Run Perl6 Now! | pugscode.org ('Overview', 'Journal') | pugs.kwiki.org | chan now logged: xrl.us/e98m | openfoundry down. svn co svn.perl.org/perl6/pugs/trunk
Set by metaperl_ on 4 March 2005.
01:41 edgewalk1r is now known as edgewalker
metaperl erxz_paste where? 02:21
perlbot paste? 02:22
perlbot Paste your code and #perl will be able to view it: sial.org/pbot/perl dragon.cbi.tamucc.edu:8080 erxz.com/pb
erxz_paste "metaperl" at 4.30.46.9 pasted "help needed creating compatible subroutine" (16 lines) at erxz.com/pb/507
offby1 Should I be surprised if code from the Synopses doesn't work in pugs 6.0.9? i.e. ``sub numcmp ($x, $y) {return $x <=> $y;} 03:02
print numcmp (y=>3, x=>4), "\n";'' yields ``No compatible subroutine found: &y''. That just means some bits aren't yet implemented ... right?
buu offby1: Probably. 03:23
offby1 it's brand-new, so ...
buu offby1: If the synopsis's don't work, it's a bug in pugs.
offby1 wonders if he overlooked a mention of a bug list 03:24
buu offby1: That syntax doesn't look correct
wolverian pugs apparently doesn't have => yet 03:25
offby1 buu: let me see if I can find the page from which I copied it. 03:26
buu: dev.perl.org/perl6/synopsis/S06.html, about one-quarter of the way down 03:27
buu Well 03:29
if it doesn't work in pugs
it's a bug
Try adding a test case
offby1 dunno how yet
buu look in t/ 03:30
offby1 yup
buu Then email it to someone with commit privs 03:31
offby1 oops 03:32
buu Eww, emacs 03:33
offby1 :-)
jdv79 emacs rules 03:34
offby1 buu: how could you tell?
buu offby1: Your quit message.
offby1 ah 03:35
funny, I accidentally killed the buffer; I'm surprised (and impressed) that the IRC client was graceful enough to send the quit message.
(and if buu doesn't know what I'm talking about; that's OK; it's emacs-speak :-) 03:37
buu I understand your heathen lingo.
offby1 hey, I use vim too.
I'm what you call "ecumenical". 03:38
buu heh
ingy hola 03:40
buu HALLO INGY 03:41
You dead yet?
jdv79 POD hater
ingy I lived!
offby1 wow, pugs is certainly active -- half a dozen commits today. And due to the beauty of svk, I can read the commit logs even though the main subversion server is down.
offby1 states the obvious once again
ingy is the svn server back up?
buu ingy: svn.perl is
offby1 svn.perl.org is up.
jdv79 that one's always up isn't it? 03:42
openfoundry is the funky one i think
ingy is pugs mirrored there?
buu jdv79: It's open as much as openfoundry is
Whcih is to say, like 99.9% of the time
offby1 openfoundry is at least the one that's not up now.
buu Until they have a 3 day outage =/
jdv79 yup
offby1 at least they were nice enough to put up a placeholder page (in two languages, no less.)
ingy tonight I get to hack pugs 03:43
I'm excited
buu ingy: You can still be excited?
ingy I'm definitely tired 03:44
but I did pretty good in the race
so that counts for something
jdv79: I don't hate POD 03:45
buu ingy: What place?
ingy I did good because it was a very fast 72 mile race and I finished ~30th of ~120 03:47
which is not good
jdv79 72 mi of what?
ingy but I am also the fattest guy out there
jdv79 on/in what?
ingy and I'm 40+ yrs old
offby1 ingy: no.
nobody's THAT old. 03:48
offby1 is N years old, for large values of N.
ingy jdv79: lance racing ;)
metaperl igy did you quit working at activestate? why?
ingy because the boss was a Dick! 03:49
hahaha
jdv79 i don't like them
offby1 OK, I've written the simplest, dumbest failing test in the world. It's so small that I could paste it here. Should I just send it to a dev list? Paste it? Do "svn add" and pray for commit access :-) ?
I'm so dedicated that I even made a symlink to /usr/bin/pugs
jdv79 lance? 03:50
ingy armstrong
jdv79 what bike do you have?
ingy trek 03:51
jdv79 i got a cannondale
offby1 dinner time
ingy Somebody just emailed me: Many thanks for your greath code,
jdv79 no races though, just local MS tours and such
ingy cool
racing is a thankless sport 03:52
jdv79 you get personal satisfaction though, right?
ingy the best you can hope for is to get in pretty good shape
but I'm 30lbs over race weight
metaperl ingy, what happened to the Spoon docs? I dont see any in the distro anymore: cpan.uwinnipeg.ca/dist/Spoon
ingy I get more satisfaction from hacking for sure 03:53
jdv79 haha
ingy but racing keeps me young
metaperl: there are none :(
metaperl I just wanted to read the docs. I am a CGI::Prototype fan personally
jdv79 its definitely not as painful
metaperl there used to be
ingy ?
metaperl you talked about how you and Dave used Spoon at Socialtext
in the docs
oh maybe taht was spiffy 03:54
ingy Spiffy has docs
docs for Spoon is a premature optimization
;)
I want to make a Kwiki variant where people can document modules for other authors 03:55
metaperl well I hope openfoundry opens up early tomorrow
ingy I was going to use POD but who knows now
metaperl and I hope I can commit my perlmongers talk on junctions tomorrow
ingy is there another place to get pugs source? 03:56
metaperl the topic says...
ingy ohhhh
metaperl I think people are sending patches to p6c
and luqui applies them
buu ingy: What do you mean 'document modules for other authors' ?
metaperl but I have whole new files to add
ingy buu: I'll explain it later 03:57
I need a shower, then hack hack hack
jdv79 shower sounds good
ingy =) 04:00
autrijus hi. I'm @ hongkong airport without much battery :) 04:20
but I did manage to commit the fix for "@a" and "@a[]" parsing raised yesterday here, as well as many other fixes. (r581) 04:21
should be back to taipei in 3~4 hrs.
obra hi autrijus 04:22
autrijus greetings obra san. how's life?
obra today, I cleaned out my homedirs and put more of my life under svk control
not very exciting, but useful
autrijus verily. I did that when my hd was not broken 04:23
obra today was a quiet day when I should have been hacking
autrijus but evidently I should push it up somewhere else, instead of letting it stay in the same disk.
not in the hacking mood? :)
obra yep. I have .subversion-repository on my colo box
never managed to get myself to the right mental state
too many open loops
autrijus caffeine? ritalin? sleep? 04:24
obra I also need to find good things for me and kaia to do when we're in .tw
had plenty of sleep. my shoulders are too tense. and I've not had good blocks of time
autrijus ah. virtualtourist may or may not help there.
havn't seen their taiwan section.
obra *nod*
I should also try wikitravel/wikipedia 04:25
autrijus there is a wikitravel?
wow there is.
obra it was "ok" for geneva 04:26
autrijus nod. battery soon below 5%. may disconnect any time from this screen.
see you in a few hours :)
obra safe flight
buu Bye!
obra autrijus, sync the bps repo and look at clkao's new toys 04:27
ingy autrijus: hihi
obra QUEBEC branch
customizable homepage
autrijus nice!
obra also, new preferences API 04:28
autrijus lots of interesting developments indeed
obra also, more tests
ingy ritalin++
obra customers++
autrijus needs to learn to write perl5 again
obra if so, then we should talk when you're in taiwan
autrijus sure :)
obra I may even have interesting RT haxx0ring.
autrijus out of battery &
ingy hi obra 04:29
ready for taipei?
obra heya ingy
hahahahahahah.
ingy ?
buu ingy: So should: *foo*b; just render as text: *foo*b; ?
obra no talk written, no vacation plans done, lots of work here to finish first
ingy buu: yes
buu hrm
ingy I will write my talk in taipei 04:30
definitely
obra I may as well
ingy hi autrijus
buu ingy: I think he's gone
obra half my talk will be "the future of RT" and the other half will be a design session. "let's add a feature to RT"
with a progress report the next day 04:31
ingy cool
buu ingy: Eh, which characters can precede an opening * again? 04:33
Just white space? 04:34
ingy {
sorry...
/ {
buu Er, a slash can come before it? 04:35
ingy obviously you can say /*this*/
buu ah right
ingy other than that, must be a space I think 04:36
the ending tag is different
buu Right.
ingy hmmm, maybe ( too
buu Hmph.
ingy we'll need to play with this to get it right 04:37
write tests and review them for clarity
buu I've got tests
84 so far =] 04:38
ingy nice. I'd like to see them
erxz_paste "buu" at 24.206.236.187 pasted "kwid.t" (48 lines) at erxz.com/pb/508
"buu" at 24.206.236.187 pasted "test.kwid" (40 lines) at erxz.com/pb/509 04:39
buu I'm trying to make *bar*bar* parse as text
ingy I think that is right 04:40
definitely /etc/apache/ should be text 04:41
buu Yeah
ingy one would think
buu I agree.. it just makes it fun to parse
ingy yeah
ok must get off irc 04:43
&
crysflame . 05:54
buu death 05:59
wolverian satan. 06:06
buu Heh, incarnations of immortality++
Aankhen`` Is there any property or something similar in Perl 6 that allows you to restrict a hash to fixed set of keys? 06:35
buu Read the synopsis! 06:41
Aankhen`` Which one? 06:42
buu No idea 06:56
Preferably the one on hashes.
Aankhen`` Er... are you talking about the Synopses? 07:01
S09 doesn't seem to have anything on it. 07:03
nothingmuch morning 07:13
crysflame . 07:23
ingy yay 07:24
I moved Test.pm to its own ext/Test/ dir
where it belongs
and the tests pass
now i need to get modules to install in the perl6 libs 07:25
then we can write modules galore
hi lucs!
lucs Heya ingy :) 07:26
obra ingy, would it make sense to sketch out a naming hierarchy for perl6 modules?
crysflame oo, thinking ahead! obra++
hi, lucs
obra "Skud was going to do some of this 5 years ago"
crysflame (: coral)
lucs Yep. Hello :)
ingy obra: I already have
obra cool. url?
ingy well what do you mean by naming hierarchy? 07:27
have you tried `pugs -V` ? 07:28
obra WWW, DBD, etc,
ingy same as Perl5 right? 07:29
obra so that when people start creating zillions of modules, we don't end up with quite the chaos of cpan.
ingy I don't really care about that
obra I expect that perl6 might have a saner standard library
ingy sure
obra ok. then what bits were you talking about when you said you'd already done it? 07:30
buu ingy: Er, pugs -V errors for me
ingy just figured out where Perl6 modules get installed so that they don't clash with Perl5 ones
obra ahh. cool
ingy buu: cool
buu Oh wait, that's because pugs didn't build. src/Posix.hs:15: pugs_config.h: no such file or directory
Is that supposed to get generated? 07:31
ingy I get:
obra but now, sleeptime
ingy /Users/ingy/local/lib/perl6/darwin-2level
/Users/ingy/local/lib/perl6
/Users/ingy/local/lib/perl6/site_perl/darwin-2level
/Users/ingy/local/lib/perl6/site_perl
.
crysflame ta, obra
ingy so pugs has perl6 libs compiled in... and perl5 ones too 07:32
buu hmph 07:34
lucs Yikes. Is it js 07:44
Is it just me,
or does pugs evaluate int(0.8) to 1?
nothingmuch morning lumi 07:45
lumi Morning
lucs just figured out why the sample shuffle.p6 fails once in a while. 07:47
nothingmuch care to enlighten? 07:48
lucs It's because pugs evaluates int(0.8) to 1 :)
(for example) 07:49
nothingmuch aha
i reckon that is a bug?
behavior should stay the same from perl5?
or should 'rounding' be more dwimmy for newbies (but not for people used to integer arith) 07:50
crysflame or defined as posix?
nothingmuch has no man int 07:51
lucs Wait a sec. Does int() mean something different in Perl6?
nothingmuch i don't think it should
in fact, i'll be upset if it did
lucs I sure am :)
nothingmuch in that case i will pettition for a 'use int qw/sane/' pragma
lucs It looks like a bug, eh.
nothingmuch writes a test 07:52
lucs There is already a t/int.t; it could use more tests I guess. 07:53
nothingmuch uhuh 07:55
todo_is(int(0.8), 0, "floats aren't rounded up");
todo_is(int(0.5), 0);
lucs (er, t/op/int.t)
nothingmuch enough?
0.5 is not a todo
lucs If it fails, probably enough, yeah :)
scw how about int(1.5) == 1?
lucs Probably redundant, but it doesn't cost much.
nothingmuch 1.5 is 1 07:56
1.51 is not
scw so not every program cast 0.5 to 0 but 1.5 to 2 :p 07:58
nothingmuch do we have zip() yet? 08:01
oh well, that's another test 08:02
ok, big t/op/int.t is made 08:13
i would have written it as for((float,\tfloat,\tfloat)\nĀ„(expected,\texpected\texpected)\nĀ„(todo,\ttodo,\ttodo))-> $in, $expected, $todo { if ($todo){ todo_is(int($in), $expected, "int($in) is $expected") } ... 08:15
but we have no zip
lucs Do you have commit rights? 08:16
nothingmuch i have on openfoundry 08:17
but it's down
lucs Ah.
nothingmuch so i use darcs and just let autrijus handle the mess
lucs Well, thanks for building it.
nothingmuch if someone has real commit rights i can put it up wherever 08:18
as unidiff or replacement file 08:19
luqui: can you commit changes to t/op/int.t 08:21
?
lucs found a bug
luqui ahh, where?
nothingmuch i can put it up on my webspace, p6c?
luqui whatever 08:22
your call
nothingmuch ok, one minute please, i'm making sure it's going through a proper make test
ok, nothingmuch.woobling.org/pugs/ has int_test.diff, and a modified t/op/int.t, your choice 08:25
now for t/op/zip.t
luqui making test... 08:26
looks good
committed 08:27
nothingmuch is(~@a, ~@b, "arrays are equiv"); # is this good enough as a test
?
luqui you know, I think it is
there will be some cases where @a is not @b and that will still pass
nothingmuch right, but not trivially 08:28
luqui right
nothingmuch my @a = 0, 2, 4, my @b = 1, 3, 5; my @e = 0, 1, 2,3, 4, 5;
then it evals a @z, and makes sure ~@z is eq ~@e
no one will get angry, right?
luqui don't think so
they shouldn't, at least 08:29
nothingmuch btw
the semantics of @a Ā„ @b Ā„ @c are that you have @a[0], @b[0], @c[0], @a[1] ..., right?
it's stacking that way
luqui yeah
nothingmuch goody 08:30
luqui it's "list associative" as damian calls it
nothingmuch i think that having some ops behave like that is the smartest choice of language design since i god know when
luqui you can do (@a [yen] @b) [yen] @c to get @a[0] @c[0] @b[0] @c[1]...
nothingmuch ooh 08:31
wait
scw make failed with ghc-6.4, Conflicting exports for `getEnvironment'
nothingmuch a, c, b, c, a, c, b, c, right?
luqui yeah
nothingmuch i thought it was magical, but it just makes sense
=)
zip() takes many arrays, right? 08:32
(i thought you were using yen as a slice)
luqui uh huh. it's a semicolon operator
zip(@a;@b;@c)
if you use commas, it will just concatenate
nothingmuch okies
good to know
luqui nothingmuch: do you know much about parsec?
nothingmuch nothing 08:33
luqui nothing much?
nothingmuch i started with haskell on friday
luqui ahh, darn. I'll just have to wait for autrijus then
nothingmuch i couldn't get the get numbers from user and do stuff to them excercise done in under 15 mins
so i went to sleep
and haven't tried again since
nothingmuch finds "programming" in perl6 rather intuitive 08:35
i'm not sufferring too much from "oh crap, it's supposed to be different"
luqui yay
that's what we're going for
nothingmuch i think so far the only thing that has been bothering me aesthetically is colon as a list separator 08:36
my head keeps parsing it like a syntax error
lightstep what is represented as a Syn expression? a syntax tree?
nothingmuch which i usually do when i make hashes
luqui you mean the invocant colon?
nothingmuch (\nkey => foo;)
in perl5 i did that all the time
luqui oh, the semicolon multidimensional separator 08:37
(or rather, two dimensional separator)
nothingmuch so i got used to treating a ; anywhere but the end of an expression, as well as ; at the end of a one line block structure (like if () { foo; }) as syntax errors
in my head
but i'll grow out of it
luqui we should probably disallow null dimensions for that reason 08:38
nothingmuch ok, t/op/zip.t is done 08:41
lightstep is the darcs repository updated frequently? 08:42
nothingmuch autrijus has a minutely sync script
but it's one way
and there are conflicts 08:43
lightstep between what?
nothingmuch since patches you apply get reapplied, but without the logic
svn.perl.org now
svn.openfoundry.org later, i think
lightstep so i'll go learn about svn
nothingmuch uses it anyway, since his mind isn't rubbery enough to fit svk into the mess in one weekend
i'd suggest svk
it works with svn 08:44
but in a more darcs like way
lightstep i don't know darcs, except get and pull
nothingmuch that's really all you need to know about darcs
well, 'record', 'whatsnew'
it's sane
central VCS seems simpler, i think, but they really aren't 08:45
lightstep how do i get/pull with svk? 08:46
nothingmuch there's some instructions somewhere
let me find them
i'm not 100% clear on that 08:47
svk.elixus.org/?SVKTutorial
lightstep thanks
nothingmuch ok, luqui, same ordeal with the zip.t file
unidiff and file
t/op/pipe.t 08:52
ooh 08:56
spooky
eval '@c <== @a' is causing "Fail: cannot cast into a handle: VList [VInt 1,VInt 2]" 08:57
despite the eval
luqui nothingmuch: sry, was away... I'll go get it 08:59
nothingmuch thanks
any clue on what is op with evalling pipe ops?
luqui where was the site again?
nothingmuch nothingmuch.woobling.org/pugs/
t/op/zip.t or zip.diff
luqui thanks, applied 09:00
r583
nothingmuch callooh, callay 09:01
luqui: any ideas for how to test locomotive ops? 09:03
i have @a ==> @b, and @c <== @a, checking that @c = @a, and @b = @a
and i have @a ==> grep {} ==> @b
and @c <== grep @a;
luqui yeah that's pretty good 09:04
nothingmuch and again
luqui you can also test with optional args
nothingmuch explain
luqui sub foo ($x, ?$y, *@z) {...}
nothingmuch ah, user subs accepting pipings
luqui @list ==> foo $x # doesn't bind $y
pipes directly into @z
nothingmuch pretty 09:05
does that also work with foo($x, @list) in the same way?
or since @z is flattenned... ?
luqui your foo won't accept a pipe
nothingmuch wait, if it's flattenned that doesn't matter
luqui it'll just croak
nothingmuch no, not a pipe
"normal" passing of array
luqui I don't quite understand 09:06
nothingmuch if i just do eval 'foo($x, @list)'
it's the same, right?
luqui uhhh... perhaps not
nothingmuch as eval '@list ==> foo $x
luqui I think it will try to bind that to $y
maybe
p6l time
nothingmuch the first element of @list, you mean? 09:07
luqui no, the whole reference
nothingmuch hmm
luqui asks p6l
nothingmuch goody
buu Why is p6 so batshit insane
nothingmuch perl5 is too 09:08
buu Maybe I'm just used to it.
luqui yeah, the difference is that you know perl 5
buu But come on now: @x>>+<<@y
@x ==> foo $x;
nothingmuch that's one fo the nicest things ever
buu That's sheer evil.
crysflame blinks
nothingmuch hyper ops are beautiful
and piping helps readability
crysflame wow, that's interesting
nothingmuch only the invocant is concerned
buu Well, yeah
I kind of like them
But they're definately weird.
nothingmuch crysflame: the first will add elements of @x to elements of @y
crysflame @x ==> foo $x? 09:09
nothingmuch crysflame: see what luqui was just mentioning
crysflame right
luqui the slurpy parameter acts like the functions stdin
crysflame @x ==> foo($x) comes to mind
buu luqui: Er, what?
luqui same thing, far as I know
crysflame oh, jeez
luqui buu what what? 09:10
buu luqui: What did you mean by 'acts like stdin' ?
nothingmuch buu: it takes pipes
luqui when you pipe into a function using ==>, it goes directly to the slurpy parameter, bypassing positionals
(which had better be optional)
buu Oh. Good.
nothingmuch ok, 09:14
t/op/pipe.t is ready
putting up
done
lightstep recursion it really really broken!
s/it/is/
nothingmuch note that without modifying the test two tests will never untodo 09:15
luqui thanks, applied
nothingmuch because they have fatal runtime errors
the eval '' is dying
hence commented out
ok, that's enough p6 for today 09:16
nothingmuch goes to rip a cd
luqui lightstep: how so?
lightstep hanoi.p6 always gets 'S' in $s 09:17
i can't say more until i hack Pretty.hs
09:17 Aankh|Clone is now known as Aankhen``
nothingmuch are the exegis examples made into tests yet? 09:18
luqui dunno
nothingmuch i would be glad to translate them and make sure they work like they should, by running unit tests for those programs
s/running/writing/
luqui that'd be pretty cool
you know the earlier ones are wrong nowadays, right? 09:19
nothingmuch i will try to fix them up as i goi
needless to say i will be bitching here and on p6l a lot
luqui good
nothingmuch now to get chromatic to write Test::Builder like he promised 09:20
;-)
hola orb 09:32
theorbtwo Allo, nm. 09:36
nothingmuch what beith new? 09:37
nothingmuch is at work, his laptop's CD is broken, and someone wants him to listen to something 09:40
what's an x11 based cdripper usually available on redhat?
hopefully based on cdparanoia + ogg or lame 09:41
luqui grip?
nothingmuch sounds good
luqui lame encanta a luqui 09:42
09:54 buu is now known as buu_sleeping
nothingmuch loop (1 .. 20) is the "right" way to get 20 iterations, correct? 10:04
Aankhen`` Yea. 10:41
nothingmuch 100% sure it's also allowed? because for (1 .. 20) works, bug loop (1 .. 20) is a parse error, and is also not tested for 10:54
wolverian nothingmuch: then loop is unimplemented. 11:04
nothingmuch test test test
wolverian nothingmuch: there's nothing wrong with 'for 1..20 -> $x { print $x }' though.
nothingmuch without needing the counter, i'd like to just do some block 20 times
wolverian for 1..20 { ... } 11:07
'loop' is meant exclusively for c-style loops, as far as I know
(loop my $i = 0; $i < 20; $i++ { ... })
nothingmuch ok, that's what i was checking
luqui: want to look at exegesis 1's example? 11:08
the fileserver is down so i have a free 10 mins
luqui okay 11:09
luqui doesn't remember exe1
crysflame (1..20) ==> for $x { print $x } ?
nothingmuch tree example
crysflame++
nothingmuch only wants for (1..20){ do_something_unrelated_to_x }
crysflame osmosis works!
nothingmuch err, exegesis 2 11:10
it's #1, but it's 2
luqui crysflame: (1..20) ==> for -> $x { print $x }
that acutally, er, works. but maybe not
nothingmuch my int ($pre, $in, $post) --> my ($pre is int, $in is int, $post is int)?
luqui sub statement:for (*@vals, *&code)
nothingmuch my ($pre, $in, $post) is int, constant = (0 .. 2)? 11:11
crysflame luqui: heh. i was just making up syntax.
(1..20) -> for $x { ... } ?
nothingmuch for is a sub that also looks at it's code's prototype and then decides how to dispatch @vals?
luqui crysflame: sure, but I don't think you can use pipes on for 11:12
because it's statement-level
crysflame hmmm
nothingmuch in the for (@arr) -> type syntax
crysflame ok
luqui nothingmuch: hm?
nothingmuch ignore me
dev.perl.org/perl6/exegesis/E02.html
the "my int ($pre, $in, $post)" bit doesn't look right 11:13
luqui nothingmuch: I think it's right... we really haven't cared enough to define that part of the language
we'll probably allow both: my int ($x, $y) and my (int $x, int $y) 11:14
because they're both unambiguous
nothingmuch ok
and constant is a trait, not a part of the type, right?
well, duh
is rw
crysflame print (1..20) ==> int ==> join ',' ? 11:16
or is that the wrong idea
wolverian 1..20 ==> map { int } ==> join ","
luqui wolverian: precisely
==> print on the end if you please 11:17
wolverian 1..20.map:{ int }.join(",") # or is the : optional again or what? :)
luqui don't think it's optional in that form
definitely not, actually
crysflame hmm, ok
wolverian does map understand only adverbial blocks? 11:18
crysflame 1..20 >>==><< int ?
wolverian or can I pass it a regular block as .map({...})?
also, is the precedence right with the .join directly after the adverbial block?
luqui wolverian: the precedence is okay
adverbial blocks go in *& parameters
wolverian ah, okay. 11:19
luqui and I don't know how you get something into the *& parameter without using an adverb, but maybe there's a way
crysflame that is, given a func that takes one thing, can i replace any use of map { func } with >>...<< func?
luqui uhh...
wolverian multi sub map (Code *&c, *@rest) { ... } # would the slurpiness DWIM there?
luqui I think *& is a zone just like *$, *@, and *% 11:20
and it has to go at the end, because that's where it goes in for
wolverian mm. okay.
I'll need to reread that bit. :)
luqui but @Larry is thinking about removing the zone restriction
wolverian ah, okay.
luqui and I don't think *& has been written in any official doc yet 11:21
(maybe it has)
wolverian I don't remember it, but again, I don't remember the zones very well either.
luqui no, there's a passive mention in A6 about *&
wolverian crysflame: >>...<< only works with infix functions, I think. :)
luqui you can do it with prefix, but then it's ...<< 11:22
and postfix: >>...
like, @objs>>.meth
crysflame hmmm
wolverian do we have any postfix functions that aren't methods? :)
crysflame oh, neat
luqui well, @nums>>++
although using prefix there might be clearer... just maybe 11:23
crysflame (rand x 20) >> .int >> join ',' >> print?
wolverian that could easily still be a method, though. which I think is a good thing. probably.
crysflame: ==> you mean.
crysflame errr 11:24
(rand x 20) >> .int ==> join ',' ==> print?
wolverian or maybe: (rand x 20)>>.int>>.join(',')>>.print
theoretically that might work. :)
crysflame >>. >>. >>.
i feel like i'm playing zork orsomething
wolverian heh. 11:25
crysflame >> applies a list to a .method effectively, while ==> pipelines the left side into the right side's arguments, presuming the right side isn't a method (infix functions only?)
so (rand x 20) ==> .int is bad
(rand x 20) >> .int is good
nothingmuch luqui: first bits of exegesis 2 is "my $var is trait' 11:26
i'm writing t/base/traits.t
luqui (rand) xx 20 ==> .int means $_.int((rand) xx 20)
nothingmuch testing for 'my $const is constant = 2'
crysflame oooh, okay
so ==> is slurp args
wolverian luqui: is it xx nowadays?
luqui: oh, is xx list constructor and x string?
luqui x is for strings, xx is for lists
right
nothingmuch testing that you can't assign to $const later on 11:27
crysflame so when i said >>==><< i meant >>
luqui probably
crysflame since the hyper form of ==> for a single list left side and a single arg right side is >>
nothingmuch and testing that it's value remains after trying to assign
any more ideas?
wolverian @array_of_arrays >>==><< stuff; # ...that is so ugly.
crysflame yep 11:28
wolverian s/<<//
crysflame oh, gack
luqui well this really gets into testing traits
crysflame ok
nothingmuch that's the origin of the dead fish operating
or
s/ing$/or/
not object traits
that's for the 6.2 tests vs. apocalypse 12 stuff to define later 11:29
luqui container traits
wolverian hrm, this should work, right: for @aoa -> @a { ... }
nothingmuch yes
crysflame @a, @b >> sum ==> ',' -> join ==> print ?
sums each of $a[0], $b[0], etc
joins with , prints.. hm.
there's going to be some mad crazy learning required for all this
luqui wolverian: yep
wolverian -> join should be a syntax error. 11:30
luqui well...
wolverian (as -> wants a code block, probably)
luqui yes I think that works
crysflame: what's that -> doing in there?
wolverian also, you want (@a,@b)>>.sum
(are the arrays flattened there? I forget what the policy is now.)
crysflame i'm learning syntax by probing the edges of it
wolverian (if it's changed.)
crysflame will stop, as i've hit the brick wall of the moment 11:31
wolverian crysflame: >> and ==> are only two operators.
it's not _that_ complex.
luqui I don't know if they're flattened there, actually
wolverian let me see what pugs does. I think it was a bug though, regarding that.
crysflame i would think that if the rside of >> expects two operators and gets two on the left side then.. dunno
wolverian: >>, ==>, ->, and >>+<< went by int he past hour or two 11:32
luqui anyone seen autrijus
nothingmuch probably asleep
how would i test for 'is chomped'?
in perl5 i'd prolly do a pipe open
and nonblocking write
so that it doesn't die
wolverian wow, my pugs is broken.
luqui you could also open yourself
nothingmuch ah
good point
lumi Hi, sorry
luqui hi lumi
crysflame hi
luqui you've been in and out all day 11:33
nothingmuch how would you name the test for $0?
lumi has been scholded
luqui $0?
nothingmuch program name
luqui that's not the program name anymore
nothingmuch ah 11:34
$?...
luqui it's more like $?PROGNAME or something
$0 is the match object
wolverian I still think it'd be nicer to objectify that kind of variables.
$SYS.progname # or so.
nothingmuch t/magicals/progname.t?
wolverian or taking a page from Larry, $SYS<progname> :) 11:35
nothingmuch so should i just stub $SYS.progname in the test?
wolverian er, no, that's my delusions.
I'm not authoritative.
use $?PROGNAME for now. 11:36
luqui yeah that should be fine
nothingmuch todo_is($?PROGNAME, "t/magicals/progname.t");
well, (eval '$?PROGNAME', ...)
john examples/hanoi.p6 gives the wrong answers, algo is sound, probably something worng in the argument passing. Turned the example into a test, where should it go? t/examples/hanoi.t maybe?
lumi How do you test that? Anything might call the test, no? 11:37
nothingmuch someone else mentioned it too, john
perl t/test.t yields '$0 eq 't/test.t' under perl
assuming it runs under the pugs test harness, where it should, it shouldn't fail
anything else is the test runner's responsibility
john any need for the hanoi test code?
nothingmuch they must open the test, and read the comment that says "run me properly" 11:38
wolverian heh, latest pugs: 11:39
Failed 2/55 test scripts, 96.36% okay. -5/1003 subtests failed, 100.50% okay.
nothingmuch hehe
wolverian I am more than okay!
nothingmuch do we have rx// like quotes yet? 11:41
wolverian we don't have rules yet.
nothingmuch that could be todo, i mean syntax wise
qq//
wolverian yes.
nothingmuch i see 11:42
t/op/quoting is all todo
almost 11:43
wolverian well, say qq/foo/ works :)
nothingmuch but that's about it =) 11:44
wolverian qq() works as well!
nothingmuch how silly would it be to create a todo_like sub in Test
wolverian q// does not.
nothingmuch which always says 'not ok name # TODO like() is not yet implemented
todo_like(eval '$?PROGNAME', eval 'rx#t/magicals/progname\.t$#', "progname var matches test file path"); 11:45
?
beh, qw((foo bar)) does not yield '(foo', 'bar)' 12:04
Juerd It does not? 12:12
nothingmuch nope
should it?
should i add a test to t/op/quoting.t?
i will 12:13
either 'q((foo bar))' yields '(foo' 'bar)' or 'q(())' does not parse
Juerd w
nothingmuch one of them is the right behavior, i just don't know which
Juerd The former
Even qw<<foo bar>> results in '<foo', 'bar>' 12:14
only without qw, << is special in meaning ōæ½xAB
nothingmuch shouldn't that be a lone <<>>?
ah
Juerd See [email@hidden.address] in p6l 12:15
wolverian shouldn't q((foo bar)) yield _one_ string?
nothingmuch qw
i think yes
wolverian your original question had two elements 12:16
I think. :)
er, sorry, it had qw.
Juerd wolverian: Yes - we were talking about qw though :)
wolverian you just later used q(), which confused me.
Juerd So the q must be a typo
wolverian I blame it on having to use windows right now.
nothingmuch is full of typos
wtf? 12:17
google groups finds that message id
but if i click the title it says 'perl-perl6-language doesn't exist'
feh!
Juerd nothingmuch: tr/-/./ 12:18
Bug in GG
nothingmuch goody
Juerd Google's new Groups interface SUCKS 12:19
nothingmuch should i also test q\0 \0
Juerd It doesn't parse quotes in a useful way
nothingmuch where \0 is really a char?
Juerd Did they code this in PHP or what?
nothingmuch that char, not just a char
Juerd nothingmuch: Yes, that should work.
wolverian I think I read google uses python, mostly.
Juerd nothingmuch: It's a much used technique in generated code, even
wolverian: Python and C, yes
wolverian which is lightyears better than PHP. 12:20
nothingmuch a friend of mine was thinking of applying for a job
the skillset they wanted was c/c++, python, perl
eval '@quote3b = q:2/$foo $bar/'; 12:21
todo_is(@quote3b[0], 'FOO',
q:2 means q:double?
Juerd What's :double?
nothingmuch qq
Juerd omg.
nothingmuch yes =(
wtf?! 12:23
q/$foo $bar/ returns two elements, '$foo' and '$bar'?!
why god why?!
12:24 Aankh|Clone is now known as Aankhen``
nothingmuch somebody please look at lines 24 and 26 of t/op/quoting.t and tell me i'm reading wrong 12:32
wolverian ..that looks wrong. 12:45
but then again, I don't know the latest designs on quoting. :) 12:46
qmole /8
nothingmuch well, i'm off 12:58
autrijus greetings! 15:43
lumi Hello 15:46
autrijus hey.
openfoundry is back
merging everythiing back now
Run Perl6 Now! | pugscode.org ('Overview', 'Journal') | pugs.kwiki.org | channel log: xrl.us/e98m 15:48
cls_bsd greak 15:51
in such time it proofs svk's power 15:52
autrijus =)
metaperl good news 15:55
autrijus the bad news is that yesterday, the final night in beijin
we had this great idea of eating out at a "regular folk" foodplace for once 15:56
instead of eating in Proper Restaurants like we did before
result: serious and extremely painful case of diarrhea for the past 24 hours :-/ 15:57
clkao poor autrijus
cls_bsd oops, after svk up -s, my pugs is still 6.0.9, but there's 6.0.10 in CPAN
autrijus cls_bsd: rerun perl Makefile.PL ?
cls_bsd poor autrijus
scw autrijus: //pat. That always happends it China.. 15:58
autrijus nods. should've known better
cls_bsd so what's your choice next time in China? :p
clkao autrijus: great, you just hit svk's bug. log message became "1"
it was fixed last night
autrijus cool.
cls_bsd grep VERSION lib/Perl6/Pugs.pm
$Pugs::VERSION = '6.0.9';
ok, re-run svk up -s again 15:59
ok, updated... 16:00
statico whoa, did i miss chromatic on irc? 16:03
autrijus a few days ago, yeah 16:04
the interview
openfoundry svn fully synced up. 16:05
ingy hola 16:06
autrijus: I moved Test.pm into ext and got it working 16:07
now I need it to install in the perl6 paths
autrijus cool. 16:12
so tell me the plan now
is it still possible to use upload p6 modules to CPAN?
or do we drop that route?
I somehow still think reusing Perl6::lib::* is attractive, but if you have a better idea I'm all ears 16:13
ingy pugs can load both perl5 and perl6 modules 16:15
it has paths for both
autrijus not seeing that
ingy the perl6 ones come first in @INC
autrijus you seem to have moved away p5lib
ingy I am describing the future
autrijus ahh ok. go on
ingy perl6 modules use a pugs aware build system 16:16
that installs them into the perl6 trees
very simple
autrijus so we use pugs to run Makefile.PL? 16:17
ingy as for cpan
autrijus: eventually
autrijus ok. what about cpan?
i.e. what should we tell darren?
ingy but that is not the future I am talking about
autrijus nod.
ingy hold on
so say I write a perl6 verion of a perl5 module 16:18
autrijus grabs some pain killers; type ahead, I'll brb
ingy eventually cpan will need to be able to accomodate perl5 and perl6 modules of the same name 16:19
autrijus p6 modules have this "long name"
ingy name a module for instance
autrijus so ideally it should be uploaded under the long names
which includes the author etc
ingy my cpan solution is to simply change the dist name 16:20
Perl6-*
for the short term
autrijus Perl6-* is bad
see existing CPAN modules
there's loads of Perl6-* 16:21
and they are p5 modules.
which is why I proposed Perl6-lib-*
ingy decent point...
autrijus I pondered v6-*
to pun with "use v6;"
but I think Perl6-lib-* is clearer
ingy I don't grok Perl6-lib- at all
what is the trickery that led you to that?
autrijus twofold: 16:22
in CPAN we need a dinstinct prefix for distnames
in the short term
to mark p6 modules
Perl6-lib-* is the most reasonable thing I can think of.
now, the obviously easy thing to do with regular makemaker 16:23
ingy it seems rather awful
autrijus sure. better suggestion being?
ingy why cruft up the perl6 module layout ffrom the author perspective
autrijus I think v6-* is too vague.
ingy everything should remain the same
and the tools should dtrt
autrijus uh. sure. 16:24
so I agree w/ your plan
we could have a M::I extension
that simply prefix the distname with Perl6-lib-
but install things under p6 libparh
libpath ev
libpath even
and keep everything else the same
ingy I completely agree... 16:25
autrijus but the prefix should still be there.
so the extension needs to
ingy except Perl6-lib- is sooo not shiny
autrijus 1. prepend something to distname
ingy v6- is pretty shiny
autrijus 2. find pugs or perl6 executable and run tests with them
CPAN Upload: v6-DBI 16:26
CPAN Upload: Perl6-lib-DBI
I wonder if we should have a poll...
I think v6-DBI may mislead people thinking that's dbi v6
ingy I really think the prefix is not the right thing long term 16:27
cpan needs to be reorganized
thus freepan or a lot of work
autrijus agree 16:28
in the long term we are all reborn anyway
but in the short term.
ingy we can go with the prefix for now
autrijus with a clear mark that it is purely transitional 16:29
ingy but we may see it still in 20 years :P
autrijus well yeah
but without a solution
some CPAN authors may not bother to convert things to p6
which is a loss :)
deadlocking bad, makeshifting good (I think)
ingy sooo... I think Perl people adopt fast enough for v6- 16:30
autrijus so you think v6- is the way to go?
ingy definitely
Juerd There'll be an army of porters! </dream>
autrijus there will.
mm I'll ask on #perl a bit.
ingy when there are 500 v6- modules, Perl people are smart enough to know they aren't all at version 6 :P 16:31
Juerd I don't like the (or any) prefix
pjcj Has perl6 got v-strings?
ingy Juerd: solution?
we don't like the prefix either 16:32
autrijus pjcj: v-objects, yeah.
Juerd I think CPAN should be left for what it is and a new repository should be created
lukhnos /win/win 5
ingy freepan!
autrijus Juerd: I agree.
lukhnos oops sorry
autrijus however, recreating the toolchain takes Time.
Juerd This might be a directory in CPAN, as to abuse existing mirrors
But it should be differently structured
pjcj v6 won't clash with v-objects?
autrijus mm? v6-DBI-1.0.tar.gz 16:33
it's DBI.pm inside still
Juerd CPAN is really great, but every Perl person you meet can give you a list of things it should improve on, but it cannot improve for it's monolithic.
ingy autrijus: help me create freepan in taiwan
we can use foundry for much
Juerd autrijus: v6 implies a version, and that's what I don't like
autrijus that's also what I fear. 16:34
ingy bah!!!
Juerd Maybe a change like going from - to _ is enough to indicate the difference
ingy that's so silly
lumi p6- ?
Juerd Or the version number notation
autrijus mmm p6.
ingy after 3 days, EVERYONE WILL GET IT
Juerd p6 is better than v6, but still includes a version number...
ingy so it really doesn't matter what it is. just keep it as short as possible 16:35
autrijus well for that matter, "Perl6-lib" includes a version number :)
Juerd If we use a prefix, we will be stuck with that for the next 20 years, even long after perl5's use has dropped to zero.
autrijus that is not a module name prefix.
it is merely the tarball prefix.
ingy I agree with Juerd on that
Juerd Doesn't matter
Perl 6 isn't "Perl 6" years after its release
It will be "Perl"
autrijus true. but then, we'll have 6PAN/FreePAN 16:36
and it can drop the prefix
Juerd You know that will not happen
autrijus I don't know. ingy is still alive after all
I'd be happy if we can find a better short term solution
than p6-*
ingy oops, I have to run. I'll code this all up by tonight.
Juerd Something as simple as defining a format that is incompatible with p5's is enough 16:37
I'm very tempted to suggest a non-.tar.gz suffix
ingy autrijus: I have a better short term solution
Juerd For something that still is .tar.gz, though
autrijus ingy: pray tell
.par.gz ;)
Juerd As an example: Debian packages (.deb) are really just ar archives (.ar)
ingy don't upload to cpan. just have people use your pugs repos for perl6 modules for now. 16:38
that will buy time to get it right
autrijus not thinking that scales, though
ingy try it
lumi Which makes sure we won't see it in 20 years?
autrijus so hmm. so instead of applying a PAUSE account 16:39
apply an openfoundry account
ingy yes
for now
autrijus instead of being a pause uploader, become a pugs committer
instead of managing your own release
let pugs release for you
ingy and we turn that into freepan for mirorring
all of Kwiki is under SVN as well as CPAN, and I don't use CPAN anymore 16:40
it is the new way
anyway it buys us a month or two
cls_bsd ha
clkao by that time we'll have svkup for mass distribution 16:41
ingy really have to run. ttul!!
autrijus I can see that.
CDybedahl t/magicals/env.t tests for $TEMP, which is not set on any OS I have running (FreeBSD, Solaris, OSX), so that test fails. Since the test just wants a pre-existing env var, I think $PATH would be a much better choice. 16:43
autrijus I think we should just zap that test.
CDybedahl That works too :-) 16:44
cls_bsd let pugs release you # autrijus> let pugs release for you
XD
autrijus so you folks are ok with the delay tactic ingy proposed? 16:46
i.e. if you want to write pugs/perl6 modules, just apply pugs committership for now
and put your modules in ext/ 16:47
rooneg as long as it's just a short term solution that seems fine... once pugs starts getting objects and rules we'll need something better though, as module writers will start coming out of the woodwork... 16:48
autrijus yup. 16:49
pjcj yeah - I think it works short term
cls_bsd ex/t
ext/
oops
rooneg i mean for now it seems like there are a rather limited number of modules people are likely to write anyway
autrijus well except we're telling people to write imaginary modules 16:50
but yeah, I think I'll update PA01 to say that. 16:52
rooneg hey, if someone with commit access has a sec, svn:ignore in src/ and ext/Kwid/ could stand to be updated... it'd make my 'svn status' output look so much nicer ;-) 16:53
autrijus can you do it yourself? 16:54
rooneg at electricjellyfish.net ? :)
rooneg yep, if you want to give me access ;-)
autrijus invitation sent 16:55
rooneg thanks
autrijus np :)
welcome aboard
cls_bsd Failed 11/62 test scripts, 82.26% okay. 46/1071 subtests failed, 95.70% okay. 17:11
not bad, though
autrijus huh? 17:12
it's all pass for me.
what did you do? :)
cls_bsd make test 17:13
:p
autrijus weird... latest openfoundry?
what failed?
cls_bsd sure, just svk up
hmm
such as line 78,79,81 in 03operator 17:14
ok, I should svk up again@@"
rooneg doesn't get any failures either
metaperl those look like my junction tests (without checking operator.t)
maybe your pugs binary is old... make install did not overwrite my last one and I had a similar problem 17:15
do pugs -v
cls_bsd no, I thought I update t/*, but in fact I didn't :~
autrijus pebcak :p 17:16
cls_bsd re-make again 17:17
Not in MANIFEST: t/op/assign.t
Not in MANIFEST: t/op/pipe.t
Not in MANIFEST: t/op/zip.t
so I missed something?_?
rooneg nah, i think manifest just hasn't been updated in a little while 17:18
lucs I think 'nothingmuch' added the pipe and zip tests just a few hours ago.
autrijus yeah. commits welcome, etc.
cls_bsd oh, I just afraid it's my own problem :p 17:19
hcchien all tests successful for me. :)
john At revision 467. all OK
cls_bsd great, now I saw lots of "....ok" 17:21
Files=1, Tests=1, 1 wallclock secs ( 0.84 cusr + 0.02 csys = 0.86 CPU) 17:23
autrijus t/op/int all passed. 17:36
lucs !
autrijus committing
lucs Fantastic!
I love you all!
autrijus lucs++ # welcome aboard, our 20th committer
r471.
theorbtwo Going back a little, re package formats... 17:38
autrijus yes?
theorbtwo .tar.gz has the bad side-effect of needing to penetrate more levels before you can get any useful metadata.
offby1 where should I send a tiny unit test for pugs, which demonstrates a failure?
theorbtwo How tiny? Just paste it? 17:39
offby1 yup, that tiny.
theorbtwo Paste away, then.
offby1 excluding the prelude, which you can probably guess: 17:40
sub modify_readwrite (Int $x is rw) {$x = $x + 1;}
ok (8 == modify_readwrite (7), "read-write parameter can be modified");
I'm guessing I'm making a mistake somewhere, or failing to understand something, since this is so simple.
When I run it like ``pugs /usr/local/src/pugs-6/t/is_rw.t'', it says 17:41
unexpected "s"
expecting ";" or end of input
NonTerm "/usr/local/src/pugs-6/t/is_rw.t" (line 8, column 1)
rooneg .tar.gz has the benefit of not getting into a huge debate over the virtues of various archive formats ;-)
theorbtwo Is there a missing ';' at the end of the previous line?
True enough.
But it doesn't leave much for 'file' to recognize.
offby1 theorbtwo: I don't think so. If I omit the "is rw", it compiles, and then (correctly) fails at runtime. 17:42
theorbtwo I'd start by stealing the .deb format, then modifying from there.
(Quite possibly by making the metadata use yaml for starters.) 17:43
lucs I'm pretty much of a Perl6 newbie, but does it make any sense to be able to make '7' writable?
theorbtwo No, it doesn't, but that's certianly the wrong error message.
lucs Oh.
autrijus "is rw" is not parsed. 17:44
to test it, put it in eval. :)
theorbtwo Shouldn't the parser just accept is \w+ ?
autrijus I'm also think that eval_ok and eval_is works
right, that's the idea. it's just not implemented.
so, oil for food, I mean, test for code, as usual :) 17:45
offby1 But I want it to succeed, not fail. Don't I?
theorbtwo Yeah, it should be a todo test with an eval.
offby1 oh wait, I think I see the problem. I'm modifying a constant.
ahem
autrijus I'd be happy if someone can hack Test.pm to add todo_eval_is and todo_eval_ok :)
rooneg proposes tying autrijus to a chair and making him implement is rw. or perhaps given. one of the two ;-)
autrijus I'll do that tomorrow now I have respectable bandwidth :)
theorbtwo Back in .tw/~autrijus ? 17:46
autrijus yup.
I need to sleep soon.
offby1 autrijus: I take back my last comment -- the problem is in compilation, not runtime. pugs says 17:47
unexpected "s"
expecting ";" or end of input
NonTerm "/usr/local/src/pugs-6/t/is_rw.t" (line 8, column 1)
autrijus hopefully when I wake up there will be proper "is rw" tests :)
offby1: right, because parser doesn't grok traits.
it's to be expected.
offby1 ok
autrijus so it sees a "NonTerm".
offby1 is there a workaround?
rooneg starts looking at adding todo_eval_is and todo_eval_ok
autrijus which is a facy word to say it's not terminating
offby1: sure, use 'eval'
to test parse bugs, always use eval.
offby1 I don't want to test parse bugs; I want to test read-writeableness. How can I use "eval" to mean "this paramater is writable"? 17:48
theorbtwo The workaround is what you just did: tell the chan that it's not working, get a test case in, and somebody will make the test case pass eventally.
autrijus offby1: you do this, eg:
eval 'my $sub = sub ($foo is rw) { $foo = 3 }; $sub($x)';
and then test if $x is 3.
theorbtwo todo_eval_is(eval 'sub plusplus($foo is rw) {$foo++;} my $x=3; plusplus($x); $x', 4); 17:49
offby1 you're expecting that to fail at the moment, right?
theorbtwo Yep.
autrijus yup. eval'' makes it fail in a harnessable fashion
journal up. take care, folks. :)
theorbtwo todo_... tests are for things that you presently expect to fail.
autrijus zzz &
theorbtwo Sleep well.
offby1 gotcha 17:51
sorje is increasingly fascinated by pugs.. must resist.. exams next week.
offby1 what's remarkable is how short pugs' source code is. 17:52
sorje And the speed of the project is astonishing
offby1 What's the p6 equivalent of perl5's $@ -- the variable that holds the text from the last exception? 17:53
statico sorje: join the club 17:56
sorje The burden of a real life. ;-)
Odin-LAP Fuck real life. 17:57
Viva la revolucion! 17:58
autrijus offby1: afaik that's unspecced. which may mean it's the same as perl5, or it's all united under $!. ask p6l? (I'm pretty sure it's p6l this time.)
offby1 autrijus: I see some of the existing tests use `$!', but that's wrapped in an "eval" ,which implies it doesn't yet work.
autrijus right. 17:59
theorbtwo p6l
Odin-LAP autrijus: Ehrm ... exceptions are pretty well defined, I think.
theorbtwo (p6l is when you aren't sure what the language is; p6c is when you you have questions about the compilers.)
Odin-LAP I think Apocalypse 4 tackles them rather verbosely...
offby1 For playing with pugs, I thought I should confine myself to Synopses and Exegeses. 18:00
autrijus A04 covers eval? 18:01
not seeing it covering eval"".
only eval{}
but my head is dizzy so maybe I'm missing something 18:02
offby1: Apoc is actually okay for things not specced in Syn
Exegesis is actually the least normative of the 3.
offby1 all dem Greek words! 18:03
theorbtwo When the Synopses and Apocolapses disagree, it's probably because the Syn has been updated more recently... but Apocs will sometimes cover things that Syns do not, because the Syns are meant to be... well, synopsies.
Odin-LAP autrijus: No, but it specifies exceptions... :p
And I thought there was no eval {} anymore..? 18:04
autrijus yup there's try{} now
Odin-LAP Erm. 18:05
rooneg does eval have a return value? I'd love to be able to get data out of eval '$x eq $y', for example, without having to construct code and stick it in the string i'm evaluating...
Odin-LAP YOWCH!
There's lots and lots of talk about exceptions, and how exception handling and stuff should work in Apoc 4 ... almost none of it is in the synopsis...
autrijus eval does has a retval
rooneg in that case i need to figure out why my code isn't working ;-) 18:07
offby1 finds his answer in A04 18:08
rooneg sees the problem... need a way to eval in caller's namespace... 18:09
autrijus ok, $! after eval is implemented. 18:11
rooneg: hmm. is there a specced way in p6 to do that? 18:12
rooneg i have no idea
autrijus in pugs we can use caller continuation to do that
but it's not considered proper p6 now
ask around?
rooneg goes to look at the specs
autrijus r472. now really zzz & 18:15
btw, darcs and svn.perl.org should now both be synced per minute. 18:16
john r472: src/Prim.hs:284: Variable not in scope: `glob' 18:19
rooneg here's an interesting one. $CALLER::foo doesn't play nicely with := 18:27
sub foo { my $x = "foo"; my $y := $x; bar(); } sub bar { return $CALLER::x eq $CALLER::y; }
in bar when it evaluates $CALLER::y it looks for $x in the local scope, not the caller's scope 18:28
nothingmuch evening! 19:08
ah, we have a pasteling, how fun 19:15
hola stevan 19:20
stevan hola nothingmuch
did you get drunk and dance the other night?
nothingmuch no 19:22
we drove an hour
got drunk
saw the line, went back to the car
(driver did not get drunk)
drove to another place
didn't find it
drove back to the first
waited some more in line
helped an idiot who got his jeep stuck in the mud wiggle it a bit 19:23
and then went home
disgruntled and looking to make something of the night i started with YAHT
stevan sounds like lots of fun :)
nothingmuch since i was drunk i went a little slow, but i got somewhere eventually
stevan Drunken Haskell :P
nothingmuch beh
autrijus ows me big time now... 19:24
he promised try/catch
and by now i've done a bajillion more tests
i wonder if i can make him sign a contract to implement features i test for
stevan did you read his journal? he is apparently not in good shape 19:25
nothingmuch: you did contribute didnt you?
nothingmuch i did read it
many of luke's commits from today were tests i submitted
zip.t, pipe.t, fixed int.t
then the file server came back up
stevan excellent 19:26
nothingmuch so i could resume real work
stevan hmm
Prim.hs seems to be broken,.. I cant make the latest
nothingmuch now i'm back from work and orchestra
so i can get back to fun
but first! coffee beckons
nothingmuch wonders if autrijus could set up a subscription payment link 19:28
if forgot how to whip up the URL
lucs nothingmuch: "work and orchestra" Are you a professional musician? 19:35
nothingmuch no
orchestra is my hobby
lucs Ah.
nothingmuch just started playing the bass this year
lucs Cool.
nothingmuch so i get to play in a little kids thing
my teacher's old teacher organizes it 19:36
lucs What are you guys working on?
nothingmuch 3 pieces right now
19:36 edgewalk1r is now known as edgewalker
nothingmuch bach's brandenburg symphony, 19:36
mahler's 3rd symphony
and something i don't know yet
lucs NIce stuff.
nothingmuch and i'm too lazy to get up and grab the sheet music to find out what it's called
it's very silly
we have 30 violinists 19:37
3 of them boys
3-4 are advanced, and can really get it right most of the time
the rest are at my level, sort of
there's me
lucs Oh well. I'm sure it's fun anyway :)
nothingmuch and 3 cellists on rotations
and a recent addition was someone playing a recorder
he is surprisingly good 19:38
for his age he seems to understand
more than most
anywho, it's good fun, even though we sound horrible
and i try to take it a bit seriously
at work i'm an integrator and whore-geek
lucs Were you playing another instrument before the bass?
nothingmuch when there's something no one else can afford to learn how to do i'm tasked with it 19:39
guitars
which i still occasionally dabble with
and friends' drums
but not enough to say i play
lucs The fretless aspect of the bass must be interesting :)
nothingmuch and piano
i have a fretless fender jazz bass
i consider it a guitar
lucs Ah, cool.
nothingmuch it's very different techincally from the double bass
regardless of the arco which is altogether different 19:40
lucs We don't hear of Fender bows, for example :)
nothingmuch and this time i'm getting tought right too
taught?
yes, speller likes that
well, not only the bowing
the left hand works very differently
i know the notes and where they are, and knew them when i started 19:41
but the control is gained quite differently
hmm 19:42
do we have anything i can use to truncate a string in pugs?
substr doesn't seem to be there
rules aren't either
stevan chop and chomp I think are implemented 19:43
nothingmuch reverse? 19:44
stevan dont know
nothingmuch ugg
i can probably fake a substr with chop, and array ops
if it returns correctly
chop is not tested 19:45
nothingmuch writes
wow, many bugs in chop 19:51
my @array = ("Foo", "bar");
chop(@array) 19:52
@array is now eq ("foo ba")
doesn't even work as chop($str, $str2)
ugh! %hash = ( "key", "value", \n ... ) doesn't parse 19:58
Khisanth perl6 still has chop? 20:08
nothingmuch why not?
Limbic_Region but doesn't it have a pohc? (remove first character of a string) 20:09
integral character, byte or grapheme? ;-) 20:10
20:10 buu_sleeping is now known as buu
Limbic_Region code point of course 20:10
nothingmuch does anybody have the capacity to fix r472?
Khisanth nothingmuch: do people really have to remove one character that often that there needs to be an operator for it?
nothingmuch $! being set after eval fail is broke compilation of Prim.hs
i don't know 20:11
shapr ah, is thatwhat happened?
nothingmuch i think removing a feature that is not problematic is not a wise choice for backwards compatibility
i use it occasionally
although i use chomp more often
especially on arrays
especially with -F
Khisanth backwards compatibility and perl6 don't exactly go together considering how many things it breaks
nothingmuch or rather equivelently oriented programming 20:12
well, it breaks things around sytnax
or features which really shouldn't be there (formats)
but a basic, unintrusive operator...
beh
i wouldn't really mind it being put in perl6's Scalar::Util
how do i specify which tests to 'make test'? 20:14
shapr How do I run the tests? 20:18
nothingmuch 'perl Makefile.pl'
.PL, sorry
shapr ohh, make test from the line above maybe?
nothingmuch 'make' to compile
'make test'
or './pugs t/op/test.t'
or 'export HARNESS_PERL=$PWD/pugs'
and then 'prove t/op/test.t'
shapr ok, I think fixed the the Prim.hs not compiling. 20:19
nothingmuch *kiss*
for testing 'my $var is constant'
t/op/is_constant.t?
t/op/is.t?
t/op/container_traits.t?
xerox Why you haskell people are into perl6 ? 20:20
shapr Which Haskell people? 20:21
nothingmuch shapr - did you commit?
shapr No, because I'm using darcs and my local mail agent is broken. But the diff is one line :-) 20:22
nothingmuch i gave up on darcs
i'll have to make svk darcs native to be happy
xerox shapr, you and other I saw, I think.
nothingmuch using svn now, because i was getting tired of resolving conflicts from my own patches
what is the fix?
shapr put " glob <- askGlobal" on line 284 20:23
of Prim.hs
nothingmuch beh! 20:24
the last stament in a do construct must be an expression
shapr I like darcs, it's easy to start with, and easy to use. And I like hacking on Haskell apps, so that's nice too.
nothingmuch line 285
darcs is wonderful
i wish i could use it everywhere
shapr I'm using the darcs repo, maybe svn has something different?
nothingmuch but for that svk needs to be multi backend and two way
maybe i'm stupid and i mucked something up 20:25
ah
spaces vs. tabs
lets see if that fixes compilation
shapr I never use tabs.
buu tabs!
Tabs are <3
nothingmuch nonono
don't start please
i won't be able to ask questions! 20:26
buu shapr: If you would just use tabs, all your problems would go away!
nothingmuch huh?
src/Prim.hs:288: Warning: Defined but not used: glob
buu Recently I was poor and living on the streets, so I started to use tabs. Now I make 250 grand a month!
nothingmuch oh
that's from 2 lines down
shapr Yah, I dunno what's up with that, but the tests pass.
nothingmuch will remove the line and see what happens 20:27
tests still pas 20:28
i'll ci your fix with that line commented out, how does that sound?
shapr sure
Are there failing tests that I can attempt to implement? 20:30
nothingmuch grep -r 'todo_' t
stevan Is Prim.hs is still broken?
shapr ah
nothingmuch stevan: shapr fixed it
i'm committing
as soon as make test finishes
i would be very very happy if you do t/base/given.t 20:31
but i think that requires smartmatch
stevan nothingmuch: ok
nothingmuch hola orb
stevan nothingmuch: if I do t/base/given.t? the haskell for it?
nothingmuch shapr: maybe more urgent: t/op/int.t
stevan: sorry, i didn't understand what you meant 20:32
theorb Hm, has the fix for the src/Prim.hs:284 not landed yet, or is there a problem with my SVN?
stevan nothingmuch: me either :)
nothingmuch i'm committing as i type
stevan theorb: nothingmuch is still testing it
shapr tries to figure out how to enable one test at a time. 20:33
nothingmuch shapr - 'export HARNESS_PERL=$PWD/pugs'
then use 'prove'
rooneg shapr: just run ./pugs t/path/to/foo.t
nothingmuch or just ./pugs t/test.t
shapr ah
nothingmuch committed.
shapr++
shapr But, I want to run all the tests, and add one new failing test. 20:34
stevan prove t/*
nothingmuch make test will test everything
prove t/ will work with HARNESS_PERL set correctly
(make test just sets HARNESS_PERL before running Test::Harness, which prove is based on, btw)
stevan: that won't run nested tests, and will run editor backups, etc
shapr This is my first time using anything Perl since '98 or so. 20:35
nothingmuch prove -r t/ will run stuff recursively
theorb Welcome back, shapr.
shapr grins
buu shapr: I'm so sorry!
nothingmuch hmm 20:36
$! is not really set to the value of die ""
is that a problem?
it contains it
die "foo"
$! eq 'Verror "foo" (......)
anybody want to rule for "my $var is constant" test? 20:37
otherwise i'm committing as t/op/is_constant.t
as that makes the most sense for me
theorb Hm, well, makes more sense then any existing directory, but t/var/is_constant.t might make more sense. 20:38
rooneg might as well. it can always be renamed later if someone doesn't like that name.
nothingmuch i've heard of 'is' being an operator 20:39
shapr I really enjoy test-driven-development. This is the best way to have specifications.
nothingmuch what's more, since this is testing perl6 in perl6, to test we need perl6
theorb Is can certianly be an operator; however the interesting part is the behavior on the variable, not the operator.
nothingmuch so tests tend to create themselves very very rapidly 20:40
ok, good point
theorb s/on/of/
nothingmuch var/constant is an oxymoron though ;-)
theorb And? Never stopped us before... 20:41
shapr nothingmuch: both t/op/int.t and t/base/given.t appear to pass.
nothingmuch shapr - run them with './pugs t/op/int.t' 20:42
this will show you 'not ok blah blah # TODO
this means we know it will fail
so this is ok
once the feature is implemented you get 'ok ... # TODO'
which is an error
(unexpected success)
so you correct the test
shapr Hm, they all say ok. 20:43
nothingmuch so maybe it was fixed
did you update? luke committed this about 12 hours ago 20:44
shapr Ah, given has TODOs
lucs I think int() was fixed a bit earlier.
autrijus++ 20:45
nothingmuch oh my
even based on his log messagve
he deserves another donation
journal entry
does substr remain generally unchanged? 20:47
theorb Hmm? 20:52
nothingmuch should substr in generally work the same way?
theorb I don't recall anybody saying it was changing (but that doesn't mean much). 20:53
nothingmuch googling didn't help
theorb It's used in S09 without mentioning changes, so I'd assume it works the same. 20:54
nothingmuch that's the only thing i found 20:56
it does look the same
beh
someone can change the test later
theorb Ja.
shapr das machst spaƟ 20:57
theorb shouldn't use random bits of German, because it makes people asusme that he actually speaks it. 20:58
nothingmuch =)
shapr ich habe keine ahnung!
nothingmuch how are you gettting along with it?
shapr I'm from a little town in Alabama. 20:59
My german isn't that great. 21:02
nothingmuch mine is non existent 21:03
lumi What's an ahnung?
nothingmuch my grandmother still doesn't forgive me
she's austrian
orb moved to germany recently, so he should know some by now
lumi I speak twenty seven or so random words of German
shapr lumi: idea or suspicion, "aning" in Swedish.
theorb I know a very little.
nothingmuch my german was very embarracing for my parents
shapr "Jag har ingen aning." is the same thing in Swedish. 21:04
nothingmuch i spoke only hebrew all the time
except for once
at my aunts wedding
lumi Not "inte"?
nothingmuch when i pointed to the cross (a very modern, metallic kind of cross)
and said 'ima, tiri' (look mom), 'et ha flukzoig haze' (at that airplane) 21:05
airplane was, ofcourse, in german
theorb Ooops.
nothingmuch the only word i ever did speak spontaneously,
shapr Nah, "ingen aning" is "no clue" whereas "Jag har inte aning" would be "I have not clue"
lumi Fluegzeug?
nothingmuch i guess
i don't know how to spell in german either
(and you know it, lumi)
lumi I'm just guessing though
I can spell, I just can't speak :P
theorb asks Jess how to spell it. 21:06
theorb can't even spell in English.
nothingmuch goes wild with ref to lvalue in substr
shapr I wasn't exactly born on the bayou, but I was born in Selma, Alabama. I spent 27 or so years living in the USA. 21:07
theorb Flugzeug, apparently.
lumi You should've said Kreutze (I think it's spelt)
A friend of mine is always listening to Bach's passions, so he knows all those sorts of words 21:08
theorb Ah, Selma, famous for being the birthplace of much of the civil rights movement.
...kicking and screaming.
ingy which svn repos are people using now?
nothingmuch svn.openfoundry.org
theorb Kreuz, sayth the Jess.
nothingmuch i do, anyway 21:09
shapr My mother was working at the Public Library in Birmingham when the 'Race Riots' happened. She heard about it on the radio and looked out the front window. She said she saw about fifty people yelling, but that wasn't quite as many as the radio claimed.
nothingmuch anybody up to date with recent bleadperl fixes to substr lvalue? 21:10
theorb Odd... 21:24
sub foo { my $a = 2; } say foo();
...gives me bool::true
...and for some reason I seem to be stuck in ? mode. 21:25
nothingmuch t/op/substr.t committed 21:26
someone please look at it
i converted it to to perl5
then made sure it passed, and converted back
hola larsen
larsen nothingmuch: Hello.
theorb Allo, larsen. 21:27
Why all the evals? Is substr() not even parsed as an as-yet-unknown sub? 21:28
nothingmuch it's an unknown sub 21:29
it barks looking for the proto
ugg
what's $/ in p6?
s/proto/signature/
theorb Hmm, we have eval '', but not try {} ? 21:30
nothingmuch no
that was the first test i gave
autrijus promised to add it
but it requires given
which we don't have either
so i submitted a test for that to
o
that needs smart match
it needs $! 21:31
etc etc etc
theorb Being made into an attribute of the FH, I think.
nothingmuch hmm
theorb Hmm, try doesn't need any of that; catch does.
nothingmuch so when you chomp a string
theorb Objects would be a great goodness.
nothingmuch right, but the task is still daunting
i'm bugging autrijus on it daily 21:32
theorb You don't have to chomp strings generally... just read them, and they should come chomped for you.
Or at least, that was the theory...
At some point.
nothingmuch there's a chomp operator
so i'd like to test it ;-)
theorb I really need to start re-reading the synopses.
justatheory hrm?
theorb I read them all when they came out... but it's been quite a while, and there's been updates. 21:33
I no longer remember what was set in silly-putty, and what is just something Larry was throwing around in an email or two, but decided against. 21:34
nothingmuch perlbot, nopaste?
perlbot Paste your code here and #perl will be able to view it: sial.org/pbot/perl
pasteling "nothingmuch" at 192.168.2.16 pasted "chomp test" (3 lines, 136B) at sial.org/pbot/7932 21:35
nothingmuch is this worth anything?
theorb Do they pass? 21:36
nothingmuch yes
theorb Donno. I thought the API of chomp was getting changed, but that may have been completely idle fantasy.
nothingmuch i think it's valuable as just a flag that gets raised if something changes 21:37
forcing the person changing chomp around to fix the test too
although i don't know how meaningful that test is
theorb Hmm, the non-symetery of -- vs ++ is annoying to me.
nothingmuch -- vs ++? 21:38
autoinc and autodec?
theorb Comment and list concat. In haskell.
nothingmuch ah
heh
i didn't notice that
any one care to vote whether to add chomp.t or not? 21:39
theorb If nothing else, it serves as a bit of a reminder that it's underspecified. 21:40
...when somebody does a diff in the morning.
nothingmuch hmm
theorb Though if it's really unspeced, put it in the proper dir. 21:42
Hmm? One of your substr tests unexpectedly succeeded for me. 21:43
nothingmuch www.perl.com/pub/a/2004/03/p6pdiges...40314.html
see Mutating methods
huh?
crap, i forgot a plan for it 21:44
theorb t/unspecced is where tests for behavior that isn't part of any Syn, etc, are supposed to go... I think.
nothingmuch oh, crap
i see
nothingmuch was too trigger happy 21:45
wait, that shouldn't happen
lines 44, 45
the eval should just die
and $r remain undef
i'll move chop.t to unspecced too 21:49
i'll also add t/op/readline_chomped.t
based on exegesis 02 21:50
theorb Cool.
nothingmuch blammo 21:52
nothingmuch continues trying to unit test exegesis 02 example 21:54
FOAD Hi. 21:58
nothingmuch hi, FOAD! 21:59
theorb Hello, man with name not expandable in mixed company.
nothingmuch shift+f
<tab>?
oh
FOAD Hi nothing, orb. 22:00
shapr eff the ineffable
FOAD Meh, it's something else completely.
nothingmuch google define: foad says nothing
nothingmuch has no wtf program
theorb Fuck Off And Die.
nothingmuch ah
FOAD I'm shocked.
theorb I did that earlier. 240V hurts! 22:01
nothingmuch 240 is much worse than 110
when we lived in the US i got zapped once
it wasn't /that/ bad
but here i got a burn
beh 22:04
nothingmuch needs help with exegesis 2 example
anybody care for subetha fun?
i've always wanted to try it
theorb subetha?
nothingmuch google for sub-etha-edit 22:05
it's a collaborative editor thingy
lumi Okie
I'm not sure what that was
theorb Ahh.
lumi But I'd love to try it
theorb It does sound kind of nifty. Should try it with Jess some time. 22:06
nothingmuch you use macs, orb?
theorb Nope.
It's mac-specific?
nothingmuch so...?
lumi For now it seems so
theorb Sad. 22:07
nothingmuch maybe you can get them to port it to gnustep
lumi - when netstat is slow, do 'netstat -lan'
easy mnemonit
c
i should have PMed that, sorry 22:08
crysflame .g subethaedit
there's a kwiki plugin for interacting with subethaedit, iirc 22:09
nothingmuch uhuh
theorb Hmm? The other way around makes sense...?
nothingmuch anybody interested is welcome to try and edit woobling.org 22:12
theorb Hm, I think I need a dead-tree version of YAHT. 22:14
I want to hack pugs, but I'm not sure where to start.
nothingmuch just start with YAHT 22:15
it's really simple
i was drunk and i got pretty far
theorb I did, and it is.
nothingmuch bugging the eyes?
taking up screen space?
theorb But I get partway and get distracted; that ha[[ens to me a lot less with dead-tree.
nothingmuch ah 22:16
nothingmuch asks to duplex print
or print a chapter, then when done, print the next one on reverse side
theorb wishes he still had access to a university printer of some sort.
lumi What/where YAHT?
theorb (Paying for ink-jet ink)--
perlbot, YAHT? 22:17
lumi . . Haskell Tutorial?
nothingmuch laser printers++
is anyone less of a perl-6 grasshopper and more of a perl-6 chicken going to help lumi and I out?
s/I/me/
theorb perlbot, define YAHT Yet Another Haskell Tutorial www.isi.edu/~hdaume/htut/ 22:18
Hmm? What are you looking for?
nothingmuch hola luqui
Khisanth YAHT is ...
nothingmuch to actually get the exegesis 2 example sort of working
theorb YAHT is Yet Another Haskell Tutorial -- www.isi.edu/~hdaume/htut/
Khisanth perlbot YAHT is Yet Another Haskell Tutorial -- www.isi.edu/~hdaume/htut/ 22:19
perlbot added YAHT to the database
Khisanth perlbot: YAHT?
perlbot Yet Another Haskell Tutorial -- www.isi.edu/~hdaume/htut/
theorb Thanks, Khisanth
luqui hola nothingmuch
Khisanth I think it also has "is also"
too bad there isn't an universal irc bot interface :) 22:20
crysflame depends on what you consider universal :)
working on getting purl in here.. if i set up locking on her factoid databases, it might just happen. 22:21
Khisanth well they are all very similiar but different enough that commands that work with one will not work with another
lumi Which is intentional 22:22
theorb Hmm, we still need some literals: inf, nan, and '...'. 22:24
nothingmuch isn't ... pretty much a synonym for 'die "this is not yet implemented"'?
theorb Yes.
(At runtime; at compile time in the presence of a pragmata.) 22:25
Is there any document on what tests belong where? 22:26
luqui theorb: what at compile time? 22:27
theorb IIRC, there's supposed to be a pragmata that makes any use of the ... thingy die at compile time.
luqui uh no 22:28
you prototype with ...
so the pragma would be 'no prototypes'
which is just silly
theorb Um, 'no prototyping' makes more sense to me. Or 'use production'.
luqui I mean prototypes like Perl5's 'sub foo($$);' 22:29
is 'sub foo($x, $y) {...}' in perl 6
theorb I know; prototype is a somewhat overloaded term.
luqui it just doesn't make any sense to disallow ... at compile time. you wouldn't be able to predeclare anything. 22:30
theorb That's a special use of the yaddayaddayadda operator.
Er, special term. 22:31
luqui oh, so you mean anywhere but in a {...} it would be disallowed
that makes a bit more sense
theorb Bah, groups.google.com is broken (at least for me)... but the only hit it comes up with for "yadda yadda yadda" pragmata is mine, so it's probably just a combination of bad memory and wishful thinking. 22:33
luqui yeah, I didn't remember anything about it...
theorb Oh, wait, now I'm seeing one by you, but google won't show it to me. 22:34
luqui there is no group perl-perl6-language
stupid ggroups
lumi That needs tr/-/./ then 22:35
luqui ah, doing that in the address bar makes it work 22:36
afbach pardon - may I ask a question about t/03operator.t tests 22:38
theorb Yes, you may. In fact, you have permission to ask any question you like about pugs. 22:39
afbach if I change the '>' to '<' on one test (all(@) > any(@) pugs 22:40
fails to parse it 'unexepcted 'o'
Sorry, I'm really new to IRC but I hoped it was the best way to get a quick answer.
luqui afbach: I'm working on that problem right now 22:41
the problem is that it's parsing < as the beginning of a subscript
afbach Hmm, and "<=" choked it too, but the 'all infix op' test worked w/ a '<' 22:42
Thanks, glad its a 'known' bug - I have to go start the grill for dinner 22:43
nothingmuch luqui: is that tested? 22:45
luqui it's a recent introduction
so no
nothingmuch i'm creating t/syntax 22:46
theorb (Also known as eval STRING R Us. 22:47
nothingmuch hah 22:48
theorb wonders if that's the right place to stick todo_is(eval 'my Int $i; $i=42;', 42, 'my indirect object types'); 22:49
nothingmuch my int is more than syntax, methinks
theorb It is more then syntax, yeah, but the syntax is the first bit of it. 22:50
nothingmuch hmm
so you argue anything that is not yet parsed correctly should be put in t/syntax first
and then moved to t/(op|base|...) when ready?
theorb Hmm... yeah, on second thought, that doesn't make all that much sense.
nothingmuch (and possibly not yet implemented correctly?)
svn mv t/op/assign.t t/syntax/decl_vs_assign_prec.t ? 22:59
should 'make test' use 'prove -r' instead of hard coded test list generated by Makefile.PL? 23:02
theorb calls it a night. 23:04
shapr calls it a waste of brain cells and looks for the alcohol. 23:05
nothingmuch anybody want to review a test of control blocks? 23:27
buu nothingmuch: Sounds like satan. 23:28
nothingmuch uhuh
most is commented out 23:29
because it's not parsed yet
hola stevan 23:31
given array @foo 23:33
whose elements are array refs 23:34
[ $var, [ @array ] ]
+@foo[1][1] will yield the number of elems originally in @array
right?
luqui: could you please see a nasty test i wrote, tell me if it's ok? 23:36
uh 23:39
my (@array, @other);
todo_is(+@array, 0, "array has nothing in it")
if my (@array, @other) = (), it is 0
svn mv t/op/assign.t t/syntax/decl_vs_assign_prec.t 23:56
oops
nothingmuch.woobling.org/pugs/t/bas...l_blocks.t
someone please post feedback

Powered by Cro and the RakuĀ® Programming Language.

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

Thank you!