 batdog is now known as batdog|gone 03:45 batdog|gone is now known as batdog
QtPlatypus batdog: Can you test my code and see if it segfaults your pugs? 03:58
Khisanth QtPlatypus: do you have the latest pugs? 04:32
QtPlatypus: hmm that would require parrot wouldn't it?
svnbot6 r4466, putter++ | Created docs/other/rules_bootstrap 04:38
QtPlatypus Khisanth: Yes I have. 04:46
svnbot6 r4467, putter++ | hw2005.txt: 6 spelling corrections. (not UK-isms - doublechecked with google site:uk word :). 05:12
masak question: in Apocalypse_Now.spork, the word "DWIMer" is used. Isn't "DWIMmier" more correct? Or is "DWIMer" accepted usage? 05:28
both are kinda awkward :/
I guess the problem I have with "DWIMer" is that it sounds more like a noun to me 05:32
mugwump DWIMier is what I'd use
masak no double "m"?
mugwump nah
it's an acronym
masak true
should I change it, or should I leave it alone? I don't want to ruin someone's talk :)
mugwump you could hyphenate it 05:33
masak "DWIM-ier"?
mugwump sure
masak yuck :)
luqui follow english inflection rules. DWIMmier is the most correct 05:34
if you want to emphasize its double role as a verb and an acronym
masak I have vague memories of people using it that way
luqui I think I do
masak including @Larry
mugwump groups.google.co.uk/group/perl.perl...this+group
sure, "dwimmier" is fine 05:35
masak so, should I change it or leave it alone?
I always become a bit nervous with all these svn powers granted to me :)
luqui svn doesn't allow users to destroy data 05:36
masak I know 05:37
SamB perhaps this is the intent of the generous svn access?
so that people will be more carefull in their hacking?
masak but will someone become happier if I change it, or will I piss someone off?
SamB you may annoy someone, but probably not "piss off".
mugwump I'm sure no-one will care, as long as the changelog entry is amusing enough
luqui there's always svn update -r 1337 if you piss someone off
because the one who gets pissed off has svn powers too :-) 05:38
masak SamB: svn is like hacking where everyone can see you
ok, I'll change it :)
SamB I was accused pretty nicely of breaking a recent release (though it turned out it was broken before I touched it)
luqui and our project leader seems to be appreciative of the fact that we're doing work at all 05:39
even if half of it is wrong
masak luqui: yes, it's the thought that counts 05:41
luqui well, and the fact that it takes a fraction of the time to correct as to write in the first place 05:43
masak unrelated question: if i have a cvs repo, can I use either of svn or svk on it?
svnbot6 r4468, masak++ | After some deliberation, I have dared to change DWIMer into DWIMmier.
r4468, masak++ | I do believe this is the more correct form. I also believe I am a bit
r4468, masak++ | too hung up on grammar sometimes. ;)
masak one changed line, three lines of comment :)
luqui masak, you can use svk 05:44
masak neat
how?
luqui and you can use cvs2svn to convert it to a svn repo
masak hmm... it's not my repo 05:45
luqui svk mirror cvs::pserver:....
masak luqui: oki, thx
luqui I just hashed through some problems with that, so you can talk to me if you encounter them
masak my first problem will be to install svk :) 05:46
luqui ahhh
masak after using svn for some time w/ pugs, i really don't like cvs
luqui yep. after finding out about svn and reading the user manual, it took me about 10 seconds to decide to convert everything I had to svn 05:47
masak lol
luqui svn is the one of the most beautifully designed piece of software I have ever seen, internally and externally
masak if one didn't know about the order of things, one would wonder why cvs was ever created 05:48
what are the main differences between svn and svk?
luqui afaict, svk is a nice mirroring frontend to svn
(and various other rcs's)
masak a nice mirroring frontend? whazzat? 05:49
luqui it allows you to work on svn repositories while you are offline, and sync up when you come back online
plus, it fancifies working with branches even more
for branches, svk:svn::svn:cvs 05:50
but I think svk uses svn on its backside
you can ask more at #svk
I'm just getting started with it
masak sounds pleasant, in any case
luqui the only trouble is installation
masak oh :) 05:51
luqui I couldn't install it until I got gentoo, which had a nice ebuild for it
masak right now I
am svk.elixus.org/?InstallingSVK
I have FC3
luqui maybe with the release of 1.0 installation has been made easier
masak yum is happily working right now with subversion-perl
luqui yum actually had that 05:52
?
wow
masak apparently
luqui that was the main difference between FC2 and gentoo to me. FC2 had about 1/3 of the packages I wanted, gentoo had 98%
is there an "FC3 unstable" or something like that 05:53
that has more packages, but which haven't been tested as much?
masak dunno, I may have added a few repos to my list 05:54
make is chugging away happily now 05:55
this is so exciting!
luqui is currently building kdelibs
and has been for the past 30 minutes
masak ah 05:56
luqui but I suppose it's not as bad as ghc
masak perlbot nopaste
perlbot Paste your code here and #<channel> will be able to view it: sial.org/pbot/<channel>
luqui ?eval say "moo";
evalbot6 Error: No compatible subroutine found: "&say"
luqui ?eval "moo";
evalbot6 'moo'
luqui ?eval "moo".chars;
evalbot6 3
masak I just got a test failure and a question "is this reasonable? [Yn]" 05:57
hold on
luqui ?eval "foo" ~~ /foo/ 05:58
evalbot6 Parrot VM: PANIC: Out of mem! C file src/memory.c, line 45 Parrot file (not available), line (not available) We highly suggest you notify the Parrot team if you have not been working on Parrot. Use parrotbug (located in parrot's root directory) or send an e-mail to [email@hidden.address] Include the entire text of this error message and the text of the script that generated the error. If you've made any modifications to Parrot, please describe them
pasteling "masak" at 130.238.83.176 pasted "I got a test failure when trying to install svk" (10 lines, 338B) at sial.org/pbot/10870
luqui ack
masak luqui: you crasched evalbot! you bastard ;)
geoffb </lurk>masak, luqui: as I recall, svk uses *most* of SVN as its backend, but not the working copy library, which the SVK developers (clkao) don't like, and which is apparently a large part of the reason SVN is actually slower than SVK, despite the latter being written largely in Perl instead of C<lurk> 05:59
SamB how do you include the text of a script like that?
masak geoffb: ah, thx
SamB I mean, you would have to include EVERYTHING that got fed to evalbot6 up to that point, wouldn't you?
QtPlatypus luqui: regexes his briking in evalbot
SamB: I suspect it was I who killed evalbot :( 06:00
masak QtPlatypus: why do you suspect that?
luqui masak, the test failure looks reasonable, but go ask #svk 06:01
luqui & movie
masak luqui: I'll go with "yes"
Darren_Duncan fyi, r4465 gave these results with all the default settings and no parrot or perl 5: Failed 24/322 test scripts, 92.55% okay. 280/6143 subtests failed, 95.44% okay.
luqui yeah, it's not like it's going to go: whoops, a bug, system "rm -rf /"
QtPlatypus masak: I was testing if eval bot handled subs included in regexps. It was working normally before then. 06:02
masak w00t! I now have svk! :D 06:03
QtPlatypus: so what went wrong? 06:04
QtPlatypus It spat that error.
masak yes :) but why?
QtPlatypus I suspect that my code tickeled a bug in the parrot based regex parser. 06:05
Darren_Duncan question: I'm already running svn and it seems to work fine ... is there any reason for me to switch to svk? 06:08
obra do you ever work on multiple branches or work without a net connection? 06:09
do you work with Large source trees?
Of the answer to any of these questions is "yes", then svk may be an amazing win for you.
Also "do you hate .svn directories?"
luqui "do you like ketchup on fries?" 06:10
Darren_Duncan currently my answer to all 4 of those questions is 'no' 06:11
all 5 that is
that said, I may want to work on multiple branches in the future
obra Oh. do you ever want to keep local commits to repositories you don't have write access to?
Darren_Duncan I haven't needed to yet 06:12
obra ok
Darren_Duncan to be honest, Pugs is the only project I've used svn with
obra What do you use for other projects?
Darren_Duncan my own projects I don't use any formal version control with
I treat each CPAN release like a check-in in a way 06:13
and otherwise just make tarred copies of intermediate steps as something to revert to
but I'm also the sole author
I intend to use something like svn when I work on projects with multiple authors
this is just a historical thing, not necessarily a good reason; the main good reason is simplicity 06:14
my more immediate reason for the question was whether I should use svk with Pugs 06:15
but I don't yet consider that to be 'Large'
when my SQL-Routine/Rosetta project becomes good enough that multiple people *want* to author it, I'll probably move it into version control 06:16
speaking of SQL-Routine etc, I have an ambitious goal for it to be "done" by the end of June ... as in end-to-end useful in daily work 06:17
masak w00t! svk works! :) 06:18
Darren_Duncan this includes a new SQL-Routine-SQLParser and Rosetta-Emulator-DBI modules, with which any other Perl programs written to use DBI or one of its wrappers can "just work" with Rosetta etc, which means more seamless portability across databases 06:19
I'll return tomorrow, good night!
nothingmuch morning 06:20
masak morning 06:23
svnbot6 r4469, putter++ | hw2005.txt: minor grammar tweaks. mostly tenseand plurality agreement, missing commas, preposition selection, breaking ;-ed sentences. Nifty paper. 06:26
ingy hola 06:45
svnbot6 r4470, putter++ | rules_bootstrap: added "compiling rules into Parsec combinators" plan mentioned in hw2005.txt.
ingy nothingmuch: are you going to the hackathon? 06:46
nothingmuch ingy: yes
why do you ask?
ingy nothingmuch: do you know if there is broadband there?
nothingmuch got something interesting for me to try?
iirc there isn't
ingy hmmm 06:47
well that will certainly suck
nothingmuch it will
ingy grrr
nothingmuch but it will also make us seem super secret too
ingy but I already seem sooper sekrit :P 06:48
Aankhen`` curses the fact that reinstalling Windows means reconfiguring his entire development environment, which in turn means figuring out what he had done to his development environment in the first place. 06:50
ingy Aankhen``: sadness
nothingmuch_ sorry
ingy: you said "grr", then I got nothing 06:51
Aankhen`` Indeed.
But on the bright side, my system's now blazingly fast. :-D
ingy 23:48 < ingy> but I already seem sooper sekrit :P
Aankhen`` (for someone who played Doom 3 at 20 FPS at 640x480, Low Quality, 55 FPS at 1024x768, High Quality, 4xAA and 8xAF is quite fast enough =) 06:52
ingy fyi I said "grrr"
I feel grrr8
like Tony
nothingmuch_ maybe we can get a temp solution if we all chip in some $$$? 06:53
ingy well we'd better bring it up soon 06:54
Aankhen`` wonders whether HTTP::Message would be better as a role rather than a class. 07:05
Heh. I'm looking at the latest messages in the "reduce metaoperator on an empty list" thread, and Google popped up "4.1 Naming and binding" from www.python.org in the ads. 07:19
ingy sweeet
Aankhen`` Most of the time, though, messages on p6l and p6c bring with them ads for mod_perl... usually this: "Introduction to mod_perl (part 5): More Perl Basics". Freaky. 07:20
pdcawley Oh ghod. Not the reduce metaoperator thread. 07:23
I'm really hoping it'll've died out by the time I write the next summary.
lathos pdcawley: Incidentally, Buscador is up and feeding on p6*
buscador.simon-cozens.org/
pdcawley Buscador? 07:24
crysflame heh 07:25
er, ww
hi
lathos++ # buscador
nothingmuch_ lathos: what does buscador do differently? 07:27
autrijus_ greetings from APW! 07:31
QtPlatypus Greetings. 07:32
lathos Quite a lot. It keeps track of messages over multiple lists, it can arrange threads in several different ways, it keeps track of people from different email addresses, it gives you syndication feeds for everything...
It separates off attachments in a friendly way.
QtPlatypus prefs the nntp interface.
lathos buscador.simon-cozens.org/name/view/529 is an example. 07:33
ingy hi autrijus-san 07:35
autrijus_ greetings ingy-san. 07:36
ingy :)
svnbot6 r4471, autrijus++ | * ApocNow: Content's mostly there.
autrijus_ I learned about latex presentation (beamer) today.
it makes beautiful presentations.
ingy I'm totally rethinking Perldoc
autrijus_ however \latex{} is much more verbose than `Kwid`
so I'm going back to Kwid/Spork for my talk. :) 07:37
ingy ;)
pdcawley Is it easy to get at Opentype goodness in LaTeX yet?
autrijus_ no idea :)
ingy autrijus_: how do you feel about a hackathon without internet? 07:38
lathos Also buscador.simon-cozens.org/mail/lurk...looper.com
knewt_ autrijus_: committing again already eh :)
autrijus_ knewt_: :) 07:40
ingy: completely without? that'd be bad
hm. anyone have tuits to chase test failures about the undef->undefine change?
and maybe grep the t/ and examples/ a bit to change undef($x) to undefine($x) 07:41
hm. come to think about it, maybe delay this after the release 07:48
svnbot6 r4472, autrijus++ | * make http-server.p6 more readable
autrijus_ does a local branch for s/undef/undefined/
undefine, even
knewt_ ah, not too bad. only 16 commits since i synced up before flying yesterday. but then, autrijus_ /has/ been busy flying over here :) 07:50
svnbot6 r4473, autrijus++ | * hw2005 - review from Nicholas. 07:53
pdcawley Are the reviews collected anywhere? 07:54
07:56 Aankh|Clone is now known as Aankhen``
knewt_ autrijus_: do you have on a black t-shirt with a camel on it? 07:57
autrijus_ knewt_: yes! 08:01
svnbot6 r4474, autrijus++ | * further review comments from Allison.
r4473, autrijus++ | * hw2005 - review from Nicholas.
autrijus_ pdcawley: they are collected in my inbox... they are not public
but I mostly just applied their unidiff patches
08:01 integral_ is now known as integral
autrijus_ pdcawley: care to review the paper for me? :) 08:01
pdcawley wonders why feather's zsh will only use C-h to delete stuff. 08:02
autrijus_ the deadline for submission is tomorrow :-/
pdcawley autrijus: I've just skimmed it, I'll give it another read later.
autrijus_ pdcawley: danke!
knewt_: where are you?
nothingmuch_ morning 08:03
autrijus_ yo nothingmuch_
nothingmuch_ err, morning autrijus
08:03 nothingmuch_ is now known as nothingmuch
knewt_ autrijus_: orange shirt 1 back and 4 to the left 08:04
autrijus_ gotcha :) 08:06
nothingmuch autrijus: how's Vienna? 08:22
autrijus_ nothingmuch: excellent weather 08:26
despite the forecasts
svnbot6 r4475, autrijus++ | * correct getc's use of adverbial function parameters 08:29
autrijus_ sri_: you around? 08:44
svnbot6 r4476, scook0++ | * hw2005: 'et al.' fixes & some 80-char reflowing
r4477, autrijus++ | * hw2005 - more reviews from shapr.
Juerd geoffb: Yes 08:47
nothingmuch autrijus: did you get my parsec musings? 08:48
autrijus_ nothingmuch: no. where/
nothingmuch one sec
autrijus_ I'm out of battery... 0%
autrijus_ runs out to get eletricity
yay, just in time 08:49
pasteling "nothingmuch" at 212.143.92.226 pasted "using parsec to make pesto" (6 lines, 387B) at sial.org/pbot/10873
autrijus_ <- glad that I left the room before it starts to beep real loud
nothingmuch what room? 08:50
autrijus_ the apw conference room
the APW is underway
nothingmuch oh!
then don't let me bother you
autrijus_ nah, that's fine
parsec can handle non-Char types
that you know
so of course it can match trees 08:51
nothingmuch the problem is how to make it convert it's matching paradgim
so that it does backtrack,
and that many shapes of a grammer can match
but that some are better than others
autrijus_ well, you can have the matches return a weighted thing 08:52
and sort by weight
nothingmuch right
autrijus_ that won't be hard actually 08:53
just two lines or so
nothingmuch but i need to prune the match space very aggressively
autrijus_ yeah, so you need nondet
you can chase haskell.org/hawiki/NonDeterminism and links
it's a well known design
nothingmuch okay, i'll keep thinking about it
when harrorth is done I'll try to tackle that next
autrijus_ k 08:54
nothingmuch beh, fucking network 08:56
q[pdcawley] Mmm... nondeterministic programming... 09:00
clkao autrijus_: do you happen to have my svk slides? iirc you cped them for me to do a presentation somewhere 09:01
q[pdcawley] Shame that my call_cc implementation doesn't work or I'd have the bones of NonDet programming for Perl 6 implemented by now.
xerox AMB ? 09:02
autrijus clkao: can't find it easily 09:05
q[pdcawley]: uh?
q[pdcawley]: I can give you callcc for perl6 in pugs
nothingmuch is q[] a cult? 09:06
q[acme] nothingmuch: yes 09:38
nothingmuch what is the cult about? 10:14
q[acme] about non-globally-unique nicks 10:20
nothingmuch ooh, wow
i would join, but i'm not sure i want to 10:21
lathos I'm not sure you need to. :) 10:24
10:27 q[acme] is now known as acme
theorbtwo is glad to say that he's never come across another 'theorbtwo' spelt as such. 10:28
clkao q[acme]: the other acme was slashdotted yesterday 10:29
acme i noticed - i get a lot of spam too! 10:30
pjcj I suspect that's a mistake and the spammers meant to send it to the other chap 10:31
you should let them know
autrijus yay, finally got contents done 10:55
svnbot6 r4478, autrijus++ | * Content and formatting foreeze for apocnow
autrijus 30 minutes left
time to google image
xerox 30 mins left to.. ? 10:56
autrijus the Perl 6 track of Austrian Perl Workshop
10:58 Aankh|Clone is now known as Aankhen``
xerox Any live audio/video feed? :) 11:03
q[pdcawley] xerox: Just watch the SVN logs... 11:08
autrijus heh :) 11:12
no, I don't think there's one
but it's broadcasted to the FPW
so it should be recorded somewhere
autrijus_ hey chip-san 11:40
Aankhen`` Woo hoo! No more slow down while compiling Pugs.AST.Internals! 11:47
I wub my new processor. :-D
autrijus_ new processor++ 11:49
how good is it?
Aankhen`` Works great so far. =)
Juerd How bad was your previous? :)
Aankhen`` On my old processor, I could barely do anything else while compiling Pugs, especially Pugs.AST.Internals. 11:50
Everything would freeze up.
Now, there's absolutely no slowdown -- everything is as responsive as ever.
HyperThreading++
q[pdcawley] What's your new processor? 11:51
Aankhen`` Remind me to measure the time...
Intel Pentium 4 630.
(3.0 GHz., HT, 32/64-bit ready, 2 MB cache)
q[pdcawley] Shiny! 11:52
Aankhen`` Literally -- I got a new case too. =)
autrijus_ very shiny!
Aankhen`` And the guy threw in a nice 430W CoolerMaster power supply for no extra charge. 11:53
It's fun to watch the test scripts being run... I can't even see the individual tests, it just flashes a number and says "ok", then moves on to the next one. :-D 11:54
I can't wait to see what happens on the rules tests, though. :-P
autrijus_ :) 11:55
Aankhen`` goes to see what happens when he starts Doom 3.
A nice and steady 45 FPS. =) 11:58
The 1 GB of RAM probably helped, come to think of it. :-P 11:59
Ah, if only I had five bucks for every smiley I typed...
autrijus_ I hope I can have access to that as a laptop :) 12:05
Juerd 9 different users are logged in on feather, some run irssi, some use editors - the load is 0.00 :) 12:10
Limbic_Region hey autrijus_ - how is your net access ATM? 12:11
webmind Juerd, no one is using pugs? :) 12:19
Juerd webmind: Not many
webmind: And not much
autrijus_ Limbic_Region: passable 12:20
webmind Juerd, well.. I guess most of the coding is typing anyway 12:22
svnbot6 r4479, autrijus++ | * images part done for apocnow! 12:36
autrijus_ whew. 12:38
pugscode.org/
click on "APW slides" from left
Juerd autrijus_: slide12c 12:40
autrijus_: s:0/@h/%h/
integral hmm, what mac browser has a full screen mode? Opera? 12:41
autrijus right.
good catch
Limbic_Region autrijus - in that case, if you get some free time could you take a look at perlmonks.org/index.pl?node_id=464502 and perlmonks.org/index.pl?node_id=366280
autrijus I'm going to speak in 3 mins
Limbic_Region tall_man (a pretty smart guy) thinks he has found a bug
Limbic_Region just can't make heads or tails of it 12:42
oh - well have fun ;-)
Aankhen`` integral >> Opera does have a full screen mode, assuming it's available for the Mac. :-)
Limbic_Region break a leg as they say
integral Aankhen``: pity it's not very macish ;-)
Aankhen`` shrugs.
Luckily I'm not inflicted with Mac OS-itis. =) 12:43
Apple does return the favour though... have you seen iTunes on Windows?
integral hasn't
Aankhen`` It's bloated, slow, ugly, and badly behaved.
integral heh
Juerd autrijus_: slide20b - I thought there were no parens around the ;; part of a loop loop?
Aankhen`` I'm sure you could say the same about Opera on the Mac. ;-) 12:44
lumi Opera has fullscreen on mac 12:46
Aankhen`` wanders off. 12:47
Juerd autrijus_: slide30c 12:49
autrijus_: ('a'..'c') ōæ½xBBxōæ½xAB 3
autrijus_: Shouldn't that be just >>x, as there is no array on the RHS?
integral lumi++; thanks! 12:57
Limbic_Region oh - I think I finally get the bug tall_man was trying to point out 13:13
I have seen it myself but thought it was that way by design
if you have a named parameter but invoke the sub as though it were positional - it shouldn't automatically morph for you should it? 13:14
I assumed the answer was yes - but apparently tall_man doesn't think so
for instance
?eval sub bar ( $foo) { $foo } bar(3); 13:15
evalbot6 \3
Limbic_Region tall_man is saying that for that to do the right thing, I should have had to call bar as bar( foo => 3 )
osfameron why \3 rather than 3 ?
Limbic_Region ?eval sub bar ( $foo) { $foo } bar(foo => 3);
evalbot6 \3
Limbic_Region that part I don't know osfameron, but it doesn't have anything to do with the bug 13:16
osfameron true, sorry
Limbic_Region grrr
my plus symbol is broke
would you mind trying those two for me again 13:17
only in the sub signature make it plus$foo
osfameron ?eval sub bar (+$foo) { $foo } bar(foo => 3);
evalbot6 \3
Limbic_Region and now the other way? where bar is just called with bar($foo) 13:18
osfameron ?eval sub bar (+$foo) { $foo } bar(3);
evalbot6 \3
Limbic_Region yeah
that probably shouldn't work
osfameron why not?
Limbic_Region I thought it was by design but apparently not
$foo is a named parameter 13:19
it is magically becoming a positional
osfameron I thought that was designed
it's to prevent you from having to do
my ($foo, $bar, $baz) = @_;
Limbic_Region I am not sure you are understanding 13:20
in the sub signature - you can have, required, optional, named, positional, etc type variables
using named variables (which are optional unless marked as required) 13:21
you can place the variables in any order
so
sub bar ( plus$foo, plus$bar) { ... }
could be called as bar( bar => 3, foo => 2)
or bar( foo => 3, bar => 2 );
order doesn't matter 13:22
without the plus symbol - they are positional (meaning they have to come in order specified in signature)
what is happening in the bug
is a named parameter is becoming positional
osfameron ah
Limbic_Region if you say it is named then you really need to name it foo => 3 13:23
you can't just choose to leave off the name (3)
but pugs isn't b1tching
osfameron - do you have commit rights?
osfameron Limbic_Region: nah, I mainly lurk 13:24
Limbic_Region perlbot nopaste
perlbot Paste your code here and #<channel> will be able to view it: sial.org/pbot/<channel>
Limbic_Region ok - well I am going to come up with a bugtest and hopefully someone can commit for me - otherwise I will tonight
QtPlatypus Has anyone tested my buggy code?
pasteling "Limbic_Region" at 129.33.119.12 pasted "Can someone with round tuits please commit this test as t/pugsbugs/named_2_positional.t after performing a sanity check" (14 lines, 384B) at sial.org/pbot/10879 13:30
Limbic_Region is off to a meeting now
mj41 perlbot: nopaste ? 13:47
perlbot Paste your code here and #<channel> will be able to view it: sial.org/pbot/<channel>
wolverian Limbic_Region: I'll take a shot at it.
pasteling "mj41" at 147.229.221.107 pasted "1 2 or 2 2 ?" (16 lines, 134B) at sial.org/pbot/10881 13:48
wolverian heh. Test.pm declares the description arguments as +$desc 13:52
so that shouldn't work either as a positional
but just about _all_ the tests use it as a positional
QtPlatypus Can some people try out the code in sial.org/pbot/10867 and see if it causes there pugs to segfault. 13:53
svnbot6 r4480, wolverian++ | Named args as positionals test from Limbic_Region++ 14:03
autark-jp QtPlatypus: did not segfault here. 14:05
QtPlatypus Odd. What version are you using? 14:06
autark-jp #4470
QtPlatypus Me has r4459, I'll see if upgrading helps. 14:09
14:20 Odin-FOO is now known as Odin-
svnbot6 r4481, rootmj++ | + sub reset_params returns Void is export { %PARAMS = (); $IS_PARAMS_LOADED = 0; } 14:28
mj41 Src (pugs is 80% perl 6, so many ugly hacks, but you can change htdocs/* without server restart )wiki.kn.vutbr.cz/mj/attach/pugs/http-server/ 14:44
Slow online demo a05-0612a.kn.vutbr.cz:8080/page.p6
domm sorry, quick question 14:47
osfameron ?eval sub avg { [+] @_ / @_.elems } avg(3,5) 15:25
evalbot6 1.0
osfameron !
?eval sub avg { ([+] @_) / @_.elems } avg(3,5)
evalbot6 4
osfameron phew
svnbot6 r4482, iblech++ | * docs/other/rules_bootstrap, t/pugsbugs/named_2_positional.t -- Usual svn props 15:30
r4482, iblech++ | * docs/talks/README: Added autrijus++'s "Apocalypse Now" talk.
r4483, iblech++ | docs/talks/README -- Oops, fix typo.
Juerd Hmm 15:34
How about [+: @_]
:)
xerox :+]
Juerd No
Must be prefix.
xerox I was just smileying :(
Khisanth ?eval [+: 1..10] 15:35
evalbot6 Error: unexpected "[" expecting program
osfameron ?eval [+: 1,2,3]
evalbot6 Error: unexpected "[" expecting program
Khisanth that doesn't work!
Juerd Duh
I just suggested it. Things don't get implemented within seconds from that.
Khisanth but what advantage does it have?
osfameron it's pretty
and you don't need more disambiguating parens 15:36
Juerd It makes it a non-listop
Khisanth thought you meant that was an alternative :)
Juerd (precedence wise)
Khisanth [+| @_]? :P
Juerd And I think it's clearer
No |.
| is used in infix operators that you'd reduce.
: is not, afaik.
Khisanth larry was right, everyone wants the colon
Juerd Can we keep this all a technical discussion? 15:37
It's not about wanting the colon.
It's about lack of alternatives, where a colon would work and even look nice.
Khisanth actually I do agree with all the stuff you have said
integral what if I, as a haskell user, like to name my custom infix operators things like :+: ?
Juerd integral: You'd get very pissed. 15:38
Khisanth integral: back into the hell you were in with source filters?
integral bah! :-)
Juerd =:= is a problem too.
Khisanth integral: but if that is the case you would run into that problem ... oops you can redefine the [:] as well can't you? :P 15:39
integral Juerd: is a delimiter other than whitespace needed?
Khisanth of course you would end up with your own custom languages nobody can read
Juerd integral: Yes, because [] creates arrays too.
integral err, and doesn't this have problems with [] creating arrays?
Juerd It has to be currently invalid syntax.
integral I thought the only reason [] could be used for reduce was that *only* an operator could be in it, so it's very unambiguous? 15:40
Juerd integral: There's absolutely no difference between [op] and [op:, parse wise.
Longest match wins.
That's why ~ doesn't clash with ~~.
integral hmm 15:41
osfameron but it's then visually ambiguous because it looks like it contains an array
Khisanth osfameron: doesn't seem like a problem if you are used to reading from left to right
osfameron maybe not, but if you see something ending ...1,2,3], it could be confusing. Anyway, no longer a convenient "pill" 15:43
Juerd All true.
svnbot6 r4484, iblech++ | undef(...) is now undefine(...) (see 15:51
r4484, iblech++ | groups.google.com/groups?threadm=20...wall.org).
r4485, iblech++ | ext/Test-Builder/t -- unTODO four succeeding tests.
QtPlatypus I beleave there is a metaoperator or metatinfix thing like postfix and infix. 16:00
svnbot6 r4486, iblech++ | t/rules/from_perl6_rules/charset.t -- Make it compile again.
r4487, iblech++ | Make while.t parse again, too.
r4488, iblech++ | ext/CGI/t/util.t -- Make it parse again, too. Now all tests should at least 16:07
r4488, iblech++ | parse again. :)
xerox One gets multiple `++'s writing long comments, yay ;) 16:08
PerlJam xerox: and that's why bot karma is relatively useless. 16:09
corneli hi! im new around here - is there a doc somewhere to check what features of perl6 are available in pugs? i want to help! 16:12
PerlJam corneli: Start at www.pugscode.org/ and go from there. 16:16
corneli: checkout a copy of pugs and start playing.
corneli PerlJam: i have a copy of pugs... just don't know what I can play with :) 16:17
PerlJam corneli: Well, what do you want to play with?
do you know haskell?
do you know perl5?
corneli nowhere enough haskell to help with the projects (never got to monads) - i do know perl5 16:18
PerlJam have you read the AES?
corneli not too closely - only picked a few features from here and there. 16:19
Limbic_Region wolverian plusplus - thanks for adding the test for me 16:21
PerlJam corneli: you might want to start by looking at the tests and try adding some of your own.
corneli: or maybe translate some perl5 to perl6
corneli: or reading the AES and to find things that are specced but not yet in pugs and not yet tested for and write a test for them. 16:22
corneli Great - Are the AES the formal sources for what should be working (even if it's not implemented now)? i had the feeling that pugs introduced at least some differences? 16:24
macGuy hello 16:25
if i have a foreach statement like foreach $_ (@_) and sometimes in the iteration i want to know which index i am looking at in @_ how can i tell 16:26
QtPlatypus macGuy: Are you asking about perl6 or perl5? 16:27
macGuy would there be any difference?
im working on a project that would be executed on perl 5 16:28
QtPlatypus Then you realy should ask in a perl channel. perl6 is very diffrent to perl5
macGuy k
wolverian Limbic_Region: you're welcome 16:52
Limbic_Region: did you notice that most of the .t files depend on the wrong behaviour :)
PerlJam corneli: The information path is like this (from least definitive to most definitive) Apocalypse, Exegesis, Synopsis, perl6-language.
corneli Great. THanks 16:54
PerlJam And while that last one has the most up-to-date information, you often have to wade through the random ideas of other non-@Larry people to get it. 16:55
corneli Looks like it's a very practical development effort 16:57
PerlJam corneli: Areas that are currently in-the-works are better OO support and p6rules. 17:00
rjbs Stupid question. Can one hyper a bare block?
Limbic_Region wolverian - yes 17:01
PerlJam rjbs: and what would that mean exactly?
rjbs @a >>{ $^a ~ " is with " ~ $^b }<< @b
Limbic_Region wolverian - I think that was tall_man's point
wolverian Limbic_Region: who?
PerlJam rjbs: that looks like zip
Limbic_Region perlmonks.org/index.pl?node=tall_man
rjbs I thought zip would just produce the zipped list? 17:02
I understand I could map {...} zip
wolverian Limbic_Region: yeah.
I guess one could s/// the files mostly to fix it 17:03
Limbic_Region I am going to search for the portion in AES6 that says it is bad
perhaps add it to a comment
wolverian well, it can't really be fixed until the test cases are fixed
or the whole test system goes kaboom 17:04
Limbic_Region are there really that many tests using named parameters as positional - or am I missing something? 17:08
wolverian maybe not - I only looked in pugsbugs
hmm. builtins/arrays/ is full of it too. 17:09
Juerd Who are at the APW? How are things there?
wolverian Limbic_Region: $desc in ok() is named
and about all the tests I see use it as positional
Limbic_Region Juerd - see use.perl.org/~cog/journal/25113 17:10
apparently there is an IRC channel to see what's going on
wolverian a simple fix is to simply convert +$desc to $desc, as just about no-one uses the named interface
Juerd I see 17:11
Thanks
Limbic_Region walang anuman
wolverian - not sure that's such a good idea either
wolverian Limbic_Region: probably not, but the other option is the mass s///. do you have tuits for it? :)
Limbic_Region probably should confirm that it needs to be fixed, fix it, and see what break
wolverian of course it needs to be fixed. it's wrong behaviour.
(as said by Larry) 17:12
Limbic_Region wolverian - ok - found supporting doc 17:13
Larry didn't say that it needed to be fixed
he said that tall_man's original test was flawed
in fact there was a real problem, just not the one tall_man thought 17:14
he (tall_man) later /msg'd me with what he thought the real problem was
and he is right
from S06 "They can only be passed by name"
so if you feel like updating the test with a comment from S06 - be my guest
"Arguments that correspond to named parameters are evaluated in scalar context. They can only be passed by name, ..." 17:15
it is a bummer that many people incorrectly used named parameters then 17:17
Juerd That makes sense, doesn't it? 17:19
All non-slurpy arguments are evaluated in scalar context.
And all slurpy arguments are evaluated in list context.
Named arguments are never slurpy.
It also makes sense in that the value of a pair is always scalar 17:20
Limbic_Region Juerd - it makes perfect sense - Pugs is broke though 17:21
what's worse - is people have written numerous tests (according to wolverian) using the incorrect behavior
so fixing the bug will result in many tests breaking
Juerd Interesting 17:22
What were people thinking, I wonder in such cases.
I can understand that there is a bug
Limbic_Region probably copy/paste someone else's test
Juerd But why were these tests written
Limbic_Region that test had it wrong and people just changed variable names leaving in the plus symbol not knowing what it was for
Juerd Or were they testing existing behaviour? That's always dangerous - to test after the fact.
Limbic_Region that's just a guess btw - no idea if it represents reality
Juerd Or is perhaps context not clear enough? 17:23
Limbic_Region the scalar context isn't the part that I am hung up on
basically if you create a signature with a named parameter - you can call it without naming it
Juerd Oh
Limbic_Region Pugs magically turns it into a positional parameter 17:24
Juerd Aha
That is a bug then yes
Limbic_Region I would demonstrate but my client eats plus symbols
Juerd aww
Have you tried '+', "+" and \+?
Or can you not even enter them?
Limbic_Region ' ', " ", \
it eats them
Juerd I think it's a proxy that does this 17:25
Limbic_Region cgi irc in case you were wondering
Juerd Not CGI:IRC
SamB really?
Juerd For fun, try %25 and see if you get %25 or % out of it
Limbic_Region SamB - really to what?
%25
Juerd Weird.
SamB maybe the CGI is a broken?
Limbic_Region Juerd - I don't think it has anything to do with a proxy since it works fine everywhere else
Juerd I doubt that the CGI is broken. CGI:IRC is very popular and the bug would have been noticed and fixed long 17:26
It could be an old CGI:IRC.
Limbic_Region possible
I am reading the FAQ now 17:27
Juerd Can you visit arbitrary URLs? 17:28
I've installed cgiirc on feather, to see what it'd do
Limbic_Region yes
Juerd Let's see how this thing is configured first
juerd_cgi hello! 17:31
plus:
Juerd Hm
Stupid thing!
BUT - I can debug this one :)
Limbic_Region I should debug too but I only need it at work (where I can't ssh to the machine it is residing on) 17:32
at home, I don't think to do that since I can use a real client
17:34 Aankh|Clone is now known as Aankhen``
Limbic_Region Juerd - once you get it fixed I will just use feather if you don't mind 17:34
Juerd ## The clever regexp bit is from cgi-lib.pl
Oh god.
They took something "clever" from cgi-lib.pl
No wonder it's horribly broken.
Not tomention the indenting. 17:35
Juerd removes the s/\+/ /g line.
How the HELL does this *work*, anyway?
map { s/\+/ /g; ... } split(...);
Tell me - how does that s/// mutate a read only $_? 17:36
Sure, this worked in perl 4 and early 5s
But how can this work with current perl?
Obviously it works.
juerd_cgi Hi again 17:37
plus: +
hurrah.
Limbic_Region yeah
what's the url?
oh - please please please tell me it is running on port 80? 17:38
juerd_cgi It is
But I want a way to avoid that this is used by anyone 17:39
While providing all users with access
So I'm going to ssl-ify this first
Which I've done only one time before
Can take a while.
Limbic_Region doesn't have a feather account
you doing a self-signed certificate?
juerd_cgi No, CA-cert 17:40
svnbot6 r4489, iblech++ | t/ -- s/undef/undefine/ as appropriate
juerd_cgi And that feather account will be arranged if you send me an email requesting it
Limbic_Region right - but your CA-cert, is that going to be self-signed? 17:41
juerd_cgi It's going to be signed by cacert, I hope :)
Limbic_Region ok - but I won't actually need to log in right? Since I couldn't do that from work anyway
juerd_cgi Otherwise I don't see the point.
You would need to log in via ssl, http-auth
You can visit gmail, so you can visit https
Limbic_Region oh - didn't realize cacert was an organization - since ca = certificate authority
I normally self-sign my ca and then sign what is needed since I don't do business with anyone that doesn't trust me 17:42
let me know when/if you get it set up and if you need anything from me (besides a acct request which I will send momentarily)
svnbot6 r4490, iblech++ | t/ -- More s/undef/undefine/ (this time I used a grep to find undef(...)s 17:46
r4490, iblech++ | instead of relying on all-skipped tests as shown by the smokes).
r4491, iblech++ | ext/ -- s/undef/undefine/, too
r4491, iblech++ | util/perl6.vim -- Highlight undefine.
geoffb morning all 17:47
geoffb mutters venomously about video drivers crashing his X session on a regular basis 17:48
Limbic_Region geoffb - you don't use named parameters as positional in tests you write do you? 17:49
geoffb blinks
Ummm, I don't think I'm the one to talk to. 17:50
I've seen discussions about that, but I stayed out of them
Limbic_Region svn.openfoundry.org/pugs/t/pugsbugs...sitional.t
well - when I first started writing tests I copy/pasted from an existing one
so I may have done it (initially) without even knowing 17:51
apparently there are going to be a number of tests that break once this bug is fixed
Limbic_Region is just trying to make people aware
geoffb ah 17:52
Well, I don't think so.
My big hurting place is the fragility of use ...--perl5
And unfortunately in trying to test for it, I just recurse into issues with Test.pm, and issues with use of normal P6 modules, and so on. 17:53
And the fact that for some annoying reason, code is not parsed the same interactively as from a file. 17:54
vel hi all
geoffb evalbot6, sure, I can understand that there's some sops to safety there, but pugs in interactive mode should be damn near identical to pugs working on a file.
</rant> 17:55
vel can someone help me with chomp in perl 6?
Limbic_Region vel - I believe it is still mostly unspecced
what specifically do you need to know?
ninereasons autrijus, in perlcabal.org/~autrijus/apoc/slide62b.html#end 17:56
autrijus, do you mean "those are the only two things that work now" ?
autrijus, or do you mean "those are only two things that work now" ?
vel just simple and shortest way to get rid of \n and parse csv file
Limbic_Region vel - see svn.openfoundry.org/pugs/examples/g...hangman.p6 17:57
search for chomp
apply it to your situation
holler if it doesn't work as expected
vel Limbic: oh, I did trust me. thanks for link... 17:59
geoffb I love finding cool new hotkeys by accident
Limbic_Region vel - when in doubt, checking examples/ t/ and ext/ will usually leading to finding what you are looking for (if Pugs has it yet that is) 18:00
vel Lumbic: is that yoda speaking? ;) 18:02
svnbot6 r4492, ninereasons++ | Apocalypse_Now: fix a few typos
Limbic_Region vel - it is me thinking and typing at the same time 18:03
I am currently reading autrijus's slides
ninereasons autrijus, I think it might be clearer to say, "those are the only two programs in our examples/ that work now" (compile to Parrot) 18:06
autrijus, if that's true to your meaning.
Limbic_Region I think that will come out in the presentation itself 18:07
basically it is "don't think this stuff just magically works"
"we have a very limited focused amount of functionality"
ninereasons Limbic_Region, yes, but "that's the only two things that works" doesn't work. 18:08
Limbic_Region "we are going to spend the next 10 days locked in a room with plenty of vices to knock the rest out"
ninereasons - keep reading though
ninereasons Limbic_Region, purely a syntax issue here.
Limbic_Region I agree - the word "only" conveys an absoluteness that isn't accurate 18:09
ninereasons Limbic_Region, the number of "that's" and "two things" and "works" does not agree 18:11
I think he means "those are the only two things that work, right now" . doesn't he? 18:12
Limbic_Region "those are the only two programs in our examples/ that work now" # is what I believe is meant 18:13
I honestly don't think anyone at the presentation will get it wrong though
gaal howdy
Limbic_Region salutations gaal
geoffb hi gaal 18:14
gaal L~R, remember our brief discussion of want nad caller?
Limbic_Region I saw your thread on the list
I read $larry's response
I was left wondering - WTF does that mean
gaal :)
well, here's my interpretation 18:15
caller(?$kind, +$skip, +$level)
if you say :kind<Method>, you only look at methods 18:16
then you go to the skipth Method
there are other details to work out, but so much for selection. 18:17
Limbic_Region gaal - I doubt you are going to make me understand
all that is important to me is that you (or someone else on the team) understands
gaal as for want, it's much simpler than the A said.
Limbic_Region which is tooooo bad
gaal just look at the S verison 18:18
Limbic_Region on list of things to do 18:19
gaal (btw, even without any elaborations, we need some reworking to fix pugs' want, because Ctx doesn't have, for example, the number of wanted return values.)
(nor whether they are rw)
Limbic_Region right gaal 18:20
which is the very specific thing I was trying to get fixed
so you could determine if you were in a void context or not
Limbic_Region wanders off for a bit
gaal now, s06 doesn't specify the signature for want, so i suppose we can give it the same selection interface as caller, adn have want just delegate. that's the easy bit :) 18:21
oh, L~R, that much you *can* do today.
or should be able to, at least.
data Cxt = CxtVoid | CxtItem !Type | CxtSlurpy !Type
oh wait, you do get if we're in LValue context 18:23
that much can be useful
brb
Limbic_Region waits for gaal to return 18:26
gaal back
wolverian Limbic_Region: yes, that's the problem - people use Test::ok and others incorrectly. we can either fix Test::ok and friends which is the easy route (change named to positional) or mass change the .t files 18:27
which will take some time probably
Limbic_Region pugscode.org/talks/apw/slide30c.html#end
shouldn't that be >>xx<< there?
iblech Limbic_Region: Is >>{...}<< in any of the AES? Wasn't able to find it... 18:28
wolverian shouldn't that be Ā»x, actually?
Limbic_Region I am not even going to attempt unicode with this client 18:29
gaal >>{...}<< ? the vampire bat operator?
Limbic_Region iblech - if it isn't in one of the AESs - it has definately been discussed on the list
iblech Limbic_Region: Thanks, I'll google then
geoffb wolverian, personally I'm thinking the positional usage of Test subs is a good thing. When you've got a lot of very similar tests in a test file, positional calling cuts the clutter.
wolverian geoffb: I agree. 18:30
gaal L~R, so, does want not work at all for you?
Limbic_Region gaal - want as a number and/or want.count
gaal oh, well that you'll have to wait a bit for ;-) 18:31
Limbic_Region so how else would you currently check for void context?
gaal it returns show Ctx
which should say "void".
sorry, not show Ctx 18:32
just the string "Void"
see Pugs/AST/Internals.hs:1010 ish
Limbic_Region you mean want returns show Ctx
let me try something then
gaal more or less 18:33
Limbic_Region ok - then your understanding and my example are in disagreement
gaal hmmm, except that it doesn't work :)
Limbic_Region perlbot nopaste 18:34
perlbot Paste your code here and #<channel> will be able to view it: sial.org/pbot/<channel>
gaal ./pugs -e sub x { say want } x();
Limbic_Region right
gaal says "List (Any)"
that's a bug then.
:)
Limbic_Region right - perlmonks.org/index.pl?node_id=464377 is where I first found out
that's the second bug tall_man has uncovered
or rather - that was the first 18:35
the named parameters being morphed to position was the second
gaal it oughta return an object that stringifies to "Void" in that case.
Limbic_Region gaal - there are tests for want already though
so I didn't see where writing a new one was helping?
gaal okay; i've however been working on caller, which despite what A06 says, isn't really much like want. 18:36
only the selection stuff might be
and anyway caller is used as a primitive for want
sooooo... we need tests for caller ;-) 18:37
likely, you could from pure perl code call caller, use the .sub result and query that to get all the info you could have gotten from want.
though that bit isn't 100% certain yet. 18:38
Aankh|Clone Hrm. 18:39
One-line C<class { ... }> declarations parse, but multi-line ones don't?
18:39 Aankh|Clone is now known as Aankhen``
Limbic_Region gaal - actually, looking at t/builtins/want.t all the want tests that are using want() are todo 18:41
Aankhen`` Apparently multi-line class declarations do parse. Ignore my previous statement. 18:42
pdcawley ?eval class Array { method bibble { 42 } }; my @foo = (); @foo.bibble 18:43
evalbot6 42 18:44
pdcawley Hurrah!
?eval [].bibble
evalbot6 Error: No compatible subroutine found: "&bibble"
pdcawley ?eval class Array { method bibble { 42 } }; [].bibble 18:45
evalbot6 Error: No compatible subroutine found: "&bibble"
wolverian ?eval ref []
evalbot6 ::Array
Aankhen`` ?eval method *Array::bibble () { 42 } 18:46
evalbot6 undef
wolverian hmm. do references not dispatch automatically yet?
Aankhen`` ?eval method *Array::bibble () { 42 }; [].bibble
evalbot6 Error: No compatible subroutine found: "&bibble"
Aankhen`` ?eval method *Array::bibble () { 42 }; <a b c d e>.bibble
evalbot6 Error: No compatible subroutine found: "&bibble"
Aankhen`` Bah.
pdcawley ?eval my @foo = (); ref @foo
evalbot6 ::Array
pdcawley Hmm... time to write a test.
Aankhen`` ?eval method *Array::bibble () { 42 }; my @foo = (1, 2, 3, 4, 5); @foo.bibble();
Limbic_Region ?eval method ::Array { method bibble { 42 } }; [].bibble;
evalbot6 42
wolverian pdcawley++
evalbot6 Error: No compatible subroutine found: "&method"
Aankhen`` ?eval method *Array::bibble () { 42 }; my @foo = (1, 2, 3, 4, 5); @foo.bibble(); 18:47
evalbot6 42
Aankhen`` That works.
wolverian Limbic_Region: heh, method ::Array?
pdcawley Unless someone else wants to -- I'm slightly tied up.
Limbic_Region wolverian - why not
Aankhen`` ?eval method *Array::bibble () { 42 }; my $foo = (1, 2, 3, 4, 5); $foo.bibble();
evalbot6 42
Limbic_Region WAS being silly
Aankhen`` Kewl.
pdcawley Is ::Foo a symbol constructor then?
wolverian pdcawley: it's a type.
pdcawley ?eval ref ::Array
evalbot6 ::Class
pdcawley ?eval ref ::Bibblyplop 18:48
evalbot6 ::Type
wolverian see. :)
pdcawley Heh.
integral hmm, doesn't that mean I can get silent breakage?
pdcawley Is there a 'symbol' type? I've got quite attached to them from playing with Smalltalk/Ruby
wolverian what sort of?
integral If I use ::Foo as a type variable, but then some module I load, loads a ::Foo
Aankhen`` ?eval class Foo { submethod BUILD ($self: ) { $self.baz = "quux"; say $self.baz; } method bar () { say ::?CLASS.new(); } } my $foo = Foo.new(); Foo.bar();
evalbot6 Error: unexpected "{" expecting trait, ";" or end of input
gaal you can get anything you want in alice's restaurant, integral. 18:49
pdcawley Allison's restaurant surely?
gaal :)
Aankhen`` ?eval class Foo { submethod BUILD ($self: ) { $self.baz = "quux"; say $self.baz; }; method bar () { say ::?CLASS.new(); } }; my $foo = Foo.new(); Foo.bar();
evalbot6 Error: unexpected "{" expecting trait, ";" or end of input
integral hmm!
Aankhen`` Bleh.
pdcawley Do you need a ; between } and }? 18:50
wolverian no.
Aankhen`` I dunno.
I would have thought not.
Aankhen`` reduces the size.
pdcawley And you can't 'say' anything with evalbot.
Aankhen`` Ah, right.
?eval class Foo { method bar () { ::?CLASS.new() } }; Foo.bar(); 18:51
corneli Was the zip(@a => 2, @b => 1) changed? is it only zip(@a;@b;...) now?
evalbot6 Error: unexpected "{" expecting trait, ";" or end of input
Aankhen`` ?eval class Foo { method bar () { Foo.new() } }; Foo.bar();
evalbot6 {obj:Foo}
Aankhen`` Darn.
::?CLASS doesn't work.
Aankhen`` goes to check that his syntax is correct.
pdcawley It's $?CLASS isn't it?
Limbic_Region pugscode.org/talks/apw/slide40b.html#end 18:52
s/ot/of/
pdcawley ?eval class Foo { method wibble { $?CLASS } }; Foo.new.wibble
evalbot6 \::Foo
pdcawley Seems so.
wolverian hmm, why is it $?CLASS and not ::?CLASS?
pdcawley Because it's variable like.
Aankhen`` I guess S12 is outdated then: (The current class may always be named as ::?CLASS even in anonymous classes or roles.) 18:53
pdcawley ?eval class Foo is Array { }; class Array { method bibble { $?CLASS }}; Foo.new.bibble
iblech wolverian: IIRC both ways are correct: $?CLASS is a Class object as a scalar variable, while ::?CLASS is a Class object as package name
evalbot6 \::Array
wolverian iblech: right.
gaal ?eval ref Array 18:54
evalbot6 ::Class
pdcawley ?eval class Foo is Array { }; class Array { method bibble { $?CLASS }}; Foo.new.class
evalbot6 Error: No compatible subroutine found: "&class"
pdcawley ?eval class Foo is Array { }; class Array { method bibble { $?CLASS }}; Foo.new
evalbot6 {obj:Foo}
pdcawley ?eval class Foo is Array { }; class Array { method bibble { $?CLASS }}; ref Foo.new
evalbot6 ::Foo
Aankhen`` iblech >> So ::?CLASS hasn't been implemented?
iblech Yep.
Aankhen`` OK.
gaal wait, if i write a func that accepts a class as an arg -- which serendipitously I AM JUST DOING, does hte signature have "::Class $kind" rather than "Class $kind"? 18:55
pdcawley Class $kind
wolverian gaal: Class is pretty certainly predeclared in Perl 6 18:56
pdcawley But you'd call thing(::Whatever)
gaal so why
?eval Array
evalbot6 \{obj:Class}
gaal oops
?eval ref Array
evalbot6 ::Class
gaal with the :: ?
pdcawley Easier to read.
gaal heh 18:57
ok
pdcawley You *know* it's atype.
Same as why there's still Sigils.
svnbot6 r4493, rootmj++ | docs/talks/Apocalypse_Now.spork 19:00
r4493, rootmj++ | @h{'!'} @h.method -> %h{'!'} %h.method
pdcawley ?eval sub foo { $?CALLER_CONTINUATION } ; foo
evalbot6 Error: Undeclared variable: "$?CALLER_CONTINUATION"
pdcawley Not implemented yet then. Bah.
iblech It's &?CALLER_CONTINUATION
But it seems to have some bugs
pdcawley ?eval sub foo { &?CALLER_CONTINUATION } ; foo 19:01
evalbot6 \sub {...}
pdcawley Full continuation?
iblech Dunno 19:03
pdcawley Right, I'll have a play. 19:04
wolverian are there other kinds of continuatinos? 19:05
s,no,on,
pdcawley wolverian: There's it's poor relation, the 'escape continuation' which isn't really a continuation because you can only use it once.
Personally, I prefer to think of them as chocolate teapots, but I'm picky like that. 19:06
wolverian chocolate teapots. heh. 19:08
nothingmuch evening 19:10
geoffb g'day 19:14
pdcawley Arg, can't remember the rules for slurpiness in signatures. *@foo slurps the rest of the args right?
Aankh|Clone Yes, pdcawley.
SamB wonders why in the world you would want to use the same continuation twice 19:18
svnbot6 r4494, iblech++ | t/operators/repeat.t -- unTODO a now suceeding :todo<bug> test. 19:19
r4495, Aankhen++ | * fixed a few parsing problems in HTTP::Message.
pdcawley SamB: All sorts of reasons.
I'm writing some example code now.
SamB all sorts of MUTABLE STATE THAT MAKES YOUR HEAD EXPLODE reasons? 19:20
integral thinks using the same subcontinuation twice is more useful more often
svnbot6 r4496, iblech++ | Implemented ::?CLASS, ::?ROLE (currently same as ::?CLASS), and ::?PACKAGE.
r4496, iblech++ | Aankhen``++
pdcawley Writing it, and understanding how it works, can make your head explode.
However, the kind of things you can do with full continuations can be a delight to *use*.
gaal anyone remember which yapc movie it was that had the Meta joke? 19:22
rjbs not one I've seen, I think. 19:23
gaal oh 19:24
oscon movie :)
www.perl.org/oscon/2003/movies/truth/
awwaiid yeah, reuseable continuatiosn rock :) 19:26
pdcawley ?eval my $try = -> *@choices { if 1 { 10 } } 19:28
evalbot6 \sub {...}
pdcawley ?eval my $try = -> *@choices { if 1 { 10 } }; $try()
evalbot6 10
pdcawley boggles. How do I paste an extended lump of code in here? 19:30
wolverian see /topic
gaal perlbot nopaste 19:31
perlbot Paste your code here and #<channel> will be able to view it: sial.org/pbot/<channel>
pdcawley Ta. 19:32
pasteling "pdcawley" at 217.146.110.1 pasted "Weird syntax error" (35 lines, 724B) at sial.org/pbot/10895 19:33
pdcawley HTF can it not be expecting an 'i' after the opening brace of a code block? 19:34
gaal that's just the parser not being suffiiently friendly.
the error is deeper in.
could it be you want &give_up := ? 19:35
Limbic_Region oh yeah - parser errors are a PITA to debug 19:36
pdcawley Wow. Now it's *really* breaking.
I could be wrong, but I *think* it's compiling now. But oh *my* is it throwing a spectacular error. 19:37
pasteling "pdcawley" at 217.146.110.1 pasted "Huge syntax error" (4 lines, 7K) at sial.org/pbot/10897 19:38
gaal that's a run time error 19:39
pdcawley Indeedly.
gaal not a syntax error :)
iblech It seems you can't re-bind &subs. 19:40
pdcawley And I still need 'sub' before sub blocks. 19:41
Limbic_Region but not all sub blocks
pdcawley And when in doubt, stick a ; on the end. 19:42
Limbic_Region it is weird but sometimes omitting it works and sometimes it doesn't
pdcawley Ah. Now I'm just getting a simple 'program failed' at... um... everywhere.
Ah. No, I understand that. 19:43
Seems &?CALLER_CONTINUATION doesn't work...
But we probably knew that.
pdcawley notes that the power of non-deterministic programming *still* blows his mind. 19:45
Hmm... thinking about this, I might be able to write 'choose' *without* needing a 'proper' continuation -- might be able to get away with 'return' in a pointy block. 19:47
iblech That doesn't work yet -- t/pugsbugs/return_in_anonymous_subs.t still fails 19:49
pdcawley Shame... 19:50
Wouldn't hurt to have an (in)sanity check to see if what I'm thinking of should actually work...
Aankhen`` Hmm. 19:53
Aankhen`` ponders HTTP::Message as a class versus HTTP::Message as a role. 19:54
pasteling "pdcawley" at 217.146.110.1 pasted "Continuationless ND programming" (16 lines, 377B) at sial.org/pbot/10898
svnbot6 r4497, iblech++ | Make ::("?CLASS") etc. work (symbolic dereferentiation for type vars). 20:05
Aankhen`` Does a plain ::?CLASS work? 20:06
iblech Yep, I implemented some minutes ago :) 20:07
pdcawley: Does this example work? Should it be checked in to example/?
pdcawley I think it *should* work.
But I'm not sure Larry does. might have to implement it with a continuation -- when they work. 20:08
Aankhen`` iblech++ # ::?CLASS =)
iblech Aankhen``++ # asking for it :) 20:09
pdcawley From things Larry's said on p6l it sounds like 'pointy block return' should uses an escape continuation. So that code probably shouldn't work.
The continuation version *should* work though. 20:10
pasteling "pdcawley" at 217.146.110.1 pasted "Continuationful ND programming" (19 lines, 412B) at sial.org/pbot/10900
pdcawley That should work, if 'sub callcc (Code &block) { &block(&?CALLER_CONTINUATION) }' works. 20:11
Oh. Fuck. Just worked out why it's failing. 20:13
Breaking differently now...
gaal um, how do i "return;" in p6? 20:16
svnbot6 r4498, iblech++ | t/var/symbolic_deref.t -- Tests for ::("...").
gaal and not from haskell? :)
will return () work?
iblech ?eval my $counter; sub { $counter++; return; $counter++ }; $counter 20:17
evalbot6 \undef
iblech Hrm 20:18
Err
?eval my $counter; sub { $counter++; return; $counter++ }(); $counter
evalbot6 \1
iblech Seems to work :)
gaal huh? i thought "return;" wasa syntax error now?
Limbic_Region gaal - when?
gaal well, it must have been just unimplemented when i had asked :)
oh, but crap, i'm in prelude and i *can't* return 20:19
shucks.
Limbic_Region ?eval my $counter; sub { $count = 3; return; $counter = 42; }.(); $counter;
evalbot6 Error: Undeclared variable: "$count"
Limbic_Region heh
?eval my $count; sub { $count = 3; return; $counter = 42; }.(); $counter;
evalbot6 Error: Undeclared variable: "$counter"
Limbic_Region grrr
pdcawley s/counter/count
Limbic_Region ?eval my $count; sub { $count = 3; return; $count = 42; }.(); $count;
evalbot6 \3
iblech gaal: Use if(return_condition) {} else {...rest of code...}?
pdcawley Hang on, '\3'? 20:20
Limbic_Region gaal - AFAIK, the return problem was fixed way early on in Pugs development
pdcawley - I don't know why the \
gaal must have asked a long time ago then :)
pdcawley ?eval 3
evalbot6 3
gaal iblech - no, because it want to retrun false
undef in scalar context
() in list context
pdcawley blinks.
gaal *and i can't use want* :) 20:21
Limbic_Region pdcawley - because it is tied to a variable?
?eval "foo"
evalbot6 'foo'
gaal meh, i'll just write this in haskell. it's easier there.(!)
iblech :)
Limbic_Region ?eval my $foo = 'foo'; $foo;
evalbot6 \'foo'
Limbic_Region yep
pdcawley ?eval my $foo = 3
evalbot6 \3
Limbic_Region tied to a lexical even 20:22
must be how evalbot works under the covers
pdcawley ?eval our $foo = 3
evalbot6 \3
Limbic_Region ok - so maybe it is how it gets passed up from p6 to Haskell out to the screen
Limbic_Region notices there are no more straws to grasp and calls it a night
SamB what is ()? it looks like nil! 20:23
iblech evalhelper.p5 uses eval($code).perl to display the results
Limbic_Region heads homewards
Aankhen`` ?eval 3.perl 20:26
evalbot6 '3'
svnbot6 r4499, iblech++ | t/oo/magical_vars.t -- Added tests for ::?CLASS and ::?ROLE.
Aankhen`` ?eval 3.perl.perl
evalbot6 '\'3\''
Aankhen`` ?eval [].perl
evalbot6 '[]'
Aankhen`` ?eval [].perl.per 20:27
evalbot6 Error: No compatible subroutine found: "&per"
Aankhen`` ?eval [].perl.perl
evalbot6 '\'[]\''
Aankhen`` Heh, neat.
gaal what's the haskell V type for Class? VType? 20:28
the *pugs* type that is 20:29
or is it a VObject that happens to be a class? 20:31
svnbot6 r4500, iblech++ | * Pugs.Parser -- Parse $::! etc., too. 20:39
r4500, iblech++ | * t/var/symbolic_deref.t -- Tests for this and $::("!").
Aankhen`` iblech, I was wondering -- how old are you? 20:40
iblech 16
Aankhen`` Ah.
Where you from?
iblech Germany, Bavaria, Augsburg
Aankhen`` Ohh.
iblech Ok, gonna sleep, school tomorrow 20:42
Night :)
Aankhen`` G'night.
I think my computer is trying to tell me to get off. 20:44
G'night. 20:45
shaine if ~ is concat now, whats ones complement? 20:46
20:47 sbkhh is now known as Odin-
arcady I think +~ 20:48
shaine thank you
arcady actually, no 20:49
gaal ?eval +~0
evalbot6 0.0
gaal nope :)
coaster ~^
gaal ?eval ~^0
evalbot6 Error: unexpected "^" expecting term
gaal beh, quoting borkage.
arcady ?eval +^ 0
evalbot6 Error: unexpected "^" expecting term
arcady oh right
gaal ~^ can't be right, ~-ops are stringy. 20:50
coaster i got that from a03
This is now the bitwise XOR operator. Recall that unary ~ (1's complement) is simply an XOR with a value containing all 1 bits.
[Update: ~ is now string concatenation. Bitwise XOR is +^ or ~^ depending on whether your doing numeric xor or stringwise.]
gaal well you aren't doing stringwise xor here! :) 20:51
shaine ah crap, i gotta go
Khisanth hmm got parrot 20:54
gaal i need a haskell function that's prolly in the prelude, but i don't know its name: 20:56
mkInf :: a -> [a]
integral that is in the prelude, can't remember the name though :-/ 20:57
gaal mkInf x = x : mkInf x
ah, there's 'iterate'
which i can use with 'id' 20:58
integral hmm, I thought there was something specific
gaal repeat!!
zvon.org++
integral ah, repeat :-)
gaal thanks :)
integral www.cs.uu.nl/~afie/haskell/tourofprelude.html # Tour of the Haskell Prelude, is rather useful
gaal www.zvon.org/other/haskell/Outputpr...index.html has a great "related" field :) 20:59
caller becomes so beautiful 21:00
formatFrame $ last $ take skip $ repeat fetchCaller
where fetchCaller = monadic black magic :)
sorry, forgot:
integral oh, you want replicate (replicate n x = take n $ repeat x)
Juerd So, suppose E is the euro sign, does Efoo = $foo automatically convert? ;) 21:01
integral or maybe !! ?
gaal formatFrame $ last $ take skip $ filter wanted $ repeat fetchCaller
actually this may not work, come to thing of it
integral formatFrame $ (filter wanted $ repeat fetchCaller) !! skip -- probably with a off-by-one error 21:02
gaal because this is inside a weird monad thingie
integral ah, you need a M variant of these things
gaal i take care of the off by one
before that
anyway, the problem is that i wanted fetchCaller to do local (const $ envCaller env) stuff 21:03
integral hmm, well there's things like mapM, maybe there's a repeatM 21:04
xerox integral, there indeed is.
gaal and: i don't want to replicate, because i need to actually generate more than i take, because i filter.
xerox integral, I ran a bot on #haskell while lambdabot is on vacation, if you need it :)
gaal (what i forgot the first time i gave the function) 21:05
integral oh, neato xerox++ :-)
gaal xerox: clues on whether that would work as i want it to with Reader?
xerox gaal, sorry I was not reading all the discussion :) 21:06
pasteling "gaal" at 192.115.25.249 pasted "current op1Caller impl" (21 lines, 878B) at sial.org/pbot/10902
gaal i'm adding smarts to this function, to take some filtering responsibility upon itself 21:07
so i wanted to refactor the bits that do the recursion and upstacking out of the main code
basically i want a fetchCaller local function that gives me the next stack frame every time it's called. 21:08
oh, now i realized there's the additional complication of maybe running ouf of call stack :) 21:09
i can't just take as much as i like, the generator isn't really infinite.
boo, it was elegant for a moment there. :)
xerox :)
gaal no doubt i can have fetchCaller fail when it does run out of stack though? 21:10
thatwould be nice. 21:11
GHC / typical haskell optimization question: 21:17
say i define a local function to be if (x) then id else ...complex func.... 21:18
where x is a variable of the main function
will the local function be efficient when x is True?
_metaperl what does ++before the @ sign do here: 21:19
gaal will its applications be nicely optimized away that is?
_metaperl sub NestedLoop (++@loop, +$OnlyWhen, +$code) returns Ref {
what does "++@loop" mean?
gaal i think it means mandatory named arg
_metaperl in fact, that does +$OnlyWhen mean also
the "+" in front of $OnlyWhen
gaal i'm sure that means optional named arg :)
so you can call NestedLoop :loop<a b c> :OnlyWhen<x> 21:20
but not NestedLoop()
_metaperl that's a serious hijacking on the autoincrement operator 21:21
gaal i think it's not finally blessed, but larry's considering it. 21:22
it's mentioned in s06 i think.
as an option.
_metaperl when do you think p6 is going live?
gaal it already does if you use iblech++ 's cd :) 21:23
pjcj what does that mean?
_metaperl pjcj: you mean "going live" ? 21:24
pjcj yes
_metaperl I mean when will it "ship"
when will it be the main line stable perl
q[pdcawley] Some time. 21:25
_metaperl when features are now added to p6 and p5 is just for maintenance
gaal so, is Class an object inside pugs?
if i have sub foo(Class $x)
what's the haskell type of $x?
wolverian _metaperl: ++ is an unofficial '+$foo is required' prefix
gaal sorry to repeat my Q from before, i'm needing this.
wolverian right, gaal told you that already. sorry to repeat. :) 21:26
gaal aha! looking up op1 "ref" proved useful :-) 21:27
xerox Goodnight folks :) 21:28
pjcj _metaperl: I suspect those are three different times: 1. depends what you mean by "ship" 2. 3 to 8 years? 3. when 5.10 is releaseed? 21:30
clkao pjcj! 21:38
pjcj clkao! 21:40
clkao so what about the uncoverable mark we discussed yesterday 21:41
pjcj yes - what exactly is it that you need? 21:43
clkao i want magic comments 21:44
pjcj ah - is that preferable to an external file for you too? 21:45
clkao yes. because if you change the relevant statement you will have to update the file in a painful way i presume
i mean, it should be supported that way as it's obvious, people who don't like can of course establish their own policy to use external file 21:46
pjcj if you change the statement itself, yes
ok - seems that most people want comments, so I had better support that 21:47
clkao ))
:))
now, plan for release!
limbicLR wonders if he can now type a + and not have it be eaten
cool
Limbic_Region Juerd++ 21:48
gaal pugscode.org/talks/apw/slide12d.html 21:51
didn't autrijus mean "%h" ?
wilx` Heh, is it only mine Firefox that merges all the bullets from 1. to 4. into one line? 21:53
On that page?
gaal yes. :-p
kelan would anyone mind helping me understand a section of chapter 6 of the harrorth saga? 22:03
sapp3r msg /nickserv register sapp3r fritjof 22:04
gaal hasn't read it yet
sapp3r: now sounds like a good time to change your password :) 22:05
sapp3r already have :)
dumb typo, sorry for interuption.
kelan when does the NA hackathon happen?
i think some good things will come out of that
gaal dunno, but it's sure to be on the site 22:06
google?
kelan i could probably look it up on perlmonks faster. heh
it was in the announcement section
wilx` Hmm.
Does Perl6 support multiple parents inheritance? 22:07
kelan i'd be surprised if it didn't
but i don't know for sure
wilx` I'm just wondering because I just came over #39 of the presentation... 22:08
kelan from S12: Perl 6 supports multiple inheritance, anonymous classes, and autoboxing. 22:09
gaal durn, there is no repeatM. i'm sure this is trivial.. any lambdaheads still about? 22:10
repeatM :: Monad m => a -> m [a] 22:11
kelan hmm
wilx` Huh.
gaal repearM x = do return x : repeatM x
?
wilx` repeatM = return $ repeat
Maybe... 22:12
kelan how about: map (\x -> return x) repeat x
er
how about: map (\x -> return x) repeat y
to be clearer
gaal dunno, can you do that? :)
kelan probably, but i think the type from mine would be `[m a]` instead of `m [a]` 22:13
gaal return $ repeat is really cute
22:13 wilx` is now known as wilx
gaal if it works 22:13
point-free and all.
kelan actually mine could just be `map return`, but again, wrong type 22:14
hugs says `return . repeat` is type `Monad a => b -> a [b]` 22:15
so that looks like what you want
gaal uhhhh another lame question, to do this ad-hackishly,
oh wait, i can just inline that.
heh heh heh.
wilx Wheeee!! 22:19
That presentation is long even for me as reader.
And I have been doing like a click every two seconds :) 22:20
gaal me too, the build was finished when i got to 20b and then the compiler errors started getting my attention. :)
but up till then my though about this talk was this: 22:21
wow, to be there without being involved in pugs must be mind blowing
i wish i were there like that!
but then i thought
to be involved in pugs is way more mind blowing :)
though i still wish i were there! 22:22
wilx :)
gaal hey, this is weird. the primDecl for op1 "ref" says it returns Str 22:27
but its definition looks like it returns a list of VType ?
alternatively: what's "Type"? somewhere it said data Int | Str, but what does that mean? how can a type be an int?
...and where are all the user defined classes etc.?
kelan wow poe is getting ported already? 22:28
gaal iblech++ # the mad porter
kelan i would have thought poe would be held off until event semantics have been spec'ed 22:29
poe might not even really be needed since p6 will have an event loop
integral POE isn't just an event loop really. It's provides it's own object system for example 22:30
mugwump and continuations etc
mugwump declares an apostrophe jihad on integral
integral POE::Session::YieldCC does continuations for perl5 POE leakily
err, they do that all by themselve's hone'st! 22:31
gaal so, i'm getting the tireds. should i march on past 'em and hack, or sleep and go see the antonioni+wenders film tomorrow morning? 22:35
problem is that in cases like these i invariably look to a role model and ask, what would my cat do? 22:36
and that's sortof a sold trial when *sleep* is the case.
also, i just got my second spam for today from a certain local spammer. ordinarily this would piss me off, but a week ago a new law was passed here that awards people about $200 damages per unsolicited message. maybe i should sue, if just to make them hurt? 22:38
gaal sleeps & 22:42
clkao seen autrijus 22:58
jabbot clkao: autrijus was seen 10 hours 16 minutes 59 seconds ago
23:35 _meppl_ is now known as meppl