Run Perl6 Now! | pugscode.org ('Overview', 'Journal') | pugs.kwiki.org | channel log: xrl.us/e98m
Set by autrijus on 6 March 2005.
metaperl_ I got the Algorithms book by Rabhi and Lapalme 00:08
hi cls_bsd 00:24
do you work for autrijus' company also?
ingy mauke: can you help me once more? 00:40
pasteling "ingy" at 24.19.10.51 pasted "can someone look at this? I can't get it to compile." (23 lines, 685B) at sial.org/pbot/7959 00:43
hoowa hihi 01:07
sleeping? 01:08
hihi i want to rewrite module in perl6. does pugs support all perl6 syntax? 01:16
rooneg nothing supports all perl6 syntax yet ;-)
01:17 khisanth_ is now known as Khisanth
hoowa ??? 01:17
poke
stevan hoowa: pugs is not done yet :)
hoowa: it is missing some very important features still
hoowa supports how much?
stevan hoowa: whats your module? 01:18
hoowa in pate
my pate 01:19
stevan is it on CPAN?
hoowa not i forgot my password on cpan. i want to rewrite other module ini read ? md5 ? crypt::blowfish ? ...... 01:20
i like crypt::blowfish, that one pureperl5 named crypt::blowfish_pp 01:21
stevan hoowa: Pugs is still in the early-alpha stage, it will not have enough features 01:22
hoowa: but you can write help tests
hoowa: or write some example code
hoowa example code ? about "helll world"? 01:23
stevan have you downloaded Pugs yet? 01:25
hoowa yes. but not installed. 01:26
stevan look in the examples directory 01:27
there are some working examples in there
hoowa ok thank u. 01:28
ingy is having minor monadic success :D 01:41
rooneg anyone else seeing this error? src/Main.hs:159: Variable not in scope: `config_file_sep' 01:43
scw rooneg: make clean; perl Makefile.PL; make 01:44
rooneg: I met it five minutes ago :p
ingy scw++
make purge is better 01:45
btw
rooneg weird. i had tried 'perl Makefile.PL; make', but hadn't done the make clean. i wonder why that mattered...
ingy rooneg: it thought your Makefile was up to date
rooneg hates make
ingy make is your friend 01:46
rooneg except when it's not ;-)
ingy just get in the habit of:
rooneg didn't you just spend like a day fighting with makemaker? i'd think you'd be more in the 'make can take a flying leap' club at this point ;-)
ingy svn up; make purge; perl Makefile.PL; make test
I like make more than MakeMaker 01:47
rooneg yeah, well, that's not saying much...
ingy make is actually very nice for what it does 01:48
lumi ?? Regression? 01:50
Darren_Duncan I've got a Perl 6 language question that I haven't seen the answer to in the Synopsis ... 02:17
If a subroutine variable is required (not optional), does Perl prevent passing of undefined values to it, or does it only require that any kind of value or variable is passed to it? 02:19
Khisanth I certainly hope you are ... 02:23
unless undef stops being a valid value
Darren_Duncan s/variable/argument/ - what I meant to say 02:24
so it is valid for a required argument to be given the undefined value? 02:25
buu Darren_Duncan: At a guess: yes. 02:26
I think it just checks for *numbers* of arguments. 02:27
Darren_Duncan right then, I'll continue to test for undefined input values as I did in Perl 5
thanks
buu You could probably get away with something like
sub foo( $x is not undef ); or something
Darren_Duncan if that works, great
buu Not sure really. 02:28
Darren_Duncan of course, I still have to test array elements if they all must be defined too
I'll do the explicit tests for now
Steve_p Hide? 02:49
crap, hide?
joejiang_ hello fayland 02:53
we have a channel at #perlchina at this server 02:54
welcome to join us
autrijus greetings. 04:19
is the 6.4 bug still there?
mm shapr is away
seano 6.4 bug?
autrijus conflicting export of getEnvironment on GHC 6.4 on POSIX
seano import foo hiding (getEnvironment) 04:20
autrijus sure.
attempting a fix now
seano man, handling of my is incredibly wonky now...
autrijus it is! 04:21
I really need to sit down and get eval-during-parse implemented
shapr: darcs and svn.perl.org are both synced per minute
seano My guess is that it really should be a macro that can take either a '$var' or a '($list_of_vars,...)', and produce a lexicalization of those names, plus a list for future assignment.
autrijus shapr: I attempted a fix; see if you can get it running 04:22
seano Also LHS ternary broken :(.
autrijus yup, I think so too. can you write a AST for my()?
(or did P6C.pm already do that?) 04:23
seano That was a real eye-opener into how weird Perl5 is for me. i.e. '($x ? $y : @z) = @w' correctly propagates context.
autrijus: I'll look at it, but it will take me a bit of time to understand how that code works, i.e. constructing a list-lvalue. 04:24
autrijus: and no, didn't do that. There was a hack to get both "my ($x)" and "my $x" to DTRT with respect to context, but nothing systematic.
autrijus purrs. 04:25
bbiab
seano What I really want to see is "my" completely separated from assignment, so you can then have some sort of kick-ass destructuring assignment like my ($x, ($y, $z, @ws)) = [1,[2..5]]
autrijus: why are "constructs" separated from "declarations", rather than all gathered as "statements"? 04:31
autrijus I thought the magic that does binding is not my but :=-^H 04:34
seano: I don't know, I merely followed pmichaud and luqui's naming convention
it's just names tho. 04:35
not really reflected on the AST level
the AST level is just Syn and App.
mugwump 71 nicks!
autrijus where Syn means special forms.
hi mugwump!
when are you arriving?
mugwump oh... I've got it right here
autrijus thomasyeh++ # pugs works on AIX 04:36
mugwump hmm, 21:50 on 23 Mar
autrijus good. now I can really deploy it to $real_work.
seano autrijus: re decl/construct -- fair enough. re bind/assign -- it seems like they should always be the same in terms of how they tear apart the RHS.
autrijus that is something I don't know about.
I know := does the unpacking the same way as sub params
I have no idea what = does.
is it specced somewhere? 04:37
seano on AIX? Poor pugs...
autrijus mugwump: cool. that's taipei time?
seano no -- I meant
no -- I meant "should" in the "ought" sense. It would just be painful for them to act differently.
autrijus so it's just like := but destructive?
mugwump yes, taipei time. I'll drop an e-mail too 04:38
seano Just like := but non-aliasing, i.e. updates to the lhs vars don't update parts of the rhs construct. But I could be completely on crack.
autrijus my $a = (1,2,3);
seano note it's been >~ 1yr since I've read through the apocageses.
autrijus my $a := (1,2,3);
and it's synopses now :)
seano that one actually bites me, because I expect "@x=1..3;$x=@x;$x" to return 3. Man, I feel obsolete. 04:39
Khisanth 3? but $x is a reference ... 04:40
seano yeah, but not in P5!
Khisanth notes this is extremely confusing :)
seano is too confused to notice....
Khisanth hmm ghc actually requires perl5
seems quite "circular" 04:41
autrijus yup. so it's bootstrapping in a roundabout way.
seano haha. I get a kick out of how the functional-lang folks who tend to dump on Perl still use to glue GHC together
Don't get me wrong, Haskell's a treat, but the whole "we're so pure, you can't insert a print statement" thing can get tiresome (and yes, I know the GHC extensions get around this). 04:42
Khisanth by Perl can be functional too
oops s/by/but/
seano and sysfunctional!
s/sys/dys/
tmoertel` thinks a lot of fp folks *like* Perl
seano has yet to see them admit it. 04:43
scw odd... eval 'my %z = 0, 10000; %z{0}++' makes pugs stop. 04:44
seano autrijus: re decl/construct, I guess my point is that they should all be "macro ...", with the macro controlling how the rest is parsed. Kinda like Lisp.
Khisanth still has to get a copy of HOP
scw instead of return something like undef.
seano scw: hashes can't be named 'z' 04:45
Khisanth I hope you are kidding
seano should have added a smiley or something, I suppose...
Khisanth seano: with all the changes to perl6 that actually sounded believable!
scw eval 'my %hash = 0, 10000; %hash{0}++'
autrijus seano: I completely agree. 04:46
I also think the macro bit is not specced at all.
scw leads the same result..
autrijus if you can help getting that done, great; I think I'll punt till 6.28318
Darren_Duncan I just got back ... now reviewing last 36 minutes of log ... 04:47
jesse_ autrijus: what are you deploying pugs for in production?
autrijus jesse_: as a scripting language for OpenAFP.hs and PDF::FromHTML
jesse_: basically a reasonably complex DB2->AFP->PDF distribution engine that is powered by several REST web services.
jesse_ Wow. fearsome 04:48
autrijus and it's all on AIX ;)
seano autrijus: help get the macro part done? hahaha! That's some major hair. I guess I'd start by getting '$param is Expr' and 'is parsed ...' to work, then replace my, for, etc. by P6 subs.
04:48 jesse_ is now known as obra
obra yeah, I'm truly afraid 04:48
autrijus seano: that sounds good, and much more practical :)
but it's not as fearsome as what Dan did with parrot 04:49
obra hah
autrijus and its register-spilling monstrosity
obra Pugs NCI
seano The next two weeks will probably be a bit crazy for me with grading etc. for finals, but hopefully I'll have quality P6 time during spring break.
autrijus woot woot.
seano++ # in advance
Darren_Duncan FYI, I'm mostly finished writing my Perl 6 LKT now; still have a few things to figure out ...
seano autrijus: no fair playing the expectations game ;).
autrijus Darren_Duncan: wonderful
Darren_Duncan for example, with multi-dimensional structures, when an assignment copies just a reference, or the whole array 04:50
seano well..., I'm more or less comfortable with/resigned to the register part, but objects, mmd, etc. are just scary.
autrijus seano: let me up the ante and send you an invitation mail to become a committer
seano autrijus: thanks, but I'd like to at least be comfortable enough with the code to not auto-kick my own ass before doing that. 04:51
autrijus Darren_Duncan: I think the idea is "you can't easily tell the difference."
Darren_Duncan sure but ...
it would be nice to know when, if I modified a value passed in as an argument, whether the caller's context changes
I know that the base arguments are copied
seano One of the pains of 'is parsed' is that it requires fixing of the internal P6 grammar production names, and hence the structure of the grammar, since user programs will embed them. 04:52
autrijus *nod*. paste some example code?
seano I'm not sure I have a good enough idea what Larry's idea of the language is to do such a thing.
autrijus or nopaste, really; or post on p6c :)
seano brb
Darren_Duncan I'm looking for relevant portions in the Synopsis, currently I'm reviewing "Data Types" in #2 to get a different question answered re sigils 04:53
autrijus greetings flw. 04:55
Darren_Duncan FYI, if you want to see the current state of the module, look under darrenduncan.net/p6c/KeyedText.pm
Ignore the SYNOPSIS, which isn't translated yet
autrijus cool! 04:56
"under the same terms
as Perl 6 itself."
except there is no such a term :D
(j/k)
Darren_Duncan for contrast, add a '5' before the .pm to see the Perl 5 version
autrijus ok 04:57
mugwump Do laywers care about chronological consistency?
autrijus probably not.
hmm
$message.msg_key~': '~
buu Release under perl5 terms =/
Darren_Duncan FYI, I started to consider using kwid, but then forgot where the detailed docs were
buu Darren_Duncan: Flee from the horror that is kwid!
Darren_Duncan sure
autrijus "{ $message.msg_key }: { join(',', ...) }"
Darren_Duncan I haven't worked out the replacement for join and map etc yet, if any 04:58
autrijus it's just join and map.
Darren_Duncan same old then
here's one ...
what's the proper way to know if a module or class has been loaded already? 04:59
I'll explain why I want to know
its re translate_message()
autrijus Darren_Duncan: it's doc/perlkwid.kwid at this moment, btw
well, defined() maybe? not sure 05:00
obra watches #perl not get it
autrijus ask on p6l? :)
Darren_Duncan does p6l have a chat room, or do I use the mailing list?
FYI, I plan to translate as much as I can without asking too many questions, then commit fixes later when I get them answered 05:01
autrijus p6l has a chatroom, and that is here.
but the mailing list is canonical :)
Darren_Duncan sure enough
at least here, answers can be fast
autrijus yup. basically we formulate questions in an appropriate form 05:02
Darren_Duncan just answer the following if you know them without having to look up ...
1. what's the p6 version of UNIVERSAL::isa ?
autrijus and made sure it's not one of the FAQs
then throw them to p6l for larry & co :)
Darren_Duncan ... actually, nevermind
autrijus it is .isa
$foo.isa
is always okay
unlike perl5.
Darren_Duncan my explicit parameter list requires the argument is of said class, no need to test 05:03
here's a different one I asked before you signed on ...
if a sub argument is mandatory, then does that prevent undefined values, or just that the caller gives any kind of argument? 05:04
In Perl 5, an undefined arg is the same as no arg
currently I assume that undefined values are valid and test for defined()
autrijus undef() ?
sure, undef is a proper arg. 05:05
Darren_Duncan ok
FYI, the only reason a few of my code lines are commented out is because the new // operator causes messed up syntax coloring on my text editor; they are actually valid code lines 05:07
I'll uncomment before commit
autrijus your text editor being vim?
Darren_Duncan nothing so geeky - its BBEdit
currently set to display Perl colors
Interestingly enough, it colors the word 'method' blue; I didn't realize that word was valid in Perl 5 05:08
seano Darren_Duncan: dude, Emacs runs on OS X. Why wait?
buu Because emacs sux!
Darren_Duncan the coloring issue was a minor thing, and I *like* BBEdit
seano contemplates war...
Darren_Duncan BBEdit is one of the best text editors there is, except for not being fully programmable, which I don't care about
very powerful multi-file search and replace features for one thing 05:09
seano Darren_Duncan: Having imbibed emacs first, I always found it was almost configurable enough for me to make it into Emacs ;).
autrijus contemplates eclipse :)
Darren_Duncan I'm sure the colorer will be updated for Perl 6 soon enough
seano notes that he will have plenty of time to do so as it starts up.
Darren_Duncan The editor is also fully Unicode aware 05:10
autrijus well it starts up in 2 seconds here... same as emacs :)
Darren_Duncan I'm saving all my files in UTF-8 format, in fact
autrijus Darren_Duncan++ # The Way To Go
seano Darren_Duncan: re multi-file -- true. I wish M-x occur worked across files...
autrijus stresses again that Pugs will not grok latin1 << >> or Y.
Darren_Duncan not just multi-file, but multi-file with regexes
seano Darren_Duncan: people who use more than 127 characters piss me off...
buu autrijus: Why not?
autrijus buu: because it's ambiguous. 05:11
utf8 << may either be a latin1 << prefixed with \xC2
Darren_Duncan Perl 6 is supposed to be natively Unicode anyway
autrijus or parsed as utf8.
buu oh
Darren_Duncan but meanwhile I "use utf8" in my current Perl 5 stuff
higher characters only show up in my example French characters that have accents, though 05:12
some parts of CPAN display them properly, others don't
but even when not, only those few accented characters are messed up, since utf-8 is a superset of ascii
seano autrijus: I'd give it a try, but I deleted it to save disk space. The poor kids in the class I TA this quarter are all being forced to use Eclipse, so I've had to start it up a couple of times. 05:13
Darren_Duncan I'm not bothering with any Unicode encodings except utf8
seano In any case, I remember >> 2 sec. Maybe my laptop drive's slow.
autrijus: re macros & grammars, S02: "Intra-line comments will not be supported in standard Perl (but it would be trivial to declare them as a macro)." makes me think something like this needs to be shoved in at a very low level. 05:14
Darren_Duncan here's a Q: ... 05:16
My message.as_string() was a guess re sorting 05:17
autrijus I'll bbiab. 05:18
Darren_Duncan Eg, if I feed %hash.pairs to map {} but want to sort it first, will a plain 'sort' implicitly sort on the keys by string value, or do I have to say that explicitly as I did?
... using sort { $a.key cmp $b.key }
arcady I think you have to sort them explicitly 05:19
seano Darren_Duncan: I guess the question is what 'infix:cmp(Pair, Pair)' does
Darren_Duncan well, that seems safer for now
seano, yes
seano So brutally define it to your will :).
Darren_Duncan sort of
assuming that $a cmp $b is what sort does implicitly 05:20
all the time
seano There was a thread awhile back on P6L on which comparison -- <=>, cmp, or something else -- is used by default between objects. I don't think there was a resolution then, but maybe since... 05:21
PerlJam I think you mean $^a and $^b there as $a and $b don't magically appear
Darren_Duncan er, yes, if that's the new names of those variables
arcady it's the new implicit variable syntax thingy 05:22
Darren_Duncan I changed my code to use that
I just noticed the .as() method in Perl data types; I could use that instead of manual joining ... 05:29
the .perl() method also looks cool ... it could possibly replace the likes of Data::Dumper 05:31
PerlJam not possibly, definitely.
well, probably with a little help to traverse the data structure if .perl() doesn't do that by default. 05:33
Darren_Duncan .perl() at least handles arrays and hashes of one dimension 05:34
PerlJam I like as :lang<perl> though as at least other languages can add their own as-handlers 05:39
arcady that doesn't seem too useful though 05:41
Darren_Duncan Oh, oh, S02 answered my question ...
"Sorting a list of pairs should sort on their keys by default."
so that takes care of that
PerlJam arcady: why not?
arcady because how does any other language deal with things like $foo vs. @foo and so on? 05:42
unless I completely misunderstand the issue
PerlJam arcady: other languages are free to deal with those as best they can :) 05:43
arcady well, you can just write a module to dump to whatever format 05:44
and in 99% of cases, you'd be using as :lang<perl> anyway
PerlJam I suppose the exact form the feature takes doesn't matter too terribly much. 05:47
arcady feel free to write the module, if you want it so badly : ) 05:48
theorbtwo Mornin, all. 05:49
PerlJam $some_big_data_structure but= python_representation_role; say $some_big_data_structure.python;
arcady I wonder how you'd serialize functions then 05:51
Darren_Duncan Mornin, one 05:52
theorbtwo src/Main.hs:159: Variable not in scope: `config_file_sep' ?
PerlJam arcady: parrot byte code. 05:55
arcady: then you use your handy-dandy parrot-to-<insert language> translator to get the appropriate language representation.
ingy ola 05:57
hola even
theorbtwo: make purge
silly rabbit
autrijus: I learned a bit about monads today 05:58
PerlJam ingy: enough to build a perl6 rules parser? ;) 05:59
Darren_Duncan I could use a bit of that 06:00
though only very basic regexps in my current mod
sorry, "rules"
PerlJam Darren_Duncan: for those modules you mentioned on p6c? (or was it p6l?)
Darren_Duncan I haven't written anything to p6l ... 06:01
All the modules I mentioned on p6c have fairly basic regexps, if any ...
in SQL::Routine, I think I only use them to test if input looks like a number 06:02
more specifically, test if its anything except an integer
a positive integer
theorbtwo That fixed it, ingy, sorry. So what dependency are we missing? Makefile.PL wasn't modified, which is the usual culprit...
Darren_Duncan In Locale::KeyedText, 2 uses; 06:03
tmoertel` is the darcs repo up to date?
Darren_Duncan 1. see if it has any non-alphanumerics
2. replace one literal with another
the latter each being defined in scalar vars 06:04
no fancy lookahead or lookback or whatever stuff
theorbtwo Oh, presumably util/PugsConfig.pm.
Darren_Duncan Here's a Q that's strictly about optimization ... 06:07
ingy theorbtwo: Makefile.PL and util/PugsConfig.pm generate src/Config.hs
Darren_Duncan I put most of my code at the top of my module, and most pod at the bottom, with a __END__ in between, so the parser stops there and doesn't waste time below, and I have lots of POD ... is this sort of thing useful with perl 6?
PerlJam Darren_Duncan: perl6 doesn't grok __END__. That's changed to =begin DATA 06:08
Darren_Duncan If I stick a "=begin END" (the perl 6 equiv) above all my pod, will that make Pugs compile it faster?
ingy Darren_Duncan: I think that's an unnecessary optimization
Darren_Duncan okay, out it goes 06:09
ingy not worth talking about
Darren_Duncan I won't bring it up again
PerlJam sleep & 06:12
theorbtwo Hmmpf, what's the URL for this pastebot? 06:18
ingy perlbot: nopaste? 06:20
perlbot Paste your code here and #perl will be able to view it: sial.org/pbot/perl 06:23
ingy theorbtwo: ^^
Aankhen`` considers whining about how he wants #perl6 to view things.
Khisanth that should probably be sial.org/pbot/perl6
or the output will end up in the wrong chan :)
pasteling "theorbtwo" at 84.245.186.254 pasted "Makefile.PL patch for dependency on util/PugsConfig.pm, and -O3 on build while I'm at it." (18 lines, 729B) at sial.org/pbot/7966
theorbtwo Thanks.
I did ask pastebot for help, in my defense, but it was not very forthcoming about it.
tmoertel` is Main.hs being hacked on right now? 06:26
ingy tmoertel`: I am hacking on it heavily
why?
tmoertel` I was going to take a cleanup pass at it 06:27
ingy pleas wait
obra hey ingy
tmoertel` no problem
ingy hi obra
once I get the test harness working again, it's all yours
theorbtwo Anybody have time to review that patch? It's pretty trivial. 06:29
tmoertel` didn't know GHC had an -O3 optimization level 06:31
theorbtwo Aye. I haven't benchmarked, but it certianly feels like it does a 'make test' run much faster.
tmoertel` theorbtwo: faster than -O2? -O2 rarely is faster than plain -O in my experience, but -O is much faster than w/o any opts 06:32
it looks like GHC takes any optimization level (-O9999 works) 06:33
theorbtwo I haven't tried comparing the various -O levels; will do that now.
tmoertel` I don't think anything beyond -O2 is implemented in the compiler 06:34
scw wow, the compilation is really slow with -O3 06:35
but takes only 1 second to generate each generation in life.p6, which took 6 seconds before. 06:37
hoowa hi autrijus 06:41
theorbtwo Hm, surprising -- my measurements say that no -O option is actually faster then -O1 or -O2 for a complete make all test cycle. 06:51
Darren_Duncan question: if I wanted to create an alias for a package name, do I use "let" or "::=" or do something else; eg "let Message be Locale::KeyedText::Message"? 06:55
reverse last 2 operands in example 06:56
I don't want to create a new class
I'll try the "::=" that aliases variables and see how that goes 06:58
or skip it
ingy tmoertel`: I am close to checking in Main.hs 07:12
tmoertel`: are you a lambdafolk?
KwikPoll: Who here uses OS X? 07:13
theorbtwo 0;
ingy theorbtwo: no mac? 07:15
theorbtwo No mac.
ingy it would be cool to work with someone like tmoertel` using SubEthaEdit
theorbtwo SubEthaEdit sounds interesting. I, of course, am not somebody like tmoerel`. 07:16
ingy perl people pairing up with haskell people
theorbtwo Hell, I'm stuck on exercise 3.10 of YAHT at the moment.
Darren_Duncan okay, I'm going to commit the first Locale::KeyedText for perl 6 now. doubtless it has bugs, but it should be at least 90% correct 07:17
ingy Darren_Duncan: as long as it's in modules
Darren_Duncan It is in modules; I followed the example of Sample-Module
ingy coolness
Darren_Duncan quick question; I forgot to use the command-line option to spec a commit message, so was dumped into Vim; I wrote a message there, but what is the keystroke to save and quit Vim? 07:24
scw :wq
Darren_Duncan clarify that; what does ":" mean? 07:25
ctrl-:?
scw Err, after pressed <esc>
Darren_Duncan that seems to have done it, thanks 07:26
nothingmuch morning
Darren_Duncan morning nothing
I just committed the first LKT a few seconds ago; update to see it 07:27
I haven't tried compiling it of course, but it should be 90% correct
its in /modules
scw nothingmuch: Hello, several tests in not.t, auto.t and assign.t failed, how about using todo_* ? 07:28
nothingmuch i didn't change not.t
and luqui untodoed some tests
he said since they were bugs, and not unimplemented features, they should not be todo
auto.t is not mine either 07:29
sorry, back 07:31
rt.openfoundry.org/Foundry/Project/...n/?rev=512
scw oh, I didn't see the commit log :) 07:38
nothingmuch i agree with him in theory 07:40
but i think it confuses users
i.e. "what's wrong? i get pugs 6.0.11 and tests failed!"
but then again, bugs *are* bugs
autrijus fwiw, I think luqui is right. 07:42
tests should never fail to compile, though.
and before a release I'll TODO all the failing tests
nothingmuch on that i agree
autrijus but inbetween releases, it's okay to have tests failing on trunk
ingy autrijus: want to see pair with me?
autrijus not yet, at $client's placed 07:43
and using their computer
autrijus is trying to get gpg running on AIX.
ingy ok
nothingmuch wouldn't mind joining a see session
autrijus you can pair with nothingmuch :)
nothingmuch i'd like to see what you guys do
ingy autrijus: yuck
nothingmuch not that i'd know how to do anything ;-)
aix... oh my
which reminds me
ingy big patch to Main.hs incoming... 07:44
autrijus woot.
nothingmuch i can try and compile pugs on aix, hpux, solaris, and osf when that box comes back to life
autrijus nice. aix is known to work.
nothingmuch given GHC on those platforms
autrijus but GHC is notoriously portable
so we don't have a pressing need
nothingmuch so we actually don't worry about it
ok
All the same for me =)
ingy: mention my name in chat or something if you are starting see and don't mind me joining. that way i get a colloquy bubble. i'm going to do some real work now that i've eaten and stuff 07:46
ingy nothingmuch: sure
nothingmuch: so you have OS X?
nothingmuch uhuh
so does lumi
ingy that's 4 so far 07:47
nothingmuch he just started and might enjoy see as well
ingy we should keep a list
nothingmuch but he starts work at around 11:00, so he's not up yet
i think there's more
ingy nothingmuch: what tz are you?
nothingmuch Israel
ingy cool
I have a friend there 07:48
the inventor of yaml
so i am used to that particular tz
nothingmuch but normally i'm in sync with europe central as far as getting up, and greenland as far as going to bed
ingy for pairing
nothingmuch i saw on the website
didn't know he was from israel though
many israeli's or people with jewish names live in NA 07:49
ingy Oren Ben-Kiki
I almost wrote Oren Ben-Kwiki
nothingmuch hah
ingy Oren and i went skiing in Canada this winter
nothingmuch skiing... mmmm 07:50
ingy his son and ex-wife live near me
nothingmuch been too long
a bit more than a year since I last went
ingy he is a genius imo
nothingmuch where do you live?
ingy Seattle
nothingmuch is he a perl guy?
ingy not really
nothingmuch i don't recall his named being mentioned otherwise than in YAML context
ah, that explains it =)
autrijus I see the name "Oren" and I keep having flashbacks from the Kill Bill movie... 07:51
nothingmuch o-ren
sort of
ingy can't wait to write a perl6 yaml parser in haskell
nothingmuch the pronounciation of Oren is much less detaché
ingy tmoertel`: Main.hs checked in 07:52
hmmm. I bet I broke Pugs::MakeMaker
I need tests :P
nothingmuch i'm creating OSXPeople on pugs.kwiki.org
clkao ingy: svk didn't treat you well?
ingy clkao: I love you 07:53
clkao: we can hack the hateful svk in Taiwan
clkao right
ingy and maybe the more hateful YAML.pm
Darren_Duncan since this was going around 10 mins ago -- I use Mac OS X, I live in Canada, and tz -800 08:02
its midnight here now 08:05
oh, the Perl 6 summaries just came out
thesaboteur heh im unfortunately on mac os x as well
I also unfortunately am south of the border
Darren_Duncan what do you mean "unfortunately"; what would you prefer
for an os
thesaboteur I would prefer a debian or Fedora system
Darren_Duncan you can dual-boot Macs with a half dozen plus Linux distros
nothingmuch especially debian or fedora
though on macs i would prefer gentoo
thesaboteur Im a newbie with perl (Just learning) and working on getting my perl fixed so I can use modules from cpan.
nothingmuch since it's a bit less trouble
thesaboteur: the #1 place to do that is www.perlmonks.org/
thesaboteur eh dual booting is to much work.
nothingmuch well, suite yourself... If you don't want linux enough to work for it ;-)
Darren_Duncan in my understanding, the most popular Mac distro is Yellow Dog Linux (YDL), which I think is a version of one of your prefs ... have to check
thesaboteur nothingmuch: ah thankye
nothingmuch YDL == redhat, sort of 08:06
thesaboteur er
Darren_Duncan well, you can sole-boot a Mac under Linux if you want
thesaboteur it's not really "my mac" hence the not trying to transform it into linux to much.
im at work
Darren_Duncan from what I've heard, its a lot easier to dual-boot with the Mac os than it is with Windows; I've heard the latter tends to screw with things
nothingmuch pugs.kwiki.org/?OSXPeople
thesaboteur I really hate macs but if I can install enough linux stuff onto it then it makes it bearable :) 08:07
autrijus try Ubuntu on PowerPC :) 08:08
theorbtwo Debian should run just fine on PPC, dual-booting.
thesaboteur yeh but I can't change the os
nothingmuch thesaboteur: if you have another mac, you can boot your mac in fw disk mode
theorbtwo wonders if it's worth the effort to set up an emulated OSX in order to pair with somebody.
nothingmuch and use ipartition to resize the partition of your root 08:09
or just copy everything over
repartition
and copy everything back
Darren_Duncan should we sign up to Perl6People too?
nothingmuch theorbtwo: it'll probably be very very slow
Darren_Duncan re the qwiki?
autrijus thesaboteur: MacOnLinux?
nothingmuch but pearpc can do it
autrijus it seemed to work fine the last time I tried
thesaboteur I dunno I got confused a few lines ago
autrijus pearpc = ultra slow.
I'd suggest moonedit ;) 08:10
Darren_Duncan I've heard that the Mac hardware is a preference among Linux people
hoowa haha tang
autrijus greetings hoowa!
Darren_Duncan any main complaints are with the OS
theorbtwo It was almost usable on my 700MHz P3; this is a 3GHz P4/HT, so should be barely usable.
thesaboteur Darren_Duncan: yeh don't get me wrong the processors are bad ass but I have a very deep hatred of macs that has been with me for years.
autrijus 3ghc sounds indeed borderline okay.
3ghz, even. bad autocompletion 08:11
thesaboteur Im attempting to install xtools now, hopefully it works and makes my perl stuff work
Darren_Duncan thesaboteur: that's unfortunate
I don't use xtools, even though I hear they're great
just the BBEdit text editor and the standard Unix perl 08:12
nothingmuch uses carbon vim
thesaboteur Darren_Duncan: I think it stems from when I was a little kid, using a school mac, I think the mac tried to molest me and I have repressed the memory.
Khisanth ...
Darren_Duncan that could be it ... surely there are no real reasons to hate macs
nothingmuch isn't xtools some kind of configuration management thingy for OSX server?
thesaboteur er I thought xtools was what gave me gcc and stuff like that? 08:13
nothingmuch ah
xcode
thesaboteur er yea thats it
nothingmuch if your mac is new, it's in /Applications/Installers
or something like that
Darren_Duncan XCode is it
nothingmuch perl works without it, btw
thesaboteur oh I downloaded it from the mac site
Darren_Duncan It comes on a separate CD with the Mac OS X Cds, or you can download it
for free
nothingmuch and XCode the IDE is not very perl friendly without camelbones
thesaboteur yeah my perl works but I couldn't install/configure a module I got from cpan
nothingmuch ah, that makes sense 08:14
thesaboteur yeah so im working on trying to figure out what I have to do, so I can compile modules from cpan without doing a bunch of magic tricks.
Darren_Duncan Apple's bundle of developer tools gives you gcc, but you can install that by itself and without XCode
XCode does require gcc though, as it uses gcc behind the scenes 08:15
thesaboteur Darren_Duncan: yea but then I wouldn't have all the library's I need though right?
Darren_Duncan libraries for what?
nothingmuch i'd just install it all
thesaboteur I dunno really
Darren_Duncan The standard bsd libraries are separate installs from carbon/cocoa etc
nothingmuch and from the web is prolly a good choice, since it's more recent
Darren_Duncan still, I just install everything, its about 500M
thesaboteur dunno but xcode install is almost done now anyways
Darren_Duncan the stuff I don't use is there to play with later when I want to try that cool Cocoa stuff 08:16
nothingmuch Cocoa is mui fun
thesaboteur as long as I can compile stuff i'll be happy
nothingmuch when i have little bits of time i try to hack in it
thesaboteur: if you want linux like goodness, fink.sf.net 08:17
Darren_Duncan once you've done the general dev tools install, you can run gcc on the command line as usual
that's what I used to compile my Perl
thesaboteur nothingmuch: yeh I have that :)
nothingmuch doesn't it want the dev tools?
Darren_Duncan I don't use Fink -- its a great tool, but I didn't need it
nothingmuch doesn't like fink because if something breaks, it all breaks 08:18
thesaboteur nothingmuch: to do anything usefull it does yea :)
nothingmuch i use it very very very sparingly
Darren_Duncan I compile perl on the command line, basic "make Configure" etc
thesaboteur fink works good for small packages that im to lazy to compile myself, like ping and whois. 08:19
Darren_Duncan Fink was more useful in the older days, by providing access to a lot of unix software that otherwise know about Mac OS X to install the normal way
nothingmuch you have ping and whois on OSX
thesaboteur yea
Darren_Duncan these days, everything I need knows about OS X, so fink isn't needed
ingy thesaboteur: do you know about Kwiki?
nothingmuch did you install the BSD subsystem?
thesaboteur ingy: never heard of it
Darren_Duncan oh yes, the BSD subsystem is required for the dev tools
thesaboteur nothingmuch: I dunno I just kept clicking agree/ok
nothingmuch huh 08:20
bummer
Darren_Duncan that is part of the standard Mac OS X install, and is installed by default with Mac OS X, unless you uncheck the option
nothingmuch lots of std unix tools are installed on OSX
but i don't remember if by default or not
ingy thesaboteur: it is a wiki framework with plugins... about 150 modules from various authors
Darren_Duncan the BSD tools includes the Perl bundled with OS X
5.8.1 was bundled with Panther 10.3
ingy thesaboteur: I have a method to install and maintain all of those without going thru cpan
thesaboteur Darren_Duncan: im not doing an os install, perl was/is installed. Im simply trying to get gcc and a perl install that will work with cpan modules. 08:21
Darren_Duncan sure, and it was simple for me
thesaboteur ingy: hmm sounds kool.
ingy Darren_Duncan: I just checked in a new Pugs::MakeMaker 08:22
could you test it out
autrijus fayland++ # volunteering to take over zh-cn maintainership
thesaboteur ingy: Im trying to get a fairly uncommon module to work though I suspect
autrijus fayland: you are a pugs committer now. welcome aboard!
ingy thesaboteur: ok
Darren_Duncan simply installing the whole dev tools set on top of a machine with the whole OS X
includes all the tools you need to do perl stuff, or compile with gcc
ingy, I hear you, will check it out
fayland autrijus ,thanks.i'll do my best
nothingmuch fades out 08:23
ciao
thesaboteur sweet the new xcode tools thingy fixed my perl :)
Darren_Duncan great to hear , sab 08:24
thesaboteur it installed to, yay.
Darren_Duncan gcc 3.4 is what currently comes with the os x dev tools 08:25
Mac OS X Tiger 10.4 is now being built with gcc 4, so its tools should have that too 08:26
okay, I see problems with the latest Pugs 08:27
earlier today, it compiled and tested without issues
ingy Darren_Duncan: your module? 08:28
theorbtwo 33/1281 subtests failed, 97.42% okay? That's expected.
Darren_Duncan now 'make' is fine but 'make test' is borked
ingy Darren_Duncan: did your module pass its test?
because it fails here 08:29
cls_bsd flush Failed Test Stat Wstat Total Fail Failed List of Failed
Darren_Duncan it says Failed 77 test scripts, no output seen ... before that it says 'unexpected "$"' and "expecting ";" or end of input
cls_bsd t/op/assign.t 2 2 100.00% 1-2
t/op/auto.t 1 256 37 44 118.92% 11-37
t/op/not.t 14 4 28.57% 1-2 4-5
ingy but I don't think it is my fault
Darren_Duncan ingy, I expect that my module will not pass its test, since it uses features Pugs doesn't support; it shouldn't even compile ...
ingy oh 08:30
Darren_Duncan but the test suite doesn't run /modules, does it?
ingy ok, then I won't attempt to fix it :P
Darren_Duncan those are "extra"
ingy no
yes they are extra
Darren_Duncan I expect to try running what I committed when Pugs has its features 08:31
meanwhile, its presence can give developers an idea what features to put in possibly earlier
I only use the simpler parts of the missing features
no overloading, no multimethods, etc
ingy ok 08:32
autrijus the failed tests is not Officially Okay
but is Acceptable during releases
it is my plan to make them TODO before each release
for that release tarball alone
ingy aye
Darren_Duncan meanwhile, people can just critique the code in my mod if they see what doesn't look like proper perl 6 code
autrijus but let them fail (as long as it's not fail-to-compile) during releases
Darren_Duncan: wonderful work!
Darren_Duncan++
autrijus is still stuck at $client_site waiting for a particularly slow GCC install. 08:33
AIX--
Darren_Duncan note that I ran 'make clean' prior to these last tests; might that have borked it?
ingy AIX-- indeed
theorbtwo You can't ssh home?
ingy hasn't worked on AIX in 10 years though
decay_ is sitting at a aix workstation right now ;) 08:34
Darren_Duncan This is an example of what was borked with the tests;
t/01basic........................
unexpected "$"
expecting ";" or end of input
NonTerm "/System/Library/Perl/5.8.1/Test.pm" (line 31, column 1)
ingy Darren_Duncan: use make purge instead
and svn up 08:35
Darren_Duncan will try that
come to think of it, my "perl Makefile.PL" mentioned something missing, but I thought it was unimportant
okay, I did 'make purge' and 'svn update' then what followed started out the same 08:36
Warning: the following files are missing in your kit:
modules/Sample-Module/blib6/arch/auto/Sample/Module/.exists
modules/Sample-Module/blib6/lib/auto/Sample/Module/.exists
modules/Sample-Module/blib6/lib/Sample/.exists
modules/Sample-Module/blib6/lib/Sample/Module.pm
Please inform the author.
ingy hrm 08:37
one sec
svn up 08:38
that was likely my fault :(
Darren_Duncan continuing, 'make' gives no errors, 'make test' same as before ... 08:39
ingy, I'll wait for you to do something, since you seem to recognize it
ingy that's inconceivable 08:40
it works fine over here
Darren_Duncan if it makes any difference, I could throw out my whole working dir and do a fresh checkout?
just in case
ingy you don't have to throw the old one out 08:41
are you using svk or svn?
Darren_Duncan I didn't mean that literally
ingy well yeah, try it
Darren_Duncan I'm using svn, same as the host server seems to be 08:42
ingy svk talks to svn servers.. but anyways...
please try a fresh checkout
Darren_Duncan I noticed something curious ... apparently there are a couple blib folders in the folder that is the parent of my pugs root dir; one says 'blib' and the other 'blib6', the latter just updated. 08:43
is it normal to have blib outside the current working directory?
ingy hmmm 08:45
Darren_Duncan svn checkout svn.openfoundry.org/pugs/ in progress
done ... revision 524
ingy It orks for me on two different machines
works
Darren_Duncan This time, "perl Makefile.PL" gives no errors 08:46
ingy osx and linux
Darren_Duncan however, while make was fine, 'make test' did what it did before 08:47
ingy delete the spurious blib dirs 08:48
Darren_Duncan eg ...
ingy what type of os are you on?
Darren_Duncan t/op/substr......................
unexpected "$"
expecting ";" or end of input
NonTerm "/System/Library/Perl/5.8.1/Test.pm" (line 31, column 1)
t/op/substr......................dubious
Test returned status 1 (wstat 256, 0x100)
I removed everything before checkout, including all the blib dirs
autrijus hrmph.
it's using the perl5's Test.pm.
not pugs'. 08:49
hence, the problem.
ingy yes
Darren_Duncan I'm on Mac OS X
ingy I noticed
Darren_Duncan and it did work fine earlier today
ingy trying to figure out why
Darren_Duncan: run `./pugs -V`
Darren_Duncan here ya go ... 08:50
Summary of pugs configuration:
archlib: /System/Library/Perl6/darwin-thread-multi-2level
privlib: /System/Library/Perl6
sitearch: /Library/Perl6/darwin-thread-multi-2level
sitelib: /Library/Perl6
@*INC:
.
cls_bsd '% pugs assign.t' and failed
ingy yow!
Darren_Duncan that wasn't all of it, my IRC client cut off
the rest ... 08:51
.
ingy Darren_Duncan: you need to use nopaste
Darren_Duncan what's nopaste?
ingy perlbot: nopaste?
perlbot Paste your code here and #perl will be able to view it: sial.org/pbot/perl
ingy make sure to select #perl6 08:52
pasteling "Darren_Duncan" at 24.109.93.98 pasted "the result of pugs -V" (14 lines, 435B) at sial.org/pbot/7969
Darren_Duncan how's that? 08:53
scw cls_bsd: yes, it fails and should fail.
Darren_Duncan I'll bookmark the 'bot for next time
ingy now paste all of the build output
autrijus I wonder if we should add a notice to Makefile.PL
that says "failed tests is okay inbetween releases"
or something like that
to reduce fear undertainty and doubt
ingy I don't really think they are so ok, but maybe I'm thinking wrong 08:54
pasteling "Darren_Duncan" at 24.109.93.98 pasted "the build information, I think" (49 lines, 2.7K) at sial.org/pbot/7970
autrijus okay, as in, we except them to happen
we will get around to it soon
et cetera.
ingy it certainly is nice to have failing tests shared amongst us
maybe we need unit tests vs regression tests 08:55
t/ and unit/
autrijus maybe. sounds bdufing though. 08:56
ingy true
cls_bsd hmm
ingy but then again I'm the bduf master :p
anyway I'm not sold on it
autrijus IGNY
ingy I think the message idea is fine 08:57
autrijus "I'm gonna need you", or something :)
autrijus thinks ingy's open source project manager points are well-stated, and that igny != bduf generally :)
ingy my motto was always YAAGNI 08:58
You Actually *Are* Gonna Need It!
Darren_Duncan: the failing test output man! 08:59
Darren_Duncan am I the first one to have this? 09:00
ingy Darren_Duncan: what directory do you run perl Makefile.PL from?
Darren_Duncan I run it from inside the root of the pugs directory, the directory that contains /lib and /t and etc 09:01
ingy cp kwid ../../../blib6/script/kwid
Darren_Duncan this is standard for CPAn distros, and the Perl distro itself, I think
ingy that is definitely messed up
Darren_Duncan at least that's how I always did it
ingy that's correct 09:02
Darren_Duncan after all, an unqualified file name, "Makefile.PL" you have to be in the same dir as the file
ingy where is your test output?
Darren_Duncan I'll paste the lot if you want, but its 77X of the same more or less
ingy well i think the problemo is the ../../../blib6 09:03
Darren_Duncan: just paste the top stuff
cp lib/Test.pm ../../../blib6/lib/Test.pm 09:04
Test.pm is not where pugs testing can find it
pasteling "Darren_Duncan" at 24.109.93.98 pasted "result of make test" (30 lines, 2.8K) at sial.org/pbot/7971
ingy thus the problem
Darren_Duncan I snipped the middle
ingy ok I got it 09:05
Darren_Duncan goodstuff 09:07
ingy can you paste ext/Test/Makefile
Darren_Duncan just a sec 09:08
ingy and also paste inc/Module/Install/Pugs.pm and ext/Test/Makefile.PL 09:09
pasteling "Darren_Duncan" at 24.109.93.98 pasted "/ext/Test/Makefile" (521 lines, 14.6K) at sial.org/pbot/7972
"Darren_Duncan" at 24.109.93.98 pasted "/ext/Test/Makefile.PL" (14 lines, 316B) at sial.org/pbot/7973 09:11
ingy autrijus: ping
autrijus: look at sial.org/pbot/7972
pasteling "Darren_Duncan" at 24.109.93.98 pasted "/inc/Module/Install/Pugs.pm" (74 lines, 2.2K) at sial.org/pbot/7974 09:12
ingy autrijus: especially INST_LIB
that is messed up!
Darren_Duncan as inane as it sounds, I'm going to try something I should't have to try ... back in a few minutes 09:13
ingy omg, is he rebooting?
Dear Darren, if rebooting doesn't work try touching your head and then your feet alternately for about 15 minutes... 09:16
Darren_Duncan: if you ever reboot to solve a perl problem again, I am never helping you
thesaboteur Then repeatedly chant "I am the walrus" while clicking your heels together.
Darren_Duncan I totally didn't expect that to work 09:17
but who knows about strange background issues
ingy obviously not you :P
MakeMaker is behaving strangely 09:18
what version is your perl?
Darren_Duncan I just figured, might as well try everything
the one I was using with Pugs was the bundled one, 5.8.1rc4
ingy 01:16 < ingy> Dear Darren, if rebooting doesn't work try touching your head and then your feet alternately for about 15 minutes... 09:19
Darren_Duncan I also have 5.8.6 installed in a separate dir; could run through that instead
ingy ok...
5.8.1 is a POS
but I'm not certain that is the problem
but i wouldn't be surprised 09:20
Darren_Duncan yet 5.8.1 worked earlier
ingy well this isn't earlier
Darren_Duncan no
ingy I get the same thing with my /usr/bin/perl 09:21
autrijus: ping... have you seen this? 09:22
Darren_Duncan currently I'm trying without a new re-checkout
make skipped everything, make test failed as before 09:23
NonTerm "/Volumes/Programming160/Perl/lib/perl5/5.8.6/Test.pm" (line 31, column 1)
So its using the newer version now
ingy sial.org/pbot/7972 ... INST_LIB is clearly set to ../../blib6/lib but ../../../blib6/lib gets used
Darren_Duncan: so is it working with the new perl? 09:24
Darren_Duncan I could try sitting in the parent of the 'pugs' directory and see what happens
no, see my above comment
ingy nononono
that will work as well as rebooting did 09:25
Darren_Duncan sitting in parent failed
ingy where is your perl-5.8.6 located?
Darren_Duncan see above comment 09:26
ingy what is the path to your perl-5.8.6?
Darren_Duncan test 09:27
du-oh, a leading "/" doesn't work here 09:28
how about: /Volumes/Programming160/Perl/bin/perl
ingy /Volumes/Programming160/Perl/bin/perl -v
/Volumes/Programming160/Perl/bin/perl -v
Darren_Duncan but I have a softlink to it at /Volumes/Programming160/perl58 , which is what I actually invoked 09:29
ingy do this...
checkout a new version
Darren_Duncan I'm on the -v
ingy cd to into it
/Volumes/Programming160/Perl/bin/perl Makefile.PL; make test 09:30
pasteling "Darren_Duncan" at 24.109.93.98 pasted "the -v output" (10 lines, 463B) at sial.org/pbot/7975
Darren_Duncan my 5.8.6 is just all the default settings, except for its location 09:31
so I don't clobber the Perl that came with the os
will do, just a sec
ingy also try running 09:32
Darren_Duncan checkout in progress
ingy /Volumes/Programming160/Perl/bin/perldoc -l ExtUtils::MakeMaker 09:33
what is the output for that?
Darren_Duncan: what is the output for /Volumes/Programming160/Perl/bin/perldoc -l ExtUtils::MakeMaker 09:34
Darren_Duncan didn't get to that yet, on 'make test' 09:35
separate window ...
the tests seem to be working now 09:36
dada hola
nothingmuch hola 09:37
ingy Darren_Duncan: ok.. you probably want /Volumes/Programming160/Perl/bin in the front of your path
Darren_Duncan your next question resulted in ...
... /Volumes/Programming160/Perl/lib/perl5/5.8.6/ExtUtils/MakeMaker.pm
ingy ok fine 09:38
the problem is solved 09:41
perl-5.8.1-RC1 sucks!!!
Darren_Duncan now it says failed 3/77, which sounds like what you expected
ingy yes
ingy never uses the system perl
nothingmuch neither
Darren_Duncan still, I'm going to test a bit more on my own to see why 581 worked before but not now, maybe
nothingmuch but then again, i'm an idiot in that respect
nothingmuch works in 5.9
Darren_Duncan to be honest, I prefer to use 5.8.6, and into there is where I install any mods ...
ingy ok I am leaving....
Darren_Duncan I just use the system perl as a shortcut when I think it won't make any difference
thanks for your time
nothingmuch ciao ingy!
Darren_Duncan see you ingy
ingy laters
Darren_Duncan and now it's 1:41am here
good night brave souls 09:48
theorbtwo G'midmorning, Darren. 09:49
Darren_Duncan ya, that too
very early in the morning
kolibrie well, I made it here! first irc in about ten years 10:40
pugs is very cool
I'm a little worried by the development pace... I can barely follow it 10:41
I have an idea for an addition to Test.pm
I wrote Test::Debugger in perl5 because I wanted some way to output the status for failed or TODO tests 10:42
I'm thinking it would be nice to have some kind of logging available in Perl6's Test.pm 10:44
If I write something (before someone else writes and commits it), how do I go about sending changes 10:45
I'm using the darcs repository 10:46
(hope I didn't miss a FAQ)
flw D:\MoChou\pugs>perl makefile.pl 10:54
*** Readline support disabled. If you want readline support,
please install Term::ReadLine::Gnu from CPAN, as well as
the GNU Readline headers and shared library.
Writing Makefile for Kwid
Writing Makefile for Pugs::MakeMaker
Writing Makefile for Test
Writing Makefile for Perl6::Pugs
D:\MoChou\pugs>dir makefile
􏿽xC7􏿽xFD􏿽xB6􏿽xAF􏿽xC6􏿽xF7 D 􏿽xD6􏿽xD0􏿽xB5ľ􏿽xED􏿽xCA􏿽xC7 􏿽xB9􏿽xA4􏿽xD7􏿽xF7􏿽xB7􏿽xD6􏿽xC7􏿽xF8
􏿽xBE􏿽xED􏿽xB5􏿽xC4􏿽xD0􏿽xF2􏿽xC1􏿽xD0􏿽xBA􏿽xC5􏿽xCA􏿽xC7 F8C0-4091
D:\MoChou\pugs 􏿽xB5􏿽xC4􏿽xC4􏿽xBF¼
2005-03-08 18:54 26,259 Makefile
1 􏿽xB8􏿽xF6􏿽xCEļ􏿽xFE 26,259 􏿽xD7􏿽xD6􏿽xBD􏿽xDA
0 􏿽xB8􏿽xF6􏿽xC4􏿽xBF¼ 56,369,152 􏿽xBF􏿽xC9􏿽xD3􏿽xC3􏿽xD7ֽ􏿽xDA
D:\MoChou\pugs>nmake
Microsoft (R) 􏿽xB3􏿽xCC􏿽xD0􏿽xF2􏿽xCE􏿽xAC􏿽xBB􏿽xA4ʵ􏿽xD3􏿽xC3􏿽xB9􏿽xA4􏿽xBE􏿽xDF 7.00.9466 􏿽xB0􏿽xE6
􏿽xB0􏿽xE6􏿽xC8􏿽xA8􏿽xCB􏿽xF9􏿽xD3􏿽xD0 (C) Microsoft Corporation􏿽xA1􏿽xA3􏿽xB1􏿽xA3􏿽xC1􏿽xF4􏿽xCB􏿽xF9􏿽xD3􏿽xD0Ȩ􏿽xC0􏿽xFB􏿽xA1􏿽xA3
NMAKE : fatal error U1052: δ􏿽xD5􏿽xD2􏿽xB5􏿽xBD􏿽xCE􏿽xC4􏿽xBC􏿽xFE􏿽xA1􏿽xB0Makefile􏿽xA1􏿽xB1
Stop.
NMAKE : fatal error U1077: 􏿽xA1􏿽xB0cd􏿽xA1􏿽xB1 : 􏿽xB7􏿽xB5􏿽xBBش􏿽xFA􏿽xC2􏿽xEB􏿽xA1􏿽xB00x2􏿽xA1􏿽xB1
Stop.
D:\MoChou\pugs>
Pugs Ϊʲô􏿽xB1􏿽xE0􏿽xD2벻􏿽xB9􏿽xFDȥ􏿽xA3􏿽xBF
autrijus 􏿽xD4􏿽xDA􏿽xB2􏿽xBB􏿽xD4􏿽xDA􏿽xA3􏿽xBF 10:55
thesaboteur erm stop using windows :) 10:57
autrijus hi flw. 10:58
that is strange.
can you try using an older version of nmake?
the free downloadable version of 1.50?
instead of 5.00.9466
I'll be back later 11:03
going away from net access & 11:04
dada it seems to work with nmake 7.10.3077, though 11:07
flw 􏿽xCE􏿽xD2􏿽xCA􏿽xD4􏿽xCA􏿽xD4 nmake 1.50
􏿽xD3􏿽xC3 nmake 1.50 Ҳ􏿽xB2􏿽xBB􏿽xD0􏿽xD0􏿽xA3􏿽xACͬ􏿽xD1􏿽xF9􏿽xB5Ĵ􏿽xED􏿽xCE􏿽xF3􏿽xA1􏿽xA3 11:16
dada flw: ? 11:22
flw ================================================ 11:26
D:\MoChou\pugs>svn up
􏿽xD3􏿽xDA􏿽xD0􏿽xDE􏿽xB6􏿽xA9􏿽xB0􏿽xE6 524􏿽xA1􏿽xA3
D:\MoChou\pugs>perl makefile.pl
*** Readline support disabled. If you want readline support,
please install Term::ReadLine::Gnu from CPAN, as well as
the GNU Readline headers and shared library.
Checking if your kit is complete...
Looks good
Writing Makefile for Kwid
Writing Makefile for Pugs::MakeMaker
Writing Makefile for Test
Writing Makefile for Perl6::Pugs
D:\MoChou\pugs>nmake
Microsoft (R) Program Maintenance Utility Version 1.50
Copyright (c) Microsoft Corp 1988-94. All rights reserved.
cp lib/Perl6/Pugs.pm blib\lib\Perl6\Pugs.pm
NMAKE : fatal error U1052: file 'Makefile' not found
Stop.
NMAKE : fatal error U1077: 'C:\WINNT\system32\cmd.exe' : return code '0x2'
Stop.
D:\MoChou\pugs>
================================================
D:\MoChou\pugs>svn up 11:27
􏿽xD3􏿽xDA􏿽xD0􏿽xDE􏿽xB6􏿽xA9􏿽xB0􏿽xE6 524􏿽xA1􏿽xA3
D:\MoChou\pugs>perl makefile.pl
*** Readline support disabled. If you want readline support,
please install Term::ReadLine::Gnu from CPAN, as well as
the GNU Readline headers and shared library.
Checking if your kit is complete...
Looks good
Writing Makefile for Kwid
Writing Makefile for Pugs::MakeMaker
Writing Makefile for Test
Writing Makefile for Perl6::Pugs
D:\MoChou\pugs>nmake
Microsoft (R) Program Maintenance Utility Version 1.50
Copyright (c) Microsoft Corp 1988-94. All rights reserved.
cp lib/Perl6/Pugs.pm blib\lib\Perl6\Pugs.pm
NMAKE : fatal error U1052: file 'Makefile' not found
Stop.
NMAKE : fatal error U1077: 'C:\WINNT\system32\cmd.exe' : return code '0x2'
Stop. 11:28
D:\MoChou\pugs>
================================================
dada uhm 11:29
what does 'dir Makefile' say again? 11:30
flw ================================================ 11:33
D:\MoChou\pugs>nmake clean
Microsoft (R) Program Maintenance Utility Version 1.50
Copyright (c) Microsoft Corp 1988-94. All rights reserved.
cd ext/Kwid/ && D:\Perl\bin\perl.exe -Iinc -MExtUtils::Command -e test_f
Makefile && nmake clean
clkao nopaste please
flw dada: please receive. 11:35
dada flw: I'm firewalled 11:40
oops
nothingmuch salad++ 11:43
perlbot: nopaste 11:45
perlbot Paste your code here and #perl will be able to view it: sial.org/pbot/perl
nothingmuch flw - you should use that
flw ok 11:46
sial.org/pbot/7977 11:47
nothingmuch err, set the chanel to #perl6 11:49
pasteling "flw" at 221.218.15.49 pasted "make pugs failed!" (78 lines, 2.9K) at sial.org/pbot/7978 11:52
nothingmuch ok, now find someone who knows more than me about windows/nmake ;-) 11:53
huh! 11:54
the makefile works 11:55
cp lib/Perl6/Pugs.pm blib\lib\Perl6\Pugs.pm
odd
dada flw: I don't know
nothingmuch does windows have the equiv of strace?
dada nothingmuch: I don't think so
nothingmuch that way you can find out rather easily exactly what makefile path it's trying to open
by looking for chdir and open syscalls
well, that's as far as I go
except that you should try going into ext/* 11:56
and see if there's a makefile in each one of those
dada the only thing that I can think of is some misunderstanding due to unicode, probably
nothingmuch yah, that happens a lot with make stuff
many perl modules' compilations break unless you set LC_ALL=C
if your username has funny chars 11:57
mine, for example, is יובל קוג'מן
make doesn't like that
thesaboteur Im really new at perl, I want to tell perl to only print out parts of a string and to discard the other parts. 12:01
print "From:\e $_[0]\e : $_[2]\n"
$_[2] this part here has a huge string in it 12:02
I want to chop out stuff like <html><body ichatballooncolor=
flw 􏿽xB2􏿽xBB̫􏿽xC3􏿽xF7􏿽xB0􏿽xD7􏿽xD6􏿽xEE􏿽xCE􏿽xBB􏿽xB5􏿽xC4􏿽xD2􏿽xE2􏿽xCB􏿽xBC􏿽xA1􏿽xAD􏿽xA1􏿽xAD
dada thesaboteur: what exactly do you want to chop out? a specified number of chars, or html tags, or what? 12:03
thesaboteur well here's the full string: <html><body ichatballooncolor="#ACB5BF" ichattextcolor="#000000"><font face="Helvetica" ABSZ=12 color="#000000">j</font></body></html> 12:04
I want to chop out everything except for "j"
but j will be different everytime
and their are other variables like "#ACB5BF" ="#000000" etc. 12:05
dada $_[2] =~ s/<[^>]+>//g;
wolverian thesaboteur: see HTML::Parser, and the various higher level interfaces to it 12:06
thesaboteur arg this more complicated then I thought it would be 12:08
metaperl HTML::Tree is the way thesaboteur 12:09
you just want the content of the font tag
my $tree = HTML::TreeBuilder->new_from_file($file);
my $font = $tree->look_down ('_tag' => 'font');
my ($content) = $font->content; 12:10
print $content
done
thesaboteur er yea but I just started messing with perl yesterday so.
nothingmuch thesaboteur: if the html is the exact same thing all the time you can also just snip that out, either like dada did, or match out the text with the whole strings on either side
metaperl++
thesaboteur: that doesn't mean you shouldn't do it right =) 12:11
thesaboteur nothingmuch: yeah I don't know how to put what dada gave me into my current setup though.
metaperl nothingmuch: thanks
dada thesaboteur: right before your print line
nothingmuch well, it's really simple, just take that chunk, and put it in, but put a 'use HTML::TreeBuilder' in front... Let me explain line by line
my $tree = # this declares a variable named tree 12:12
HTML::Tree::Builder->new_from_file($file)
and assigns it the return value of the method new_from_file called on the class HTML::TreeBuilder
that is to say
it tells HTML::Tree::Builder to create a new tree object, from the file whose path is in the strig $file
thesaboteur yea I know roughly how it works (I think). 12:13
nothingmuch ok, so we have the HTML tree of the document inside $tree
we ask it to find us an element named font
which is another object
we take that
and ask it to give us it's content
which is a string
we print that
thesaboteur yeah I see 12:14
nothingmuch search.cpan.org/perldoc?HTML::TreeBuilder <-- where to find documentation, and installation links
or perldoc HTML::TreeBuilder if you have it installed
if you don't have it installed, 'perl -MCPAN -e shell'
it will ask you to configure some options 12:15
use the default for what you don't care about
and then when you see a prompt, type 'install HTML::TreeBuilder'
thesaboteur ok kool, thanks
nothingmuch you can also use the download link from search.cpan.org
untar that
thesaboteur I didn't know I could do that
nothingmuch and do 'perl Makefile.PL' in the untarred dir
thesaboteur yeh I installed 2 modules already 12:16
nothingmuch but that won't automatically tell you about/install dependancies
thesaboteur that's badass it's installing now 12:19
it's like apt but for perl :)
nothingmuch you can also try CPANPLUS 12:20
it's the next version of the CPAN shell
a complete rewrite
thesaboteur kool
once I run perl -MCPAN -e shell do I have to configure everytime or can I run this a different way? 12:23
I ran it as non root the first time, so trying again now. 12:27
nothingmuch no, it saves it either in /Library/Perl/..../CPAN/MyConfig.pm 12:29
or something like that
or under ~/.cpan
i forget the diffs between CPAN and CPANPLUS in that respect
metaperl CPANPLUS works well for me for non-root installs. CPAN did not
nothingmuch CPAN is hard to trick to work as nonroot 12:30
but it can be done
CPANPLUS is better thought out for that
thesaboteur yeh it's working ok for me now 12:31
nothingmuch beh, what an annoying machine 12:35
thesaboteur dada: thanks what you gave me seems to work ok
nothingmuch has been running coverage analisys for a product on a box
for 2 weeks
err, it'll take 2 weeks
i've been at it for 1
and all the trouble that could happen is
now it's not powering up
=(
thesaboteur er that sucks 12:36
nothingmuch uhuh
thesaboteur what happens when you turn it on, any fans spin or anything?
nothingmuch nada
we had a power failure
thesaboteur eh 12:37
nothingmuch seems like the power supply took some punishment
thesaboteur yeah I would start there, what kinda box is it.
nothingmuch our sysop will deal with it promply, though
some mega super x86
fast
powerful
but not professional hardware
and if not our sysop, our vice president of R&D
thesaboteur heh
nothingmuch he's an EE guy 12:38
and loves nostalgic trips
nothingmuch wants sun/ibm to produce $20,000 x86 boxes
with reduntant power supplies
and little 'attn' leds
thesaboteur Now I just gotta get this thing to stop logging me off when I send lol.
nothingmuch so that we wouldn't spend so much time fixing them 12:39
i think that's a bot doing that
lol
hmm
shapr My box right next to me has redundant power supplies, and it only cost me 2500euro.
thesaboteur sun's new server the v40z er something like that has a mini computer inside of it, to give you detailed stats if it has problems, it's own little bios and stuff, hella kool.
nothingmuch we have sunfires here
an age old sun workstation 12:40
which only recently had keyboard trouble
also, a 192mhz aix box
i don't know how it's still alive
thesaboteur we used to have tons of old ss5 pizza boxes. 12:41
nothingmuch these things are just higher quality hardware
we have two x86 pizzaboxes by sun
but i don't know how good they are
thesaboteur lol
nothingmuch and 5 by asus
i hope they'll live
but these are servers
mail, dns, pdc, etc 12:42
the QA/development are the ones giving us crap
and they're cheap
thesaboteur yea we got tired of running servers on pizza boxes so we got an E4500 one day, big improvement for the cheap price.
i've never really been a big sparc fan though, haven't tried sol10 yet though. 12:43
nothingmuch sounds reliable
they're cheap now, too
thesaboteur that e4500 gave us alot of processor problems at first, that's fixed and it still prints goofy power supply errors to the console every now and then. 12:44
nothingmuch we have an e3500 as a dev machine at work
8 cpus
all slow, but still, 8 of them =)
thesaboteur yeah that's what we had
5gb ram
nothingmuch anywho, we don't have performance trouble
thesaboteur 2 18gb external drives
nothingmuch we have reliability trouble
and the sun/aix/hpux qa machines are good quality
thesaboteur yea that sucks
nothingmuch because you can't buy cheap one 12:45
s
but the linux/windows qa machines are crummy
thesaboteur heh
nothingmuch they're faster, but they break 12:46
thesaboteur I've never really had to many major problems with x86 stuff.
nothingmuch and the one i worked on just dies
died
so i'm whyning
whining
thesaboteur yea
nothingmuch ok, well, it was fixed
nothingmuch likes the sysop team =)
good think sqlite is transaction happy 12:47
thesaboteur I get to go home in an hour 12:48
nothingmuch where is home? 12:49
shapr Home is where the heart is.
thesaboteur dc metro area
nothingmuch it's very early there, no?
shapr Happily, I keep mine in my chest, so wherever I am, there you go.
thesaboteur yea, nearly 8am here
nothingmuch 1 hour day? or is there some catch 12:50
thesaboteur heh I work the 1-9
nothingmuch ah
at programming?
or do you babysit boxes at night?
thesaboteur I'm supposed to yell "fire" in the event of one.
yeah "Pat's the boxes" 12:51
nothingmuch they are that common, huh?
;-)
we have the backup tapes in a fire safe
and everything else is insured
thesaboteur so while im here making sure they don't blowup im trying to learn perl
just cause it sounded good at the time 12:52
nothingmuch well, in that case don't have too much fun
thesaboteur nothing else better to do
nothingmuch otherwise you might not notice and lose your job ;-)
thesaboteur it would be hard to miss it
lots of loud buzzers and bells and loud sound things anyways 12:53
nothingmuch for nonfires i hope you use nagios
or equiv
thesaboteur heh I had mentioned nagios when they interviewed me and they didn't know what I was talking about :/
shapr aieee 12:54
run away!
nothingmuch beh!
thesaboteur they use some proprietary closed source glorified "ping monitor" *cringe*
shapr Reminds me of a job where I interviewed and they didn't know what recursive descent parsers were....
nothingmuch if they want something propietry there's IPSentry which is very simple, or BB which is popular 12:55
thesaboteur eh I don't know what that is either.
shapr thesaboteur: do you do parsing for a living?
thesaboteur shapr: nope :)
shapr see, that's okay then ;-)
These guys did...
thesaboteur that's kinda like this database company I knew, they refused to use linux for any of there database stuff, something about not point and clicky enough for them. 12:57
so anyways afk for a few gotta go check on the boxes.
shapr Worst thing yet was a job where I got hired to increase the average cluefulness of fifteen or twenty other programmers. All of their java code used static methods, no object instances. All of their if then calls had an empty "else {};" because they didn't know you could leave it out. 12:58
nothingmuch oh crap 13:02
[root@safta yuval]# swapon -a
swapon: /dev/hdd3: No such device or address
no wonder tests are failing
uck
nothingmuch pays sysops another visit
thesaboteur well thanks for the help everyone, Im gonna go, gotta go do stuff. l8r 13:05
flw fayland 􏿽xD4􏿽xDA􏿽xC2􏿽xF0􏿽xA3􏿽xBF 13:20
shapr was that koi-8?
hcchien fayland: hola 13:21
fayland hcchien,hi
hcchien fayland: do you use utf-8 for the p6l10n-zh_cn? 13:22
fayland i'm not sure. maybe gb2312
need uft-8?
hcchien It would be better. :) 13:23
fayland i'll take care. :)
hcchien thanks.
nothingmuch utf-8 is the way perl6 is looking
theorbtwo utf-8 is the way perl6 is looking to prefer, anyway. 13:24
autrijus has stated that pugs will not accept latin1 « and », but only utf8, and it only seems fair to hold the zh_* people to the same standard. 13:25
nothingmuch i think i annoyed him 13:26
i normally go to great length to use utf8
because hebrew is very unreliable otherwise
but when i submtited t/op/zip.t and t/op/quoting.t i forgot
hcchien hmm..., theorbtwo can type chinese? :)
shapr I really prefer utf-8 myself. 13:27
theorbtwo Type, possibly. Understand, no.
nothingmuch now my vimrc has set encoding=utf8... I hope that doesn't hurt other things too bad
theorbtwo ﬒
hcchien I really prefer utf-8 myself either. 13:28
theorbtwo 万
There we are. Chinese. And I even know it means 10,000, because the unicode character browser told me so. 13:29
hcchien anyway, is anyone else interesting in p6l10n? 13:30
nothingmuch i might be tempted to say i will do hebrew 13:32
but i know i won't
=)
theorbtwo I suspect it'd be a very good way to force yourself to read the document you're translating hard. 13:35
But then again, I'm monolingual, and thus have little experience trying it. 13:36
nothingmuch well, true that 13:37
but i'm very bad with either language
hcchien now we have only .tw and .cn.
nothingmuch and languages in general
so... =)
you might be able to get lumi to do it
theorbtwo You seemed interested in l10n earlier, when you wanted to write p6 code in Hebrew. 13:38
theorbtwo sees nthListEl' (l:ls) 1 = l in Haskell Tutorial, and thinks "wow, it's already in Hebrew!" 13:39
nothingmuch ah 13:42
no, i didn't want to do it
i just showed how it might look like
i will only program in english, or the lingua franca of the period depending on what it is
and we don't see that changing anytime soon, do we? 13:43
shapr Doesn't lingua franca necessarily imply French?
nothingmuch anyone can learn perl more easily than hebrew, and most people already english, so...
yeah, that always bugged me, shapr... =/
shapr I always purposely misquote and say "lingua fracas"
nothingmuch the STM article autrijus gave me yesterday is really nice 13:45
shapr Software Transactional Memory?
nothingmuch uhuh 13:46
shapr that's some powerful stuff
SPJ is hoping to use that to make GHC SMP-happy.
nothingmuch indeed it is
shapr I'm pretty sure it will work out nicely.
nothingmuch really sees haskell becomming a major part of his work someday soon
err, play
shapr I do wish I could work region allocation in there somewhere.
I get paid to do Haskell =) 13:47
Amazingly, because I wrote an open source Haskell app and one of my clients wanted to use it.
nothingmuch ooh
what app was that?
shapr It's a webapp for searching mailman mailing list archives. I wasn't able to find an app that did that, so I wrote one for my own use. 13:48
It took me about four evenings of spare time to make it work, and it was about 200 lines of code.
It's gotten a bit bigger since I've had to handle weird special cases like duplicate message-id and broke RFC dates. 13:49
nothingmuch hopes he lives to see the day when computers will be tought what we want of them
ad they will design extensible, opaque protocols to handle the mess themselves
i have a question 13:52
about how something might look in haskell
this is a project that i'm sort of working on right now
you could call it a testing automation thing
but really it is something that must run jobs on abstract resources (remote machines, mostly) 13:53
in a certain order
and collect results
shapr Have you seen QuickCheck? 13:54
nothingmuch the resources are objects which normally look like semaphores
tmoertel's mentioned it many times on his website
i went through it once
anywho
shapr you run a job on a remote machine and want a certain result? 13:55
nothingmuch not only that
for example, to run a test you need 'resource produc daemon running', and resource '100mb of ram free'
and resource 'the target platform is xyz'
a group of resources that can be granted is typically sliced by machine
i.e. machine X is a sun box, with 8gb of ram, etc 13:56
it also has some funkier resources, like 'product is running'
which is a resource that when claimed will start a new job, whose resource is 'machine isa same as where i started'
which requires 'product binaries are available'
and starts the product 13:57
and so on and so forth
shapr sounds nifty
nothingmuch it's now trivial to make nightly builds with this, load balance tests, and so forth
theorbtwo Sounds an awful lot like apt-get.
nothingmuch theorbtwo: simpler example: make
shapr sounds like an interesting cross between design by contract, STM-optimistic locking, and unit testing.
nothingmuch ooh, it's 16:00
shapr it's 15:00 here.
theorbtwo Not sure about simpler. 13:58
nothingmuch time for cigarette & coffee
anywho, the protocol between resources is 'you were granted', which gives you a grant
you can claim the grant, or forefit it, and it might be revoked
this allows jobs which depend on each other's serial execution to share a resource
and not cause deadlock
theorbtwo But yeah, same thing.
nothingmuch when you got a grant, you have a token 13:59
which you return when you're done
the allocator api is for resources and requests
can you grant this now?
can you grant this ever?
etc
it's very oop
and i wondered how it might look in haskell/fp
but i've got to go for 5 mins
otherwise Uri's timeline is ruined ;-) 14:00
shapr sounds a lot like STM, but you might also want to look at QuickCheckM
viirya autrijus: ping 14:08
nothingmuch anywho, 14:15
i'm sorry to say this
but some guy who left work a week ago to study is having his goodbye party
and he brought sushi
so i'm introducing further delays, or perhaps a rollback of the conversation ;-)
shapr heh
sushi!
nothingmuch to reimburse you a bit 14:16
i'll give you a tip
brahm's first cello sonata
go get it, it's good for you
autrijus viirya: pong 14:17
greetings, folks, btw.
autrijus finished backlogging
shapr hiii autrijus!
autrijus finally got back from $work
shapr boings cheerfully
autrijus convinced gpg to look like pgp for a complex large application on aix. 14:18
painful stuff.
shapr I've seen some wrappers that do that, is that what you used?
autrijus yeah, I used pgpgpg
but it's not a complete emulation
so I made it more complete.
autrijus still backlogging 14:19
shapr cleans out the backlog
theorbtwo Allo, autrijus.
autrijus yo theorbtwo. how's life? 14:20
viirya autrijus: 'require' prim searchs incs not recursively. you know? 14:21
autrijus why should it search recursively?
perl5 does not
viirya autrijus: causing make test failure now.
autrijus really.
tmoertel autrijus: do you want darcs patches sent to [email@hidden.address] ? 14:22
autrijus tmoertel: yes, that is my intention
so when I'm not available other committers can freely commit them after some reviews
theorbtwo Not bad. A bit worried that some of my solutions to the exercises on YAHT aren't that hot.
autrijus viirya: what make test failure?
theorbtwo: well, they work, that's the important bit :)
viirya autrijus: make test could work, I mean.
/could/couldn't/ 14:23
autrijus trying to duplicate.
viirya pugs doesn't find out Test.pm.
theorbtwo Most of them. Some I am thus far unable to slove. 14:24
autrijus it all worksforme. 14:25
viirya: osx?
viirya autrijus: linux
autrijus weird. let me boot into fbsd.
viirya How it works? Now requireInc just search ./Test.pm in code. 14:27
autrijus are you working on the darcs repo? 14:28
viirya autrijus: no, svk. 14:29
autrijus weeird.
rebooting to fbsd now, bbiab.
rebooted. 14:39
clkao you seem to have a slow computer?
Limbic_Region that's a neat trick - rebooting without losing IRC connection
;-)
autrijus Limbic_Region: it's called "GNU Screen"
clkao: I have a semifunctional freebsd.
got no bandwidth in .cn to set it up
clkao autrijus: meanwhile, be glad that you use asus.
www.beareyes.com.cn/2/lib/200503/05...305003.htm
it seems very suitable for, uh, heavy duty users like you
autrijus viirya: make test worksforme on fbsd. 14:40
statico <
clkao: as long as it's asus with via chipsets, not nvidia...
autrijus viirya: can you start from a clean tree?
I'll do the same.
viirya autrijus: All I do is 'make test'. It doesn't work. 14:42
autrijus viirya: did you do "make purge"?
then "perl Makefile.PL" etc?
viirya why doing make pruge?
autrijus because a certain ingy broke the PugsConfig.pm process twice 14:43
viirya /pruge/purge/
autrijus I seem to remember that there's patch somewhere to hook PugsConfig.pm as part of "make" dependency
viirya autrijus: I check out a new copy of pugs.
autrijus yeah, theorbtwo had a patch. 14:44
theorbtwo: are you not a committer?
theorbtwo: would you like to be one and commit it by yourself?
mm -O3. 14:46
I think the gain on making tests is little
and the loss on recompilation is larger by comparison
as has been shown here
so I'll refrain from making optimization the default. 14:47
I will, however, add an "optimized" target.
so you can do "make optimized" and be happy.
hi fayland! 14:48
I understand you failed making pugs?
Coke_ seen obra
autrijus hi Coke_ :)
Coke_ whoops. wrong window. Hi, autrijus. 14:49
fayland ya,let me do it again
hcchien fayland++ # convert the doc to utf-8
fayland wait a moment
:)
Coke_ hey, autrijus - can you add <a name> markers to svn.perl.org/perl6/pugs/trunk/docs/...rview.html ? 14:50
autrijus Coke_: I'd gladly accept patches that does that :)
Coke_ (I'd like to link to "What was the revised bootstrapping plan?" from parrotcode.org)
fayland autrijus: NMAKE : fatal error U1052: file 'Makefile' not found 14:51
Coke_ is pugscode.org in the same repo as parrotcode.org ?
autrijus mm gugod volunteered to kwidify PA01
fayland autrijus, just 'svn up' and perl makefile.pl nmake test,and return that
autrijus Coke_: no, it is not.
fayland: rename your C:\Program Files\Microsoft Visual Studio*
or just remove that from your path
and run perl Makefile.PL again. 14:52
fayland ok.
autrijus viirya: so it works now? 14:53
I've committed theorbtwo's patch as r525
you no longer need to do "make purge" again.
also, for people who's into fast pugs, "make optimized" works now.
greetings stevan-san. 14:54
fayland autrijus, it doesn't work. the same result. 14:55
autrijus got some time to help me to redo the PA01 chunk of "Is there a CPAN for Perl 6 modules?" ?
fayland: so nmake is still in your path?
stevan morning autrijus (at least morning for me that is)
autrijus fayland: try to find out where is nmake.exe and remove it from path.
hcchien hmm...., test failed on op/not
fayland I have two nmake, one at c:\usr\bin,another just removed 14:56
Steve_p Yeah, op still needs to be changed per Larry's comments
stevan hcchien: I think luqui un_todo-ed the tests which are bugs and not un-implemented
Steve_p thought he TODO'd those
autrijus fayland: ok, remove c:\usr\bin from your path as well.
stevan autrijus: are you asking if I want to help on PA01?
autrijus fayland: or rename nmake.exe
stevan: yeah. if you don't feel like doing it, I'll ask for ingy's help :) 14:57
autrijus is in code mode not docs mode
hcchien so I should ignore the failed and make install? :)
autrijus hcchien: yes, that's the plan
nothingmuch yummy
stevan autrijus: I can do that, sure
nothingmuch morning stevan
stevan hey nothingmuch
nothingmuch we had wine too
hcchien autrijus: not marked as 'TODO'?
nothingmuch =)
autrijus stevan++
nothingmuch (i had wine leftovers)
autrijus hcchien: luqui's idea is that TODO means todo
i.e. 14:58
missing features
and misimplemented stuff should fail
to give them higher priority.
now I'd gladly accept a patch to Makefile.PL that says that.
scw What's the easiest way to test single .t file?
autrijus i.e. "Now, type 'make test' to test -- some tests may fail; that's okay inbetween releases... etc"
scw: prove.
env HARNESS_PERL=./pugs prove t/op/foo.t 14:59
fayland autrijus, it returns cann't find nmake and lwp from microsoft, then return the same.
theorbtwo autrijus, thanks for the offer of committer bits, but I'd rather wait until I write a patch that just goes in.
stevan autrijus: where is the original PA01? docs/01Overview.html? 15:01
scw Seems the @INC is not corrected, I got: NonTerm "/usr/share/perl/5.8/Test.pm" (line 31, column 1)
autrijus stevan: yup, that is the one.
scw: "make purge" ?
scw: oh, prove? 15:02
fayland autrijus: Fetching 'Nmake15.exe' from download.microsoft.com...
scw autrijus: ya
fayland autrijus: that's right?
autrijus fayland: yeah, so Makefile is really not there for nmake. that's weird.
fayland: are you under cygwin?
viirya autrijus: certainly weird.
scw autrijus: I used 'HARNESS_PERL=./pugs prove t/op/assign.t' and got such error..
viirya autrijus: it doesn't work.
fayland no cygwin.
theorbtwo Why not just ./pugs t/op/assign.t ? 15:03
stevan scw: I am getting the same thing "NonTerm "/System/Library/Perl/5.8.1/Test.pm" (line 31, column 1)"
viirya autrijus: complaining Can't locate Test.pm
autrijus: stevan's problem is that.
fayland autrijus: but i can see makefile in the folder
nothingmuch mine too, btw
autrijus scw: add this
PERL6LIB=./ext/Test/lib
nothingmuch s/\./\$PWD/; # works under all shells, less problematic 15:04
scw autrijus: This works :)
autrijus fayland: nmake /f Makefile
nothingmuch stevan: go write yourself down in pugs.kwiki.org/?OSXPeople
viirya autrijus: after adding that, it surely works.
autrijus: but why not search all dictionaries under Incs. 15:05
fayland autrijus: NMAKE : fatal error U1052: file 'Makefile' not found
autrijus, still the same 15:06
autrijus viirya: i'd argue it's NotABug then. inconvenience, yes, but not a bug.
fayland: nmake /f c:\absolute\path\to\Makefile
viirya autrijus: It's your intention to just add module one by one? 15:07
autrijus viirya: if you do "make install"
then it will just pick things up.
before "make install", yes.
I'll remove PERL5LIB and PERLLIB from the search path now.
viirya autrijus: ok. see. 15:08
nothingmuch stevan: did you write a test for your mail titled 'Bugs' of yesterday?
fayland autrijus: no help.
autrijus: :(
stevan nothingmuch: yes, t/op/sub_return_values.t has some tests for that
autrijus fayland: that is a bit too weird.
nothingmuch stevan++
fayland autrijus: no body else met that? 15:09
autrijus nope.
nothingmuch somehow i missed it in the svn commit RSS
autrijus fayland: ok, try these commands:
hcchien should we merge t/0*.t to t/Synopsis ?
hcchien is back to pugs. :p
autrijus perl -Iutil -MPugsConfig -e "PugsConfig->write_config_modules" > src\Config.hs 15:10
perl Makefile.PL -version
ghc --make -o pugs src/Main.hs -isrc
fayland: try the three commands. 15:11
fayland ok, wait a moment
autrijus hcchien: that was the plan, but the new, much more precise layouts seems to be overlapping a lot with the original Synopsis plan. 15:12
i.e. t/syntax/ t/magicals/ etc
so I'm not at all sure how to link the tests with synopses.
hcchien autrijus: yes, I saw that. and I think it woule be better. :)
autrijus except with the new regime 15:13
we no longer know which synopsis paragraph is tested
stevan autrijus: I was thinking of writing up a test document
autrijus and which is not.
stevan: elaborate?
hcchien I am thinking about this issue, too.
stevan one that would describe the contents of each test file, so people would be able to find where a test goes better
autrijus stevan: it is best to annotate synopsis with test file names, I think.
stevan autrijus: that works too 15:14
fayland autrijus: Can't locate object method "write_config_modules" via package "PugsConfig" at -e
line 1.
autrijus because to annotate with actual tests is cumbersome
so maybe just add random =head3 lines
or other lines
to the synopses pods?
i.e. annotate paragraphs with =head3 (untested)
and for things that has tests, annotate with =head3 t/op/zip.t 15:15
or something. that's just a wild idea.
stevan autrijus: ok I will start on that after the PA01 stuff
autrijus stevan: woot!
stevan: if you do that, maybe just put a README to t/
and say "see Synopsis/*.pod for our coverage statement"
and then start annotating *.pod under there. 15:16
scw autrijus: Hope my commit doesn't break any thing and I have to go home now, 'night :)
autrijus 'nite!
nothingmuch i disagree
autrijus scw++
nothingmuch: cool. elaborate?
fayland autrijus: maybe just write_config_module ; no s
nothingmuch that is hard to make a table like analysis
featurex- yes
autrijus ooh. wow!
nothingmuch feature z- no
autrijus "Assigning to slice of array now works and some tests on this is added.
fayland: right!
nothingmuch: but "feature x" is complex 15:17
stevan nothingmuch: so are you thinking a feature matrix is better?
nothingmuch i don't know
autrijus it is best if we can cross reference.
stevan nothingmuch: I have a deal for you
fayland autrijus: now it's compiling
stevan I will start with annotating the Synopsis
nothingmuch i think we should make grep -r on *.pod easily recover:
a. what's covered
b. what's to cover
c. the things name, for further, or location in file 15:18
autrijus scw++ # slice array assignment. fantastic
nothingmuch for further grepping, that is
stevan nothingmuch: how can we get that out of grep?
nothingmuch maybe grep -B or -A depending where you put that
or =head1 Smart Matching [unimplemented, tested] 15:19
fayland autrijus: now i got '== Version: 6.0.10 (r525) ====================================='
autrijus: done.
autrijus fayland: ok... I guess that works.
fayland: but you still need to fix the nmake issue to run the test suite. hrmph.
nothingmuch perhaps [pugs status: unimplemented, tested] for more uniquy extraction
autrijus wonders if it's too late to switch to Build.PL
15:19 Aankh|Clone is now known as Aankhen``
fayland autrijus: i have no idea how to fix it. :( 15:20
nothingmuch i doubt it is
except you might make people angry
autrijus just "[pugs: todo, t/op/zip.t]" is fine
fayland autrijus: it works in other modules.
autrijus fayland: bizzare.
nothingmuch autrijus: i agree
autrijus =pugs todo t/op/zip.t 15:21
or, less bizzare:
=head3 pugs todo t/op/zip.t
array.t is now all pass. 15:22
fayland autrijus: see u tomorrow. i must go sleep cause i must get up early to have lessons
autrijus fayland: okay. we'll figure out the problem later. 15:23
fayland autrijus: okay. bye.
autrijus hi ninereasons. 15:25
ninereasons hi
nothingmuch beh, sorry 15:26
discussion about something interrupted
autrijus: i think it should be on the original =head$x of the docs
and perhaps preceding blocks of example code 15:27
autrijus maybe. we need to mark it though.
B<pugs todo t/op/zip.t>
works for you?
nothingmuch i think so 15:28
stevan autrijus, nothingmuch: why don
t
I just start going through it
nothingmuch i can help, but not for a few hours... when i'll get home
stevan because I am thinking that we may have plenty done which are not covered in the Synoposis
autrijus hey.
stevan since the Synopsis are not 100%
autrijus let us abuse S<>, ok? 15:29
S<todo t/op/zip.t>
and just grep for 'S<' ;)
stevan is S<> valid POD?
nothingmuch heh
autrijus it is.
stevan nice
autrijus it means <nobr>.
nothingmuch that's like mucking html with &nbsp;
and grepping for that
;-)
stevan LOL
autrijus bwahaha.
stevan ok S<> it is
autrijus yeah. and I think for things we have done and synopsis talks nothing about 15:30
maybe we add something like
nothingmuch we could also do Z<>pugs blahZ<>
autrijus S<unspecced t/op/foo.t>
nothingmuch: ugly!
nothingmuch oops, i forgot the ;-)
autrijus :)
theorbtwo Why not put comments next to the tests, rather then the other way around?
nothingmuch that is mostly done, i think 15:31
except where tests are two synopses at once
autrijus theorbtwo: both needs to be done.
stevan theorbtwo: I wanted to make a test overview basically
nothingmuch see for example t/op/smartmatch.t
autrijus the Syn->Tests directly is the currently lacking one though.
stevan so people can easily see what is in where
nothingmuch it's pretty verbose that way
autrijus and like nothingmuch said, tests are less structured than syns 15:32
so it makes sense to link from the structured one to unstructured ones.
I think, anyway
nothingmuch perhaps for further greppability, and goodness in general, S<unimplemented, F<t/op/blah.t>>
err, awkability
autrijus S<unimplemented t/op/blah.t> is good enough, really :)
nothingmuch i'm just getting carried away 15:33
autrijus wants to keep the entry barrier as low as possible
stevan I agree, if we need more we can figure it out then
autrijus stevan: the map can then be generated by a inverted index script.
nothingmuch we could expand that, to S< blah F<blah»
autrijus the tests -> syn map that is
stevan dashes off to get some more coffee 15:34
nothingmuch should shut up and get back to work, or not get back to work but actually do something
autrijus :) 15:35
nothingmuch autrijus: did you also stub !~? 15:39
15:39 _metaperl is now known as metaperl
autrijus nothingmuch: no. want to do that for me? 15:40
just do what I did with ~~ and replace eq with ne.
nothingmuch i'll look at your diff and try to repl
btw, is my %hash = (
autrijus no need to diff
nothingmuch "foo", bar", 15:41
autrijus just look at Prim.hs
and copy the two lines that contains ~~
nothingmuch "blah", "blah",
);
still allowed?
(what about parsing it?)
(it = !~)
autrijus it's already parsed. 15:42
nothingmuch goody
autrijus nothingmuch: I think it's still allowed, aye
nothingmuch it doesn't parse
new test =)
autrijus let's see if I beat you to it :) 15:43
r530
should be now parsed.
nothingmuch sticks tounge out
no fair
autrijus smiles
dada autrijus: which time plan do you have to implement unary * (the "splat" operator)? 15:44
theorbtwo ponders...
autrijus dada: I had not thought about it much, to be honest. definitely before 6.2.0. 15:45
theorbtwo What's the time-plan for implementing "Haskell is the XS of pugs"?
autrijus theorbtwo: Kwid.hs?
it should happen as soon as ingy and I gets some pairing time.
which is, at latest, on YAPC.
dada autrijus: well, if it's more than 2 weeks from now, I will try to give a stab at it. if it's sooner, I don't even try :-) 15:46
autrijus the final week of march
dada: please give a stab at it!
I don't plan to see 6.2.0 before YAPC::Taipei.
theorbtwo Kwid is trying to muck with the parser, though... I was thinking more in terms of exporting a couple of functions.
autrijus theorbtwo: nono, kwid has its own parser...
nothingmuch isn't 6.2.0 two weeks from now? ;-)
autrijus nothingmuch: who said that? ;)
theorbtwo: so it's just exporting the usual functions.
nothingmuch just kidding, autrijus =) 15:47
nothingmuch twiddles fingers, waiting for pugs to finish compiling 15:48
autrijus != in haskell is /=
btw.
think math. :)
nothingmuch i thought it was an alias for the perlop
i just saw it break =)
autrijus see "ne" above =) 15:49
nothingmuch dances the svn commit dance 15:50
Steve_p heh 15:51
Steve_p forgot to do the dance last night 15:52
stevan autrijus: I am still getting the problem with 'make test' 15:53
cant locate Test.pm stuff
autrijus stevan: "make purge" not curing things?
stevan no
autrijus most bizzare. fresh checkout?
stevan r531
I will try again 15:54
autrijus I mean, raze things
and "svn up"
rm -rf *; svn up
or, if using svk, rm -rf *; svk revert -R .
nothingmuch can you assign pairs to a hash yet? 15:55
autrijus mm !~ works now
nothingmuch++
nothingmuch yeah, that was really hard
=P
i was told what to do, and it was two full lines! 15:56
autrijus gosh!
that's almost a dissertation right there!
nothingmuch notes that ghc is /still/ compiling Eval.hs
autrijus notes that "make optimized" is bad for your health 15:57
nothingmuch it's better afterwords
my laptop is 400mhz
eval '%hash4 = ("key" => "value")'; 15:58
this is fatal
autrijus since it's a casting error.
wow bizzare.
looking into it
nothingmuch ok(eval ' 15:59
my %hash = (
"foo", "bar",
"blah, "blah",
);
1;
', "expression parsed (well, lexed)");
autrijus yeah, I'm looking into it. I think I know what's wrong. sec.
nothingmuch nok... VError "\nunexpected \"b\"\nexpecting term postfix, operator, \",\" or \")\"" (NonTerm "<eval>" (line 4, column 25)) 16:00
err, append 'say $!' after the test
it's a different thing
fixed "blah -> "blah", still error, quite similar 16:01
say ~%hash -> Fail: cannot cast as Str: VHash (MkHash [(VStr "blah",VStr "blah"),(VStr "foo",VStr "bar")]) 16:02
should i test for stringification in t/hash.t? 16:03
autrijus sounds good
nothingmuch i don't know what it looks like, i just know that it will
stevan autrijus: it seems to be looking for blib6/ which is not there 16:04
autrijus weird. it is there for me. 16:05
your ExtUtils::MakeMaker version?
stevan 6.12 16:06
autrijus mm 6.17 here. 16:07
stevan ok, I will update it
autrijus so when you do "make"
stevan make works
autrijus things are generated in blib/lib? 16:08
instead of blib6/lib/
stevan its make test that messes up
yeah looks like it
autrijus post your Makefile's
test_dynamic :: pure_all
section.
nothingmuch does anybody know what the output of hash interpolation will look like?
stevan ok one sec
very strange 16:09
it is making blib6 in the parent directory
autrijus wtf...
stevan yeah
odd
arcady nothingmuch "By default, keys and values are separated by tab characters, and pairs are terminated by newlines." - S02 16:10
nothingmuch how is this related to osx is what i want to know
stevan this is the test_dynamic stuff:
PERL_DL_NONLAZY=1 $(FULLPERLRUN) "-Iinc" "-MExtUtils::Command::MM" "-e" "ENV->{HARNESS_PERL} = q*/Users/stevan/Projects/PUGS/pugs/blib/script/pugs*; ENV->{PERL6LIB} = q*/Users/stevan/Projects/PUGS/pugs/blib6/lib*; test_harness($(TEST_VERBOSE), 'inc', '$(INST_LIB)', '$(INST_ARCHLIB)')" $(TEST_FILES)
nothingmuch i have it too
arcady++
autrijus weird. what perl has 6.12 as EUMM?
stevan Module::Install::Pugs.pm line 9 is "$self->set_blib('../../blib6');" 16:12
that seems like it might be the problem
maybe that should only be one ../
nothingmuch assumes that you subclass Hash and stringify it otherwise then it's diff 16:13
can't test till 6.2.0
eval '~%hash9'
that's also fatal
cannot cast into [VPair]: VSub..
autrijus stevan: how about taking an absolute path at that point?
nothingmuch: yup, I'm aware of it now.
nothingmuch just in case it might be unrelated, autrijus 16:14
Tom Waits++
stevan autrijus: what do you mean by absolute path? (my Makefile skills are limited)
Boogie Down Productions++
nothingmuch google++
autrijus stevan: I'll deal. sec 16:16
nothingmuch ah, hip hop
that's why i never heard of it
nothingmuch doesn't know anything about hiphop
a teeny bit about pop
metaperl stevan, you are too old to like BDP
stevan check out my hip hop album: diskjockeybyrd.iuma.com
nothingmuch knowlege stops at 1999 or so, when sister stopped watching TV that much
stevan metaperl: not true, i am too young 16:17
metaperl oops
document contains no data
my album is gone from the web... sigh
stevan metaperl: its working for me, but very very slowly 16:18
nothingmuch autrijus: is t/syntax/lf_in_list.t, which assigns (\n"foo", "bar\n"blah", "blah"\n); to %hash a todo, or a bug? 16:19
autrijus nothingmuch: I think it's a todo. not sure 16:20
nothingmuch todo then
autrijus nod. still fixing blib6 issue. 16:21
for web app writers:
cocoon.apache.org/2.1/userdocs/flow/index.html
I intend Pugs to be able to do that (and more) on 6.2.
(that's for another $project)
metaperl Perl has something called Contize 16:22
nothingmuch uhuh
autrijus and Continuity
metaperl on CPAN
nothingmuch brock/awwaiid does it
autrijus which are badly broken.
nothingmuch he is a very cool guy
theorbtwo (And to think, I knew him way back when.)
nothingmuch says he will hack pugs once he grows out of OcaML a bit
Ocaml? OCAML?
beh
autrijus Contize works only on shallow real hashes. 16:23
nothingmuch metaperl: how could they not be broken? perl is not pure
autrijus not objects or control flows in general.
well, javascript is not pure either ;)
and yet Cocoon/Rhino does that beautifully.
metaperl this cocoon thing is powerful
nothingmuch i think Coro can be used to butcher that
metaperl but it's Java based, not javascript AFAICT 16:24
autrijus nothingmuch: no, I hacked that and talked to Coro author.
nothingmuch in a way that really works
ah
autrijus metaperl: no, the continuation engine is javascript.
nothingmuch just speculating
autrijus I tried an attempt in perl5 to do continuations
with regex.
and another attempt with $^D.
both works to a degree, but are super-fragile.
nothingmuch brb 16:25
autrijus I'll spare you the details :)
err, $^P, not $^D.
but Pugs is pure CPS.
and has a small enough AST
so should be easy to contize.
hi nnunley. I'm daydreaming here about cocoon.apache.org/2.1/userdocs/flow...tions.html in Pugs :) 16:26
metaperl will you be able to hijack the javascript part of cocoon directly? 16:27
with no mods?
autrijus Inline::Rhino? sure. 16:28
but I want to write the logic in perl, not javascript!
(although javascript is actually a neat little language to write logics.) 16:29
nnunley autrijus: Nice. Have you looked at seaside? 16:31
nothingmuch sits down...o O ( now what should I do? )... ... ...( go home! )
autrijus nnunley: yeah, and wash, and borges, and lots of other things.
nothingmuch writes 'irc log ca. 18:30' on his hand 16:32
autrijus (and the LL* talks)
nothingmuch ciao!
metaperl the main wash site is down. do you have a link to seaside?
autrijus nothingmuch: mmm Memento
nnunley seaside.st/
nothingmuch oh btw, thanks for the STM article, autrijus!
autrijus beta4.com/seaside2/
nothingmuch t/syntax/lf_in_list.t comitted
autrijus nothingmuch: np :) glad you enjoyed it. I think STM is really good.
nothingmuch++
nnunley autrijus: That redirects to the seaside.st site.
metaperl I say forget YAHT. Go straight to this algorithms book
autrijus ahh. ok
nothingmuch oh, lastly
did you look at the private msgs i sent you?
autrijus the algorithm book is wonderufl. 16:33
wolverian hrm, how do I specify that a sub doesn't return anything in the sub signature?
autrijus nothingmuch: I did. It looks vaguely like doing a combinator library in perl5. I'm not sure it's my forte.
wolverian: returns Void 16:34
(not sure, just a guess.)
nothingmuch i was actually thinking of doing it in haskell and perl together
each side for it's forte
autrijus ah. that could be fun. I think you can write it up a bit and solicit suggestions on #haskell
there's several such combinator libraries in haskell already.
nothingmuch right now it's low on the task circular buffer/stack thingy 16:35
it's a dependancy of pugs hacking which is a dependancy of haskell
and it's also a dependancy on SICP
(finishing it, that is)
and prolog
autrijus mmm #perl6 is apparently full of ADD folks.
<- being one
nothingmuch add?
autrijus en.wikipedia.org/wiki/ADD 16:37
nothingmuch finished packing... ciao!
autrijus stevan: blib6 should be fixed. 16:38
Steve_p Since I just ran across it and wished for something like it in Perl, is there a function composition operator/function in Perl 6?
stevan autrijus: just svn updated, will let you know
metaperl maybe we need a thin layer between haskell and perl 16:39
autrijus perl5?
Steve_p heh
metaperl so function composition, etc. dont need re-implementation
Steve_p Inline::GHC :)
autrijus sub infix:<o> (Code &f, Code &g) { sub { g(f(@_)) } } 16:40
not quite.
in ruby it's 16:41
class Proc
def *(other)
raise unless arity == other.arity
proc {|*a| self.call(other.call(*a))}
end
end
the proposal to implement it in Ruby was rejected: 16:42
"A very few people would use this feature."
thanks, matz :)
Steve_p I was thinking a case like MY::postamble in EU::MM could easily use it 16:44
stevan autrijus: its working now
Steve_p tires of hacking the symbol table :)
autrijus cool!
Steve_p: post to the idle speculation list!
Steve_p p6l? :)
autrijus yeah :)
that's larry's words, not mine
Steve_p That's why I have my questions in test cases tied to a specific Larry email :) 16:45
autrijus heh heh.
mm I think this p6 summary is the longest to date. 16:46
clearly we're doing something right :) 16:47
Steve_p It's been a busy couple of weeks :)
Steve_p is still trying to catch up from being sick
Coke_ has yet to see a summary. 16:53
autrijus www.nntp.perl.org/group/perl.perl6.compiler/278 16:54
Coke_ I see pugs has a queue at rt.perl.org/rt3/ now.
autrijus it does. 16:55
had not announced it yet
still needs to do some mailgate foo.
(and I like using tests as tickets :)) 16:56
(but I'll announce it before YAPC.)
Steve_p remembers to checnk for spam in Perl5
17:00 dadaumpa is now known as dada
Coke_ sorry, didn't mean to steal your thunder. It's already out there, though. (Note, there's ALSO a perl6 queue. someone explain THAT to me. =-) 17:01
Steve_p I thought it was there for questions 17:04
Steve_p takes a peek
Hmmm...test patches, patches to languages/perl6 17:06
Hey, , there's ALSO a perl6 queue. someone
explain THAT to me. =-)
Crap
Thu Mar 04 17:30:00 2004 coke - Queue changed from parrot to perl6
:)
It seems like Perl 6 is all languages/perl6 related tickets 17:08
ingy hola
autrijus wow, look at #21679
that is very far along 17:09
but I can't seem to download the attachment?
it's all \x00
the .tar.gz one is good 17:10
Steve_p And it's all over two years old :-/ 17:11
autrijus and that has never landed 17:12
so is bitrotten 17:13
Steve_p :(
autrijus use inline 'Haskell' 'some code here'; 17:14
Steve_p heh
autrijus use extension 'Haskell' 'file.hs';
I'm just mimicking that patch.
not a bad idea.
ingy: what do you think? :)
that's sad. if that patch has landed and polished a bit 17:15
language/perl6 could've been bootstrapping :-( 17:16
Steve_p I'll have to check to see how bitrotten these patches are
autrijus is language/perl6/ still working?
I mean, compiles, passes tests and all
Steve_p Good question :-/ 17:19
Coke_?
ingy well I think you should hust be able to write modules in haskell with no .pm file required and just `use` them like any other
I plan on implementing the plumbing during yapc 17:20
autrijus woot!
let us totally do that
starting from Kwid!
theorbtwo Much coolness.
ingy and then yaml
autrijus and then yaml.
that way haskell and pugs will have YAML in no time
ingy does the happpy dance
autrijus too
Steve_p and then the breakfast cereals...Ooop, this isn't the "Holy Grail" ;) 17:21
autrijus I also happen to think YAML can help us fix the AST.
fix, as in, enrich the types.
especially references and co. 17:22
ingy: YAML can't serialize infinite data structures, right? :)
ingy autrijus: sure it can :P
autrijus ok... give me a YAML serialization for the fibonacci series...
ingy as long as it can be expressed in terms of a scalar sequnce or mapping 17:24
autrijus predices explosive productivity during the 2 weeks of YAPC::Taipei
Steve_p YAPC::Taipei is two weeks long?
ingy but you could come up with a way to do it
autrijus Steve_p: the conf itself is 2 days; but the prehackathon and posthackathon is 2 weeks in total.
ingy I hope I have time to write my presentation on Kwiki 17:31
:p
hcchien ingy: do you send me your schedule? :) 17:32
Limbic_Region I parsed that wrong ingy - I read it as "I hope I have time for a quickie"
ingy hcchien: I think so 17:33
ingy always has time for a kwiki
hcchien is checking 17:34
ingy hcchien: you mean my flight schedule, right?
Dear Brian, 17:36
Congratulations! You have been accepted as a presenter for
the O'Reilly Open Source Convention 2005 at the Oregon Convention Center,
Portland, Oregon, August 01, 2005 - August 05, 2005.
The following has been accepted as a 45 minute session for the event:
"Apocalypse Now! - Perl6 is Here Today"
sorry :(
autrijus woooot!
ingy++
dada ingy** 17:37
autrijus mmm ICFP deadline is 13th april
larsen I humbly add my ingy++
obra ingy: woot
ingy I guess that means rewriting Spork in Perl6
autrijus "Practice and experience: Functional programming in education and industry." 17:38
somebody remind me to submit the proposal before YAPC::Taipei hackathons are over :)
hcchien ingy: yes. so I didn't see that before. :) 17:39
obra autrijus: will try
autrijus wonders when will euroscon start to call for papers
dada autrijus: well, the CFP for the Italian Perl Workshop is already out :-)
obra does euroscon have a date yet? 17:40
obra still hasn't heard back from ORA about OSCon
autrijus not sure.
when is italian perl workshop?
rgs obra: october 17-20 it seems
dada autrijus: 23-24 june
autrijus icfp is September 26-28
so I'll shape my schedule around that
AOSS is early september @ srilanka or cambodia 17:41
obra ponders early october eurohackathon
autrijus so euroscon fits just fine
obra: count me in :)
clkao whoot
autrijus june is a bit early. 17:42
obra autrijus: was hoping for that
clkao round world hacking tour again?
obra we'll need tshirts
hcchien hehe
shapr autrijus: you're going to ICFP05?
hcchien I want to go to Ams, too.
autrijus shapr: I am.
if my proposal is accepted, that is.
if not, I'll try to. 17:43
shapr hmm
autrijus I got mail from wadler yesterday :)
asking me to list pugs on his fp-in-realworld page.
shapr heh, cool 17:45
dada I'm afraid I MUST pose a few stupid questions now 17:47
shapr does muscle poses
autrijus I just submitted autrijus.org/tmp/pugs.html :)
dada: fire away, sire
dada am I right in suspecting line 190 of AST.hs (vCast (VArray (MkArray l)) = l) being responsible for the non-working of *@array? 17:48
autrijus why do you think that?
I can't see a connection :)
dada because op1 "*" = return . VList . vCast
autrijus I thought it's the "*" op being wrong. 17:49
"list" op does what it should do.
* and ** needs to do more
but they don't
dada oh
ok, ok :-)
autrijus the "," Syn prolly need to take account of App "*"
and flatten them
alternatively, make "*" a Syn.
I'm not sure whetier if prefix:<*> is rebindable. 17:50
or whether it's implementable in perl6.
dada slowly, very slowly
I'm just hacking around, I'm totally clueless in Haskell
autrijus or both. or neither. or both.
autrijus starts resembling larry.
ok. let me explain :) 17:51
dada Syn == syntax
autrijus right. Syn is "special form" so to speak.
it means things that are inherent to AST
and are not rebindable.
i.e. can't be reimplemented in pure perl6.
and you can't generally take references to Syn stuff. 17:52
until we have macros.
which I'll not go into.
dada ok, and they are implemented in Eval.hs (mostly, at least)?
autrijus luqui: hey, is prefix <*> ?
luqui: hey, is prefix <*> a Syn?
i.e. what will happen if I redefine prefix:<*> and prefix:<**>? 17:53
luqui are you using <> as a quoter in your question?
autrijus yes I am.
dada: yes. see line 270
luqui well I wonder what it would return if it weren't
autrijus I wonder too.
also, "is returns Void" sane?
luqui yeah, that's been mentioned befor 17:54
wait
"is returns"
autrijus is returns == returns
dada autrijus: ok, I follow
autrijus iirc.
luqui oh... is returns(Void)
autrijus yeah.
luqui traits don't behave like verbs
autrijus <- about to sleep, not in the tiptoe typing shape
luqui so the parens are required
autrijus yup.
dada: so see line 365
nothingmuch autrijus: tomorrow i'll start writing tests for OOP
autrijus nothingmuch: woot 17:55
nothingmuch and rules
dada autrijus: seen
nothingmuch i wonder if one could test macros
dada autrijus: and also line 448-455
autrijus you see there an implementation for flattening.
luqui nothingmuch: with 'eval', I don't see why not
autrijus yup.
dada autrijus: in 365?
nothingmuch without starting a separate interpreter
luqui i want macros to be implemented
autrijus in 367 in fact but yeah.
nothingmuch well, supposedly they can go farther than that...
dada I see, I seeeee 17:56
luqui farther than what?
autrijus dada: ok. now see Bind
mm wait a bit
nothingmuch maybe i just dunno 17:57
i thought they weren't limited to anything
that is, they're as powerful as source filters
autrijus still in Eval 17:58
line 559
nothingmuch well, ok, why not
autrijus it is calling bindParams prms invs args
so either bindParams or doApply needs to analyze invs and args
dada autrijus: ok
luqui so why can't you test that with eval?
autrijus and see if they contain Syn "*"
nothingmuch luqui: i dunno 17:59
=P
autrijus and then unpack them away.
that is one approach
but I had not thought of it in detail
and also "*" is currently an App not a Syn
dada gosh 18:00
autrijus to make it a Syn, change Parser line 402
, preOps " = ! + - ~ ? * ** +^ ~^ ?^ \\ " -- Symbolic Unary
nothingmuch i doubt splat will change any
autrijus into something like
, preSyn " * ** " ++ preOps " = ! + ......"
nothingmuch oh, Syn vs. app is not synopsis vs. apocalypse =)
*blush*
what is an App?
autrijus and then add a line on 498 18:01
shapr I have an App in windows!
autrijus preSyn = ops $ makeOp1 Prefix "" Syn
that's all it takes.
nothingmuch shapr -= 0.01
shapr snickers
autrijus nothingmuch: App means "apply", it means calling a function.
shapr deserved that
autrijus Syn means "syntax", it means a special form.
I need to write them all in PA02 ;) 18:02
I wonder if I can convince Bestian to fly to OSCON with ingy. 18:03
shapr I'd like to meet ingy, I've heard fascinating things about that person. 18:04
ingy: do you have a website?
ingy shapr: zombo.com
shapr ingy: oh, you know Steve Howell? 18:05
stevan autrijus: question about Perl6/lib
shapr fooey, flash.
stevan is that gone away now?
autrijus sighs.
yeah.
stevan my @*INC is : /System/Library/Perl6/darwin-thread-multi-2level
autrijus and v6/ is also gone away.
stevan ok 18:06
autrijus they may come back in the future.
we don't know.
stevan so what should I tell people re: perl6 CPAN?
other than to submit to Pugs
and not CPAN
autrijus "wait for Pugs::MakeMaker / Module::Install::Pugs on CPAN ;)"
"once they are available individually on CPAN, you can start submit things to CPAN, because that means we've figured out the distribution problem" 18:07
"suggestions welcome etc"
stevan ok
autrijus (which shouldn't be long; YAPC::Taipei should also tackle that.)
stevan sounds good
autrijus is piling lotsa things on Y::T
dada uhm 18:09
pugs> ? *(1,2,3)
*** Evl: App "&prefix:*" (Syn ","
it's still an App :-( 18:10
autrijus right. what's what I mean by * is App, not Syn
have you adjusted the Parser?
dada autrijus: yeah
autrijus and removed * from Prim?
(it's no longer a primitive)
sorry I didn't mention that.
dada oops
remove the op1 or remove it from initSyms? 18:11
or both?
or neither? :-)
autrijus hehe.
both.
hi Schwern! 18:19
nothingmuch noticed the avg population of #perl6 is growing 18:20
dada AAAATGH
autrijus dada: sorry for the pain :-/ 18:21
dada autrijus: no no
I was just modifying the wrong files :-)
autrijus roflmao!
nothingmuch stevan->ping 18:22
Limbic_Region well if the H had been a C I would have thought dada was playing "guess the amino acid"
dada Limbic_Region: :-) 18:23
stevan autrijus: whats up
or rather nothingmuch: whats up 18:24
nothingmuch stevan: how sane would it be to subclass Tree::Simple objects?
stevan takes a big swig of coffee
nothingmuch: I do it all the time
nothingmuch i'm not very satisfied with setNodeValue et all
i'm not very satisfied with setNodeValue et al
eck, shell reflexes
stevan what do you want it to do?
nothingmuch colloquy should have an option to disable history =)
well, i need objects with types
and i need them to have weird content 18:25
and i don't want the tree nodes to delegate to another object,
stevan nothingmuch: I usually subclass it to do just that
nothingmuch or contain a hash
because that is annoying
stevan hmmm, so you want to store extra info right in Tree::Simple then?
obra sene autrijus
nothingmuch goody
Schwern autrijus: Hey
obra oh. 3 minutes. ok :)
nothingmuch sort of 18:26
perhaps stringification, too
obra autrijus-san: your isp hates you
nothingmuch i want to be able to feed them to petal templates easily
autrijus obra: how much?
obra [email@hidden.address] host eforward2.name-services.com[216.52.184.242] said:
550 5.7.1 Invalid recipient: [email@hidden.address] (in reply to RCPT TO
command)
stevan nothingmuch: I am sure you could add more fields to it, just watch out for clashes
autrijus what the... 18:27
obra "a lot"
autrijus I don't understand.
autrijus.org. 3601 IN MX 5 mail.autrijus.org.
autrijus.org. 3601 IN MX 10 eforward4.name-services.com.
nothingmuch uhuh, that is the usual perl5 oop mess.. i know =)
autrijus obviously it should try mail.autrijus.org first.
nothingmuch has ISP trouble too.... =(
nowadays it's relatively stable 18:28
but a month ago i was on the verge of crying for about two weeks streight
Limbic_Region autrijus - is there a A record for mail.autrijus.org or only a P record?
autrijus mail.autrijus.org. 3526 IN A 220.132.132.105
nothingmuch my friend in switzerland had pretty graphs for mailq sizes thrown off scale for years =)
autrijus I'm sure I got it correct :-/
stevan loves his cable modem
Limbic_Region tests
nothingmuch lucky for me i even had a backup mx 18:29
obra autrijus: but if your backup mx is losing, that's bad
autrijus yeah it should not be losing at all. will look to it tomorrow
Limbic_Region autrijus - you have SMTP services running on that box? 18:30
obra I wonder if it's worth taking eforward out of dns until it's fixed
Limbic_Region $ telnet 220.132.132.105 25
Connecting To 220.132.132.105...Could not open connection to the host, on port 25: Connect failed
obra better to have tempfail/queueing
autrijus it worksforme.
try again?
Limbic_Region ok
same thing - let me try from outside government firewall 18:31
autrijus don't tell me my ISP is blocking incoming 25.
ingy Schwern: did you submit an oscon proposal this year?
obra it works for me now.
autrijus shudders.
obra I wonder if there's a cable cut
ingy shapr: *you* know Steve Howell?
nothingmuch autrijus: that's just what happenned to me a month ago
yuck yuck yuck
Limbic_Region autrijus - working here 18:32
nothingmuch they said AOL was complaining
Limbic_Region you should probably disable vrfy btw
nothingmuch i asked why the blocked incoming and not outgoing
they said 'just cause'
autrijus how do I disable vrfy?
Limbic_Region I didn't test some other potential no-nos but if vrfy is enabled they are likely as well
Limbic_Region is a sendmail weenie not postfix
in fact, you probably shouldn't have postfix identify itself as such 18:33
I need to check one more thing to rule out it being a firewall issue on my side
shapr ingy: via TheWiki and #python, yes. Early into YAML I was considering writing a YAML parser to help out.
But I got distracted, as usual. 18:34
Limbic_Region: Want to try a postfix replacement written in Haskell? =)
dada yeah!
Limbic_Region autrijus - nothing wrong with incoming 25 on your end - must be a firewall issue on my end - go figure ;-) 18:35
dada pugs> ? *(1,2,3)
*** Evl: Syn "*"
shapr ingy: Sadly, I have all the focus of a gerbil. methylphenidate upgrades my focus to praire dog level, so I can now stay employed.
dada (and later...)
autrijus dada: woot!
dada *** Error: "Unknown syntactic construct"
autrijus dada: sure. that means you need to implement Syn "*"
preferably just above Syn ","
Limbic_Region shapr - not especially. Just wanted to let autrijus know of potential security risks
dada I'll look into it
Schwern ingy: Nope. Thought of something interesting to say too late. I'm writing up one for YAPC though. 18:36
autrijus dada: I don't think I've made you a committer, did I?
dada nope
autrijus dada: email to send invitation to? :)
shapr Limbic_Region: I suspected not, therefore the smiley at the end attempting to imply the non-serious suggestion part. I just had to wave my lambda pom-poms, sorry :-)
dada autrijus: dada at perl dot it
stevan autrijus: commiting the PA01 change now
autrijus stevan: yay! 18:37
Limbic_Region autrijus - it is in your main.cf file - www.postfix.org/postconf.5.html 18:38
dada it is possible that ghc doesn't like \t in source files? 18:39
ingy Schwern: I got one accepted
autrijus Limbic_Region: vrfy closed.
ingy so prepare a place for me 18:40
nothingmuch dada: it needs to be consistent
and consistency of haskell code in pugs is spaces
autrijus dada: what nothingmuch said. when in doubt, try expandtab :)
I've seen tabs in tests.
I don't quite care :)
but .hs files wants spaces.
feel free to add -- vim: magic lines
or emacs, etc
as you see fit, at the end of .hs files.
dada nothingmuch: "Be consistent. Be nice" vs. "Be consistent. Or be dead"
autrijus the latter. 18:41
haskell, I mean. not pugs :)
dada one more thing learned :-)
autrijus stevan++ # excellent answer to PA01!
nothingmuch haskell likes tabs too
but it's either or, really 18:42
nothingmuch did yaht stuff (ooh, so many) with tabs
autrijus viirya / hcchien: can you re-translate the "Is there a CPAN for Perl 6 modules?" chunk to zh-tw/PA01 ?
zh-tw/01Overview.html, even.
dada which is the current status of tests with the svn version? 18:43
I have several failures, but I don't know which ones are mine and which not :-)
nothingmuch dada - list the test files 18:45
perlbot nopaste
perlbot Paste your code here and #perl will be able to view it: sial.org/pbot/perl
nothingmuch only set #perl6 as the channel
autrijus see? that's the problem with making tests fail :)
(but yeah, one should be able to count delta in head)
nothingmuch well, otoh if they all pass, then you are ignoring problems =)
unless you explicitly search for them
pasteling "dada" at 193.203.230.22 pasted "failing pugs tests" (6 lines, 374B) at sial.org/pbot/8003 18:46
autrijus which I do anyway :)
nothingmuch dad- not yours
s/(?=-)/a/
theorbtwo nothingmuch-- # useless use of (?...) in RE. 18:47
nothingmuch uh, fine
dada nothingmuch: ok, thanks
nothingmuch s/(?:?=-)/a/
doesn't that help readabilty?
theorbtwo s/-/a/
Much better.
nothingmuch uh? 18:48
i'm finding a place, that has no width, followed by a -
and replacing it with 'a' to make 'dad-' into 'dada-'
i could have just used tab completion, but then i would not show off my mad regex skillz
theorbtwo s/-/a-/ 18:49
nothingmuch then no one would know i know zero width lookahead assertions =)
theorbtwo Yes, but everyone would know what the RE did without checking perlre.
nothingmuch beh, regexes followed by typos correct the typos, everyone knows that =) 18:50
theorbtwo *= nothingmuch
there, that's more than --
i've got to go put my sister to sleep
theorbtwo nothingmuch++ # karmic balance. 18:51
ingy nothingmuch: hopefully that's not like putting your dog to sleep 18:52
PerlJam nothingmuch: just remember that in perl6 your RE would be : s/<before: ->/a/, while theorbtwo's would still be s/-/a-/ 18:53
autrijus wow. today's journal is really long. 18:57
zzz. :) &
ingy autrijus:it seems that test failures in the pugs tests prevent the ext tests from running 19:00
theorbtwo
.oO(Woo, I'm a "helpful soul")
 19:03
autrijus yow, luqui's post on p6l kept me awake. 19:04
ingy: really. hrmph.
theorbtwo Hm, on second thought, I think I'll take that committer bit, if it's still on the table.
james at mastros dot biz. 19:05
autrijus sent. 19:06
welcome aboard!
theorbtwo Thanks! I'll try to not make too much of a pest of myself. 19:07
dada autrijus: is there something dangerous in the "??????" text I'm about to agree? :-)
ingy my oscon talk will be ... interesting 19:08
autrijus: I only have 45 minutes so far, so it might be hard to integrate Bestian
autrijus dada: wagner.elixus.org/~hcchien/termtouse.html is the en-us.
ingy of course I would be up to the challenge
autrijus ingy: I mostly think just to send him there to chat with people.
dada autrijus: thanks
autrijus and avoid the hefty entrance fee 19:09
ingy autrijus: good idea
autrijus dada: it's sf.net boilerplace.
boilerplate, even.
luqui++ # pattern matching subs makes sense!
I think my evil plan of inflicting haskell-sense on luqui is working.
obra hah
theorbtwo Well, I can read the TOS exactly enough to tell that it looks real pretty. 19:10
ingy how hard will it be to implement p6 rules?
autrijus ingy: given PGE milestone 1, not hard at all 19:11
trivial almost.
without PGE milestone 1, I think a week of concentration.
ingy is there an expected timeline for milestone 1? 19:12
obra ingy: I'm working on it 19:13
ingy: have you seen the milestones doc? comments would be appreciated
Coke_ hey, obra.
obra Coke_: did you get my mail?
ingy obra: url?
obra svn.perl.org/perl6/doc/trunk/plans/...stones.pod 19:14
I've been working on fleshing it out with design team members. more questions to have it answer would be useful
Coke_ I'll comment on the parrot stuff, and how that might realte to the other projects as well as how leo's most recent 0.2.0 roadmap might integrate.
obra "send patches"
Coke_ (and several nits)
ingy obra: what's pod?
obra Coke_: please do
ingy: :P
ingy :P
Coke_ btw, =head3 isn't supported by my pod reader. =-) 19:15
obra when I can run kwid with TPF perl6, I'll switch
ingy no more pod after yapc
obra Coke_: it's supported by pod2html, which is the use case here ;)
Coke_ actually, that's what I used.
obra huh. I know there are version skew issues
some pod2htmls deal and others don't 19:16
but really I'm much more concered with getting the raw information written down at this point
autrijus yup. 19:17
ingy obra: you just gave me an idea
obra Coke_: can you point me at leo's plan when you reply?
ingy actually this was thought of by autrijus long ago, but...
well not quite... 19:18
obra what is it?
ingy a kwiki plugin such that you can use svn/svk to check out and edit pages
and svn hooks to store the metadata
obra isn't that just kwiki::archive::svk? 19:19
but yes, I'd like that. a lot
ingy not quite
so would I
obra :)
Schwern ingy: What's your talk about?
ingy Schwern: Apocalypse Now -- Perl6 is Here Today!
I don't think my Kwiki tutorial got accepted, but I don't know yet 19:20
Schwern: basically I'm playing god\Wautrijus, since she won't be there 19:21
dada ouch
src/Eval.hs:368: Variable not in scope: `retval' 19:22
ingy broken code!?!
Coke_ leo's plan is spread over the mailing list in about 6 emails. =-)
ingy I like haskell because generally when it compiles it works
getting to compile is the hard part ;) 19:23
pasteling "dada" at 193.203.230.22 pasted "offending snippet from Eval.hs" (6 lines, 257B) at sial.org/pbot/8014
19:24 DrSockHunter is now known as DrHyde
obra Coke_: If you can tell me what to search for, synthesis is my hob 19:24
job
ingy autrijus: why was PERL5LIB removed from the path? 19:25
Coke_ obra: groups-beta.google.com/group/perl.p...52e2fbac32
Limbic_Region Coke_ - IIRC, some of Leo's spread out proposal extends back to when Dan was around and said N-O
autrijus ingy: because I was appending Perl6/lib/ to there
ingy: and you removed it
so it now sees Test.pm 19:26
which it can't parse
and bokred.
nothingmuch beh
wifi problems
autrijus if you want proper p5 interop, we can add it
later
but not like this, not now
ingy I see
makes total sense
but Darren's problem was different
autrijus: did you see that btw? 19:27
Schwern: in perl-5.8.1-RC1 when I set INST_LIB to ../../blib/lib MakeMaker kindly changed it to ../../../blib/lib 19:28
dada ohhhhh shame on me
I didn't see the difference between retval and retVal
autrijus ingy: what problem again? I thought it's the hardcoded ../../blib 19:29
which broke bad MM
and I fixed.
ingy let me review patch
autrijus ingy: I just used FindBin::Bin. 19:30
dada how do I get only the first elem of a list in Haskell?
autrijus head
clkao not car ?
clkao giggles
dada it compiles! it compiles!
I needed a $ before the head 19:31
ingy autrijus: that will work although it should use M::I's base
dada feels like playing with lego toys
ingy will fix
theorbtwo sometimes thinks haskell needs a good case of parens and commas. 19:32
autrijus ingy: I'll leave it to you :)
Khisanth but it has both!
autrijus theorbtwo: you can add parens liberally :)
commas is harder :)
dada still, my splat operator doesn't do what it should do :-( 19:34
autrijus as long as it doesn't break more things, feel free to just commit it in. 19:35
I'll take a look tomorrow.
really sleep &
ingy autrijus: done
obra Coke_: mailed to myslef for later raiding 19:36
autrijus I take it back; even if it breaks more things, as long as it doesn't break the whole test tree catastrophically, still commit it.
Limbic_Region wonders how autrijus is able to get any sleep 19:37
after having as mentally stimulating as he must have on a daily basis my mind won't shut off to sleep
theorbtwo L~R, try noting the time now, and when he next says something. 19:38
Limbic_Region [theorbtwo] - and? 19:40
dada autrijus: were you talking to me about committing?
Limbic_Region yes dada
theorbtwo Yeah, he was, dada.
dada ohhhh 19:41
theorbtwo Subtract. Ain't that long.
dada ok, I will commit then. it seems that I didn't broke anything
(yet :-)
Limbic_Region theorbtwo - any is 0 and the time autrijus is away is longer than that 19:42
ingy r536 20:00
nothingmuch hola lumimies 20:09
lumimies Morning
theorbtwo G'localtime, lumi.
lumimies Sorry I'm multiplying
I mean morning in the loosest possible sense 20:10
crysflame heh
dada dammit 20:16
I need a way to debug what's happening in pugs 20:17
does ghc have a debugger?
seems not 20:19
rgs there's ghci
dada can I run pugs through ghci?
wait, wait, doesn't ghc have all that +RTS stuff on the command line? 20:20
theorbtwo It does. You can run pugs with +RTS ... -RTS options. 20:23
But I'm not sure it's useful.
dada no, it is not 20:25
since my ghc is compiled without DEBUG support, I can't use -D 20:26
I wonder how does autrijus does debugging of pugs source
20:36 Coke_ is now known as Coke-afk
lumi That's so annoying 20:39
20:58 saorge` is now known as saorge
tmoertel ingy: any reason why getLibs in returning an IO action? 21:05
s/in/is/ 21:06
Steve_p Good afternoon stevan :) 21:13
stevan howdy Steve_p
ingy tmoertel: what would you suggest? 21:15
tmoertel ingy: if it's not dependent on IO, then "IO a" could become "a"
ingy tmoertel: I fought with it for 2 hours. Feel free to refactor 21:16
tmoertel otoh, if you are expecting that in a later rev, we'll need to perform IO like doesDirectoryExist on the paths, we ought to leave it in IO
ingy tmoertel: I was hoping to pair with you sometime
btw
tmoertel cool
ingy do you have a mac?
tmoertel ingy: I'm on i386-linux-thread-multi right now, but I have a mac or two in the office 21:17
mauke I've removed the IO part 21:24
tmoertel ingy: take a look at community.moertel.com/ss/space/tmoertel/Hackery
mauke it seems to compile
dada I go, goodnight goodfellas 21:43
joodie waves
ingy tmoertel: can you commit that or should I? 22:51
tmoertel ingy: I can send a darcs patch out
ingy how do darcs and svn get resolved? 22:52
tmoertel ingy: it looks like svn is the master repo and autrijus syncs from svn->darcs periodically
ingy: my guess is autrijus can go darcs->svn, too 22:53
ingy: when I commit darcs patches, the go to the perl6-compiler list
nothingmuch ingy - they don't
syncup is just svn pull 22:54
darcs recrod -l
that way conflicts are mad
e
ingy well go ahead and submit the patch 22:59
tmoertel ingy: maybe I'll just post a normall diff -u patch and you can apply it?
ingy I don't need the diff 23:00
sure I'll do it
tmoertel ingy: you'll need to make a few related changes, too 23:01
ingy crap. cut/paste from that site borks newlines
tmoertel ingy: see community.moertel.com/ss/space/tmoe...Libs.patch
ingy can you email the patch to ingy at ttul dot org
oh
ingy should learn to read... 23:02
tmoertel I commented out the "PERL5LIB" and "PERLLIB" line to match your more-recent version
ingy: at one point I saw a doesDirectoryExist test. do you still want to do that? 23:03
ingy tmoertel: applied as r539 23:07
tmoertel: no
because relative paths should work 23:08
perl5 doesn't do exists. neither should we
(read, it already bit me ;) 23:09
tmoertel that's sound logic, afaics :)
ingy laters&
nothingmuch good night 23:10
tmoertel 'night
safrican Hello 23:36
arcady hi 23:43
mugwump so, who has an opinion on the junction vs eigenstate thing ? 23:50
metaperl_ mugwump, can you expand with examples? 23:51
I know eigenstates were in Q::S
I am presenting tonight at our perl mongers meeting on Perl 6 Junctions
metaperl_ groks URL
mugwump ($a = any(1..9); $b = any(1..9); $c = any(0..9); if ( ($a + $b) == ($a * 10 + $c) ) { say "A + B = AC"; say "$a + $b = ".($a*10+$c) }
metaperl_ actually tomorrow: www.hcoop.net/~terry/perl/talks/p6-junctions/ 23:52
mugwump see also examples/sendmoremoney.p6 in pugs svn 23:53
the above is the simplest form of this problem I could find 23:54
obra mugwump: what are your .tw dates?
mugwump 23rd midnite - 31st midday
obra oh. so no hackathon? :/ 23:55
mugwump what date does that start / run to? 23:56
obra not sure. hoas
mugwump > For now, it would be on 3/28 - 3/31 (4 days) in Hualien.
- 簡信昌 23:57
blast :)
obra go for 2 days 23:58
or 3
it's so worth it
hialien is fucking gorgeous
mugwump yes, it seems a little early to leave doesn't it 23:59
metaperl_ is hualien where the wulai people live?

Powered by Cro and the Raku® Programming Language.

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

Thank you!