pugscode.org <Overview Journal Logs> | r1773/winxp: 587/4256, smoke/win2k (occasional): xrl.us/fqum (214/4453, r2198) Mac OS X (62/4459, r2116) | pugs.kwiki.org
Set by theorbtwo on 21 April 2005.
00:35 nothingmuch_ is now known as nothingmuch
nothingmuch ♥ Set::Object 00:39
00:42 sekimura_off is now known as sekimura
nothingmuch *yawn* 02:27
nothingmuch goes to sleep
mugwump tucks nothingmuch in 02:33
nothingmuch changes his mind actually
mugwump: Class::ISA isn't necessary in the spork fix
we only need to traverse one namespace level
mugwump ok, well that makes it a lot simpler :) 02:34
nothingmuch indeed =)
otherwise i'd have used Devel::Symdump->rnew(ref $self)
btw, if Spork want Spiffy and Spoon and Kwiki and whatnot
why not Devel::Symdump too?
mugwump only a probably irrational tendency to see a module name like that and think "that's for development, not 'production' code" :) 02:35
nothingmuch agrees that it's wrongly named
it should be Symbol::Traverse
mugwump Yes - introspection isn't development only 02:37
nothingmuch ponders writing Class::Methods::Suicidal 02:43
methods that happen once per instance
and then then call SUPER::
nothingmuch really goes to sleep 02:55
gantrixx so I'm just starting to fool around with Perl6 04:06
I've downloaded and build parrot
how do I get Perl6?
this channel has a lot of people not talking 04:08
mugwump we're thinking
obra we're also sleeping 04:09
gantrixx: www.pugscode.org
gantrixx so is there a daily build I can dowload?
mugwump we're not doing binary releases of pugs yet... after 6.28, perhaps 04:10
some people are though
gantrixx: you'll also need ghc 6.4 or newer 04:11
gantrixx ghc? 04:12
I don't even know what that is
Here is my concern folks
I've been using perl for a long time
I love perl, but ther is a stong move toward python advocacy 04:13
I believe that Perl6 will address all the shortcommings of perl5 and hopefully will obsolete python
obra gantrixx: perl6 is still in the early development phase. It has a way to go. 04:14
a long way, even
gantrixx I'm getting ready to start a contract with a 10 year development cycle and I have to make an arguement to the powers that be to stick with perl5/6 and not use python
mugwump gantrixx: that's a fairly long cycle :)
gantrixx DoD, c'est la vie
obra do you expect to switch languages midstream? 04:15
(perl5 to perl6)
gantrixx so they don't want to invest in a lot of perl development if in 5 years everyone is using python and perl5 is not supported
yes, I do expect to start off in perl5 and then migrate to perl6
mugwump hmm, start in ruby perhaps. 04:16
ruby -> perl 6 will be easier (IMHO) than perl 5 -> perl 6 04:17
Sure, Larry's doing his Perl 5 to Perl 6 converter and all, but I think it will suffer from the same problem as the shell to Perl converter
gantrixx if it can't be done in C/C++, Perl, and HTML then I don't think it needs to be done 04:18
mugwump sure, I understand that.
gantrixx I'm tired of relearning the tools of the profession every 3 years
so do I spend my time learning Perl6 or Python 04:20
because there are a lot of the software engineers that are advocating python
mugwump That choice is a no brainer really, after all... 04:21
obra It does have a big advantage over perl6.
gantrixx I guess they are trying to predict 5 years in the future
mugwump "Not having a choice streamlines the thought process" -- GvR
obra It exists now.
gantrixx seriously? you would choose Python over Perl6?
obra If I had to write production code in the next six months? Hell, yes. 04:22
gantrixx that is why I'm investigating Perl6 now to see if I should stick with the Perl advocates or jump in with the Python guys
obra I'd prefer perl5, but if I was asked to pick one or the other.
perl6 is a development project. Autrijus runs pugs in production. But he's special ;)
gantrixx well of course we would use Perl5 now, but 5 years from now we would be writting in Perl6
mugwump there are lots of programming constructs that can't be expressed simply in python due to GvR's insistence not to include any higher order stuff 04:23
gantrixx so do you think Perl6 will really be ready Q3 2005? 04:25
or is this another mozilla project?
mugwump gantrixx: have a look at the pugs codebase and examples, a serious amount of the language is already working
obra gentrixx: how much time can you volunteer to help perl6 along? 04:30
gantrixx They will allow me to donate some of my Perl6 work on the job back to the community 04:34
"some"
it is a DoD contract
but if I write an snmp module that is generic, they don't mind if I donate that back to the community
mugwump the challenge will be writing the Perl 5 code in a way that can be easily ported to Perl 6 04:35
And making sure your test suite is *comprehensive* enough to spot interpreter bugs
gantrixx well, to be honest, it isn't uncommon to completely rewrite some stuff so I'm not that worried about it
what is the glasgow haskel compiler? 04:37
stevan gantrixx: Pugs is written in Haskell
mugwump haskell is a functional language, a dialect of ML
stevan so you need it to compile pugs 04:38
gantrixx why not, we are short on reinvented wheels
stevan gantrixx: I think you will have a hard time future-proofing a 10 year project 04:39
mugwump For instance, if you were to use Class::Tangram as an accessor generator package, it would be fairly trivial to mechanically convert the Class::Tangram class definitions to Perl 6 Class definitions 04:40
stevan and I would be surprised if you didn't re-write major portions of it regardless of the language you use
mugwump (not to say that Class::Tangram is as comprehensive as Perl 6 Classes,Roles,etc)
gantrixx perl5 will be obsolete by the time this thing goes live, so the question is will Perl6 be the way to go or is it Python
stevan gantrixx: when is it going live? 04:41
perl5 will be around for many years to come
perl 5.005 is still on many prod machines
gantrixx the development will be done in like 5 years or so
but it doesn't actually reach full maturity for 10 years
at least that is what the "plan" says 04:42
I worked on Iridium too and in 10 years when we delivered it, it was already obsolete
so the plug could be pulled, hell this will span 2 more presidential administrations
stevan gantrixx: that is a rather long schedule
gantrixx so was putting a man on the moon 04:43
it's a 10 year contract, what else can I say
stevan gantrixx: thats a good thing, I am just not used to anything that long 04:44
stevan tends to work in months or weeks
gantrixx DoD stuff is like that
they will spend 5 years testing it
stevan gantrixx: I won't ask you any details,.. top secret and all
gantrixx probably part of that contract is the initial maintainance phase
then they usually bid out continued maintainance
it does require a clearance 04:45
and believe it or not the DoD is very accepting of open source technology
autrijus rehi.
stevan well if you prototyped in perl5, with an eye towards perl6 ... 04:46
by the time you were ready Perl6 might be done
morning autrijus :)
gantrixx I think they are tired of companies going out of business and being left with no support
autrijus I just read the backlog of yesterday :)
stevan gantrixx: it makes a lot of sense
autrijus it's curious how NathanJY happened to stop speaking altogether when I joined the channel :)
gantrixx I wish I could tell you some of the stuff that is out there and the old old unsupported technology it is built on 04:47
autrijus stevan: btw, great work on force_todo
stevan gantrixx: my father worked for the DoD, I am familiar with how messed up in can get :)
autrijus: thanks :)
I actually think it is speeding things up (or maybe it is my imagination)
autrijus that may very well be the case
gantrixx it can be a mess of inefficiency and red tape, but my experience is that that is the maintainance part
autrijus since there's no extra fileread/processing for each test 04:48
gantrixx the development part is pretty much a ball buster
stevan autrijus: yes, what bothered me was that the greatest penalty was for files which did not use it
gantrixx: I can imagine, I recently read about FAA dev standards, crazy stuff 04:49
gantrixx oh now that is something I can talk about
I've done some work for the FAA too
stevan autrijus: I am also in the process of removing the todo_* functions, and added a todo named param
it should slim down Test.pm a bit more
autrijus stevan: yup, I saw that. great
I'm fixing parsefails left and right
stevan autrijus++ 04:50
autrijus but the cxt/typ patch looks like a keeper
stevan good
mugwump is that the GADT-based parser change, autrijus ?
gantrixx there are strict requirements on mission critical parts on an airplane. now adays much of "parts" are software, so the FAA requires that all code be hand reviewed by and outside group
stevan gantrixx: I have always been interested in embedded work, but my C skills are pretty sad
gantrixx so companies like Honeywell that write autopilots subcontract this obligation out a bunch of flunkies 04:51
autrijus mugwump: no actually... it's a separation of Cxt (context) and Type from String
mugwump: previously I got them all messed up. Juerd enlightened me
mugwump: one good thing from this is that we cleanly support junctive types 04:52
stevan junctive types... mmmmm
I still need to find a place to use them,... maybe somewhere in hangman, just to be gratuitous
mugwump junctive types, is that like 1`Oz|1`m|1`orange|etc ? 04:53
autrijus mugwump: no, more like
mugwump or just junctions
gantrixx nice talking to you guys
stevan sub foo (Str|Bool $bar)
autrijus sub foo (Int|Str $a)
gantrixx I have to go
autrijus ciao gantrixx :)
sub phash (*@a) returns Array&Hash { ... } 04:54
stevan autrijus: I didnt get that haskell book, I decided to just order Algorithms instead
autrijus shudders at the thought of phash
stevan: cool
stevan: hm, I'm seeing 04:55
mugwump has a little moment stomaching that concept
autrijus $ ./pugs -MTest -e 'ok(1)'
ok 1 - pugs: cannot cast from VRef <Scalar::Const> to [Char]
fixing
stevan autrijus: it worked for me (I think I have the latest code too) 04:56
hmm, even with this: pugs -I ext/Test/lib/ -e 'require Test; ok(1)' 04:57
autrijus that works for you?
stevan yes
autrijus weird. 04:58
stevan but you know,.. I may not have the latest build
either way I have to go to sleep as I am starting to see double
autrijus that's fine... I'll figure it out
see you :)
stevan be back in ~8 hours :) 04:59
bsb Is phash pronounced /fash/? If so, I'll shudder too 05:06
autrijus bsb: it's pronounced "pseudohash"
bsb or p6opaque? 05:07
autrijus one of the gravest peril known to camelkind
bsb I get it now, Array&Hash
If I make pugs segfault, how do it tell? 05:13
s/how/who/
autrijus me
do you have a one-line test case?
bsb pugs> ?
Segmentation fault
is there and out of band way to send build info and context? 05:14
mugwump points bsb at nopaste.snit.ch:8001
autrijus bsb: is this 6.2.0?
bsb yes
autrijus bsb: it's a known issue that has since been resolved
bsb great 05:15
obra Hey glasser 05:17
Huskie Hi I need some help with some parsec code 05:18
bsb Another one: pugs> none(1).pick # returns 1 05:19
autrijus bsb: do you have svn installed? 05:20
Huskie any lambda folks that could help me..thx
autrijus bsb: and may I make you a committer so you can add the failing test to t/junctions? :)
Huskie: yes?
Huskie how can i paste code to show u 05:21
autrijus perlbot: nopaste
perlbot Paste your code here and #<channel> will be able to view it: sial.org/pbot/<channel>
Huskie pasted to sial.org/pbot/9574 05:23
my ? is how i would call my eval function on the AST
bsb yeah to svn, I'm rebuilding. Ok, I'll fix the test 05:24
autrijus bsb: are you a committer? if not, can you give me your email addr so I can send an invitation to you? 05:25
Khisanth given the rate of change is there actually a reason to use pugs releases instead of from SVN? :)
bsb bsb at bereft dot net 05:26
What prodecure do I follow to ensure I'm not making things worse? just make test?
Huskie autrijus, do you see my pasted code? The last line is what I would like to use but gives me an error 05:27
autrijus bsb: you can't possibly make things worse. but yes, "make test" works; after "make install" you can also simply run pugs on selected tests
Huskie: what's the error msg?
Khisanth: yes, the trunk may occasionally receive huge patches that kills the tree :)
Huskie Couldn't match the rigid variable `a' against `CT' 05:28
`a' is bound by the type signature for `run'
Khisanth chainsaw patches 05:29
autrijus Huskie: your signature should not be 05:31
Parser a
it should be
Parser CT
Huskie: when in doubt, remove the signature and let GHC figure it out for you
bsb: invitation sent. welcome barod! 05:32
err, welcome aboard!
Huskie thanks a lot autrijus! It solves the problem. 05:33
and thanks for the inspiration with your work
autrijus no prob :) 05:34
05:58 chady_ is now known as chady
Huskie autrijus, any pointers on how to output code from an AST 06:03
autrijus Huskie: depends on your target code 06:04
target language, rather
Huskie I have an idl compiler that successfully builds an AST and i need to generate stub C code from it 06:05
for ex, interface { int func(int a); }
autrijus oh. you need to write an AST reduction module 06:06
look at src/Pugs/Compile/Parrot.hs for an examples
autrijus needs to run. bbiab 06:07
Huskie cool.. looks like i have to understand PrettyPrint 06:08
bsb What should "(sub {3} | sub {2})()" produce in scalar context? 06:13
Warnocked on p6l...
.. warnocked, everywhere .. :( 06:19
autrijus bsb: put it in as a test. 06:22
bsb OK. is( (sub {3} | sub {2})(), (3|2)) and fix if the answer changes 06:23
autrijus not sure if you can compare junctions using is
probably flatten both sides using .values or something.
bsb yeah, that's what the other tests do 06:24
autrijus add yourself to AUTHORS too, if you're not 06:25
already there
* `(sub {3} | sub {2})()` implemented.
enjoy.
work &
bsb implemented faster than the test, ++ 06:27
06:37 castaway_ is now known as castaway
gaal kungfuftr: awake? 07:04
ingy hola 07:16
ingy just finished a flippin sweet test framework module 07:17
gaal hey ingy 07:18
i'm cleaning up the makefile a bit
i was bothered by how 'make smoke' tried building the targets again and again
my findings:
ingy pause.perl.org/incoming/Test-Chunks-0.12.tar.gz
gaal 1. the smoke target depended on util/run-smoke.pl, that looks bogus to me 07:19
2. that script always runs make explicitly :)
ingy ouch
do you need help or something/
gaal 3. 'make optimized' and 'make unoptimized' are what the lovely manual call "double colon" rules, whcih basically means "always run this rule (even if its deps are fulfilled)". 07:20
www.gnu.org/software/make/manual/ht...html#SEC50
i love GNU docs. not.
anyway, run-smoke should probably no longer be a script; it should be entirely a make target. rihgt? 07:21
ingy well I haven't looked at the Makefile in sometime
gaal question is, is there any reason for the double colons? i wonder who instated them.
ingy double colon rules are for when you want two or more rules with the same target
foo :: this
foo :: that 07:22
gaal they just waste time relinking, for me, because they also rerun the rule always.
ingy double colons are crucial
gaal well, they aren't used that way for the (un)?optimized targets.
ingy to be able to add on to rules genned by makemaker
gaal aha! 07:23
ingy :)
gaal i knew there was a reason i was asking you :)
ingy otherwise just use single colons for non MM targets
=)
I need to port Test::Chunks to Perl6
gaal okay, i'll finish fixing this up then. 07:24
other than environment, is there a way to pass an option to a rule? e.g. the smoker script used to 'make optimized' hardcodedly; i want it to choose the pugs target in a saner way. 07:26
autrijus gaal: "make smoke TARGET=optimized" ? 07:53
gaal yes, that's what i ended up doing.
i hope it's portable though :/
actually i was doing something slightly different; because if you do the above you can't omit the TARGET setting (i think?) 07:55
i added an ifdef in the $pugs target to check $(PUGS_OPTIMIZE), expecting 1, 0, or no def at all; and build -O1, -O2, or nothing from it. 07:57
r2208 - i hope i didn't break anything :) 08:14
mj WinXP, r2207 - 159/4468 subtests failed, 96.44% okay. 08:17
autrijus gaal: you broken everything ;) 08:27
ifdef is a GNUism.
gaal that's too bad :)
autrijus please fix :)
gaal sorry, reverting. 08:28
autrijus no prob
gaal is there a preferred svn command for reverting a patch? or should i just svn up -r2207 and ci the old versions? 08:30
autrijus gah, that again
src/Pugs/Eval.hs:226:48: My brain just exploded.
gaal: you can apply the reverse change via svn merge
gaal: but just svn up -r2207, cp somewhere, svn up, mv somewhere back, ci
gaal done as r2209. now to figure out how to do this portably. 08:33
autrijus :) 08:34
kungfuftr gaal: moo? 08:43
gaal moose!
kungfuftr lo
gaal i was doing some changes to the smoke target, but they turned out to depend on non-portable features of gnu make. 08:44
that is,
the way i did it depended etc.
kungfuftr ah
gaal i'm sure it can be done in an even uglier and more portable way.
kungfuftr $^O is horrible, but handy 08:45
gaal the thing is, make smoke relinks pugs needlessly.
no, the unportability was in make itself.
there is one improvement i'd like y'all to try though.
can you edit your Makefile.PL and change "optimized :: " to "optimized :" ? 08:46
(remove one colon)
on my systems that alone is a big win :)
..after doing that rebuild the makefile and do make smoke.
it shouldn't relink pugs if you alreadt had it built. 08:47
kungfuftr um... it'd have to be in a few hours time... have to do urgent work on search engine indexer
gaal k later. 08:48
pasteling "bsb" at 203.214.67.82 pasted "has_ghc_package using $ENV{GHC_PKG} || $ENV{GHC}" (20 lines, 710B) at sial.org/pbot/9581 08:55
bsb does that patch look reasonable?
I'm trying to get started here, bear with me 08:56
autrijus I'd suggest to use 08:57
/\bghc(?=[^\\\/]*)/
in the 08:58
+ $ghc_pkg =~ s/^ghc/ghc-pkg/ # ghc-6.5 => ghc-pkg-6.5
line
err
/\bghc(?=[^\\\/]*$)/
magnetic hi 09:02
autrijus yo magnetic 09:04
bsb ok. svn ci comes next, here we go
autrijus I wonder if I should spend any time replying to BrowserUK on perlmonks.
gaal autrijus, what thread? 09:05
autrijus gaal: www.perlmonks.org/?parent=448319;node_id=3333
gaal he's quite the intelligent/thoughtful person; in my experience discussions with him were always worthwhile.
autrijus gaal: shuffling was one of the few cases that GHC performs much much better than perl.
gaal that doesn't mean he's always right :) 09:06
nothingmuch_ good evening 09:07
gaal now i'm interested in how to do this efficiently in haskell myself :)
so i hope you do reply :) 09:08
autrijus maybe you can write it as an exercise :) 09:11
gaal :) 09:12
autrijus I replied.
gaal autrijus++
nothingmuch_ autrijus: some guy came in here yesterday, demanding that we switch to C and scrap parrot 09:19
maybe we can meet in the middle?
we could implement with only the most monadic C like constructs, and everyone is happy
09:19 nothingmuch_ is now known as nothingmuch
autrijus nothingmuch: the joke is no longer funny :) 09:20
nothingmuch beh!
so how is everyone? 09:21
castaway nm!
gaal is listening to Abbey Road and pondering Autrijus' challenge.
nothingmuch si!
castaway Happy Friday! 09:22
nothingmuch happy friday to you too!
autrijus I'm just fine :) trying to get more infinite loops fixed
nothingmuch huraah
gaal at some level, obviously we can do in-place shuffle because we operate on perl types in place. 09:23
so waht i need it sems is to wrap a haskell array in monadic.. thingie,
and shuffle that.
my problems are straightforward so far:
1. how to wrap the arr^H^H^Hlist in a monadic thingie 09:24
integral an array of IORefs?
castaway :)
gaal 2. how to unwrap it afterwards
1.5. how to shuffle elements in it.
integral: IORef is in the language? or is that a pugsish type?
integral it's in the language 09:25
GHC also has Data.Array.IO and Data.Array.ST
nothingmuch darcs get nothingmuch.woobling.org/Algorithm-...cy-Objects
gaal trick is to do it w/o copying over the whole list.
otherwise we didn't win much. 09:26
nothingmuch tests are not fully converted yet though
autrijus gaal: try to think it this way
nothingmuch but it may be useful for someone
autrijus gaal: you read a bunch of Bugs wBufs via readBuf
gaal: you read a bunch of Bufs via readBuf
each is a Ptr
then you shuffle those Ptrs randomly
then you print them out.
the easiest way is perhaps to read the entire file into mem as a CString 09:27
gaal well, okay, but then i have to have had my data in a particular format, not a native @array.
autrijus gaal: uh, you are doing this in pugs? 09:28
not haskell as I thought?
gaal no, in haskell; i'm thinking of the comparison to perl that BrowserUk was making.
though i guess in that case we can take the problem on his terms and indeed read the file *as* anything we like. 09:29
autrijus well, you can do the naive thing and see how slow it is.
use getLine repeatedly until eof
into IO [String]
err, I mean [IO String]
gaal of course it'll be slow, that was his point, no?
nothingmuch not IO [String]? 09:30
autrijus hrm. 09:31
okmij.org/ftp/Haskell/perfect-shuffle.txt
try merging it with the most naive 09:34
lines . getContents
I think it will be sufficiently fast.
just try it and see :)
Juerd GOOD MORNING EVERYONE 09:36
:)
autrijus hey Juerd-san
Juerd Hello 09:37
What is -san?
castaway Good nearly-midday Juerd
autrijus Juerd: -san is like Mr.
Juerd autrijus: In what language?
In any case, hello autrijus-san :) 09:38
castaway: That is what is so good about this morning - it's already almost over.
kungfuftr (search indexing)-- # takes too long and too many resources 09:39
09:39 joepurl_ is now known as joepurl
castaway Juerd, not mine Im afraid, I'm running out of time, and I actually wanted to go home early.. 09:40
Juerd I'm going to the office late 09:41
castaway lucky
Juerd That works much better than going home early
autrijus Juerd: in japanese :)
Juerd It's one of the benefits of not having an employer telling you the times to work
autrijus: I see 09:42
castaway I dunno, I prefer to relax when its over and done, and not before I get started .) 09:44
Juerd I prefer to jump right into bed after work ;)
castaway each their own :) 09:45
castaway goes back to DB grumbling
Juerd My addiction won't let me though
Juerd now has domed labels with his company logo
hurrah
Let the name branding begin.
They made the 1" x 1" ones 22mm x 22m, so those are useless. 09:46
But the rectangular ones are just the perfect size.
bsb Reading from L<S03/"Junctive operators"/"Junctions are specifically unordered">
castaway domed ? 09:47
Juerd castaway: A layer of plastic, as on computer case badges
It makes the labels look a little like Mac OS X aqua buttons :) 09:48
kungfuftr Juerd: do we get free ones?
bsb test: for all(@foo) { %got{$_}; };
Juerd kungfuftr: What would you do with labels that have my logo on it? :)
bsb: Do junctions in list context flatten, then? 09:49
kungfuftr Juerd: use them as rubber feet for my desktop unit
=0)
Juerd Didn't you mean to say for all(@foo).values { ... }?
kungfuftr: Heh, they were a bit too expensive to be used out of sight :)
But they'd work well as rubber anti-slip feet.
bsb Juerd: Don't think so
I'm going from L<S03/"Junctive operators"/"Junctions are specifically unordered"> 09:50
Which seems to be saying something unusual
Trying to check my reading of it
gaal autrijus, the doc you linked to still doesn't do in-place ordering, does it? if you have a list in memory it won't shuffle it in place. if you have the data on disk, you read it into a tree and extract a shuffled list, which is a copy of the data - sure you could do a list of pointers to the data, but you stil have to allocate a new list. 09:52
s/ordering/shuffling/
castaway oh, nifty
bsb It looks like `for all(1,2) { ... }` can run the block in parallel 09:53
Juerd bsb: Could be made to run the block in parallel.
autrijus gaal: uh, not sure what you mean.
gaal: sure it doesn't do inplace ordering
if you want inplace ordering, try IOArray :) 09:54
gaal think of it from the perl starting point. you have an @array, and you can do shiffle \@array.
autrijus IOArray of String
gaal looking up....
autrijus sure; if you have an IOArray, you can shuffle IORef IOArray
it's in Data.Array.IO
gaal Data.Graph.Inductive.Monad.IOArray ?
autrijus no, IOUArray in Data.Array.IO 09:55
IOU is extremely fast
bsb Juerd: Yes, so in the non-parallel case should `for all(1,2) { %got{$_} = 1; };` set %got{1}=1
autrijus because it's unboxed, so is quiv. with raw C code
gaal they should have named it EFArray then.
bsb ?
autrijus gaal: heh.
IO means it's in the IO monad; U means it's unboxed
but STUArray is probably enough for you.
STU may or may not be even faster
gaal that's for STUpendously fast (maybe) array, yes? (just checking) 09:56
autrijus you can use Data.Array.MArray as the interface code. 09:57
no... ST means it's in the ST monad.
ST monad maintains state, but does not do input/output.
Juerd bsb: I don't know how for $junc should behave. 09:58
bsb Anyone else? for all(1,2) { %got{$_} = 1; }; 09:59
It's not clear in s03 10:00
gaal I want... MArray (STUArray s) (Ptr a) (ST s) ?
autrijus gaal: something like that 10:01
gaal: instead of Ptr, use CStringLen
I think a STUArray of CStringLen is probably fast enough. 10:02
you can read the entire file as a [CStringLen], convert it to a STUArray, sort it in place, then output it
s/sort/shuffle/ 10:03
castaway waves at broquaint.
autrijus still I don't think inplace shuffling is the best way to go
broquaint waves back
autrijus I think just randomising the indexes is good enough.
gaal autrijus, that depends on what you're trying to do, of course
autrijus gaal: I thought we are optimising for speed 10:04
gaal yes--
Juerd bsb: +all(1,2) is 1
gaal but maybe the thingie receiving the shuffle needs an actual list and can't be changed?
Juerd Oh, no it's not. 10:05
gaal oops, i downmoosed one of my favorite bands. yes++ then.
autrijus gaal: hm? 10:06
gaal suppose you have some function you dontcontrol that
wants an ordered list
you can't change the api; you have data and you want to feed it to that function. 10:07
autrijus sure; so you just pass it to sort()
gaal sort? shuffle() you mean? 10:08
autrijus I'm confused
18:07 < gaal> wants an ordered list
Juerd autrijus: Why with %hash{all(4,5,6)}, is all(4,5,6) a string? 10:09
autrijus Juerd: that is a bug. 10:10
write a test and commit it?
bsb Juerd: it seems like for handles Junctions specially in s03
gaal i meant it expects a list where the order matters, not that it expects sorted input
brb
Juerd I don't know what it should do - stringify as ~all(...) does, or use the junction, or result in all(%hash{4,5,6}) 10:11
autrijus Juerd: I'm of the opinion that it should autothread.
deref is just conceptually a method call
gaal: I think I'll speak with code :) 10:13
brb too
Juerd autrijus: What does autothread mean for a hash key? :) 10:15
autrijus %hash.circumfix<{}>(all(4,5,6)); 10:18
same as
all(%hash{4}, %hash{5}, %hash{6})
I'm pretty sure that's the way to go
Juerd Ok.
autrijus &
Juerd And how does assignment to a junction work?
Since %hash{$conjunc}++ is kind of likely to be used, I think. 10:19
Assign to each of the elements, and autothread around that again?
all(%hash{4} = 1, ...)
broquaint So does %hash{4,5,6} (hash slice?) behave the same as %hash{all(4,5,6)} (junction of a hash slice?) ? 10:22
Not quite, I guess. 10:23
nothingmuch hola theorbtwo 10:38
theorbtwo Hola, nothingmuch!
nothingmuch what's new? 10:39
theorbtwo Not too much.
My eyesight is starting to go bad.
nothingmuch in which sense?
castaway too much looking at computer screens .. 10:40
broquaint Get a projector?
nothingmuch projectors are hard to read from 10:41
theorbtwo My monitor looks very fuzzy.
castaway .. and find your glasses..
nothingmuch theorbtwo: LCD monitors help
theorbtwo hmms.
nothingmuch very sharp
so you can lower brightness a bit
good contrast
theorbtwo Yes, but are an expense we presently cannot afford.
castaway nothingmuch: you donating one?
nothingmuch has one on his el-cheapo laptop
and one on my desk at work 10:42
also, i'd like to note that IMHO health comes before see hacking, etc ;-)
theorbtwo I haven't been doing much SEE hacking, that's mostly Jess.
I should probably up my resolution and take breaks and all that jazz. 10:43
castaway (and thats stalling on lack of decent ways to test it atm)
theorbtwo Er, lower my resolution I mean.
nothingmuch breaks are important
nothingmuch coughs
castaway sighs and mutters "zealots" .. 10:51
theorbtwo BRB, restarting X. 10:53
Odin- castaway: What sort?
castaway the annoying sort (daft convo about what "free" means, just because I quoted "db2 is free for personal use") 10:54
Odin- Ah. 10:56
The sort I hang around. ;)
castaway feel free to join #eamcs :) 10:57
Odin- agrees that some people take it too far, but you can see where they're coming from, can't you? :> 10:59
castaway yes, and no.. to survive in this society, you need to learn to figure out people mean, without correcting them all the time.. (since 99% of the time, its actually quite clear) 11:00
being picky wont fix anything :)
castaway wonders if he's given up :) 11:01
Odin- True.
Which is what I meant by taking it too far. ;)
castaway nods. 11:02
castaway smooches theorbtwo 11:06
theorbtwo Miss me so soon? 11:07
castaway yup 11:08
theorbtwo wonders what happened to his panel.
castaway hasnt got it 11:09
I paid for my muffin with the 19 5-cent pieces today, and he didnt even blink an eye, twas most disappointing.. 11:11
(didnt count them either, though :)
kungfuftr (binary)-- # not human readable 11:12
can perl6 make binary readable please? 11:13
=0)
theorbtwo pugs -e 'print 0b1001110' 11:14
Er, s/print/say/
kungfuftr theorbtwo: heh, having to debug binary search index files, very fun 11:15
anywhos 11:16
castaway sounds like great fun 11:17
cognominal arf, I used reply to p6l in mutt and I forget to delete the In-Reply-To and change the subjet. Not using aliases was false lazyness :( 11:23
sorry about that
Juerd cognominal: [email@hidden.address] is not that much typing. 11:26
I don't use aliases.
nothingmuch_ stevan: ping 11:27
kungfuftr ($shit,$voodoo) := ("perl6","php"); 11:31
kungfuftr hides
bah, wrong usage 11:32
Juerd: i'm incorrect in usage... right?
Juerd Yes, Perl 6 isn't shit. 11:33
castaway :)
Juerd Apart from that, I wouldn't know why the syntax would be wrong.
kungfuftr Juerd: swap melarky
Juerd As binding to a string literal makes sense, and probably makes it kind of like the "is const" thing.
kungfuftr $shit = "perl6"; $voodoo = "php"; ($shit,$voodoo) := ($voodoo,$shit); # more likely correct 11:34
again, could be wrong 11:35
Juerd: wrong? 11:36
11:36 chady is now known as chady_
nothingmuch_ ? 11:36
11:36 nothingmuch_ is now known as nothingmuch
Juerd kungfuftr: It'd be easier if you said what the hell you want it to do. 11:38
wolverian language!
kungfuftr Juerd: it's to do with swapping the names of the variables without the copying... end result... $voodoo is perl6 and $shit is php
theorbtwo Anglais, por favor.
kungfuftr theorbtwo: voulez-vous couchez avec ta mere, ce soir? 11:39
theorbtwo
.oO(Huh?)
castaway say no :) 11:40
theorbtwo No.
castaway I *think* "Do you want to have sexs with my mother, this evening" 11:41
kungfuftr castaway: "will you sleep with your mother this evening?" 11:42
castaway oh "your" .. even worse ;)
kungfuftr =0)
castaway ah well, right answer anyways..
cognominal kungfuftr: either "voulez-vous coucher avec votre m􏿽xE8re" or "veux-tu coucher avec ta m􏿽xE8re?" 11:43
you can't mix "vous" et "ta"
castaway good point cognominal
,) 11:44
(language discussions on #gaia, btw :)
kungfuftr cognominal: you can mix them, since you might be talking to more than one person 11:48
etc.
hhhmmm... though
cognominal that would be "voulez-vous coucher avec votre m􏿽xE8re" then :) 11:51
btw you are welcome on grou.ch #perlfr if you want to speak French. There is a #perlfr on freenode, but they are lamers 11:52
nothingmuch stevan: gong 11:53
nothingmuch wishes he'd wake up
nothingmuch tries to speak french occasionally
i started learning twice 11:54
and forgot nearly all twice
=(
kungfuftr speaks far better french if he spends a day or so in paris, etc. 11:56
cognominal if you do come to visit the mongueurs... 11:58
or come to Marseille for the French Perl Workshop 11:59
Juerd kungfuftr: ($a, $b) := ($b, $a) swaps without copying. 12:02
kungfuftr: When there's no repetetion on the RHS, you can't speak of swapping.
nothingmuch kungfuftr: what is your opinion on that test model doc? 12:09
kungfuftr nothingmuch: which one 12:12
nothingmuch nothingmuch.woobling.org/test_result_model.pod 12:13
kungfuftr Juerd: ah ha! so $voodoo would be "perl6"
nothingmuch: not bad, though there should be no computated data within the model. 12:18
theorbtwo Hm? 12:19
How do you figure / why not?
kungfuftr & # lunch!
theorbtwo: ie: ratio... just give total, passed, failed, etc.... and to the computation elsewhere... would allow for weighting, etc. 12:20
theorbtwo If you want to compute it differently, you can compute yourself. 12:21
nothingmuch i've thought about that long and hard 12:22
and thought that perhaps it is a good idea
because the model must appear inconsistent in a simple way
if it is inconsistent
one could argue that a protocol with missing plan members has such cases which are failed
theorbtwo If you don't, the default is easy to use and good for most uses.
nothingmuch and a protocol could stub that
i think that it's a simple concept
and could be centralized into the model
and the model would have the 'default' interface, which has stubs
and perhaps a purer base class
or a purer set of methods
that don't stub
but i think stubbing is very important for writing good display tools easily
theorbtwo So do I.
nothingmuch kungfuftr: you don't have to use ratio
it's just that even Test::Harness gets it wrong some times
these computations should be standardized 12:23
and be put in one place
and should be available by default
theorbtwo There's nothing wrong with having both an ->expected and ->unexpected method, which are defined to be the negation of each-other. There's a reason perl has both if and unless.
nothingmuch ... and that reason is readabilty 12:25
it's easier to read something that sounds more like english
and there's a reason why people make fun of double negatives in english
OTOH english has many logical ambiguities
kungfuftr nothingmuch: pugs.kwiki.org/?TestModel *shrug* 12:33
nothingmuch sort of 12:34
why are you against computations, btw? 12:35
kungfuftr nothingmuch: blame mugwump... he enstilled a sense of pure data modelling into me
nothingmuch hmm
normally I agree 12:36
and I think that a pure data model *should* be accessible
but i think that normally it shouldn't be the case
kungfuftr nothingmuch: a method to do it would be nice, but the actual data be pure?
nothingmuch yes, the underlying data is pure
kungfuftr ah... k
nothingmuch Test::Model isa Test::Model::Pure
and Test::Model has data query methods that append or mark sub objects 12:37
after taking them from the Test::Model::Pure superclass
kungfuftr yar, sounds good... encapsulation++
autrijus rehi 13:27
Limbic_Region lo
autrijus so I wasted some time writing shuffle.hs ;)
Limbic_Region autrijus - the last smoke Corion did showed 178 or so tests still failing - any idea what the number is now? 13:28
Additionally, did the refactor of context parsing raise any bizarre edge cases of non-dwymery?
autrijus Limbic_Region: something around that.
Limbic_Region: yes, lots and lots of. 13:29
Limbic_Region those cases reported to p6.l ?
clkao autrijus: want to help make win32 build for beta2? 13:30
theorbtwo Allo, autrijus!
Limbic_Region was just hoping to get a summary
autrijus Limbic_Region: uh no, so far it's been cleared up by lwall and juerd 13:31
so, nothing language-shaking per se.
although I did find out some weird cases with := semantics. still thinking about it
theorbtwo <autrijus> That doesn't look right. <larry> But it is! 13:32
autrijus :D
Limbic_Region right - what I meant by cases of non-dwymery were rules that made sense at first and second glances but given situation X made no sense at all
kungfuftr autrijus: that operators/binding.t ? 13:34
autrijus kungfuftr: yeah, it fails, I'm investigating 13:35
kungfuftr autrijus: proabably the ($x, $y) := ($y, $x) i added
Juerd juerd.nl/filesize.png # I sometimes *love* KDE 13:37
(Not going to show you my pr0n archive, but I know now exactly which parts to clean up (hm, dubious)) 13:38
theorbtwo Autrijus, I'm starting to look at a more perlish (IE return whatever you want) eval_haskell, and I'm not sure how to implement it, and I'm also not sure I understand your Evallable design from meme.b9.com/cview.html?channel=hask...te=050416. 13:39
I was looking at possibly a largish valFromDynamic function, that takes a Dynamic (which is required to be a Typeable already) and makes it into a Val. 13:41
stevan pugscode.org <Overview Journal Logs> | r1773/winxp: 587/4256, smoke/win2k (occasional): xrl.us/fqum (214/4453, r2198) Mac OS X (160/4483, r2216) | pugs.kwiki.org 13:41
theorbtwo pugscode.org <Overview Journal Logs> | r1773/winxp: 587/4256, smoke/win2k (occasional): xrl.us/fqum (214/4453, r2198), Mac OS X (160/4483, r2216) | pugs.kwiki.org
stevan nothingmuch: BANG!! 13:42
autrijus theorbtwo: that looks promising enough
theorbtwo: the Evalable design is orthogonal 13:43
it's designed to make
eval_haskell "'x'"
eval_haskell "return 'x'"
both work.
by essentially adding the proper amount of lifting
theorbtwo Right. You could do that in valFromDynamic -- one is a Char, the other is an IO Char. 13:44
If you've got an IO a, then unsafePerformIO it, and get an a, and then do the normal thing on that.
(Well, from a Dynamic IO a to a Dynamic a, hopefully.)
(This assumes, of course, that I can figure out how to do that... I think it's dynApp, but I'm not sure, and dynApp seems to be undocumented.) 13:45
autrijus uh, valFromDynamic should probably not unsafePerformIO it.
valFromDynamic :: Dynamic -> Eval Val 13:46
theorbtwo Oh, that also makes a lot of sense.
autrijus you can do typecasing based on typeOf.
theorbtwo Righ. 13:47
Right.
autrijus perl5 shuffle: 14:15
4.619u 1.010s 0:05.89 95.4% 10+79736k 0+0io 0pf+0w
GHC shuffle:
3.007u 0.038s 0:03.10 97.7% 323+359k 0+0io 0pf+0w
autrijus proceeds to reply to BrowserUK
Juerd And pure perl 6 shuffle? :) 14:19
autrijus I don't know :) 14:20
maybe we should compile it to parrot :)
it may be even faster.
Juerd What do you think of Bool +$read|r to allow both the name 'read' and the name 'r'? 14:21
:read and :r
Same thing.
It doesn't play nice with precedence in the signature, but normal precedence plays no role there anyway.
autrijus I think it's a weird idea. 14:22
Limbic_Region apparently gcc 4.0 has a brand new optimizer - wonder if that would affect the Parrot results *grin*
autrijus Bool +$read|$r maybe?
I know!
Juerd autrijus: I had that before, and then I didn't like the redundant $ :)
But I agree that it looks much more natural
autrijus hm
what happens if the user gives both
Juerd Same thing as when the same name was used twice. 14:23
:read :read
:read :r
:r :read
autrijus k.
Juerd All exactly the same thing
autrijus I think the extra sigil is a win.
Juerd (And I don't know what happens then :P)
autrijus: The problem with it is that aesthetically (because of precedence), I want to repeat the + too there 14:24
And that leaves you with +$read|+$r, which is a lot of typing and linenoise
And begs for +$read|?$r, which is impossible to work with :)
But perhaps it's just my sense for aesthetics that needs adjustment
autrijus can you have 14:25
+$read|@r
?
Juerd I'd have no idea how to interpret that 14:31
This is a good reason to either not repeat the sigil, or to enforce that it is the same
s:2nd/ to// 14:32
autrijus gaal: perlmonks.org/?node_id=450426 14:34
Limbic_Region print rand 10, $/ for 1..1_000_000; # succinent 14:41
autrijus yeah, and I could've used perl -wle or something :) 14:44
it's a bit of stretch to call the haskell version as a "port" though. 14:45
Limbic_Region just realized who he was suggesting a shorter version too 14:46
autrijus because instead of sorting string refs, it sorts offsets
Limbic_Region hangs his head since his golf'ing handicap is astronomical
autrijus but problem is perl5 doesn't even have that option; substr() into a multi-megabyte string is not a win
theorbtwo Hometime already, Corion? 14:57
Oh, I guess it is. 14:58
theorbtwo doesn't have a clock on his screen at present.
castaway sighs 14:59
Corion waves from home. 15:01
Another bad job well done. Cpansmoking can be such fon 15:10
s!fon!fun!
pugs.exe: cannot cast from VList [] to Handle -- I think I remember that bug ... It happens in t/pugsrun/02-dash-n.t and 03*.t - they both try to read from =<> 15:11
autrijus indeed. fixing 15:13
fixed
ok, infinite loop gone. 15:14
may I request another round of smoke? 15:15
(r2217)
Corion Just smoked 15:16
r2216 - datenzoo.de/pugs/win2k.html
autrijus cool. then no need to 15:17
Corion (and no infinite loop there, although abs.t still looks weird :) )
autrijus ok. I think Pair wants to be Pair objects. 15:21
instead of VPair.
sadly I need to crash now. gotta wake up real early tomorrow. :/ 15:29
aiming for a weekend release, then 15:30
autrijus waves &
Corion good night then :) 15:31
Yay. async() and system() now work together it seems! 15:37
So there is little preventing me from writing a small HTTP::Proxy and testing it via subprocesses. Except that I don't have time this weekend :)
nothingmuch: BTW, if a test crashes, it would be nice to have the remaining (planned) tests not as plain fails but somehow as different fails. A darker red, for example :) 15:44
nothingmuch Corion: that can be done 16:02
want to patch? it's a simple fix
Corion nothingmuch: No hurry - just an idea :)
nothingmuch i just can't do it right now
what needs to be done to the model, to do it cleanly: 16:03
Corion nothingmuch: np - I won't be at home for the next few hours either
nothingmuch grep -r for 'stub'
find the code that generates them
and add a property
Corion I only wanted to dump the idea before I lose it again ;)
nothingmuch then make a proper method for it
and a css class for the display
and add knowlege about the class to Test::TAP::Model::Subtest::Visual
okay
Corion bbl 16:04
kungfuftr kungfuftr.com/pugs-smoke.html - back fer now 16:36
cognominal Larry has answered my mail about functional extension to Perl6. I don't get these two lines about already supported/speced features 16:53
autogenerating arrays
[] pattern matching arguments
is the first, only our classic map and grep?
where is the second specced? 16:54
nothingmuch autogenerating arrays: anything infinite, i think 17:23
and much more flexibly: gather { take }
i think that it will pause when taking
and then when a new element is needed it continues after the take()
and reiterates the gather block
or the nested loops inside it, more accurately 17:24
anyone here a getopt guru? 17:54
i need some advice
cognominal I still don't get what he means by "[] pattern matching arguments" 18:02
crysflame context? 18:08
PerlJam cognominal: sub foo ([$head,@tail], $second_arg) { ... } 18:09
cognominal: foo would be called as foo(@array,$blah);
nothingmuch: I'm not a guru but I can play one on IRC. :-) What advice do you need?
ingy all I know is that every programmer writes their own version of getopt at some point 18:10
PerlJam ingy: indeed. That was one of the first things I wrote in python when learning the language. 18:11
cognominal Perl_jam: thx, this is indeed a form of pattern matching 18:16
boy, I will love Perl6
PerlJam cognominal: You don't love it now? ;) 18:18
PerlJam notes that perl5 was released 2 years before it was stable enough to be generally useful. (give or take) I imagine that perl6 will be the same. 18:20
nothingmuch sorry, back
sister was exploiting my licensed status
PerlJam: Is there a fun module that lets me convert args to something Params::Validate might like to eat? 18:21
basically, i want @ARGV to be treated as a positional + flag + named param list
and I want to name positionals and flags myself
and be able to say $config->field();
and get the CLI value
castaway GetOpt::Long ? 18:22
PerlJam nothingmuch: Getopt::Declare perhaps? (I don't really know, but look on CPAN in the Getopt:: namespace)
nothingmuch castaway: i am sort of hoping to do it without writing code
and guessed someone might have something for the job
castaway pigs may fly ;)
nothingmuch castaway: tsk tsk
reminds me of yet another discordian quote:
PerlJam castaway: hey, that's not an unreasonable expectation. Almost everytime I think to do something cool, I find that someone has already done it.
nothingmuch "Eagles may soar, but cows don't get sucked into jet engines" 18:23
nothingmuch wonders what shapr is doing instead of appreciating my references
nothingmuch feels guilty all of a sudden
i just realized that I make about 300-400 queries a day on search.cpan.org 18:24
i should donate something
search.cpan.org/~abw/AppConfig-1.56.../Getopt.pm <-- hmm
nothingmuch thinks the only sort of sportish thing he can do better than others, aside from throwing knives, is running up the stairs with coffee 18:27
for some reason it never ever spills
crysflame nothing: fwiw, google site:search.cpan.org works 18:29
nothingmuch crysflame: usally i type search.cpan.org/perldoc?Some::Module 18:31
i do this a lot for modules I don't have installed
or for modules with long doc pages 18:32
where the terminal doesn't give me enough perspective
(no scroll bars)
ingy nothingmuch: have you seen my new baby, Test::Chunks? 18:41
nothingmuch nope
ingy is a proud mother
castaway wonders whether it needs Spiffy/Spoon/IO::All ,) 18:42
nothingmuch hmm... looks very useful
ingy castaway: :p 18:43
nothingmuch ingy: you're good at speccing things
castaway SCNR ;)
nothingmuch nothingmuch.woobling.org/test_result_model.pod
ingy nothingmuch: how so
nothingmuch the model isa Test::Results::Model::Pure
ingy: YAML is pretty good, kwid is pretty good
Kwiki is fun to work with (except in hebrew, that was a pain) 18:44
ingy I didn't write the YAML spec
nothingmuch (but then again, CamelCase is not possible in hebrew)
nothingmuch thought ingy was part of the spec
anyway, so kwid is pretty good
and that's it
ingy I helped
nothingmuch but despited the fact that you suck at speccing
ingy but Oren did most of it
nothingmuch i'd like you to look at that pod ;-)
s/despited/despite/; 18:45
ingy ok
nothingmuch is the 'all' target the default for makefiles? 18:46
i.e., is that what's implied when you run make with no args?
castaway think so
ingy nothingmuch: we should talk about this at the hackathon
nothingmuch ingy: hackathon is far away 18:47
gaal wants to do Test::Harness in p6
i said it should be split up
ingy it fits well with my sidelined babeltest framework
nothingmuch because IMHO Test::Harness sucks for reusability
babeltest?
ingy it is a descendant of Ward Cunningham's fit testing
a defined object model would be really useful 18:48
a well defined one
babeltest has an adhoc object model
but it is a object tree based testing system
nothingmuch what is FIT? 18:49
ingy have you ever looked at FIT?
nothingmuch no
ingy guess not
nothingmuch 10 word limit ;-)
ingy it is very cool
nothingmuch boy, you're wasteful
ingy you should check it out
I have a Test::Fit on CPAN 18:50
?
nothingmuch 40% of what FIT is by that def is 'very cool'
if you consider the 10 word limit
ingy I'm not considering it though :P
I never really finished Test::FIT 18:51
nothingmuch yeah yeah
fit.c2.com/ <-- good enough start?
ingy FIT is cross language testing framework
yes
you can spec a project with FIT and implement in any language
nothingmuch hmm 18:52
is it sort of like a test plan
and the second argument to is()?
ingy Babeltest is my fork of it
you should read about it
before comparing it to Test::Harness
nothingmuch okay
i'll do that
ingy I actually made Test::Fit output Test::Harness output 18:53
nothingmuch my grand scheme is that Test::Result::Model is generic enough to be useful
so that most test protocols can fit into it
and those that can't should easily be modded
ingy it should be programming lanuage independent
nothingmuch or are just too dumb, and that's their problem
it is
ingy good
nothingmuch i'm already using Test::TAP::Model for a C++ framework
with modified TAP output
so the Test::Result::Model should have an interface through which it can be constructed easily 18:54
it should know to subscribe to a standard event parser interface
and then you have Test::TAP::Parser
so you smudge the two together
you can also subscribe to Test::TAP::Parser if you want to display running results, for example 18:55
but the point is that after words, for aggregated display
you have a very stable
very standardized object model
that you can feed in a cross language way to display tools
if it's in p6 then theoretically through parrot this is shared
ingy I haven't looked at TAP yet
nothingmuch so reporting tools in ruby could use it
ingy do you think test chunks could benefit from TAP integration? 18:56
nothingmuch tap is nice
but i think a bit too flat
base it on Test::Builder
ingy base what?
Test::Chunks 18:57
nothingmuch yup
ingy it is already built over Test::More, so yeah
cognominal I try to compile hs-plugins and I get : ./Plugins/ParsePkgConfCabal.hs:280:49: Not in scope: `hugsOptions'
nothingmuch ah, okay
so it is integrated
ingy you can use it just like test more
I did one cool trick... it adds strict and warnings to your test files automatically 18:58
stole that from Spiffy
nothingmuch source filters? aaaah!
ingy yeah, why not. just forget about it 18:59
it is truly useful
nothingmuch =) 19:00
ingy the test suite for Test::Chunks is mighty clean
nothingmuch does it use Test::Chunks?
ingy but I got bit by not using strict
theorbtwo cognominal: Try asking in #haskell? 19:01
ingy anyway a test model sounds lovely
let me know when it gets more formal
nothingmuch i wish I had more time to code it
but Test::TAP::Model is a good start
what I would like to do is purify it a little 19:02
ingy I'll study that
nothingmuch reimplement Test::Harness::Straps so that it's not so icky to work with
Juerd www.guardian.co.uk/online/news/0,12...73,00.html
This means that in reality, we're even much smarter!
nothingmuch make Test::TAP::Model not be isa Test::Harness::Straps just because that saved me some coding
crysflame nothingmuch: hmm, nice.
Juerd (Weird how it can differ 10 "iq points", as IQs are normally-distributed quotients...)
nothingmuch cannot mentally perform on marijuana 19:03
i slow down
to a point where I'm too lazy to formulate sentances
ingy my idea for babeltest.org was to have a wiki where anyone can add tests to define a project. and then implementors can use the test specs against their project in their language of choice. and the community could always see what tests pass. 19:04
cognominal Juerd: every month, psys find a new addiction to enlarge their line of business
ingy which is what fit really should have been
but ward is concentrated on business not open source 19:05
nothingmuch ingy: this reminds me of wheat
ingy and then ward went to work for M$ and fell off the planet
wheat? 19:06
crysflame ingy: whoah!
ingy++ # community testing wiki
nothingmuch and how do you do complex tests?
ingy make them simpler ;)
nothingmuch www.wheatfarm.org/
ingy: simpler might mean rewriting
i think a test with loops is better than a test that is completely imperative 19:07
if the loops can save some writing
i normally write tests with 0 control flow
so that they are always the same
ingy nothingmuch: what do you think of this style of test? www.rafb.net/paste/results/eO2TDJ84.html 19:10
Juerd: are you behind www.rafb.net/paste/ ?
nothingmuch why is the second run where it is? 19:11
i don't really understand
mauke no, unless Juerd is really Jacob Cohen
ingy the 2nd run?
nothingmuch after __END__ there is the data 19:12
oh
it's just text
silly me
ingy just a piece of text
happen to be lying around ;)
nothingmuch so how does $chunk->(decoded|encoded) know to compare?
i like the fact that you don't "intelligently guess" the plan but leave it for the user 19:13
ingy the 'encoded' chunks are given a base64 filter
nothingmuch lets backtrack
ingy ok
nothingmuch the test is supposed to test a base64 decoder?
ingy the test is testing that the base64 filter feature of Test::Chunks works properly 19:14
nothingmuch or is base64 encoding an integral part of Test::Chunks
ah
pretty
ingy you can specify any number of filters on a test chunk
normalize and trim are the defaults
castaway Juerd is the tnx.nl one 19:15
ingy er, norm and trim I think
nothingmuch ingy: sounds very useful
ingy you can turn it off with
nothingmuch ingy++
ingy --- foo -norm base64 whatever
you can concentrate on one test with the pseudochunk 19:16
--- ONLY
that's why I base my plan on chunks
nothingmuch are 'decoded' and 'encoded' arbitrary names?
ingy yes
and they become methods of the Test::Chunk object 19:17
nothingmuch Algorithm::Dependency's test suite could use this mapping
want to have a go at translating? ;-)
it has a big matrix every once in a while
and i'm too lazy to finish the suite for Algorithm::Dependency::Objects
(nothingmuch.woobling.org/Algorithm-...y-Objects)
Juerd ingy: No, I have nothing to do with rafb.net 19:19
ingy oh ok. it looked somewhat like a pastebot you used I thought...
castaway tnx.nl/scribble.plp ? 19:20
ingy I found it by typing 'nopaste' in the firefox location area
castaway bookmarks++
Juerd ingy: I own the domain pastebot.org
ingy: But didn't create pastebot
And don't maintain the pastebot linked to that domain.
castaway ah
Juerd tnx.nl is also mine 19:21
Which I still like a lot for its lack of length.
ingy I love typing free text as urls in firefox. it will usually dwim
Juerd Oh, I completely forgot I was going to add [AES]\d\d shortcuts.
ingy: Google loves you.
For Great Statistics
ingy rafb's paster lacks a couple polishing features though 19:22
at least someone loves me
theorbtwo (Crack for Statisticians!)
cognominal for my slides I want colorized perl6. I am thinking of using vim for that to generate colorized html. 19:26
do you know the correct invocation from a script? or a better way? 19:27
Juerd My experience is that vim's highlighting sucks for Perl 6. 19:29
castaway highlighting perl6 sounds like a nightmare ;) 19:31
cognominal this is not for every day use, just for generating a few slides. I will not use exotic quoting...
Juerd cognominal: It fails for more trivial stuff than that too. 19:32
You'll see
cognominal what I see so far is good enough
Juerd tnx.nl/S03 now works, btw
castaway cool
cognominal I just want to automate the generation of slides
Juerd /bind meta-- insert_text tnx.nl/ 19:33
in irssi
PerlJam castaway: hilighting perl6 will be *easy* if we have the perl6 rule engine and the grammar at our disposal.
Juerd then it's alt+- S03
castaway hmm?
where does that insert what to?
Juerd castaway: It makes alt+- a shortcut for typing tnx.nl/ 19:34
tnx.nl/http://tnx.nl/http://tnx.nl/...://tnx.nl/ :)
castaway ah, I see
Juerd So if I want to point someone to perlop, I just type alt+- perlop, resulting in tnx.nl/perlop
castaway nifty
Juerd Or for suffering from buffering, tnx.nl/sfb
cognominal I just want a quick fix, not a general solution...
Juerd Or when someone has problems with understanding Perl 6 operators and rules, tnx.nl/S03,S05 19:35
Etcetera :)
castaway PerlJam: build those into elisp ;)
sorry cog, wrong $EDITOR
PerlJam castaway: sure ... as soon as I have a parrot backend to elisp.
cognominal castaway, I would not mind an emacs solution 19:36
castaway grins at cognominal
I dont have one handy tho, sorry
cognominal btw. are the synopsis on dev.perl.org up to date? or should are I trust the Perl6::Bible instead? 19:37
ingy the bible is up to date
PerlJam cognominal: the synopses are the most up-to-date official docs by the cabal
ingy 0.15
cognominal PerlJam: yes, but are the synopsis online on dev.perl.org up to date? 19:38
obra c1;dev.perl.org should be current as of last night.
and that should always be true.
PerlJam ingy: is the bible generated from the svn repo?
obra dev.perl.org pulls nightly from the canonical svn repo 19:39
cognominal so why the last modified so out of date?
ingy PerlJam: yes
PerlJam ingy: excellent.
cognominal Last Modified: 9 Dec 2004
Number: 5
that convey the wrong message 19:40
obra cognominal: when was syn5 last updated?
cognominal obra: I don't know. my understanding was that it was recently updated 19:41
castaway "21:38 < theorbtwo> echo foo | md5sum will sum the trailing newline that echo pugs there" :) (pugs on te brain..) 19:42
PerlJam cognominal: it had some minor updates in Feb I think, but hasn't been changed since. 19:45
But the last major change was in Dec.
obra svn log svn.perl.org/perl6/doc/trunk/design/syn/S05.pod 19:46
there have been what appear to be typo fix class updates in feb, yeah.
actually
given those dates, I think that's the date it was loaded from cvs 19:47
PerlJam obra: still, there are no modifications since that date (that have been checked into svn) 19:48
obra PerlJam: right
cognominal: what changes did you think there were? 19:49
cognominal I thougt you were just talking changes, so I was alarmed to see the changed date on dev.perl.org. so my questions? 19:50
nothingmuch oh crap: Set::Object::_(self, ...) 21:00
prepend: Usage:
from the likes of: my $found ( @prospects ) { 21:01
if ( -f $found && -x _ ) {
nothingmuch pulls mugwump up by the ear
what time is it there? around morning, no? 21:02
Qiang 5pm here. weekend starts :) 21:10
stevan uh oh,.. someone went to sleep and broke the build :) 21:11
theorbtwo Odd, most people do it the other way around. 21:13
stevan hmmm I wonder if src/Pugs/Types/Object|Pair.hs are on autrijus's machine 21:14
and he forgot to svn add them 21:15
I commented out the imports and so far it is compiling
theorbtwo It'd hardly be the first time somebody made that mistake.
stevan we shall see
:)
I do it all the time too :P
sleep deprivation has that effect
theorbtwo I don't have that excuse, though. 21:16
stevan theorbtwo: do you want me to make that OS 9 image?
theorbtwo Yes.
Thanks.
stevan BTW _ I realized that I already threw out the jaguar images
theorbtwo Do you have rsync? 21:17
stevan no idea
never used it
theorbtwo Do "which rsync" from a command line? 21:18
stevan got it 21:19
theorbtwo OK, now I have to figure out how to do this. 21:20
stevan ok 21:21
theorbtwo Pick a username. 21:29
stevan?
stevan yup
21:48 BigBear is now known as BigBear_hiding_o
castaway hugs Limbic_Region 21:59
theorbtwo stevan, can you put the jaguar 1 CD in your drive, then do "sync --progress -vz --rsh='ssh -p2222' /dev/cdrom desert-island.dynodns.net:/usr/src/mac/jaguar-1.iso" please? 22:01
Er, s/sync/rsync/
Limbic_Region hugs castaway and apologizes for not reciprocating sooner 22:03
castaway no worries
hows tricks over the pond? 22:05
Limbic_Region pretty good 22:06
castaway nifyt
Limbic_Region just found out the original closing cost estimate was inflated
$5200 versus $700
nothingmuch apropos prices, castorbway: did you guys find some proper flights? 22:07
castaway proper?
no.. getting-from0a-to-b flights, yes (more or less)
QMario How do I make Perl print out random numbers in a certain range? 22:08
castaway perl6 ?
theorbtwo (Low closing costs)++
nothingmuch proper like not too expensive
Khisanth castaway: perl5 but since he got banned from #perl I guess he is trying here :) 22:09
castaway grins
QMario How do you know that, Khisanth?
castaway print rand(6) -> random number from 1 to 6, I think 22:10
perldoc -f rand
QMario Thank You.
nothingmuch QMario: perl -ne 'BEGIN { $/ = \1 }; $_ = unpack("L", $_); print "$_\n" if $_ > $min and $_ < $max'
perl -ne /dev/random
mauke haha, nice
castaway: you're wrong, btw :-)
castaway nothingmuch: in perl6 pls!
nothingmuch err, $/ = sizeOf, ofcourse
Khisanth QMario: because I was around when you decided to go preaching religion in the wrong place 22:11
nothingmuch say rand(6);
theorbtwo From 0 to 6, including 0 but not including 6.
castaway (please dont bring it here, hatever it was)
castaway was close!
Limbic_Region wonders why BrowserUk made this statement www.perlmonks.org/?node_id=450588 ? 22:14
castaway he needs a reason?
Khisanth Limbic_Region: ask him/her!
Limbic_Region Khisanth - I was wondering if there was something obvious that I was missing
that they were comparing apples and oranges or something 22:15
that autrijus cheated
dunno
castaway presumably he was thinking that haskell was inline in perl there? and not underneath 22:16
castaway shrugs
QMario How do i print random numbers as integers(whole numbers).
?
nothingmuch how childish
that's haskell
Limbic_Region QMario - this is NOT #perl or #perlhelp
nothingmuch had much more respect for BrowserUK 5 minutes ago
Limbic_Region asked the question in a /msg 22:17
theorbtwo QMario, it isn't perlmonks.org either.
castaway QMario: first you get a perl book ..
nothingmuch QMario: my code would do that
it reads data from /dev/random
QMario Can you send it to me?
nothingmuch and then unpacks it as a long
castaway or you do perldoc -f int
nothingmuch i pasted it here
perl -ne 'BEGIN { $/ = \1 }; $_ = unpack("L", $_); print "$_\n" if $_ > $min and $_ < $max' /dev/random
mauke QMario: do you know what perl6 is?
nothingmuch just set $min and $max 22:18
QMario I don't think pastebot works for me?
nothingmuch it's not very efficient though
QMario No.
That was for you ,mauke.
mauke nothingmuch: hmm, don't you need $/ = \4?
Khisanth Limbic_Region: well judging from that person's post, he doesn't seem very familiar with Haskell so ...
nothingmuch mauke: yes
sorry, i mentioned that, but didn't fix this paste around
anyway, it will take (4 * (2**32))/abs($max-$min)/2 bytest on average 22:19
Limbic_Region Khisanth - well, I am pretty oblivious myself to Haskell and only a margine more educated about Perl internals - so that's why I asked (I figured I was just being dumb)
nothingmuch depending on how fast your /dev/random is, it might take about a minute
castaway ;)
nothingmuch to read the GB or so of entropy
castaway QMario: type: perldoc -f int
QMario Okay.
nothingmuch however, if you read from /dev/urandom, it will be faster, but the numbers won't be as cool 22:20
castaway and perldoc perlfunc, for all the other functions you may need somewhen
QMario Thank You.
What is perl6 for? 22:21
nothingmuch like perl 5 was for system administrators
stevan QMario: to replace perl5 :)
castaway its the next version after perl5, eventually
nothingmuch perl 6 is for farmers
it's got all sorts of features to simplify the language
like junctions, and chaining comparisons
castaway laughs
QMario So you guys are just getting ready for it? 22:22
theorbtwo No, we're creating it.
Juerd Oh, we're ready
castaway creating it, kind of
Juerd I know I am
Perl 6 will magically solve all problems in my life.
Limbic_Region Juerd - that's not true
it still won't get you laid 22:23
QMario Have you made any useful programs so far, Juerd?
Limbic_Region :P
Juerd Limbic_Region: Then what do you think the dwim operator does?
stevan QMario: we have an implemenation of hangman
Juerd QMario: No, never.
theorbtwo It's all about the yaddayaddayadda operator. 22:24
castaway gggles
theorbtwo Every program can be written as #!/usr/bin/perl6 -e '...'
Juerd QMario: The only thing I made that's semi useful is a program that prints a line on a matrix printer every time a certain mailbox receives new mail, so the rattling of the pinwriter indicates new mail. It's somewhat more comforting than a dingdong sound or beep.
castaway ah, thats "every program" reduced o its sigle bug, right?
theorbtwo Exactly! 22:25
Limbic_Region every program reduced to a single line containing at least 1 bug
Juerd reports numerous bugs for yadayadayada and dwim().
nothingmuch replies to BrowserUK 22:26
Juerd A recent PM thread makes me wonder
How old are you, on average? 22:27
My average age is 10.
castaway is an average of 8.
nothingmuch eh?
what is 'old' in this context?
castaway (oh, actual calculation?)
Juerd nothingmuch: What's your average age? :)
castaway mental age
nothingmuch behavioral like?
Juerd sum(@ages) / +@ages
And then the int of that 22:28
Limbic_Region Juerd - depends on if you say your average age at your first birthday is .5 or not (0 years old is valid or not)
theorbtwo Nah, she's an average of around 18, but it's on a binormal distribution.
nothingmuch Limbic_Region: in the first year of our lives i think we exist
Juerd Limbic_Region: 0 years is an age, so it counts.
nothingmuch is there some kind of reference table?
i don't know what to say
Limbic_Region nothingmuch - then it is real easy
Limbic_Region likes his reply in that thread better anyway 22:29
nothingmuch sometimes i feel like i'm 12
when i'm being annoying because i'm bored
and i only realized i'm annoying because i'm bored after a while
at times i feel like i'm 6, if i'm really enjoying myself on something stupid
Juerd use List::Util qw(sum); int(sum(0..$current_age)/$current_age+1)
Perl 5 feels so primitive
nothingmuch at work i feel 30, because everyone else is 22:30
with friends i feel +=5 because they're pretty goofy
ah
castaway Juerd: you mean p6 would DWIM on the $current_age bit?
Juerd Average age has nothing to do with feeling
castaway: Haha
nothingmuch 11 22:31
Juerd castaway: No, get rid of List::Util and have all that built in
nothingmuch i thought that's what you guys meant
castaway ;)
castaway thinks the result is as meaningless as ones actual age
Limbic_Region Juerd - no need 22:32
Juerd castaway: Duh, it's based on that and only that :)
Limbic_Region sum of all numbers is (n^2 + n) / 2
castaway exactly.
Juerd Limbic_Region: I'm writing it like this so that someone who skipped a certain age can more easily modify the code to match his circumstances. 22:33
I always code with maintenance in mind.
Limbic_Region or someone who is 29 multiple years in a row?
;-)
theorbtwo So when you write difficult-to-maintain code it's on purpose?
Juerd Or other weird ages lists.
theorbtwo: Yes. 22:34
nothingmuch goes to make tea, and then hack some spork
Juerd I don't use complex mathematical solutions, like Limbic_Region does, just because they happen to match the current data ;)
castaway wonders how to skip an ag ;)
Jued++
oops +r
theorbtwo
.oO(Hey Jued...)
 22:35
Juerd castaway: I hope I find out before I'm almost 30, because that's one age I'd like to skip entirely. 22:36
Limbic_Region 25 was the big one for me
Juerd castaway: When I do find out, I'll let you know.
castaway It wasnt all that bad, iirc ;)
Limbic_Region quarter of a century
castaway BTDT
Limbic_Region 1/3 of your life
Juerd btdt?
Limbic_Region: optimist ;)
Limbic_Region doesn't think 30 will bother him at all
castaway only going to survive to 75, Limbic?
theorbtwo I'm going to be a quarter-century soon.
castaway Been There, Done That
castaway is 31 22:37
Limbic_Region castaway - more or less
theorbtwo is planning on living to 120.
castaway but feel free to relate your impressions once you get there
Limbic_Region well - I plan on trading this body in so 75 is a high end estimate 22:38
theorbtwo As in reincarnation or as in brain transplant? 22:39
Limbic_Region forgets that theorbtwo is jewish
neither of those two
castaway limbic, you forumla gets me 496 !?
what are the alternatives? 22:40
Limbic_Region castaway - that's the sum - you still need to divide by the number of elements (current age) to get the average
castaway oh, you didnae say that!
16!
Limbic_Region castaway - I didn't say the n^2 + n / 2 formula was how to get average age, I said it was how to get the sum
castaway cant read obviously ;) 22:41
Juerd re 22:45
theorbtwo L~R, still don't know what you mean by trade it in.
Juerd 00:41 < Limbic_Region> castaway - that's the sum - you still need to divide by the number of elements (current age) to get the 22:46
average
No, not current age...
Current age plus one
By code was wrong because it had precedence wrong. Sorry.
Limbic_Region Juerd - again, that depends on if you consider 0 a valid age
Juerd (Wow - such a simple thing and we're both getting it wrong)
Limbic_Region: Unless you skipped it, it should be counted. 22:47
Limbic_Region in my code I caveat the 0 so I'm not wrong :P
Juerd In that case you were 1 when you were born, I think.
Limbic_Region: You were wrong by saying one still needs to devide by the current age.
Limbic_Region no, it means your age doesn't begin until you reach 1
Juerd No, I think 0 is a valid age.
Limbic_Region and that's your right
Juerd I know the first year of living has age 0 in some databases. Not NULL, but 0. 22:48
Limbic_Region but for the purposes of my code it isn't - and I said so - so I am not wrong
you may disagree with me all day long
Juerd That's another 23:10
No thanks.
Limbic_Region heh 22:49
all I mean is that given my stated assumptions the code is correct - the assumptions themselves may be invalid/incorrect, but they are in agreement with each other
castaway *g*
Juerd The code is correct, but your explanation involving the code was not. 22:50
theorbtwo shrugs.
I think that 0 is a valid age.
nothingmuch the code may be correct, but argument is redundant
Juerd And I repeat that your code is only correct for objects that have never skipped an age or had a certain age twice.
Although if you have the same age twice in a row, I don't know if it then counts as one.
theorbtwo If you ask someone "how old is your (son|daughter)", then you will often get ages 0 > $n >= 1. 22:51
nothingmuch . o O ( the same age twice in arrow? )
Juerd It certainly counts as one if I have a certain age today and the same tomorrow, but I don't know how this goes if the age of an age is more than one.
theorbtwo Counting each discrete year only once is a good estimate. 22:52
Juerd So if your age is 10 for two years in a row, it counts twice?
You know - normally, this is the area of biologists 22:53
But we programmers are expected to know everything
For otherwise we couldn't create programs except programming languages.
theorbtwo And note that I am not the same age as I was yesterday -- I am one day older.
Juerd Age, in this discussion, is strictly measured in years, rounded down. 22:54
theorbtwo So my average age during the year from my 10th birthday to my 11th birthday is 10.5.
Juerd I should have mentioned that.
castaway depends how granular you want to go ;)
ah ok
Juerd theorbtwo: Yes, but the average age in turn is int()ed, as you see in my example code.
Hm, from age 9 to age 11, my average age was 10. It's 10 now too. I didn't age in 10 years! 22:55
theorbtwo You're only 20? 22:56
Juerd Even though I was neither 21 nor 20 at age 10.
sorje weeeh, revision 2222 :-)
Juerd Weird.
theorbtwo: 21.
sorje: Congrats :)
Hi, broquaint 22:57
broquaint: What is your average age?
broquaint Hey, Juerd.
I don't have an average age as such. 22:58
Juerd broquaint: perl -MList::Util=sum -le'my $current_age = shift; print int(sum(0..$current_age)/($current_age+1))'
Oh. An interesting new possibility.
broquaint: Are you skipping every single age, then?
nothingmuch Juerd: how old are you, unaveraged?
Juerd nothingmuch: Still 21, as I said and was 3 minutes ago. 22:59
Limbic_Region Juerd - you count years how you want, I count them like the calendar 4, 3, 2, 1, -1 (there is no 0)
broquaint I try, Juerd, I try.
nothingmuch oops, sorry
nothingmuch is not really into chat
broquaint Apparently that code isn't favoured by the windows command line.
nothingmuch with all this fascinating discussion going on
Juerd broquaint: Sorry - In geek channels I always assume sane shells.
castaway bro!
theorbtwo Hm, Limbic, interesting... you are -1 from birth until your first birthday? 23:00
theorbtwo thinks you are 0.
If you ask a proud new parent how old their child is, they will often answer something like "8 months".
Back-extrapolate.
broquaint Hey, c :)
Limbic_Region wanders off
Juerd Limbic_Region: Hm - I never really gave this much thought, but this does mean the proverbial "year 0" never existed, doesn't it?
theorbtwo Between AD and BC? Nope. 23:01
Juerd read AC and DC. Hm.
theorbtwo If an event lasted from Jan 1, 1 BC to Dec 31, 2 AD, it lasted 3 years.
(Approx.) 23:02
Juerd theorbtwo: Why the inconsistency in that, by the way?
broquaint Assumption is the mother of all fuck ups, according to the bad guy in Under Seige.
Juerd Before christ, but anno domino...
(assuming that is what AD stands for)
theorbtwo Histerical Rasons. 23:03
broquaint It would seem my average age is 11.
theorbtwo In east-asia, it's common to consider someone 1 at birth.
Juerd broquaint: You're only one year older than I am, on average :)
theorbtwo: Then east-asians have a somewhat higher average age (or lower, based on equality of current ages)
theorbtwo Somebody in my HS got their driver's license after having been out of the womb only 15 years because of that.
Juerd theorbtwo: That at first read is funny for someone who lives in a country where driving age is 18 :) 23:04
theorbtwo grins. 23:05
I'm American, and more specifically Pensylvanian.
Also, I can't spell.
nothingmuch where in Penn? 23:06
theorbtwo Lancaster -- Manheim Township.
nothingmuch doesn't know it 23:07
couldn't bother with geography when there was Sim City 2000 and the great outdoors
oh well, i was young and wreckless
reckless?
nothingmuch used to live in Pittsburgh
castaway ooh 23:08
theorbtwo Oh, I lived in Pittsburgh for a while as well.
castaway so did he
theorbtwo When were you there?
nothingmuch has very fond memories 23:09
93-94
theorbtwo Oh, well before me.
nothingmuch no, sorry
94-96
theorbtwo has some fond memories and some not so much.
nothingmuch 93-94 was in Princeton
broquaint Doing what? 23:10
theorbtwo Going to Pitt.
nothingmuch mum and dad were on subsequent sabbaticals
i think those were approx. their post doc years
theorbtwo Well, mostly being enrolled in Pitt and not going.
broquaint Cool. 23:11
nothingmuch has never been on the Pitt campus, even though mom tought there
CMU was very nice though
the robot tracks were cute
i always wanted to see one in action but never got the chance
Juerd Wow, I received spam on a domain that hasn't existed for at least 2 years. 23:20
Some dns cache they use.
(And perhaps I should remove that domain from my server... :)) 23:21
broquaint What was the average age thing about? 23:22
Juerd Inspired by perlmonks 23:23
It was a very interesting conversation about the possibilities surrounding ages.
We take too much for granted.
broquaint Of course we do, we're civilised. FSVO civilised ... 23:24
Juerd civilised 23:25
adj 1: having a high state of culture and development both social
and technological; "terrorist acts that shocked the
civilized world" [syn: {civilized}] [ant: {noncivilized}]
If social development is a requirement, I'm not civilised.
broquaint Although it doesn't state what civilised need be applied to ..
Juerd 2: marked by refinement in taste and manners; "cultivated 23:26
speech"; "cultured Bostonians"; "cultured tastes"; "a
genteel old lady"; "polite society" [syn: {civilized}, {cultivated},
{cultured}, {genteel}, {polite}]
Even more vague.
nothingmuch ingy: ping
broquaint Ah the dictionary, is there anything it won't tiptoe around? 23:27
ingy nothingmuch: pong 23:28
nothingmuch i want to subclass Sprok::Slides 23:29
ingy: want to SEE some spork hackings?
i need a bit of hand holding
and make Spork::S5 produce with class="inc"
ingy SEE is fine 23:30
location?
nothingmuch see://woobling.org
Limbic_Region nothingmuch - BrowserUk replied
Juerd Good night 23:32
zz
castaway morning juerd
nothingmuch good reply
ingy: one sec, it's pointing to the ether ip 23:33
ingy nothingmuch: seem to be having troubles connectinf
nothingmuch i'm on the wifi
try again
revdiablo Limbic_Region: heh, it seems like B-Uk's reply just changed the subject, rather than addressing the meat of the post 23:39
nothingmuch revdiablo: but he doesn't bite
he explains he prefers it
instead of arguing pointlessly
revdiablo nothingmuch: *nod* but he could have at least acknowledged that his earlier assertions were wrong 23:40
theorbtwo What is the correct device name, BTW?
nothingmuch true that
theorbtwo: eh?
Limbic_Region well - his mistake was associating Haskell with a pure FP language
theorbtwo For the whole CD-ROM drive.
Limbic_Region but I for one make mistakes like that all the time
theorbtwo It is, but it redefined "pure". 23:41
Well, at least it's mostly pure.
Limbic_Region *shrug* - he should have just replied with s/Haskell/side-effect free language/g and been done with it 23:42
revdiablo Limbic_Region: everyone makes mistakes, but some people don't ever seem to admit it. they just change the subject. :) 23:43
theorbtwo I like his last post. 23:45
450609, that is. 23:46

Powered by Cro and the Raku® Programming Language.

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

Thank you!