103 subtests fail | pugscode.org <<Overview Journal>> | pugs.kwiki.org | logged: xrl.us/e98m
Set by pjcj on 24 March 2005.
kungfuftr moo? 00:04
jabbot pugs - 1118 - Ported CGI::Lite 2.02 (complete with doc 00:07
kolibrie using the darcs repository, I'm getting ghc errors (which I don't see with the svn repository) 00:20
my eyes don't tell me what is wrong
/usr/bin/ghc --make -H200m -L. -Lsrc -Lsrc/pcre -I. -Isrc -Isrc/pcre -i. -isrc -isrc/pcre -static -Wall -fno-warn-missing-signatures -fno-warn-name-shadowing -o pugs src/Main.hs src/pcre/pcre.o src/UnicodeC.o
Chasing modules from: src/Main.hs
src/Main.hs:
Can't find module `Compile'
anyone have insights as to what that means? 00:21
Alias_ So regular expressions work yet?
kolibrie I'm pretty sure perl5 rx works 00:23
no perl6 rules yet
Alias_ wonders whether it's worth porting Config::Tiny to provide basic .ini-style config files 00:25
pjcj looks lilek iblech already did it 00:30
I'd guess the darcs repo didn't sync properly with the main svn repo. That's one for autrijus I suppose. 00:33
kolibrie when abouts is he active again? 00:34
pjcj three or four hours, I'd guess
Alias_ It's about 8:30am Taipai time? 00:35
kolibrie yep, that's what the Internet says 00:39
stevan Alias_: we have basic regexp matches with rx:perl5{} 01:08
and we have subst regexp s:perl5{}{} 01:09
and optional g fla
g
Alias_ Take a look at search.cpan.org/src/ADAMK/Config-Ti...ig/Tiny.pm
stevan s:perl5:g{}{}
Alias_ Tell me if you think current regex can support that
in the method read_string
stevan Alias_: iblech has already ported Config::Tiny 01:10
Alias_ he did?
Alias_ didn't quite get what pjcj meant earlier
stevan svn.perl.org/perl6/pugs/trunk/modul...nfig-Tiny/
Alias_ updating my svn now
stevan of course it does not work yet
because we dont have OO yet
Alias_ :)
Not to mention that Config::Tiny works a bit unusually... 01:11
in that it is used like both an object and a hash
stevan the only part of read_string that we could not support is the split with the regexp
kolibrie welcome back from dinner, stevan 01:12
stevan looks like iblech has implemented the hash part as a rw class attribute
kungfuftr you can do split with pure regex though
stevan kolibrie: thanks :), my belly is quite full 01:13
Alias_ wow, lots of new code since I last updated
stevan Alias_: yes, iblech is our current mad-porter
Alias_ oh dear... MIME::Lite?
stevan he is responsible for most of whats in modules now
Alias_ The horror!
:)
stevan Alias_: you want to see horror ... 01:14
svn.perl.org/perl6/pugs/trunk/ext/FileSpec/
Alias_ ah, he's doing Email::Simple at least
stevan a (mostly) working File::Spec
Alias_ so ext is "new Perl6-specific modules"?
and modules is "P5 modules ported to P6"\
stevan Alias_: ext/ is working perl6/pugs stuff (can be haskell too)
yes 01:15
what is in modules will eventually serve a pseudo integration tests for pugs
kungfuftr is anyone working on rules engine stuff? or is it being left until someone else with more of an idea works on it?
kolibrie stevan: I saw you were talking with autrijus about ghc 6.4 and inlining
stevan because they will be modules with test suites
kolibrie are there examples of that?
Alias_ stevan: Gotcha
stevan kolibrie: talking about 6.4,.. not inlining
kolibrie oh 01:16
stevan kungfuftr: right now it's PCRE, I have no idea about future plans
kolibrie: what are you looking for? I may be able to point you there
kungfuftr stevan: yar, i think only larry and damian actually understand the concept of the new rules system 01:17
=0)
kolibrie I'd really like to wrap up the cairo library
stevan kungfuftr: I think you are right
cairo is haskell?
kungfuftr thought cairo was svg engine
theorbtwo I think the idea is that perl6 rules will be implemented by integrating with the PGRE stuff that the p6 compiler pumpking was doing. 01:18
kolibrie no cairo is C
it's a low-level drawing engine
www.cairographics.org
stevan kolibrie: I think autrijus was using the haskell FFI to wrap PCRE (which is C)
IIRC he liked it (the Haskell FFI) 01:19
kolibrie I wanted to write a wrapper in perl5, but I got lost in XS
stevan kolibrie: at this point you will need to write a Haskell warpper
kolibrie there is a haskell port already, I believe 01:20
stevan kolibrie: then you are half way there :)
kolibrie might have to learn haskell?
has @larry written anything about binding libraries from other languages? 01:22
stevan nothing I have seen, I expect that is to be done in Parrot though
Alias_ parrot people are just starting on the structures for loading modules from multiple languages now (IIRC)
kolibrie ok 01:23
Alias_ multiply dimensional namespaces, or some such
[ language, lang-namespace, version ] or something
I only saw it in passing
kolibrie The haskell bindings are avilable via darcs: darcs get ofb.net/~abe/darcs/cairo 01:24
bbiab 01:31
01:51 metaperl is now known as _metaperl
stevan anyone know the Win32 equivalent of 'pwd'? 02:00
Alias_ 'cd' (no args)
?
stevan I am looking to do something equivalent to Cwd? 02:01
for File::Spec
the Win32 one
kolibrie been a long time since I used win32
stevan me too
ninereasons Alias_ has it. 02:03
Alias_ stevan: try just calling cd without any arguments
stevan so system("cd") will work
ninereasons ysp
Alias_ trim off trailing $stuff
stevan ok, I have no windows box to test, so I will have to wing it
kolibrie how scary - just seems like you'd go /home
Alias_ stevan: chomp($cwd = (`cd`)[0]); 02:05
WORKSFORME 02:06
stevan cool
Alias_ it isn't quoted though
just raw 02:07
quoted/escaped
kungfuftr stevan: `chdir` is the win32 command 02:08
C:\Documents and Settings\Scott>chdir 02:09
C:\Documents and Settings\Scott
=0)
ninereasons equivalent to cd
theorbtwo chdir and cd are synonymous.
kungfuftr ah
Alias_ chdir returns nothing for me
oh wait 02:10
(was in perl shell at the time)
ninereasons heh
it's very nice having the -I switch. I never use it with perl, but it's sure handy for playing with pugs 02:19
Alias_ I've only ever used it for screwing around with test scripts for other people's modules 02:21
ninereasons do other people have a problem building with Embedded Perl5? 02:25
kolibrie haven't tried it 02:27
pasteling "ninereasons" at 127.0.0.1 pasted "r1118: build fails when using Embed" (29 lines, 2K) at sial.org/pbot/8599 02:31
jabbot pugs - 1120 - more work on File::Spec::Win32; I need s 02:47
pugs - 1119 - re-adding the bug from a few days ago wh
Khisanth adding bugs? 02:49
stevan :) 02:53
readding bug report
its getting late
02:58 scw_ is now known as scw
autrijus greetings lambdacamels! 03:07
Alias_ hail phearless leader 03:08
Alias_ was pleased to see that Config::Tiny got ported without him even knowing
autrijus the mad porter scares me :)
Alias_ Config::Tiny is probably quite a nice little regexp test 03:09
autrijus aye
Alias_ at least for most of the basics
autrijus I want PGE :-/
Alias_ PGE?
autrijus the perl6 rules parser and compiler thing.
Alias_ ah
autrijus pmichaud was saying last week that he is devoting the week to that 03:10
and should show up with a working implementation... now or something
because of the way our ext/ works now, if PGE is there, it can be turned into Rule.pm in no time
and/or Rule.hs
theorbtwo Oh, good-morning, autrijus. 03:16
autrijus hey theorbtwo :) 03:17
ninereasons I see that there's still a lot of perl5 in many of the modules. to be expected at this stage, I'm sure. 03:26
autrijus yes, pleas hack away. 03:28
I'll bbiab :)
ingy hola 03:37
kungfuftr moo 03:45
jabbot pugs - 1121 - * $*EXECUTABLE_NAME and $*PROG_NAME 04:17
chip why not $*PROGRAM_NAME ? 04:23
or $*EXEC_NAME? 04:24
chip anticipates "ask Larry"
kungfuftr moo
autrijus it's PROGRAM_NAME.
I typo'ed in my commit log.
I'm using perlvar.
chip OK. not to complain, I thought everyone was zzz 04:25
autrijus nope, everyone just woke up :)
chip where are you in meatspace, anyway? 04:26
autrijus Taipei. YAPC::Taipei starts tomorrow, with me, obra, mugwump, acme, hcchien, clkao, ingy, miyagawa & go
s/go/co/
chip yowie
(teh intarweb)++
kungfuftr buggery... i'm going to be doing svn updates every 2 mins tomorrow 04:27
Alias_ autrijus: If you happen to encounter any bitching about lack of decent Perl jobs, you might casually mention for me that Sydney is currently in a decent Perl coder drought.
chip just don't get carried away and declare independence from the P6 cabal
Alias_ autrijus: There were 7 senior "real" Perl jobs posted to the main jobs site here in March so far... by far a record 04:28
kungfuftr Alias_: could do with some in oxford too
Alias_ So if anyone wants to move to Sydney for a while, it's a good time for it. They are mostly 3-6 month contracts
autrijus chip: er. but pugs was never part of cabal... 04:29
chip: although we maintain the "one perl6" policy firmly
chip alerts Larry to trigger the invasion
obra pondders the perl strait
chip calls to cancel
autrijus it's one perl6, multiple teams
please be vague on this soverignty issue and don't pass anti-separate-compilation laws 04:30
otherwise we'll be forced to buy more weapon from the haskell camp!
autrijus grins
kungfuftr autrijus: you now need inline BF! 04:31
Alias_ oh dear lord know... Human waves of first year university comp sci lecturers!
s/know/no/
theorbtwo is glad things aren't as bad as CNN was making them out to be a week or so ago.
chip increases the Parrot budget 04:32
theorbtwo
.oO(Parrot has a budget?)
04:33 mandarin is now known as Qiang
kungfuftr i thought parrot just had a budgie... and had a hidious offspring called pugs 04:34
kungfuftr should really shutup when he's ill... make no sense at all
Odin-LAP Hm. Didn't the ROC government drop its claims to mainland China at some point?
autrijus yeah.
not the reverse, sadly 04:35
Odin-LAP Too bad, really. The duality was amusing. :D
Odin-LAP doubts the PRC is ever going to give up.
Too extensive bureaucracy. They'll never get all the forms filled out. 04:36
Alias_ China has always reformed as a single country. That is the way it has always been, for X thousand years... right?
mugwump sans all the invasions afaik :) 04:37
kungfuftr mugwump: hey sam mate
mugwump heya scott
kungfuftr mugwump: finally got there i see 04:38
theorbtwo China has always been one conuntry! Even if everybody else thought Tibet was a different country until 1907 or so. 04:39
Er, 1950.
Odin-LAP Heh.
The ROC also claimed Tibet, AFAIK. 04:40
theorbtwo Reading the wikipedia article, there's hardly a clean history.
Odin-LAP For China?
kungfuftr claims ireland in his own name... since GB keep fucking it up
Odin-LAP claims Russia, just for fun. 04:41
theorbtwo In 1907, Brittian, China, and Russia recognized Tibet as being part of China... but the Chinese didn't actually control it until 1959.
Odin-LAP theorbtwo: Yes. Note that the Tibetans didn't recognise that themselves. ;) 04:42
theorbtwo Yes, well, my original specification was rather open -- "everybody else".
Odin-LAP True. 04:43
But of course, the British are notorious for recognising things that are quite ... odd. Like granting Palestine to zionists in 1917. ;) 04:44
(A promise they broke, admittedly.)
theorbtwo Well, the British had other things on their mind at the time. 04:45
kungfuftr still has to find a way he can revoke his british citizenship
theorbtwo Why, kungfuftr? 04:46
Odin-LAP kungfuftr: Eh ... make sure that doesn't leave you with no citizenship. 04:47
kungfuftr theorbtwo: dual citizen... politics... much rather be irish
theorbtwo Ah, didn't relalize you were irish/british dual.
Odin-LAP Hm. Irish/British law prevent you from being both, then?
theorbtwo Since both are part of the EU, I'd imagine there's little benifit. 04:48
Qiang british is part of EU now ?
theorbtwo And always has been.
It's not part of the european monetary union, but is part of the EU. 04:49
Qiang heh. i thought they try to stay back and watch
oh. that's what i meant
theorbtwo For that matter, they every nearly became part of the EMU.
kungfuftr Odin-LAP: Anglo-Irish agreement says that by law i am both an irish and british citizen. as far as i'm aware it's the only exception to the rule
Odin-LAP kungfuftr: Huh. I see. 04:50
theorbtwo The british lobbied hard for a country-specific-use section on Euro paper currency, and got it -- then decided not to become a part of the EMU. 04:51
So should they ever do so, there's a nice spot on the bills for a picture of the Queen.
(Or King, should they take that long.)
theorbtwo sighs... hey, I liked my #perl6 on-topic! 04:52
We should take this to #emacs.
autrijus :D
it's all chip's fault.
chip I heard that
theorbtwo ponders -- take on a project, or go to bed. 04:53
autrijus both!
go to bed and hack there.
theorbtwo It's 6 AM, and I've been up since 1 yesterday; if I got to bed, I'm going to sleep. 04:54
autrijus sleep then.
theorbtwo (Yes, that's a somewhat short day; I'm like that.) 04:55
Odin-LAP theorbtwo: Or even #ideologies!
Qiang your brain shell won't rest on perl6 ;)
kungfuftr get back to work minions!
Alias_ Sleep is for the genetically inferior, NOW KEEP CODING!!! 04:56
theorbtwo G'night, people with names that need unicode to write correctly, and people that can't stop being british. 04:57
Oh, and Alias_.
kungfuftr caffeine glands... only $5... implanted for free
nn
Odin-LAP kungfuftr: I'd prefer amphetamine.
kungfuftr: Fewer side-effects. 04:58
Alias_ passes Odin the dex
kungfuftr hhhmmm... fresh out i'm afraid... mugwump bought the lot
mugwump sniffs loudly
kungfuftr mugwump: so how are we my crack-fuelled perl monkey friend? 04:59
Odin-LAP Alias_: I think I have a bit of methylphenidate left, actually.
Alias_ I'll stick to what's legal for me, thanks
mugwump mmm, good ... not even suffering any withdrawal !
Depends which law you look at, Alias
Alias_ $local 05:00
mugwump Given that the Magna Carta guarantees that the crown will never restrict traditional ways of life, I'd say that pretty much makes all plant ethenogens legal all over the empire
finding a lawyer with the balls to push that in court is another matter :) 05:01
Odin-LAP mugwump: Most of the Magna Carta is deprecated. :p
Alias_ except all new ones discovered since the MAgna Carta
mugwump & # travelling to clkao's place
Odin-LAP Eh. According to that, you could ban cocaine, but not chewing coca leaves. 05:02
Which, according to what I've read, would be quite reasonable. ;)
autrijus stevan: asleep? 05:12
stevan: your bug has been fixed 05:59
jabbot pugs - 1122 - * s:perl5:g with captureing vars now wor 06:07
gaal theorbtwo, are you awake? 07:38
jabbot pugs - 1123 - add diag() info from tests 07:57
flw whoami 08:07
whois perl6
lightstep try / first
flw thank lightstep 08:08
kungfuftr 114/2907 subtests fail | pugscode.org <<Overview Journal>> | pugs.kwiki.org | logged: xrl.us/e98m 08:16
lightstep is there a reason to use fromMVal of readMVal instead of fromVal? 08:24
jabbot pugs - 1124 - fix number of tests in t/builtins/grep.t 08:27
lightstep and should @list.join($str) work, or should it be $str.join(@list) ? 08:31
Corion The topic is wrong - on Win2k, I get 110/2913 subtests failed as of 1124 08:32
kungfuftr lightstep: i would assume it would be a method on the list 08:33
Corion: running 1124 now
lightstep kungfuftr, i really wanted to ask if @list.join($str) should do join($str, @list) 08:34
Corion BTW, what is the customary way to have in Haskell what I'd implment in Perl via local variables? Is there anything like a hash that I can pass to the recursive functions and set entries there?
lightstep since the test assumes this
Corion, you can use records
Corion lightstep: I would assume this too, yes (but I don't know anything of the formalities of Perl6)
kungfuftr lightstep: unless it's been stated somewhere, autrijus assumes perl5 behaviour
Corion lightstep: Thanks - will search for that. Searching for "map" or "dictionary" with "Haskell" didn't produce too many useful results 08:35
Darren_Duncan I just updated and tested Pugs again -- 104 test failures now, which is 10 fewer than the IRC header says 08:36
lightstep Corion, but you can use `where' or `let' which lets you access the variables from the current scope
kungfuftr Darren_Duncan: shush
Darren_Duncan is it supposed to be a secret?
kungfuftr Darren_Duncan: no, i just have a slower build machine
=0)
Corion lightstep: Hmmm - "current scope" - that's a nice idea - I could make all these recursive invocations local to a surrounding function and set the new values there (instead of passing them around all the time) 08:37
Darren_Duncan: Weird - I got 110/2913 subtests failed (win2k), so we have platform inconsistencies. Yay.
Darren_Duncan my test is based on revision 1123 08:38
lightstep how do i print an array in perl5?
gaal print @array
Corion Or print "@array", if you want the items separated with spaces 08:39
Darren_Duncan my 104 failed is on Mac OS X 10.3.8
kungfuftr i get 120/2913 here
gaal print join ", " @array
Corion (or set $" to separate with something other than spaces)
kungfuftr: Bah. I gotta go - I'll nopaste my test output
Darren_Duncan out of 2907 here
Corion nopaste?
nopaste
gaal Hey win32 folks, please say what variant you're using - mingw, straight windows...
Corion Grr - is the infobot not here?
gaal perlbot nopaste
perlbot Paste your code here and #<channel> will be able to view it: sial.org/pbot/<channel>
kungfuftr Corion: paste.husk.org 08:40
Corion Aah - it's not automatic...
Darren_Duncan the build/test took awhile since I'm currently without my full RAM compliment, so don't plan to try again today ... i should bump the RAm back up in a couple days though
kungfuftr i'm running on freebsd 4.11 here
Darren_Duncan anyway, in case anyone's wondering about my SQL::Routine port progress ... 08:41
I haven't started it yet ...
but expect to have the whole thing done over 3-4 hours tomorrow
gaal summons nothingmuch
pasteling "Corion" at 217.86.56.202 pasted "nmake test output on Win2k (Rev. 1124; 110/2913 subtests failed)" (37 lines, 2.4K) at sial.org/pbot/8606
kungfuftr pugscode.org <<Overview Journal>> | pugs.kwiki.org | logged: xrl.us/e98m | FBSD_4.11:R1124(120/2913)
pugscode.org <<Overview Journal>> | pugs.kwiki.org | logged: xrl.us/e98m | FBSD_4.11:R1124(120/2913) 08:42
Corion Anyway - gotta run again - the extended command line handling will have to wait some more, but it really needs an overhaul. 08:42
Darren_Duncan that's about 250K of code
gaal Corion, when we get a chance we need to talk about the command line anyway
there were some ideas
autrijus I just realized Perl 5 took almost 4 years.
gaal last night i forgot.
kungfuftr pugscode.org <<Overview Journal>> | pugs.kwiki.org | logged: xrl.us/e98m | FBSD_4.11:R1124(120/2913) win2k:R1124(110/2913) 08:43
gaal s/5/6/; s/years/months/ 08:43
Darren_Duncan the port will be based on the yet unreleased 5v0.58
Corion gaal: I (no)pasted my prototype yesterday night - maybe you can look at it - otherwise, I'll be back in about 6-8 hours
gaal k
but the basic idea was to use perl to parse the pugs command line 08:44
Darren_Duncan don't know if other people have different thoughts, but I started using a simple system to differentiate perl 5 and 6 versions of things in a terse fashion, especially as they both come to be updated in parallel
simply put 5v or 6v in front of the version nums
so I say in my common changelog, for example, that SQL::Routine 6v0.0.1 is a match to 5v0.58 08:45
lightstep is there any casting going on before the primitive operators are called? or should each operator cast by itself?
Darren_Duncan or that Locale::KeyedText 6v0.0.3 aligns with 5v1.03
short and sweet
shorter than saying "perl 6 version of ..." or etc 08:46
Corion gaal: Ah, that would be a much more convenient solution, true!
(because I know Perl, but I don't know Haskell well enough)
... still that means I should try to commit my todo_is patch for -e :) 08:47
gaal tests are always welcome :) 08:48
Corion ... still no workee
gaal corion, i wrote a command line spec for perl5 with getopt::declare, which was the only *module* we could think of that could do it 08:49
lightstep i don't understand Prim.hs: how are primitive operators like `sort' and `join' that get a value and a list differentiate between op(@list), op($val,@list) and @list.op($val)?
s/are/should/
gaal it's not finished yet -- i got distracted -- and maybe it's not even possible to use that w/o some minor hacking
i should contact dconway about it i suppose
pasteling "Corion" at 217.86.56.202 pasted "todo test for multiple -e switches" (40 lines, 735B) at sial.org/pbot/8607 08:50
Corion gotta run now, see you later, all
Darren_Duncan I will also away 08:51
kungfuftr (IRC bots)-- # all made for interactions, not information
nothingmuch is summonned 08:53
i have a lesson soon though, so be swift 08:54
kungfuftr a lesson, at easter? 08:55
nothingmuch bass lesson
and we don't celebrate easter
kungfuftr ah
nothingmuch it's still holiday though
purim ;-)
gaal
gaal ah, hello. i was making coffee 08:56
nothingmuch so what did you summon me for?
gaal i wanted to chat about the server side of the smoke tests. 08:57
nothingmuch ah
today i should have time
we have delayed cleaning till tomorrow
kungfuftr ah... smoke tests... man with a plan
nothingmuch so i only have a bit of laundry
gaal i won't :( anyway the output of the report is richer now 08:58
enough to start a demo implementation of the aggregator
nothingmuch i'll shove in my hack to always print $?CALLER::POSITION in test output
goody 08:59
(nobody panic! it's only with a special env var)
and then we can really link stuff up
gaal i remember you mentioned you had a clear idea of the database side of things there, and i don't, so, well, moose.
nothingmuch ok
DB schemas are a bit of a hobby 09:00
so i don't mind ;-)
maybe we can slap a maypole frontend on the DB, just so people have access to the raw data
gaal hobby eh? that's what they call it these days
no no no
don't say that word :)
nothingmuch so just post sqlite db file?
gaal well, to2 has a nice htmlizer for a single report 09:01
nothingmuch is that stuff all checked in?
gaal i don't knwo what arch you want to do for the server side, so it's prolly
yes
nothingmuch ok, then i'll find it =)
gaal ...prolly betst to be as decoupled from the web server and be a daemon 09:02
nothingmuch yes, it'll be decoupled
content will be static for the most part, too
gaal whenever somebody POSTs a new report, the daemon aggregates it and generates a new agg report
nothingmuch exactly
gaal the apache -> smokerfilter bridge is simple enough 09:03
but i wasn't sure about the rules for aggregation. Unless we have some, this is Too Much Data
nothingmuch i'll find something that is an ok balance
and ofcourse 09:04
gaal certainly for the user, who wants to see a bottom line,
nothingmuch you guys are here for critique
stevan is probably good at this stuff
gaal and possibly for the smokefilter itself
(generating a single report for all of t/ brings the harness to about 10mb) 09:05
say we have 50 reporters, it becomes iffy - need to do things like collapse tests that everyboy agrees are OK
and just keep a set of failure messages with a reference to who saw those failures 09:06
is this talk ok on #perl6? should we move elsewhere? 09:07
kungfuftr just report failures and work out the platform inconsistencies at the serverside?
nothingmuch i think i'll try to whip up anomality detection
gaal every reporter reports their entire test, sure
nothingmuch if a test is failing everywhere, but suddenly passes somewhere
it's interesting
same vice versa
kungfuftr true 09:08
nothingmuch that way we can sort of graph the flow of pugs in terms of how tests change, across Rs and platforms
kungfuftr but you would get that from the fact it wasn't reported, etc
gaal the server side maintains a *set* of results, which should be normalized
nothingmuch anyway, teacher should be here any moment
so ciao
gaal rock the house, nm 09:09
i want to add some build info to pugs -V. does svn have something like cvs $Id$ I can use for embedding revision number? 09:17
jabbot pugs - 1125 - build fix for Eclipse
lightstep gaal, see src/pugs_version.h 09:19
gaal thanks
lightstep is 1=>2=>3 equal to (1=>2)=>3 or to 1=>(2=>3) ? 09:22
oh, nevermind, i found out it's right-associative 09:24
gaal um, is there introspection in Haskell? I have Config.hs that exports N values, and i want to query all of it. 09:36
lightstep gaal, yes and no. you'd better bundle them into a record or a procedure than use the broken introspection capabilities of haskell 09:37
gaal hmm. but i don't want to break the existing code that uses it. 09:38
lightstep you can have the inspector thingy on addition to the regular exports 09:39
gaal i could have a record built that copies them (this file is autogenerated anyway), but that's butt ugly
hmm, looks like only Main and Run use these values anyway. maybe i could refactor them to use a record? 09:41
are autrijus and co at the conf already? 09:42
lightstep why do you need it for?
hcchien gaal: no, it is tomorrow 09:43
gaal config info should be exposed to the perl6 user; and also to pugs -V
hcchien and I guess we will announce the broadcast about tomorrow morning (GMT+8?) 09:44
gaal wishes his old teleporter weren't broken
lightstep what's wong with the manual code in Run.hs and Main.hs? 09:46
gaal so, lightstep, what does work in haskell introspection? i need something simple, really. just a way to access a var by its name, plus a list of names in a moudule
nothing
i just want to programmatically get all of the data in Config
so thinking in Perl terms, i'd have wished for there to be a %Config::Config 09:47
and not a $Config::config_somevar, $Config::config_some_other_var etc.
lightstep you really don't want to use haskell introspection. there are modules for it - several packages. it requires lots of annotations on both sides of the code. you cound just convert all those variables to tntries in a Data.Map 09:48
gaal that's what i was looking for (i don't know haskell yet - just read yaht :)
lightstep Data.Map.fromList [("key", value), ...] is the syntax 09:49
gaal okay, i'll take shot at this. 09:50
lightstep, there's a FiniteMap in Internals/Map, which seems to give an implementation even when it's missing in old haskells. should i use that rather than data.map? 09:58
lightstep yes 09:59
gaal k thanks
lightstep why are there both SymVal and SymExp ? 10:01
gaal lightstep, newbie layout question: 10:07
my list is long, so i wanted to break it out to pairs on separate lines
config = listToFM [ 10:08
("key", val)
]
this doesn't parse
lightstep you need commas between the pairs
(since liftToFM expects a list of pairs)
gaal er, true 10:09
but is having ] on a line of its own ok
?
lightstep i guess so
one common indentation is:
var = fromList
[a
,b
... 10:10
]
gaal hmm, okay.
thanks :)
autrijus hi Jouke :) 10:18
Jouke hey autrijus 10:19
yes, I'm really considering starting to get involved :) 10:20
autrijus yay!
read apocryphon, write tests, port modules, think of examples
if you have extra time, learn haskell :)
Jouke hehehe...well, I'll be concentrating on getting prerequisites for pVoice working...that probably gets me involved for quite some time since it uses a lot of modules 10:22
I think that's good enough for now
autrijus yup :) 10:23
_metaperl has Jouke and Juerd confused
Jouke we're both dutch and have names from the same region in NL
but Juerd is the absolute wizard, I'm nothing like that
gaal how do i, uh, deal with VStr wanting a VStr, and my map lookup returning a Maybe v? 10:24
autrijus fromMaybe?
gaal tried using lookupWithDefaultFM, but it doesn't seem to work
not the weay i expeceted anyway :)
autrijus I think you want fromMaybe :) 10:25
elmex yaho
gaal okay :)
it worked, but i don't understand what i was doing wrong when i tried lookupWithDefaultFM, which seems to be intended for the same sort of thing 10:29
Jouke Are there modules that do inline-haskell/c stuff already or are there examples of such things? 10:30
autrijus Jouke: there's only ext/SHA1/ for now
that's a very new development.
lightstep is there still a reason to keep VThunk? or should i remove it?
Jouke ah...I was looking in /modules
autrijus lightstep: eh, VThunk is used for short-circuiting || 10:31
lightstep: how would you like to remove it?
lightstep oh. i thought it was only used for ::=
autrijus err, := too 10:32
it's also essential for the correct currying implementation
and serializable continuation
so unless you thunk of something better
it's here to stay
gaal autrijus, i'm changing Config.hs to use a Map instead of many values. it is acceptable to add an accessor Config.getConfig :: Str -> Str (I' hope i'm getting the terminology right) 10:35
autrijus sure, go ahead
gaal sorry if this is trivial: is it okay for this accessor to assume a "" default on things, rather than Nothing, seeing as the data is being generated from a perl script which curently never puts undef anywhere anyway? 10:37
Jouke should I use darcs or svn to check out the code? 10:40
autrijus gaal: sure. 10:41
Jouke: svn.openfoundry.org is the main repo.
you are free to use darcs
but currently only svn accepts direct commits.
Jouke ok, schwern says on the wiki that svn is not in sync
autrijus that is incorrect. fix the wiki. 10:42
Jouke k
_metaperl I was looking for a free svn host. openfoundry did not appear to be soliciting new members
autrijus _metaperl: applying a new project on openfoundry is like, instantly
just go to your personal page
click "register project"
_metaperl personal page?
www.openfoundry.org is where I went
autrijus rt.openfoundry.org/Foundry/Home/ 10:43
you want rt.openfoundry.org.
www.openfoundry.org is beyond my control :)
18:43 < tsee_> autrijus: Is there a guide for module authors on how to port 10:44
their modules to P6? That might get you even more tests.
18:43 < autrijus> tsee_: darren has written a summary on p6c
18:43 < autrijus> but that's not a guide per se.
maybe we want modules/README
similar to t/README
_metaperl omg, the user agreement in Chinese :)
I always hit accept anyway, but I hope I am not offering my testicles up for sacrifice
Jouke :) 10:45
autrijus _metaperl: that's a faq. :) wagner.elixus.org/~hcchien/termtouse.html 10:47
_metaperl yes, a bit of a drawback is the project id cannot have dashes...
autrijus oh? I thought we fixed that.
hm, apparently not. weird policy. 10:48
_metaperl no, I just attempted to submit dbix-dbh
and it kicked back with that error
autrijus does dbix_dbh work?
_metaperl * ID must be 3-15 characters of lower-case letters or digits, beginning with a letter 10:49
I will try
no
autrijus well then. sorry for that policy :-/
dbixdbh then
_metaperl * ID must be 3-15 characters of lower-case letters or digits, beginning with a letter
ok
gaal that looks like a product
autrijus and it is 10:51
gaal lol
Jouke is afraid he'll have to learn Haskell to achieve his goals 10:57
autrijus haskell is easy and fun to learn
and will make you write much better perl5 code :)
some people likes haskell.org/tutorial/
some people like YAHT mentioned in PA01
Jouke I tend to have real problems learning new languages...I picked up Perl in '97 and haven't learned any new languages since 10:58
_metaperl THe Algorithms book is really pretty good, but I'm buying Simon Thompson's book
autrijus ah. you can start by hacking src/Prim.hs :)
Jouke has plenty of ideas for this weekend
_metaperl so you still need the other math operators added there? 10:59
sin, cos ,etc?
Jouke unless I'm very wrong, and unless I don't succeed learning a little bit of haskell, it won't be too hard to implement Win32::OLE using HaskellDirect
autrijus right. you can do that easily. 11:00
you can also try to see if you can do that on the parrot side.
doing it at either side is sufficient.
Jouke and it's one of the basic things I need, so I think I'll be concentrating on that first
on the parrot side? 11:01
autrijus see parrot/examples/sdl/
for a sdl binding
if you do the OLE binding there
(there may already be one)
then pugs can just use that binding.
the 3rd way is to use an embedded perl5 interpreter.
so you can just use Win32::OLE. 11:02
pugs has limited support for that.
nothingmuch well
autrijus but it's kind of not very clean.
Jouke I'd rather go for a parrot or haskell implementation
autrijus ok. then you can choose to learn a very very low level assembly that is PIR 11:03
/PMC
or a very very high level language that is Haskell :)
surprisingly, writing C-level bindings in both are very easy.
Haskell in particular already has Win32.DLL.* 11:04
if you install GHC 6.4.
and it has bindings to most of the APIs.
it helps that GHC is developed by Microsoft Research.
Jouke ok, ok, I'm convinced...I need to learn some Haskell first :)
autrijus ok... lambdafolk first, birdfolk later :)
Alias_ autrijus: Does that mean what I think it means? We can do Perl6.NET relatively easily? 11:05
Jouke sounds like it
autrijus Alias_: er, it's one thing to be able to call NET and back
Alias_ I'm assuming of course, that MS Research has some form of Haskell.NET linkup for GHC
autrijus it's another to compile to native NET
the first is totally easy.
the second is a bit trickier (you need to compile to F# perhaps)
but totally doable.
since F# is really just ML 11:06
and Haskell<->ML is not hard.
Jouke ML?
autrijus ML is this metalanguage thing.
it's the canonical eager functional language. 11:07
Jouke ok...too advanced talk for me right now
autrijus (while Haskell is the canonical lazy one)
rgs F# ?
the functional side of C# ?
jabbot pugs - 1126 - added the numeric functions which were e
pasteling "gaal" at 192.115.25.249 pasted "low-level question about high-level language" (12 lines, 461B) at sial.org/pbot/8610
autrijus rgs: F# is basically ML ported to .net 11:08
rgs: it's a remarkably complete port.
rgs wow
_metaperl research.microsoft.com/projects/ilx/fsharp.aspx
rgs cool languages which have full specs can be ported to cool environments.
nothingmuch microsoft seems to invest a lot of effort in FP
rgs Perl 5 can't :-(
_metaperl they say it's Ocaml ported to .net
yes, they fund research on Haskell 11:09
gaal can someone help me with the error nopasted above?
rgs a coworker confirms that F# is ocaml
autrijus right.
only not full ocaml
gaal: it worksforme on ghc 6.4
gaal: typo.
I think
rgs this coworker once wrote a perl 5 parser in ocaml. 11:10
autrijus I want that.
rgs not as good as PPI though.
autrijus doesn't matter :)
convince that coworker to release it?
that may give us Pugs.Parser.Perl5.
nothingmuch awwaiid is an ocaml user
gaal autrijus, what's the typo? i can't see it
autrijus gaal: autrijus.org/test.hs 11:11
$ runghc test.hs
["one: 1","two: 2"]
worksforme.
rgs autrijus: that's a kind of "lint" for perl sources. cvs.mandrakesoft.com/cgi-bin/cvsweb...ecker.src/ (undocumented of course) 11:12
autrijus danke!
license?
rgs GPL
gaal thanks.. weird. 11:13
autrijus can it be relaxed to perl? :)
(otherwise I can't include it in pugs.)
nothingmuch should so many p6 modules be in the core repo? 11:14
right now it's still a convenience
autrijus nothingmuch: no. we're working on it. :)
nothingmuch but 10 modules from now i think it'd be hard
autrijus yes.
nothingmuch freepan is starting to happen?
autrijus yes.
we got bandwidth and machine and a roadmap
rgs autrijus: there's a plan to turn it into a publicly releaseable shape.
nothingmuch yay!
autrijus need more hackings tomorrow
rgs: yay.
rgs it's (c) mandrakesoft, thus GPL
nothingmuch autrijus: any help will be gladly offerred 11:15
nothingmuch used to sysop, so that stuff is all good
autrijus nothingmuch: oh. cool. yay
will keep you in touch
nothingmuch will freepan be implemented in p6? ;-)
sure thing
Alias_ freepan?
Some sort of massive all-languages CPAN? 11:16
autrijus Alias_: yapc.kwiki.org/taipei/index.cgi?FreepanProposal
"Free Programming Archive Network"
Alias_ I see some dangerous things in that list 11:17
Release via svn tag is dangerous
pasteling "kcwu" at 61.70.142.187 pasted "let script/pugscc know $ENV{GHC}" (28 lines, 1K) at sial.org/pbot/8611
autrijus Alias_: uh the idea is to use tools. 11:18
not letting people using svn propset
Alias_ Sounds like a hack waiting to happen
autrijus sure.
kcwu: looks good, commit it 11:19
welcome aboard!
integral it sounds a bit harder to mirror too: harder than just rsyncing, and serving FTP/HTTP
Alias_ Part 2. Why do I have to use your repository? What about mine?
autrijus integral: svn repo is also http repo.
so that answers your question.
Alias_: you can just use yours and send pings.
the idea is that of blogspheres ;)
Alias_ My CVS repository doesn't do that...
autrijus as long as your svn is publicly reachable.
Alias_ You mean I have to use svn?
autrijus then you need a gateway.
foundry can mirror svn from cvs. 11:20
you'll just hand it an anon acccount.
and a cvs :ext:/:pserver: line.
Alias_ But svn is such an old repository, everyone has moved on to NGVC (Next-Generation Version Control)
autrijus and it figures out the rest.
integral I always liked the separation between PAUSE, CPAN and search.cpan.org, seemed quite useful
f0rth pugscc is perl5 powered?
autrijus SVK is that NGVC.
f0rth: yes.
integral: right, and we intend to decentralize PAUSE
Alias_ eep
integral hrm, but centralisation seems to be implied by a normal mirror system where each mirror talks to just one upstream 11:21
Alias_ No way in hell you want a decentralised module source
autrijus I thought bloglines works pretty well. 11:22
Alias_ bloglines?
autrijus bloglines.com
kolibrie I'm getting this error on 'make':
src/Prim.hs:76:
Could not deduce (Floating a) from the context (Num a)
arising from use of `atan' at src/Prim.hs:76
Alias_ autrijus: That you want to design any technical system based on the structure of blogging scares me the most :)
kolibrie src/Prim.hs:76: 11:23
Could not deduce (Floating a) from the context (Num a)
arising from use of `atan' at src/Prim.hs:76
autrijus Alias_: pugs uses the structure of wiki :)
for committers.
kolibrie I'm getting this error on 'make':
autrijus it seems to have worked.
kolibrie src/Prim.hs:76:
Could not deduce (Floating a) from the context (Num a)
arising from use of `atan' at src/Prim.hs:76
src/Prim.hs:76:
Could not deduce (Floating a) from the context (Num a)
arising from use of `atan' at src/Prim.hs:76
autrijus er.
Alias_ autrijus: pugs isn't important enough yet :)
autrijus Alias_: neither is freepan :)
I don't see a "atan".
autrijus syncs
Alias_ Sure, but you can easily tighten up the commit bits later
Don't sommit to an architecture you can't control 11:24
err commit
autrijus talk to ingy not me :)
Alias_ The current structure of CPAN is ok, but the database and PAUSE itself are a mess
_metaperl oh tath is my fault
autrijus kolibrie: fixing
_metaperl kolibrie, that is my fault
kolibrie autrijus: thanks - and sorry I pasted that so many times, my buffer was stuck 11:25
autrijus _metaperl: oh ok. will you fix it?
_metaperl perlfunc has atan2
autrijus _metaperl: basically don't use op2Numeric for Floating
or you'd like me to do ti?
it
_metaperl i used op1Numeric
autrijus right. same
I fixed it. 11:26
kolibrie autrijus: thanks for fixing the darcs repo, too. Weird that "-r ." wasn't working
autrijus _metaperl: (**) is not op1. 11:29
anyway, committing
please fix ** yourself :)
_metaperl oh
autrijus remember to make before commit :)
_metaperl ok I will see what it is 11:30
ok
the bad ting about Haskell is I cannot grep for "sub op1Numeric" to find the definition of the function
autrijus ^op1Numeric 11:31
is much better.
_metaperl ah yes
now I have merge conflicts 11:33
I want to just get the current repository version.it looks like you fixed it 11:34
gaal useful Haskell resource: www.zvon.org/other/haskell/Outputgl...index.html 11:35
_metaperl paste? 11:36
gaal perlbot nopaste
perlbot Paste your code here and #<channel> will be able to view it: sial.org/pbot/<channel>
_metaperl that thing isn't working for me
i'm using a different one
autrijus &
gaal works fine here, _metaperl. are you setting the channel? 11:37
_metaperl yes, maybe firefox is having issues
anyway, my Prim.hs is borked
pastebin.com/262508
jabbot pugs - 1130 - * fix building for math funcs
pugs - 1129 - Config is now a FiniteMap; elaborate -V
pugs - 1128 - hopefully fixed atan issue
pugs - 1127 - let script/pugscc know $ENV{GHC}
_metaperl i'm removing Prim.hs and doing svn update 11:38
hallelujah.
back to normal
autrijus tsee: it's fixed a few mins ago 11:45
svn up?
gaal if i'm in the middle of a list definition in haskell: [ "1", "2", "3", >>I am here<<, "6", 7" ] -- what functional expression can replace '>> I am here<<' with several elements? 11:57
malaire do you mean a sublist or what? 11:58
gaal i don't want [ 1, 2, 3, [ 4, 5 ], 6, 7 ] 11:59
i want [ 1, 2, 3,4 , 5, 6, 7]
nothingmuch gaal: [ "1", "2", "3" ] : (foo) : [ "6", "7" ]? 12:00
malaire probably ++ not :
[1,2,3] ++ [4,5] ++ [6,7,8]
gaal instead of [4,5] i have someFunc 12:01
nothingmuch err, right
malaire++
no, it is also broken
in my example you'd have to concat all that
malaire [1,2,3] ++ someFunc ++ [6,7,8]
[1,2,3] ++ (someFunc with args) ++ [6,7,8]
gaal [1,2,3] ++ (someFunc params) ++ [6,7] ok?
nothingmuch yup
gaal thanks, malaire++ :)
if i weren't allowed to throw in the ] and [, i'd have had no other way of doing it? 12:02
nothingmuch well, you could splitAt the initial list
twice
and then concat it
but that's probably not what you mean 12:03
gaal right
nothingmuch syntatically it doesn't make sense in the same way
because every single thing inside a [] is a single element
and every function returns a single element
which could be a list
gaal i guess i was looking for @{ }
nothingmuch i don't there is such a concept in haskell 12:04
gaal k
nothingmuch you could foldr with a weird unwrapping ++ sort of thing
to make [1, 2, 3, [4, 5], 6, 7] into [1 .. 7]
gaal nah nah 12:05
i was just cleaning up my first haskell code
should we expose the Config data as %?CONFIG ? (why not? but it's not specced) 12:11
nothingmuch larry has some weird ideas about that, i think 12:12
gaal interesting, yes: groups-beta.google.com/group/perl.p...amp;rnum=2 12:17
jabbot pugs - 1131 - Ported URI::Find 0.15 and URI::Find::Sch
Jouke for one who hasn't used subversion before and just installed svk... how do I checkout pugs? 12:26
gaal pugs.kwiki.org/?FrequentlyAskedQuestions svn co svn.openfoundry.org/pugs 12:27
gaal isn't using svk though
Jouke checking out now :) 12:29
thx
step 2: installing GHC :) 12:34
gaal :) get version 6.4
Jouke yup....was downloading that one 12:37
jabbot pugs - 1132 - cleanup
dada hi Jouke! 12:45
jabbot pugs - 1134 - Ported URI::Find::Simple 0.7 (complete w 12:47
pugs - 1133 - do we like point-free?
Jouke hey dada 12:48
have you ported Win32::API already dada? ;-)
dada Jouke: errm...
Jouke if you won't, I probably will after Win32::OLE 12:49
dada oh, but Win32::API will probably die
GHC has its own FFI library, and parrot too
Jouke right, but that needs some Inline Haskell still I guess 12:50
dada yes, but you can pack that Inline as a module, rather than rewriting Win32::API
actually, I don't think pugs expose such functionalities 12:51
are you really trying to implement Win32::OLE in pugs? 12:52
Jouke I'll try to mimick the behaviour, probably mostly using Inline stuff to Haskell
by using the same API, porting stuff will be much easier
dada but 12:53
I wouldn't spend *too* much time on it 12:54
parrot has NCI, which will probably be _the_ way to write XS-like stuff in Perl6
Jouke will look at that
damn...I first chose to get all revisions, which took too long, so I pressed ctrl-c, and now I can't check out anything anymore 12:55
12:57 cnhackTNT|away is now known as cnhackTNT
Jouke ahh...fixed 12:57
nothingmuch oh my 12:58
nothingmuch is really ashamed of himself 12:59
but we have little sup links to tests with regex skips
theorbtwo Mornin. 13:02
nothingmuch hola theorbtwo 13:06
html question:
i have a <pre>
Jouke mmm....win32 ghc doesn't have a ghc executable ... and thus pugs won't build... any ideas?
nothingmuch which contains a table
in the syns
and i'm putting little <sup> links in there 13:07
how would i cause it not to break the formatting too much?
theorbtwo pre can't contain anything but text, by definition. You'd have to end the pre, put in the sup, and close the pre.
nothingmuch it does work
jabbot pugs - 1135 - Continued porting MIME::Lite, now pendin
theorbtwo (tt, OTOH, can contain anything...)
nothingmuch i don't care about validity yet
hmm
so basically we want to have better syn htmls 13:08
Jouke ah, never mind
nothingmuch theorbtwo: is your SEE ok?
i would like you to look over the sup insertion code
theorbtwo No, my emulator is completely messed up, and seemacs isn't anywhere close to usable yet. 13:09
nothingmuch either way, generating backlinked HTML for now 13:10
bummer
as a preview
well, i'll do my best to clean it up, and then check in
theorbtwo Great!
Hm, can you upload some HTML somewhere for me to take a look at?
nothingmuch yes, that's what i meant
theorbtwo Oh, also, generating <a name>s in the test HTML would possibly be nice, so I could make the test runner link to them. 13:11
nothingmuch test runner?
theorbtwo Er, <a name>s from the test numbers or names.
nothingmuch the top part of catalog_tests?
i'll do that
theorbtwo util/testgraph.pl 13:12
One thing at a time, though!
nothingmuch ah
we'll have to rethink the 0+$link id generation thing though
theorbtwo Not really; no OS I know of uses page 0 for user links, so the two won't collide.
nothingmuch that's not what i mean 13:13
Jouke does this error from nmake.exe make any sense to anyone:
makefile(413) : fatal error U1087: cannot have : and :: dependents for same target 13:14
Stop.
nothingmuch there's no way to determine where to link from testgraph.pl without knowing what the link will be
or looking
so we could use sequences
or something else
we'll see
dada Jouke: which nmake?
Jouke 1.5
theorbtwo That's what I said -- test numbers or test names.
Jouke perl Makefile.PL detected I hadn't installed nmake yet, so it fetched it for me, then 'nmake' spit this out 13:15
nothingmuch i think testgraph should link to the test file
ah
that doens't map
run <-> code is not necessarily consistent
anyway, later
halting problem and all that ;-)
nothingmuch.woobling.org/pugs_test_status/
search for 't:
S02 has quite a few
gaal nm, to2: question about the harness. 13:16
i want to capture stderr
so that if a test blows up hard, the info is at least stored somewhere.
(we already report diag())
but unfortunately this isn't captured by T:H:Straps 13:17
it isn'
t hard to hack up,
but has to be inside a long function. so my question...
patch Test::Harness locally? talk the Test::Harness people to apply my patch? 13:18
theorbtwo Hmm, both? 13:19
gaal what does it mean that T:H is in inc and not in ext/ ?
theorbtwo ext is perl6 modules that we consider "core"; inc is perl5 modules needed to install pugs.
(modules is perl6 modules that we don't consider "core", but don't yet have a better place to put.) 13:20
nothingmuch theorbtwo: how would you express a range of content, in HTML::Element?
gaal okay, q#2 then. i said "not hard to hack up", but i meant that i either depend on Test::Output for it, or slap on the part of Test::Output (T:O:Tie, really) that i need. Test::Harness being such an important module, i expect they wouldn't want to add a dep, but this is feature creep 13:21
theorbtwo nm, I could have sworn that that wouldn't work, but it looks like it does. As to breaking layout, I guess just rely on the person writing the link...
gaal why do we need a p5 test::harness, actually?
theorbtwo It's not really feature creep...
gaal by the time we run the tests, we have p6 13:22
theorbtwo Is pugs up to porting it?
Jouke ah, don't try to install it in a directory containing spaces
nothingmuch i could also s/\ {$num_of_chars_in_sub}\ //;
out of the next bit of text
but i don't want to do that yet
nothingmuch wants the synopses to have proper <divs> 13:23
gaal how ready are the file io and redirection bits? i don't know t:h intimately - yet - but that's basically the hard stuff.
hmm, there's also platform stuff like Win32::GetShortPathName 13:24
13:24 cnhackTNT is now known as cnhackTNT|away
gaal so maybe pugs isn't up to it quite yet. 13:24
theorbtwo Oh, pre doesn't have quite the semantics I thought it did. 13:25
gaal that is, pugs is, but there are deps
scw Wow, my pugs core dumped
theorbtwo scw: OS, arch, GHC version?
(We no longer support 6.2.2.) 13:26
scw theorbtwo: Linux 2.6, i386, GHC 6.4
theorbtwo Hm, odd.
scw theorbtwo: Using shell of pugs and type '? say 12'
theorbtwo Segfaults for me too. 13:28
scw Err, what's up.. 13:29
theorbtwo Odd, ./pugs -e '? say 12' doesn't give a syntax error.
scw no, '?' cast the result to boolean and say returns true? 13:30
theorbtwo D'oh!
13:31 Lynx_awy is now known as Lynx_
scw But ? at the first charactor of the shell command should mean "evaluate the perl 6 expression and give me the debug info!" 13:33
theorbtwo I know, scw.
Jouke is unable to build pugs with perl5 embedded :( no more time now, gotta work for $$$ 13:35
later people! 13:36
theorbtwo Eeep: Can't use an undefined value as a HASH reference at util/yaml_harness.pl line 88, <FILE> line 2. 13:49
$self->latest_event->{diag} .= $line if $Test::Harness::Verbose; 13:50
gaal what test?
theorbtwo t/examples/examples.t
gaal i was afraid that might happen :)
but couldn't find sometihng that did it. apparently the test is dying early 13:51
hold on, i'll fix it
theorbtwo Oh, looks like I need to make testgraph do something sane with non-test events. 13:52
gaal yes, the fix i'm doing now is exactly for when that happens.
theorbtwo Ah, yeah, the first line after the plan is "# Running under linux". Are you going to put that on the top level, next to the filename of the t file? 13:53
gaal yes, as diag elem in a node that isn't a test 13:54
theorbtwo Oh, so not what I was thinking.
gaal hmm, why does this still not break on my machine?
theorbtwo Hmm... yeah, I think I see how to do it.
gaal ah, of course, i wasn't -v'ing. silly me. 13:55
wait, you prefer it one node up the hierarchy?
i can do that too.
theorbtwo No, actually, the way you were going to do it is better. 13:56
gaal easier for me too :) one sec
jabbot pugs - 1136 - Ported Email::Envelope 0.01 (complete wi 13:57
gaal r1137 fixes it. 13:58
stevan wow,.. iblech is going crazy!!!! 14:07
morning all 14:08
gaal hey stevan!
stevan hey gaal
I need to catch up to you guys on the yaml thing
gaal there's plenty more metadata now, and i'm wokring on catching stderr from dying scripts 14:09
stevan autrijus++ # for adding $1 support to s:perl5
gaal theorbtwo is ever making the html better
stevan gaal: maybe we should fix those scripts to either just fail() or use eval
gaal++
theorbtwo++
gaal nothingmuch said he'll work on the db+server side of this, but wanted to consult you for aggregation 14:10
nothingmuch display, mostly
stevan yes I saw that
nothingmuch: what did you have in mind?
nothingmuch what I think would be cool: 14:11
we have synopses, with links to tests 14:12
we have tests, with links to synopses
and with notes at the end of the line
for pass, or fail
with the number of the case
in color
like: todo_ok(blah blahb blah) 1, 4, 5
the numbers are the cases, and they have the color codes
then i'd like the graph to link to the line
stevan so 1 is fail, 4 is pass and 5 is todos? 14:13
nothingmuch as for smoked display
stevan like that?
nothingmuch i don't understand
lets say the TAP outputed
ok 1 (file.t at blah, column x)
(i have a patch to Test.pm that does that if an env var is set)
it will place a green 1 after line blah 14:14
gaal stevan, re:fail/eval for dying tests: the problem is hard parsefails. maybe we shouldn't have any, but if they do happen, of course we'd like to know the error
nothingmuch in the htmlized test
stevan gaal: we can probably just use fail() and comment out the hardparsefail
nothingmuch: I understand now 14:15
(I think)
gaal the author of the test could.. but what if it's not a parsefail for him, or wasn't when the test was written?
nothingmuch gaal: then someone will commit a fix, i think 14:16
stevan gaal: i think if it parsefails for me, it is very likely to parsefail for you
gaal but what if it passes for you in r1000, but somebody b0rked it in r2000?
stevan this brings me to my other idea 14:17
:)
nothingmuch then r2000+x should fix it
gaal :)
stevan I made up that TODO list the other day foro autrijus
i am sure it is out of date again
with all our test parsing
it would be nice to be able to create that TODO list automagically
jabbot pugs - 1137 - handle diag messages occurring before an
nothingmuch really wants $?SUB in p56
stevan if only just to provide a short list of failing testing and files
nothingmuch s/6$//
gaal grep -r todo_ t ? :) 14:18
stevan gaal: we un-TODOed them
they fail now
TODO is only for unimplemented features
gaal so i'm not sure i understand
you want to retodo them because they were intenionally broken e.g. for a reimplementation? 14:19
stevan no
all the failing tests serve as a TODO list
of things which should work, but dont 14:20
gaal so far so good :)
stevan I would like to use the YAML harness to generate a file (or email or something)
which would list each file which had failures in it
use nothingmuch
's patch to Test.pm 14:21
and list the line numbers for each failed test
gaal actually the yaml should have a todo attrib on tests whether in verbose mode or not
stevan so whenever anyone says "is there a TODO list??"
we can say "run the util/make_todo_list.pl" 14:22
or something like that
gaal right. that's easy on my side of things. yuval, is your coordinate patch in?
scw theorbtwo: seems the '?' in interactive mode has been renamed to '!'? It give me the result I want.
stevan gaal: re: being easy; thats what I was hoping you would say :)
gaal :)
stevan I think I am going to write up a modules/README too 14:23
autrijus mentioned it
and maybe have a chat with our mad porter :)
and get him to write a How-To 14:24
theorbtwo :h lists both.
nothingmuch wants a mad prefix 14:25
gaal does TAP define /# TODO/, or is that an ad-hoc way of telling a test is todo? 14:26
nothingmuch it does
gaal in that case 14:27
nothingmuch it's like skip
gaal svn up :)
jabbot pugs - 1139 - todo property on todo tests.
nothingmuch gaal: yaml thing
jabbot pugs - 1138 - * Added test for the Proxy object.
gaal nothingmuch, the mad prefix
nothingmuch it should not have -v 14:28
that should be the default
yaml is data, not display
and as much as possible of it should be available
gaal true.
since typeglobs have gone away, perhaps you can use their ol' sigil, *nothingmuch. 14:29
nothingmuch no 14:30
that does not imply what i do,
unless i'm 'the mad splatter'
and i don't know what it does
gaal no, the prefix itself is mad.
which is what you said you wanted. :p 14:31
nothingmuch ok, i'll rephrase
nothingmuch wants to be called 'the mad %s', where %s is filled in by some cool quality he has
;-)
theorbtwo the mad tester?
nothingmuch actually i haven't written tests all week 14:32
bzzt!
stevan I think "the mad %s" is good
nothingmuch next please
beh
stevan the mad ADD driven jack of all trades with insomnia? 14:33
nothingmuch: face it,.. you un-classifyable
(which is a good thing IMO)
nothingmuch beh, fine, what ever
*sob8
nothingmuch just can't find a clean way to do what he wants to HTML 14:34
it's such a crappy format
gaal you don't say :)
stevan nothingmuch: let me try, i used to do HTML for a living
nothingmuch stevan: see? 14:35
stevan uhm 14:36
yeah let me re-install
nothingmuch okay
theorbtwo So did I; that simply leaves me with more creadance to say it's a bad format.
larsen find ./ | xargs mail yoursister
oops
stevan ooh I can just run it off the dmg :)
larsen sorry. wrong window 14:37
xerox ahah
jabbot pugs - 1140 - always be verbose
stevan theorbtwo: I totally agree, however, in a sick way I like HTML 14:38
or at least parts of it
nothingmuch theorbtwo: how do i get all the text content out of a HTML::Element tree?
stevan nothingmuch: ready and waiting for see
theorbtwo $tree->as_text;
nothingmuch no, seperate 14:39
i have a regex that matches $tree->as_text
but does not match sub elems
even though it really really should
i'm trying to figure out who is at fault 14:40
see://woobling.org
stevan: i selected the icky part 14:41
stevan ok
nothingmuch the problem is that the if is not always working, when it shoul
and i suspect the traversal is bad
i think the next step is to rewrite it
and not use ->traverse
since the docs say it's bad anyway 14:42
brb
jabbot pugs - 1141 - Ported Algorithm::TokenBucket 0.2 (compl 14:47
theorbtwo Oh, updated testgraph as r1143 14:55
Forgot I hadn't uploaded it yet...
jabbot pugs - 1143 - testgraph updates 14:57
pugs - 1142 - Port of Algorithm::FloodControl 1.00 (co
gaal yaml_harness's -v is retired; should it default '-o tests.yml -r t/' ? 14:58
... -X Disabled ? 14:59
theorbtwo I think so. 15:00
(Change testgraph to defualt to tests.yml instead of .yaml if you do.)
gaal heh - i didn't realize you were using that :) 15:01
i'm saying .yml simply because that's what i saw on META.yml
schoch autrijus you around 15:02
theorbtwo Shrug. This stuff is the first time I've ever used yaml.
gaal gotta go for ~5hrs 15:11
bye&
jabbot pugs - 1145 - forgot -X 15:17
pugs - 1144 - better defaults for yaml stuff
nothingmuch stevan: i have a small bug 15:18
ci in 5 mins
stevan ok
nothingmuch the greedy <sub> appending is broken
theorbtwo pings autrijus. 15:33
schoch anyone interested in a perl6 book developed in wiki?
theorbtwo perlbot, seen autrijus? 15:34
jabbot theorbtwo: autrijus was seen on Fri Mar 25 19:45:27 2005
PerlJam schoch: Are you writing one or are you looking for some other kind of interest? 15:36
schoch i'm interested in starting to hack up some perl6 but there isn't a good beginner guide, i read someone wrote a book with wiki and it worked out well 15:37
i suppose i'm thinking out loud a bit
jabbot pugs - 1147 - Ported Algorithm::MarkovChain 0.06 (comp
pugs - 1146 - eval{} -> try{}, not catch{}. Fixed.
schoch but it would be nice
stevan schoch: that is Scott Walters
schoch: he wrote Perl6 Now 15:38
schoch ahhh, memory failed me on what the book was
stevan and the wiki you are thinking of is www.perldesignpatterns.com (i think)
schoch ya something like that but more pedestrian 15:39
and free 15:40
stevan schoch: perldesignpatterns is free
schoch PN6 isn't though
stevan you could start here pugs.kwiki.org
make a "intro to perl6 15:41
page
ninereasons seems like wiki is a good way to write a "beginners book" about a language that's still in gestation
schoch i need to cut my teeth on the language a bit first :)
stevan schoch: I recommend the synopsis 15:42
reading them that is
theorbtwo I recommend the exgeises as well.
schoch ya, i've taken a vacation from coding for about 6 months but reading all the developments has my attention. 15:43
where does the wiki live? it is a bit slow 15:47
stevan schoch: I have no idea
schoch fennel.oreillynet.com interesting 15:48
theorbtwo pugs.kwiki.org 15:49
jabbot pugs - 1149 - removing the author list from t/README, 15:57
pugs - 1148 - adding a modules/README
Corion re 16:09
jabbot pugs - 1151 - Backlink fixes. 16:17
pugs - 1150 - Ported most of Algorithm::NaiveBayes 0.0
stevan iblech++ # << the man is INSANE 16:18
iblech, if you are listening, email me
please :)
chip who's doing all these ports anyway? 16:23
stevan iblech 16:25
masak :)
stevan he is doing like 5+ a day
jabbot pugs - 1152 - Started a "How to port Perl 5 modules to 16:27
stevan nice
that last commit was the HOWTO i asked him to write :) 16:30
iblech++
schoch: if you are looking to learn perl6, in particular the OO stuff, I would look at iblech's work and modules/PORTING_HOWTO 16:31
Corion 110/2935 subtests failed on Win2k on 1152 16:35
... maybe I should set up an hourly smoke for Pugs/Win2k :)
stevan Corion: sounds about right
Corion: nothingmuch, gaal, theorbtwo and working on some smoke testing stuff 16:36
Corion stevan: Ah - cool. My "smoke testing" stuff would mostly consist of C<< svn up; nmake clean && perl Makefile.PL && nmake test > log >> and some POE::Irc magic to send a new message about it :) 16:37
... but if others implement it for me, all the better :)
stevan Corion: they are working on something using YAML output from Test::Harness
Corion stevan: Ah. Well - I would just run everything under TEST_VERBOSE, but I guess they also want to improve coverage/blame 16:38
stevan I think that is part of the idea
I was in on it earlier, but have not been keeping up in the current plans
my $work got in the way 16:39
Corion In theory, my openfoundry.org login should work for svn too, shouldn't it? 16:40
schoch stevan thanks for the pointer
stevan Corion: I have no idea
schoch: no problem
Corion stevan: You're no committer?
stevan schoch: reading the tests is also a good thing
Corion: I am, but I have been for a while, and to be honest I dont remember ever using a login for svn 16:41
Corion stevan: Ah :)
stevan try it,.. if it doesnt work, let me know
Steve_p Corion, you have to have a loging the first time you commit
s/loging/login/ 16:42
stevan after that it never asks again ( did I mention how much I love SVN )
Corion Steve_p: Yeah, except that it doesn't seem to work. But I'm now checking out the tree with my credentials, and maybe that is the magic I've been missing it :)
Steve_p You don't need the password to checkout :) 16:43
Corion Steve_p: But I need it for commit, but it still doesn't work. Oh well. 16:44
stevan Corion, what is your openfoundry login>
Corion stevan: corion 16:45
stevan I am not seeing you on the list rt.openfoundry.org/Foundry/Project/?Queue=270
theorbtwo Odd... 16:46
Corion, did you do the registration process from the email? 16:47
You still show with no username and status=unregistered from the admin screen.
Corion theorbtwo: Yes, I did (I can even log in)
jabbot pugs - 1153 - Finished porting Algorithm::NaiveBayes 0
stevan Corion: you are listed as unregistered in the admin
Corion Weird. I'll register again
stevan for [email@hidden.address]
Corion I have the "registration complete" email here. Aaaah - but I changed the email to [email@hidden.address] - that was my error I think! 16:48
[email@hidden.address] that is, even ...
theorbtwo Oh, then I'll add that user to pugs.
stevan theorbtwo: I got it, I am right htere 16:49
Corion theorbtwo: I'm sorry for the confusion there - I didn't know the email address was tied to anything other than the point of contact :)
theorbtwo Already done.
stevan ok Corion try now
Corion Yay! Thanks all! :)))
theorbtwo Oh, well, apparently we both did it, but he is listed as a user now. 16:50
Cool. Put yourself in AUTHORS, please.
stevan theorbtwo: its a John Woo moment
Corion r1154, already done :)
theorbtwo Nifty.
Corion I guess putting oneself into the AUTHORS file is the intelligence test that everybody needs to pass when becoming committer :)
I want to add a subdirectory to t/ called "pugsrun" (like "perlrun"), in which I place my todo_is tests for the command line options (-e , -p and -n ) - any better ideas ? 16:51
theorbtwo Not really. 16:52
stevan sounds good to me
theorbtwo Test also echo "? 12" | pugs, if you can figure out how.
Corion theorbtwo: I cribbed most of the test set up from t/examples/examples.t, which uses system() and output redirection, which is OK at the moment I think 16:53
Ooops. Committed before doing a make test run ... 16:57
jabbot pugs - 1155 - Added todo tests for multiple -e argumen
pugs - 1154 - Added CORION to AUTHORS
chip Perl6::Subs is on PAUSE (yay!) but hasn't propagated to CPAN just yet (boo!) 17:01
anybody who wants it, msg me your address and I'll mail it
stevan chip++ 17:02
theorbtwo chip++ indeed 17:04
theorbtwo wonders...
chip C< sub foo ($a of Int where { $_ > 0 }, SomeClass +$b is required, OtherClass +$c) { ... } >
yes, that works today 17:05
theorbtwo Can it be used to progmatically generate signatures for funky things, like operators?
C< sub circimfix:{'>><<'} ($a, $b is Operator, $c) { ... } > ?
chip heh. This is a Perl 5 source filter, so it can only implement the features that work in Perl 5 17:06
theorbtwo Oh.
chip I can imagine an extension of it where you can define operators and P6::Subs can create mangled names and generate the C<use overload> statement for you, though. 17:07
theorbtwo: For example, you can have optional parameters or named parameters, but not both
theorbtwo Hmm, that should be possible with p5... shouldn't it?
chip optional-positional, rather. Named parameters come in optional and required flavors.
theorbtwo: where do you start grabbing pairs? 17:08
theorbtwo Oh, but you can't figure out if it was foo => 'bar', or 'foo', 'bar'.
Gotcha.
chip yeah
on the other hand, you can have named params with a slurpy hash for others, if you're inclined.
method new ( $class: +$name, *%opts ) {} is pretty much like sub new { my ($class,%args) = @_; my $name = delete $args{name}; ... } 17:09
Corion Ooops - in (current) Pugs, is C<< my @examples = ('foo bar', 'baz bar'); >> a syntax error? 17:12
theorbtwo It shouldn't be... 17:16
Works for me.
H, util/yaml_harness.pl was missing a ;... naughty people, checking in without running it first! 17:17
jabbot pugs - 1156 - * Beginning of a Pugs Class Metamodel, c
autrijus mugwump++ # class metamodel. 17:19
metametamodel, really.
stevan very nice
theorbtwo Hmm, actually, has more problems then that... gaal? 17:20
Corion 110/2938 subtests failed on Win2k/r1157 17:23
theorbtwo Look at the end of the /topic, Corion. 17:24
Corion Ooops. 17:25
Hmm - I found where my error was - and I think it's a bug in Pugs. 17:26
jabbot pugs - 1158 - * Change poetry for Prim.hs because it's 17:27
pugs - 1157 - Added TODO test for -p
pasteling "Corion" at 217.86.56.202 pasted "Syntax error with list parsing and trailing commas" (12 lines, 262B) at sial.org/pbot/8614
autrijus training comma is a known todo. :) 17:28
Corion Autrijus: Ah - good that I didn't work on a test for that then :))
autrijus check if one is there :) 17:31
Corion autrijus: Yep - in t/data_types/array.t (where I should have checked first ...) 17:32
autrijus that's fine :) 17:37
jabbot pugs - 1159 - Ported File::Basename 2.73 (complete wit
autrijus Jouke: you can do NCI and pugs will be able to inline that NCI into haskell. 17:38
or you can do Haskell FFI and I'll make Parrot call it via NVI^WNCI.
they are isomorphic.
stevan iblech is at it again 17:39
autrijus journal up. zzz &
dada bye all 17:54
chip Say ... if I use Lexical::Alias, I could implement C<is rw> 17:56
urqe, but not for named
schoch hey autrijus 17:58
just wanted to say par++, what a great tool 17:59
chip Anybody know if C<is ro> is supposed to be allowed on parametes? It's redundant with the default, but still.... 18:04
stevan chip: i think it is, but dont quote me on it 18:05
theorbtwo I think so too... 18:07
chip OK 18:09
Lexical::Alias rocks .. and it occurs to me that I could use it to spell arrays as C<Array @a> even in Perl 5, that's cool 18:10
too many hacks, too little time
ninereasons it doesn't seem right that it works to say "my %h=<a 1 b 2> ; my @a = %h; say @a.keys;" 18:22
if it's not going to work to say "my @a=<a 1 b 2> ; say @a.keys;"
theorbtwo tcts.fpms.ac.be/synthesis/mbrola.html is pretty good, combined with festival. 18:38
Oh, wrong chan. 18:46
nothingmuch wonders how good/bad would a list context join be 18:49
join(",", 1 .. 2) -> (1, ",", 2, ",", 3)
Corion nothingmuch: Isn't that C<< zip( "," x @list, @list) >> ? Not really needed IMO, and you can write it like that. 18:51
nothingmuch wonders if you can hyper pair
theorbtwo I can't think of a shorter way to write it when just writing the join...
Corion nothingmuch: But it's convenient if you're reimplementing the HTML generation of CGI.pm :)
nothingmuch map { *$_ } (@list >>=> ",") 18:52
but then you have the dangling "," that I want to avoid
zip("," x (+@list - 1); @list); 18:53
(note greater list comma, since it's multi dim)
theorbtwo still wants to know what's so bad about join(',', 1..10) 18:54
Corion theorbtwo: That doesn't return a list :)
nothingmuch and if it theoretically returns a list, i know at least my code will break
i use join in list context a lot
since i know it's always scalar
then again, join could always be translated to ~join 18:55
theorbtwo Oh, I see what you mean.
You want to create a list such that when you join it with '', you get '1, 2, 3, 4'
I'm thinking that you really don't want to do that. 18:56
stevan hey naddamucho, Dah-Orb Numbah 2
nothingmuch ?
theorbtwo Having a good Good Friday, are ya?
stevan File::Spec is done :)
theorbtwo Stevan++
nothingmuch yay
stevan++
stevan theorbtwo: not yet
I am running make test right now to make sure all is well
and I talked to the mysterious mr. iblech 18:57
he is going to direct some of his boundless energy on the t/oo stuff
nothingmuch hopes iblech is not wasting vacation time or something like that on computers 18:58
stevan yes he is :)
theorbtwo Implementing it, or completing the tests?
stevan he said he is going skiing next week though
theorbtwo: writing tests for it
rgs chip: there is an idea floating in the air, about implementing lexical aliases in perl 5.10
stevan I figure he knows as much about perl6 OO as the Damian/larry now
you know one thing I love about perl6 19:01
with function signatures, writing docs becomes so much eaiser
its practically self documenting at times
rgs that's java all over again ! 19:02
sorry.
stevan rgs: kinda :)
19:03 khisanth_ is now known as Khisanth
rgs the PORTING_HOWTO is cool 19:03
chip rgs: I remember ... I was one of the pushers. I proposed making C<\$a> an lvalue. 19:05
rgs chip: aah. like this: my (\$alias) = \$variable;
chip rgs: yeah
rgs send tuits :)
chip my ($a,$b) = \(@_)
rgs er, missing \ ? 19:06
my \($a, $b) = \(@_)
jabbot pugs - 1160 - Added tests for subtypes. 19:07
chip oopsie. yes
nothingmuch what is a nice way to split a string into chunks 19:10
with the splitting done on numeric offsets?
@offsets = (2, 4, 7); "the quick brown" -> "th", "e ", "qui", "ck brown"; 19:11
Corion nothingmuch: Unpack ? 19:12
nothingmuch with unpack i can do 'th, 'the ', 'the qui', etc with backtracks and a$offset
ofcourse, i could do this in a loop
for (my $i = 1; $i < @offsets; $i++){ 19:13
unshift @offset, 0 first
Corion nothingmuch: Ah, now I see ... Well, you could convert this into a sorted list and then use the differences between the items as length items to A.
unpack will become soo much better as we have Unicode chars as templates for pack/unpack :)
Khisanth U? 19:14
theorbtwo Or, for that matter, with substr.
Corion Anybody can tell me how I match (in Haskell) in the function signature against a Named Field of a parameter passed in
(I'm passing around a "hash" (constructed out of named fields), and I want to test some values in the signature directly) 19:15
nothingmuch errm? 19:16
Corion nothingmuch: pasting ...
chip nothingmuch: I think a for loop with substr will do it 19:17
jabbot pugs - 1163 - making Makefile.PL into Makefile.PL.loca
pugs - 1162 - * Submethod tests landed.
pugs - 1161 - File::Spec is done, tests and all; howev
awwaiid nothingmuch, in perl5 sub grab {if($_[2]){$a=shift; (substr($a,shift,$_[0]), grab($a,@_))}
chip my $copy = $str; map { substr($str, $_, length($str), '') } reverse sort @positions 19:18
awwaiid then call grab("my string here",0,5,7,10)
nice, chip
nothingmuch $_[2]?
pasteling "Corion" at 217.86.56.202 pasted "Using Named Fields in a function signature" (41 lines, 1.1K) at sial.org/pbot/8617
chip make that sort { $b <=> $a } positions
awwaiid third argument (in @_)? 19:19
Corion I really like function signatures - they take out a lot of "if" cases. But they make the "program flow" harder to follow, as you now have to look across several functions to see the state machine behind it again...
gaal good morrow
nothingmuch oh, oops 19:20
i see
nothingmuch didn't notice awwaiid recursed 19:21
now if only i had a proper zip
oh, well, it doesn't help either
Corion gaal: Did the param-parsing in Perl go forward / is there anything to do ? I'm getting no love from Haskell there :-) 19:22
gaal ah
er, no thanks for the nudge, i'll email thedamian
Corion gaal: Heh - I got a small step forward with the parsing/reworking, but then I tried to generalize my small multiple-dash-e success and got caught up in this ugly Named Fields thing... And I'm looking at all the wrong documentation ... 19:23
nothingmuch Corion: i don't see what you're trying to do 19:24
but in general 19:25
both must be changed to
_run args env { oneliner = arg }; etc
"values" in haskell are not raelly that
they're immutable
if you take env and pass it { oneliner = arg } then it returns an env which is the same, except oneliner = arg
Corion nothingmuch: Yeah - that's what I want to do on the rhs. But I also want to check for a value of env{oneliner} on the lhs - I'm fine with copying stuff around :) In Perl I'd want C<$arg{oneliner} eq ''> 19:27
... but how to do that on the Haskell side of things ?
nothingmuch ah
_run (("-e"):frag:rest) env | oneliner env = "" = foo| otherwise 19:28
sorry
| otherwise = bar
chip Having read the translation of URI I have two comments. First, :scheme is misspelled as :sheme in one place. Second, I WANT TO USE THIS LANGUAGE.
nothingmuch | lets you put a case at the pattern matching level, sort of 19:29
Corion nothingmuch: Ah - I'll stumble in that direction then :)
chip: Then kick the pumpking so he pushes the development further :)
Khisanth but don't kick too hard or you might put him out of commision! 19:30
gaal commandline = GetoptDeclare.parse <- getArgs :)
Corion gaal: :))
nothingmuch stevan: ping 19:31
stevan nothingmuch: hey
chip Corion: oh, right, like I need to speed Leo *up*
er, yeah
Khisanth gaal: Haskell?
nothingmuch SEE session realy quick
stevan nothingmuch: I am actually just heading off to lunch 19:32
gaal Khisanth: embedded perl6 :)
stevan really quick?
nothingmuch i just want to show you what i ended up doing
stevan ok
nothingmuch and wanted to make sure it stayed sort of readable
stevan one sec
Khisanth gaal: if that snippet is perl6 ...
nothingmuch but it's not that important
Khisanth: that's not p6 19:33
gaal yet
stevan nothingmuch:
nothingmuch it's how autrijus wanted the haskell side of pugs to parse it's options
stevan :P itchy trigger finger
nothingmuch stevan: ok, this is what's going on: 19:34
for every synopsis we collect the backlinks
find them with the regex
remember where they matched
and then only after we're done, destroy the text data
splitting it up
oops, wrong file 19:35
jabbot pugs - 1165 - Fixed typo in call to Test::write_log 19:37
pugs - 1164 - Added Roles tests.
nothingmuch one last todo: 19:40
make regexes match nested <code>
and across pre, or whatever
stevan I think we should not link to the code
only to the text
nothingmuch see S06 19:41
L<S06/"The &?BLOCK routine" /\$\?BLOCKLABEL contains the label of the current block, if any/>
The &?BLOCK routine
that won't ever match 19:42
because it's <code>$?BLOCKLABEL</code> contains...
nothingmuch suddenly wants to listen to Lamb
stevan I see
Lamb?
nothingmuch What Sound
google for it
nothingmuch will hand out mp3s if you promise that you buy it if you keep listening 19:43
stevan lampstar.net?
nothingmuch i don't think they're successful enough
stevan s/p/b
nothingmuch or heck, you can just decide if it's fair to buy
nothingmuch shouldn't allow himself to be so self-righteouous 19:44
and yes, tr/p/b/ and the site is correct
stevan is listening to the clips on the site
stevan thinks its ok, but not his style so he goes back to listening to Public Enemy :) 19:46
nmcfarl is listening though
lamb was a good idea
stevan ok lunch time,.. 19:47
jabbot pugs - 1167 - missing semicolon, oops
pugs - 1166 - * Added Traits tests.
stevan have fun all &
nothingmuch ciao stevan... 19:50
Khisanth hmm "First, :scheme is misspelled as :sheme in one place." <- still the same old @EXPORT stuff? 19:51
nothingmuch nothingmuch.woobling.org/pugs_test_status/
superscripted regex skip backlinks
r1168 19:52
Corion Schmorp !? Hi :) 19:53
gaal nothingmuch: it strips whitespace /indendtation?
nothingmuch no
gaal nothingmuch.woobling.org/pugs_test_...undef.html 19:54
nothingmuch ook
gaal nothing's indented
nothingmuch that code is theorbtwo's fault
gaal <g>
nothingmuch i think it needs to unexpand
gaal "unexpand"? 19:55
nothingmuch ... at 4 spaces, naturally
;-)
errm, expand
gaal oh, it's just matter of s/\t/" " x $ONE_TRUE_TAB_WIDTH/ ?
theorbtwo Oh, there's some missing pre tags.
I don't do anything special with tabs. HTML4.01 suggests (but does not require), that they use 8-char line-up tabs. 19:56
gaal nothingmuch: i don't believe that topic almost started on perl-il *again*!
nothingmuch hasn't read perl-il for a while 19:57
migo vs. yuval?
gaal in that case i'd suggest we do one of the following:
1. recommend one tab style in t/README
jabbot pugs - 1168 - Backlinks skip with regexes now. A few q
nothingmuch the reason i use tabs is that it's less work on a stupid editor, that doesn't have indent/outdent
and the reason i use 4 is that it's the default most places 19:58
aside from that i don't care
gaal 2. ask people to put in either cperl or vim signatures of the tab style they like in their tests, and parse it
nothingmuch so should it be 8?
gaal: nay nay!
gaal i think i prefer #2, if only to avoid having that discussion again
nothingmuch too much metadata overhead
maybe if we use Vim coloring it could obey it 19:59
that's what i want to do to tests next
gaal if we do, it would.
nothingmuch ok, so that's for free
and it'll do the default
but for now i'm using Text::Tabs @ 4
gaal fine by me, if only becaise that's my personal preference too :) 20:00
theorbtwo Don't forget to put in the missing pre tags!
gaal nm: what's the status of your test coordinates patch?
asking mostly out of curiosity, also wondering if i need the harness to do anything special to activate it 20:01
i'm doing a clean up of tests, replacing parsefails with clean fails(). what's the status of Dialects/perlego/sigilless? it's very todo, right? 20:10
ping Juerd 20:11
nothingmuch test coordinates?
his own perl6, that is differen
Dialects is something Juerd is going to do
maybe better
gaal nm, i just parsefailed on your last four lines. 20:12
chip Khisanth: "Same old export stuff"? What's that?
Khisanth chip: I mean is @EXPORT and @EXPORT_OK going to remain the ways things are exported? 20:13
chip Oh, no, that's all C<is export> etc. now. C< :foo > is Perl 6 for C< $self->_foo > if leading _ means "private" 20:14
gaal Khisanth, see modules/README
er
moduiles_PORTING_HOWTO
i can't spell :)
theorbtwo perlmonks.org/?node_id=442402
Khisanth gaal: that is EXACTLY the reason I asked :p
gaal lol
Khisanth sometimes hunting that one error where you have two letters flipped is quite a pain 20:15
chip I look forward to warnings on :missppelled
theorbtwo looks forward to a highly intelligent spell-checker for code. 20:16
nothingmuch darn it
aspell knows how to mind it's own business
theorbtwo (One that knows that the name of a declared sub cannot be misspelled -- except in the declaration.)
gaal Eclipse JDT is pretty good for JAva
nothingmuch but won't work for runtogether words 20:17
chip nothingmuch: "*its* own business"
nothingmuch uses apostrophe for his/her etc automatically
hard to stop myself
sorry
theorbtwo: you fix it 20:18
i don't know the semantics of tt
and Text::Tabs isn't DWIMing in that context
i'm too tired
and set the charset to utf-8 while you're at it ;-)
theorbtwo ci what you've got. 20:19
nothingmuch 2 lines:
err, 3
use Text::Tabs;
theorbtwo Er, wait a sec -- we're working with the .jmm. version, right.
nothingmuch and then expand($text); and expand($rest);
yep
theorbtwo OK.
We probably ought to do something about that, like rm'ing the other one and mv'ing the jmm one on top of it. 20:20
But I'll leave that to people who care more.
But .jmm. is rather inaccurate.
gaal jmm?
theorbtwo James Michael Mastros. Me. 20:21
gaal ah :)
theorbtwo Anyway: sleep well. 20:22
nothingmuch hola lightstep 20:23
nothingmuch is not tired sleepy
theorbtwo Ah.
nothingmuch tired of sitting
but thanks
svn blame says about 40% is yours
and i'd say lots of the lines that are mine are just modified 20:24
lightstep good night
nothingmuch theorbtwo: please do a code review 20:26
last time my code was a bit too spaghetti for you
i think this isn't much better
nothingmuch would like it better abstracted, perhaps 20:27
anyway, i'm off to get cleaned up 20:28
gaal r1169 makes all tests *run*, sticking to the "FIXME parsefail" convention 20:30
nm+to2+stevan, you're right about not needing to go through hoops to capture tests blowing up 20:31
theorbtwo t_index/t/builtins/strings/length.html looks horrible, but that's because of the BIDI, I think... 20:35
nothingmuch utf8 is your friend
gaal what editor are you using? 20:36
looks fine to me
theorbtwo The HTMLification, under firefox.
gaal ah :/
jabbot pugs - 1170 - Oops, missing parentheses. Fixed. 20:37
pugs - 1169 - replaced all remaining hard parsefails w
theorbtwo It looks... umm, acceptable... under xemacs.
gaal um 20:38
nothingmuch.woobling.org/pugs_test_...s/strings/
what is this showing?
hahaha
index.html
got me for a moment :) 20:39
Qiang neat. perl2html ? 20:42
gaal Qiang: t/catalog_tests.jmm.pl 20:44
s/t/util/ actually
theorbtwo r1171 fixes some bugs in that. 20:45
Allo, Qiang, lukhnos, lightstep.
lightstep olla
Qiang :)
theorbtwo nm, whats your best guess as to what's causing all the regex L<>s that don't resolve? 20:47
nothingmuch bad regexes 20:51
title case is funny in synopses
theorbtwo Anyone else working on it would have to have an interest extending from the insanity of Perl 5 parser internals all the way to various deep Perl 6 design issues, and I don't think anyone else besides me has the requisite multiple personality disorder. 20:52
Oh.
nothingmuch there's also my bad regex
es
in smartmatch
there's "Pointy subs"
which can't be linked
gaal escape them? 20:55
nothingmuch i tried that
didn't work naively
gaal escape them, not naively?
jabbot pugs - 1171 - yaml_harness: fix usage with no params 20:57
nothingmuch gaal: errmm, later maybe 20:58
ninereasons S29 is interesting.
nothingmuch s/later maybe/patches welcome/
gaal nothingmuch: example of something that fails please? 21:05
nothingmuch ?
gaal do you have a failing test? :) 21:06
nothingmuch colored test thingies?
gaal no no
for the regexp
nothingmuch uhm...? 21:08
i don't get it
gaal since i don't knwo the code you're talking about
can you supply me with a test case, so i can work on a patch
nothingmuch ah
errm
t/data_types/anon_block.t 21:09
it contains such a link
the bix m{}x regex in util/catalog_tests.jmm.pl handles it 21:10
s/bix/big/ 21:11
nothingmuch goes to iron a shirt 21:21
jabbot pugs - 1172 - Added tests for caller() and want(). 21:27
pugs - 1173 - set svn:ignore *.hi 21:57
pugs - 1174 - Added tests for parameterized roles. 22:17
pugs - 1178 - Found a bug in S12: 22:27
pugs - 1177 - Added tests for roles conflict resolutio
pugs - 1176 - Test correct "return value" of given.
pugs - 1175 - documentation typos
chip Any reason there's a %MY:: but not a %CALLER:: nor an %OUTER:: ? 22:41
autrijus is there no %CALLER:: ? 22:42
it's unspecced, true, but I always thought it's there.
get larry to spec all three? :) 22:43
rgs mmh %CALLER::. breaks scopes
crysflame hi, autrijus 22:45
autrijus hey. 22:46
chip tied, presumptively
or the moral equivalent
autrijus rgs: $CALLER::var already breaks scope
stevan hey autrijus
autrijus hey stevan. can you do me a big favor 22:47
and do the impossibly difficult task that is ChangeLogs ?
stevan I can if its quick or doesnt require me to do it right now
autrijus: I can do that,.. when do you need them by?
autrijus I need it in 30 hrs :)
preferably 24
jabbot pugs - 1179 - Added parameterized traits and delegatio
stevan autrijus: I will get then to you in 15
:P
BTW - File::Spec is done 22:48
autrijus my talk about it is 32 hours from now :)
stevan and tested
autrijus !!!
er. like. wow.
stevan its ugly
but it works
I talked to iblech too
autrijus pugs are supposed to be like that :)
stevan got him to direct some of that energy to t/oo/* 22:49
autrijus good.
stevan and maybe work with me to port File::basename
autrijus I think I'll call 6.2.0 right after 6.0.14
stevan I think that could work now
autrijus yay
stevan autrijus: I think that would work
autrijus we're already reasonably complete
I need to go thru your buglist after conference
stevan also,.. nm, theorb and gaal made much progress on the test catalog and yaml harness 22:50
autrijus I saw that :)
autrijus has read backlogs
stevan at some point we are going to make you a TODO list maker
ah,.. didnt know how much you had read
autrijus a TODO maker will be insanely great.
stevan :) 22:51
do you want me to TODO all the tests as well?
for the release
autrijus that'd be best.
stevan oh, and something to tell ingy when you see him,.. the Makefile.PL in Test.pm installs to my /Library/perl directory on Mac OS X 22:52
which is why I didn't activate the File::Spec one, as it overwrites my perl5 install
autrijus ew, wow 22:53
ok, will get him to fix it
ideeally after his talk
stevan ok
alright, dinner time here, I will do the changelog and TODO-ing tonight
later &
autrijus: when you have a second... look in t/oo/* 22:54
iblech is unstoppable :)
and and he wrote that modules/HOW_TO as well,.. very nice 22:55
autrijus how many persons is iblech anyway?
is he a consortium? :)
stevan one,.. on school vacation :)
he is the german equivilant of you
autrijus that is amazingly cool.
:D
stevan ok gotta go now for real, before my wife has my head :)
autrijus . o O (One day he'll learn haskell and make those tests pass)
stevan++ # thanks!
stevan autrijus++ :) 22:56
jabbot pugs - 1180 - Added enum tests. 23:07
autrijus iblect++ # Unstoppable 23:08
iblech++ # even
20 modules in 2 days and 212 OO tests. 23:12
Khisanth only 7779 modules to go? :) 23:18
autrijus only 79 to go before we have good idea how p5-to-p6 should look.
the automatic translator thing :) 23:19
Khisanth ah so it's not going to be p5 code->bytecode->some B:: module->p6 code? :) 23:20
gaal hi, i exchanged emails with Damian about Getopt::Declare and here's a summary (in the hope Corion reads the backlog): 23:25
He thinks the module is technically capable but probably much too slow for anything but a prototype. He has something planned that sould be better, but is very busy in the near future. 23:26
technically capable of parsing the perl command line, that is
looks like we can try embedding the p5 version, if we want a proof of concept, but porting it to p6 might be a bit of a big job for something that won't keep :( 23:29
alternatively we could keep doing it in haskell, though that means we have to design it first :) 23:34
Limbic_Region autrijus - why aren't you sleeping? 23:35
or have you woken up already?
jabbot pugs - 1181 - Added tests for C<is build {...}> and C< 23:37
pugs - 1182 - Added class and class|instance method te 23:47

Powered by Cro and the Raku® Programming Language.

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

Thank you!