Pugs 6.2.8 released! | pugscode.org | pugs.kwiki.org | paste: sial.org/pbot/perl6 | www.geeksunite.net
Set by autrijus on 13 July 2005.
stevan iblech: we dont need dispatch() in the JS metamodel 00:37
iblech: we can use can()..... so Class.can('foo').call(args) and instance.can('bar').call(args) 00:38
gaal hey all. you may need to make clean after this commit - r5738 00:45
night :) 00:48
svnbot6 r5738, gaal++ | config file-based build prefs 00:49
00:55 joepurl_ is now known as joepurl
ferreira Someone gets this also: "Can't open perl script "util/build_pugs.pl": No such file or directory" 01:03
svnbot6 r5739, Stevan++ | Perl6.MetaModel - (JS version)
r5739, Stevan++ | * added Perl6.Class classes
r5739, Stevan++ | - added test file for this
r5739, Stevan++ | * added POD documentation to most classes
coral mm, #perl is quiet today
ww
webfiend quiet, or just not insanely busy? 01:05
ferreira "util/build_pugs.pl": No such file or directory - is this a Cygwin privilege or someone else is having trouble with this? 01:14
coral ferreira: does make clean resolve it? 01:42
coral is guessing wildly
webfiend: quiet, as in hasn't lit up today
ferreira coral: nope. I did "make clean; svn up; perl makefile.PL; make" and it complains again. The same with "make unoptimized" 01:46
coral mm 01:47
dudley ferreira: just happened to me, too
coral must be broken in svn
i don't see any mention of build_pugs.pl in my scrollback 01:48
ferreira It must have broken with "r5738, gaal++ | config file-based build prefs" 01:53
dudley gaal: ping 01:56
coral seen gaal 01:57
jabbot coral: gaal was seen 1 hours 8 minutes 51 seconds ago
coral saying "night"
ferreira indeed, reverting to r5737, make is working now 01:58
dudley I'm guessing he just missed that file in the ci
ferreira anyone knows if autrijus who wrote the code of Pugs/Run/Args.hs?
dudley ferreira: svk blame 01:59
that'll show you who checked in each line of a file
or svn blame, if you're not using svk yet
geoffb likes that "yet" 02:01
ferreira sorry for my svn ignorance. Now I see: autrijus, iblech, bsb, theorbtwo - a team work. But most of it has been written by iblech. 02:02
dudley ferreira: hey, no worries. I just learned that last week.
geoffb s/most of it/most of the current revision/
If someone completely rewrites a file, then blame doesn't show the old author 02:03
ferreira dudley: you learned svn or you jumped right to svk?
geoffb: ok, got that
dudley I used svn for about a month before I tried svk.
mugwump I think it's a good idea to start with svn, given svk is basically a glorified SVN client with merge history tracking. 02:04
svnbot6 r5740, fglock++ | * updated ChangeLogs
r5740, fglock++ | * added unicode support to Set::Infinite
ferreira what the hell? "parse error (possibly incorrect indentation)" That Haskell is starting to look to like Python
mugwump That's glorified as in HALEEPHUKENLUJIA!!!!!
dudley ferreira: I've been wondering when I would hear someone complain about significant whitespace in haskell :-) 02:05
It's optional, though. You can use brackets and semicolons if you really want to.
ferreira mugwump: I am suffering sinner that has to live mostly with Windoze at these days. Those beautiful tools like perl, svn, etc., etc. love to make life miserable for us that are Unix-challenged 02:07
mugwump in that case, there are more windows svn clients available than svk 02:08
TortoiseSVN is cute
geoffb Tortoise(CVS|SVN|<more?>)
dudley I was just about to say that...
geoffb Great minds think alike . . . . 02:09
mugwump unless you're talking about ./foo() 02:10
dudley mugwump: ha!
ferreira dudley: I was just starting with Haskell and trying some guesses. Well, I just find out that # are not for comments and I have to use -- instead (think Lua, SQL, not Perl) 02:11
dudley ferreira: I would think learning haskell by random guesses will be fairly painful :-) 02:13
coral heh 02:14
ferreira dudley: I use Windows - I enjoy suffering. What means "Pattern match(es) are overlapped"? A pattern like ("--":xs) overlaps with ("--external":mod:rest)?
dudley I'm new to Haskell myself, so someone else may want to field this one 02:16
I think it means that the pattern that comes first will match everything that the second one will, so you can never reach the second. 02:20
There's an incredibly high chance that that's wrong, though
ferreira When a string like "--" is used in a pattern, it means the arg should be equal to this or like Perl regex /^--/? You know that? 02:22
geoffb ferreira, this may or may not help, but a string "foo" is exactly a list ['f', 'o', 'o'] 02:25
so matching ("--": ...) == matching ('-' : '-' : ...)
So yes, ("--":xs) would match anything starting with -- 02:26
dudley and ("--":xs) will match any string that begins with --
great minds again... 02:27
geoffb heh
ferreira and if I want to match exactly "--"? 02:28
dudley the :xs part tells the pattern to put the rest of the elements of the list into xs
so if you wanted to match only "--" with nothing after it, you would leave that part out 02:29
ferreira Well, now I understand why at p6c someone reported that pugs -we "@*ARGS[0].say" -foobarfluffy 02:30
outputs -f
The use of patterns in the command line parser does this.
02:31 kungfuftr is now known as kgftr|konobi
geoffb Time for a test, ferreira. :-) 02:33
dudley anyone else running ghc from darwinports on tiger? 02:35
svnbot6 r5741, Stevan++ | Perl6::MetaModel - (Perl 5 version) 02:38
r5741, Stevan++ | * find_attribute_spec() now uses the dispatcher (in BUILD order)
r5741, Stevan++ | * fixed method list in ::MetaClass POD
ferreira how do I run only a subset of tests: for example, those in "t/pugsrun"? 02:43
dudley ferreira: you should be able to just run them with pugs 02:45
svnbot6 r5742, Stevan++ | Perl6::MetaModel - removed the get_all_attributes() method in favor of the dispatcher
geoffb ferreira, I think "prove" might do what you want 02:58
looks like it
ferreira "Can't exec /usr/bin/pugs" says prove. Don't know how to tell it the pugs right there at '.'. If I try dudley's suggestion, tests are all failing miserably. My config should be screwed up. 03:04
dudley I have a symlink to pugs in my path. Maybe that's why it worksforme. 03:05
geoffb I think there's some envvar for prove that tells it where to find pugs. hmmm 03:07
ferreira Yes, it worked with the symlink and prove. Thanks, geoffb and dudley. Now I am going to try my changes. 03:12
geoffb ewww. Normally, the env var HARNESS_PERL would work, but code in the pugsrun tests forces ./pugs. bleah 03:16
That, I think, is a bug.
OK, time to take a break from Haskell tutorials for a little while, I think . . . probably a good idea to let things settle overnight before attaching chapter 7 (Input/Output) of _A Gentle Introduction to Haskell, Version 98_ 04:51
stevan *sigh* now I know how autrijus feels 05:19
$Larry is totally re-arranging how the metamodel would work
sanxiyn I checked out pugs today, but build failed with "Can't open perl script "util/build_pugs.pl": No such file or directory", any idea? 05:22
stevan sanxiyn: it seems to have been a problem today. Try checking out again now, it is likely fixed now. 05:26
sanxiyn I see no update with "svn up"... 05:30
(rev 5742)
gaal morning folks 05:31
stevan sanxiyn: did you "make clean"?
svnbot6 r5743, gaal++ | oops, forgot this for the build system. gaal--
gaal sorry about that :)
please try the latest build now.
stevan gaal: hows goes the hackathon ?
gaal we aren't hacking a lot, but we are athonning! 05:32
stevan nice
gaal it's only szabgab and me now 05:33
nothingmuch will come later
sanxiyn oops, now I see it. 05:34
gaal last night i gave the perl6 is cool spiel to szabgab and eilara
(and broke th build!)
lumi and rindolf were also here 05:35
my cat was scared but is recovering
pizza ETA... 10 hours?
stevan gaal: 2 words - baking paper ;) 05:36
gaal that's it more or les
ah!
good, you reminded me
we made crazy soft drinks
experiments with dry ice :)
stevan ? 05:37
gaal we "made smoke"
white smoke
by dropping some dry ice that i had into coke to cool it
it looks just like a canister of liquid nitro does 05:38
stevan nice
sanxiyn Now I get "Can't use an undefined value as an ARRAY reference at util/build_pugs.pl line 14." 05:39
gaal hmm! let me see
sanxiyn Must be a bad time me trying... 05:40
gaal nah, bad time me committing :)
sec
hmm, what os are you on sanxiyn? 05:41
sanxiyn Debian Linux.
gaal stevan do you get that too? 05:42
dudley gaal: I'm getting the same thing on osx 10.4
stevan gaal: I havent built pugs in a while, I use feather
gaal can you paste the perl -Iinc util/build_pugs.pl line just over that error? 05:43
stevan: clever :)
dudley, sanxyin, something like C:\perl\bin\perl.exe -Iinc util\build_pugs.pl +GHC C:\...... 05:44
pasteling "dudley" at 70.179.213.96 pasted "build_pugs.pl error output" (1 line, 1.5K) at sial.org/pbot/11904 05:46
gaal thanks
i think it's a command line length issue. darn. 05:49
szabgab hey ppl, what is the simplest recommended way to learn Perl6 now ?
QtPlatypus szabgab: Good question. 05:50
I guess from reading the Synoposes and Exengieses (spelling).
dudley Then reading the mailing list archives to see what's changed since the Synopses were last updated... 05:51
:-)
gaal dudley: can you run this exact command in your shell manually? 05:52
it'll probably give the same error
dudley I can try...
yep, same thing 05:53
szabgab thats sounds like too difficult and long and well outdated
gaal okay. now try it without the Mono -L directives
which i think aren
't required
(i'm trimming fat)
dudley no joy 05:54
gaal in line 13, just above the error, please put in this debug print: 05:55
require YAML; die YAML::Dump($opts); 05:56
and nopaste the result?
i want to confirm the cmdline gets eaten
dudley GHC: 05:57
- /opt/local/bin/ghc
- -threaded
that's it.
gaal eek :) 05:58
whoever's eating it must be hungry.
dudley heh.
gaal though it is fishy that it stops right where the first substitution takes place :( 05:59
# isn't a special character in your shell, is it? 06:00
sanxiyn # is...
# starts a comment.
gaal ah.
sanxiyn echo a # b
gaal that does put a damper on out relationship.
sanxiyn (echoes "a")
dudley gaal: what system are you on? 06:01
gaal windows.
sec, fixing.
(doing a basic sanity check first) 06:03
sanxiyn++, dudley++
r5744 06:04
please svn up and tell me it works. :) 06:05
sanxiyn gaal: Oops, I backed to r5737 and now building... 06:07
svnbot6 r5744, gaal++ | oops, ### introduced a comment on some shells.
dudley gaal++ # It seems to be working now, yay
sanxiyn dudley: Nice to hear that.
gaal thank you! 06:08
dudley np
gaal brb (shower) &
gaal pops back in 06:52
how's the build going for y'all? 06:53
sanxiyn It is built! 06:55
Now, time to play with it.
gaal cool. that's the fun part :) 06:56
QtPlatypus You might wish to look at the cookbook entries to for help coding.
jdv79 are there examples of interacting with p5 like DBI and such? 07:03
gaal jdv79: the basics are quick to summarize: 07:04
eval ($p5_string, :lang<perl5>)
use perl5:P5Module;
that's more or less it :)
jdv79 ok, i'll give it a go - thanks 07:05
gaal you can also look in t/
jdv79 is there a strict or warnings pragma in pugs? 07:06
or is that not even p6?
QtPlatypus jdv79: Not yet. 07:07
jdv79 cool
QtPlatypus jdv79: And I expect it will have.
Though they will most likely be diffrent to perl5.
jdv79 use v6 does what? 07:08
gaal jdv79: see t/unspecced/p5/
use v6 simply guards against p5 accidentally trying to run p6 code. 07:09
jdv79 my @drivers = DBI.available_drivers(); should work, right? 07:21
*** No compatible subroutine found: "&available_drivers" 07:27
weird
Darren_Duncan All ... 07:53
I have completed the first draft of my Lightning Talk for OSCON 2005, which introduces 'Rosetta'. 07:54
I request your assistence to help me make this talk the best it can be.
obra url?
Darren_Duncan You should send your replies to me at [email@hidden.address] since I don't leave IRC open all the time
coming
darrenduncan.net/OSCON/OSCON2005Lig...ngTalk.txt 07:55
this is indirectly related to Perl 6
mainly it is Perl 5 now ... but related because I will be porting it over soon
I posted my request to 3 database email lists, but thought I'd ask here also 07:56
is it worth posting to p6c or p6l, or would that be off topic (replies are private)
I think the start and end are ok; the middle 80% is what needs fixing 07:57
for one thing, this is twice as long as it has to be, taking about 8 minutes to read aloud
thanks 07:58
note that I'll be hanging up this channel in about 10 minutes, but will see any emails tomorrow ... could talk at length on channel tomorrow, though, if appropriate
obra I haven't had time to reply to you about what searchbuilder does 07:59
:/
Darren_Duncan not worried
obra Darren_Duncan: will you have slides with pretty example code?
Darren_Duncan as mentioned elsewhere, I don't plan to name any other modules, but just address the hot features that people want
I have zero slides, sorry
obra Or is the plan to just read that aloud? 08:00
Darren_Duncan the talk isn't long enough for that
I plan to read it aloud
though attempt to keep looking at the audience from time to time
obra is there a rosetta tutorial online?
Darren_Duncan while I could say all of these details from memory, I wrote it down so I didn't have to search within 5 mins to say the right thing 08:01
no tutorial yet per se
obra nod
Darren_Duncan there is a pile of documentation, but none tutorial-ish yet
this is very new, after all
I expect to have a proper tuturial within a year's time though ... maybe even give a full length talk too 08:02
of course, this is an example of something that people can contribute
obra But it's been on CPAN since 2003, no? 08:03
Darren_Duncan I also plan to have a few large updates to Rosetta between now and when OSCON hits, so certain confusing bits shouuld be gone
yes
obra But now, I need to run to work 08:04
Darren_Duncan It also didn't actually work worth a damn until around now
the text of the talk assumes I've gotten a few things working that aren't working now
jdv79 oh, a fresh build works for perl5:DBI but the latest CPAN release doesn't - even weirder
Darren_Duncan good night 08:06
svnbot6 r5745, gaal++ | allow precompiling arbitrary additional modules; setting GHC heap size via config.yml 09:50
QtPlatypus How does sub coerse:base {...} differ from sub infix:base {...} ? 11:20
?eval sub base(Int $x,Int $y where {$y>1}) {...} 11:41
evalbot6 Error: unexpected "w" expecting trait, ",", ":" or ")"
QtPlatypus ?eval want 12:28
evalbot6 'Scalar (Any), LValue'
QtPlatypus ?eval my $a = want; want; 12:29
evalbot6 'Scalar (Any), LValue'
QtPlatypus ?eval my $a = want; $a
evalbot6 \'Scalar ((mkType "Scalar"))'
QtPlatypus ?eval my Str $a = want; $a
evalbot6 \'Scalar ((mkType "Scalar"))'
QtPlatypus ?eval sub a {given want { say $_}};a() 12:31
evalbot6 Error: No compatible subroutine found: "&a"
QtPlatypus ?eval sub a() {given want { say $_}};a()
evalbot6 perl6 Error: No compatible subroutine found: "&given"
QtPlatypus ?eval sub a() {given "a" { say $_}};a()
evalbot6 a bool::true
QtPlatypus ?eval sub a() {given (want) { say $_}};a()
evalbot6 Scalar (Any), LValue bool::true
QtPlatypus ?eval sub a() {given (want) { say $_}};my $a=a() 12:32
evalbot6 Scalar (Any), LValue \bool::true
QtPlatypus ?eval sub a() {given (want(1)) { say $_}};my $a=a()
evalbot6 Error: No compatible subroutine found: "&want"
svnbot6 r5746, qtplatypus++ | A simple MIME-Base64 implimentation. 13:14
r5747, iblech++ | * perl5/Perl6.MetaModel/tests/*: Fixed the svn properties (sorry, was my fault). 13:24
r5747, iblech++ | * Added the usual svn properties to qtplatypus++'s ext/MIME-Base64.
r5747, iblech++ | * PIL2JS:
r5747, iblech++ | * pil2js.pl has a much saner API resembling GCC a bit now:
r5747, iblech++ | $ pil2js.pl -o compiled.js input.pl
r5747, iblech++ | $ pil2js.pl -o SomeModule.js SomeModule.pm
r5747, iblech++ | $ pil2js.pl --link=html -o test.js ~Prelude.js SomeModule.js compiled.js
r5747, iblech++ | $ pil2js.pl --link=js -o test.js Prelude.js SomeModule.js compiled.js
r5747, iblech++ | * Better exception handling: try { die ... } should not alert() the error
r5747, iblech++ | message. Instead, only exceptions making it to the toplevel are reported.
r5747, iblech++ | (I.e. "say $some_undeclared_variable" will now show a user-visible error
r5747, iblech++ | message instead of leaving a small entry in the JavaScript console (if
r5747, iblech++ | available).
QtPlatypus iblech: What svn propertes have to be added? 13:26
iblech QtPlatypus: See lib/pugs/hack.pm. But it's not a problem if you forget them
QtPlatypus nods. 13:28
masak ingy, ping 13:39
meppl open: config.yml: No such file or directory at inc/PugsBuild/Config.pm line 41. 13:42
known?
svnbot6 r5748, gaal++ | add usage note; some minor cleanups
gaal um, that shouldn't happen, meppl
but do this:
make config.yml
meppl okee
dont work :o) 13:43
gaal does make clean and perl Makefile.PL help?
meppl ill do
it
gaal, no it doesnt work, the config.yml is still missing 13:48
Aankhen`` I am getting the same error (Win32, MSVC++).
QtPlatypus meppl: I got the same problem. 13:49
(OSX) 13:50
Aankhen`` There is a strange circular dependency. 13:54
Makefile.PL uses PugsBuild::Config, but PugsBuild::Config needs to read from config.yml, which is generated by a Makefile target.
QtPlatypus Ick 13:56
pasteling "Aankhen``" at 61.246.147.73 pasted "Possible PugsBuild::Config patch" (25 lines, 867B) at sial.org/pbot/11909 13:59
Aankhen`` Could you please try that?
Seems to work for me.
pasteling "Aankhen``" at 61.246.147.73 pasted "Cleaned up patch" (25 lines, 880B) at sial.org/pbot/11910 14:03
Aankhen`` pokes QtPlatypus, meppl and gaal.
QtPlatypus is trying it out
Aankhen`` OK.
Bleh, I still made some mistakes. 14:04
But the patch ought to work, regardless.
QtPlatypus nods "Seems to make a diffrence" 14:07
Aankhen`` OK, then I'll commit it as a temporary workaround. 14:08
QtPlatypus Aankhen``++ 14:09
meppl it works 14:10
masak Aankhen``++ 14:13
svnbot6 r5749, Aankhen++ | * PugsBuild::Config: temporary workaround for config.yml error. 14:14
r5750, iblech++ | * PIL2JS: pil2js.pl -- Let the user choose the path to PIL2JS.js. 14:23
r5750, iblech++ | * PIL2JS: PIL::Nodes -- "my @array;" and "my %hash;" weren't working correctly,
r5750, iblech++ | because their containers pointed to the scalar undef, as a workaround, "my
r5750, iblech++ | @array = ()" had to be used. Fixed.
eric256 pastebot nopaste 14:24
pasteling nopaste
iblech perlbot nopaste
perlbot Paste your code here and #<channel> will be able to view it: sial.org/pbot/<channel>
eric256 thanks... i never remember that for some reason 14:25
pasteling "eric256" at 66.102.136.66 pasted "Make Error" (16 lines, 1.1K) at sial.org/pbot/11911
eric256 i can't get pugs to nmake..... just updated the svn and did make and thats what i got
Aankhen`` nmake seems to be working okay for me so far... 14:28
gaal BACK
oops
Aankhen`` Hiya gaal.
Hola stevan.
gaal hmm, right, that only workedforme because i already had the file in place. darn. 14:29
this build stuff is full of potholes
Aankhen`` Heh. 14:30
eric256 lol
Aankhen`` gaal, do you know where/how config.yml is used?
eric256 any idea what is going oni/ 14:31
?
Aankhen`` I see one use of it (`PugsBuild::Config->lookup('ghc_heap_size')` in Makefile.PL), but the rest don't seem to be used...
gaal look for __optimized__
Aankhen`` Where?
gaal also util/gen_prelude.pl implicitly uses it
__optimized__ is in the makefile 14:32
passed to util/build_pugs.pl
which expands it
Aankhen`` I don't see it. :-S
gaal what does your $pugs target say?
$(PERL) util/build_pugs.pl right? 14:33
Aankhen`` Um... I hate to tell you this, but util/build_pugs.pl is not showing up anywhere in the file.
Oh, one sec.
Komodo suddenly realised it IS showing up. 14:34
gaal your workaround is the correct fix, anyway
pasteling "Aankhen``" at 61.246.147.73 pasted "$pugs :" (2 lines, 2.8K) at sial.org/pbot/11912
gaal it actually used to be that way
but then i thought it was better to use make for what it's good at 14:35
Aankhen`` It's a good idea, but then how do you solve the circular dependency?
Anyway.
gaal have the module copy the defauly file when it wants it, like you do
but leave a dummy target in the makefile, so that when the config.yml fiel is edited, pugs gets rebuilt. 14:36
Aankhen`` The reason why I want to know about config.yml is that I want to figure out how to specify an installation prefix in it.
That should save me the trouble of always putting PREFIX=foo after my `perl Makefile.PL`.
gaal right. well, theoretically you could do the same kind of thing ghc_heap_size does: in the actual makefile code, perform a lookup and manipulate whatever you want. 14:37
sec
back 14:38
Aankhen`` WB.
gaal but i'm not really sure i like all this system :(
eric256 umm. did any of that have to do with my problem or was that something else? i lost the trail of thought there
gaal eric, lookin 14:39
g
Aankhen`` eric256: I doubt it's to do with your problem, since I am not able to reproduce it, so I can't really debug it.
eric256 thanks, i just didn't want to have missed the answer. lol
gaal eric, when did you last svn up? 14:40
eric256 about a minute before i posted that message
gaal made clean? :) 14:41
eric256 yep. i will agian though 14:42
gaal huh, there is something wrong there though, sec 14:43
eric256 same error after make clean
gaal svn up please
svnbot6 r5751, gaal++ | fix make unoptimized target
Aankhen`` eric256: Also, del config.yml 14:46
pasteling "Aankhen``" at 61.246.147.73 pasted "Build error: GHC heap size" (17 lines, 917B) at sial.org/pbot/11913
eric256 arg....nmake clean has a command that is too long 14:47
Aankhen`` Oh.
eric256 anyone have the link to the new nmake version?
Aankhen`` Sorry, that one was my fault.
(the build error, not eric256's problem)
gaal forum2.org/gaal/pugs/nmake.exe
Aankhen`` gaal: What variable do I set to alter the PREFIX in the final Makefile, unless it is specified on the command-line? 14:48
eric256 appears to be working...thanks gaal 14:49
gaal i don't know, it's a Module::Install / MM thing
eric256 dang i spoke tooo soon
gaal you could hack and set ENV.
Aankhen`` How does MM come in?
gaal makefile generation voodoo
Aankhen`` But where is it called? 14:50
gaal uh, if M:I doesn't use that, forget it
Aankhen`` It's not in M::I...
pasteling "eric256" at 66.102.136.66 pasted "Same nmake error, just later on." (61 lines, 5.3K) at sial.org/pbot/11914
gaal ok, then whatever it is that does honor your PREFIX today...
Aankhen`` Ah.
M::I::Makefile.
That's what uses it.
gaal fixing 14:51
svn up, eric, sorry about that. 14:52
eric256 its okay. glad to be of services ;) 14:53
Aankhen`` gaal: Sorry to keep bugging you, but do you know if EU::MM will override any `makemaker_args` calls with parameters from ARGV? 14:54
Or do the calls take precedence?
svnbot6 r5752, gaal++ | fix for make unoptimized for real
eric256 same error agian...farther down i think... 14:55
what usefull info can i give you? or do you want me to paste it agian?
gaal eric, it still doesn't work? 14:56
eric256 nope... svn up; nmake clean; perl Makefile.PL; nmake unoptimized; same error
gaal Aankhen``, not sure
do you specify PREFIX to make or to makemaker anyway? 14:57
5752? 14:58
(eric)
eric256 C:\projects\pugs\pugs>svn up 15:01
At revision 5752.
gaal worksforme :/ 15:05
ah 15:06
maybe not.
eric256 hehe
Aankh|Clone [20:24:21] <Aankhen``> Or do the calls take precedence?
[20:24:25] <svnbot6> r5752, gaal++ | fix for make unoptimized for real
[20:25:30] <eric256> same error agian...farther down i think...
-
Did I miss anything? 15:07
gaal A: 04[17:57] gaal: 01do you specify PREFIX to make or to makemaker anyway?
eric, fixing.
Aankh|Clone No... 15:08
Erm.
Wait.
gaal that's an "either or" question. :)
Aankh|Clone I don't understand.
Khisanth hmmm invisible black text :)
gaal eric256: *sigh* try again now please. 15:09
a: when do you specify PREFIX - i never use it myself and don't know - during make, or when you make the makefile? 15:10
Aankh|Clone When creating the makefile.
perl Makefile.PL PREFIX=foo
gaal sorry, i just don't know. 15:12
svnbot6 r5753, gaal++ | fix unoptimized, really for real.
Aankh|Clone OK. 15:13
15:16 Aankh|Clone is now known as Aankhen``
eric256 svn up'ing right now 15:17
eric256 begins make and holds his breath 15:18
working now. thanks 15:20
gaal cool, thanks for helping :) 15:21
Aankh|Clone I am not having a good day. 15:28
15:28 Aankh|Clone is now known as Aankhen``
gaal build systems can do that to you :/ 15:31
Aankhen`` Oh, that is just a minor niggle. 15:32
I've just been having a really bad day.
gaal oh? well in that case i have this job for you maybe you'd be interested in working on the build system? :-P 15:33
Aankhen`` Heh.
If you want help which I can provide, sure.
I'm here for a little while longer before dinner, and then I'll be back after dinner. 15:34
gaal there's not something specific - your impressions with connecting eg. the PREFIX setting to it would be valuable 15:36
Aankhen`` I am working on hooking up the `install_dir` setting from config.yml, if that's what you mean. 15:37
Just waiting for it to compile (this is the third time I'm trying, because of various related and unrelated problems).
gaal cool, good. if you see a refactoring along the road, feel free to apply or suggest it. 15:38
Aankhen`` Refactoring for what? 15:39
gaal i'm not happy with how it is now, but i don't know if it's because my code is needlessly complex, or if build systems in general are complicated
A: for build_pugs.pl + PugsBuild::Config etc.
+ Makefile.PL 15:40
Aankhen`` Ah.
I tried to understand the whole system a while back... 'twas a bad idea. :-P
gaal :)
Aankhen`` I would like to try refactoring it, but TBH, I wouldn't know where to start. 15:41
How do I escape a colon?
I want to say this: install_dir: C:\Development\Perl6
Do I just backwhack the colon?
(YAML)
gaal no need 15:42
only the first colon is significant.
Aankhen`` Ah.
vi's syntax highlighting fooled me.
Khisanth wonders if there exists any simple build system
gaal yes, the yaml mode isn't very polished. 15:44
There probably is, but it isn't portable. :-P
Khisanth between languages or platforms? 15:45
Aankhen`` Both, no doubt. 15:46
gaal seen nothingmuch 15:47
jabbot gaal: nothingmuch was seen 16 hours 15 minutes 26 seconds ago
gaal yeah, i was joking. it looks like these things get crufty but also have some legitimate concerns they can't help but have to address. 15:48
Aankhen`` Yep.
But I would think it could be refactored... after all, refactoring is not supposed to change the meaning.
gaal it's all jenga-rigged. you want to throw out something, but you find out that it provides some treatment of an obscure case in a platform you can't test on yourself, so you decide to leave it... 15:50
Aankhen`` Fun. ;-)
gaal yeah. when will nothinmuch arrive already and take my mind off this :) 15:51
Khisanth gaal: heh that seems to sound like people's description of the perl5 core :) 15:52
gaal i added my share of cruft too: I liked how ghc solved the option sorting problem: ghc +RTS -M320m -RTS --other --options 15:54
so i copied the design
but of course unluckily, i had to pass RTS paramaters myself to ghc from within an aggragated option list 15:55
so i ended up having to use a different marker for + and -
which is ugly to the extreme :(
plus, i can't escape them simply: \+RTS 15:56
because someone else "fixes" the paths: s,\\,/,g
so that gets translated to /+RTS :)
and i absolutely refuse to write code that accepts "/" as an escape character. 15:57
Khisanth this sounds extremely bad
gaal conceptually, what i need to do is serialize the various option lists 15:58
into some armor form
but the command line is a crappy transport
plus, it's insane to make command line args opaque 15:59
if anyone has any ideas for making this better, i'd love to hear
Khisanth what the heck is this being used for? 16:00
gaal pass various options around to different parts of the build system
ghc, the thing that generates the precompiled prelude... 16:01
Aankhen`` gaal: It's not parsing my install_dir... 16:02
gaal nopaste a patch? 16:03
Aankhen`` No, as in, the YAML structure, which ought to contain a string (the path), just has a 0.
gaal what's the line say? 16:04
Aankhen`` install_dir: C:\Development\Perl6
C:\Development\pugs>perl Makefile.PL
$VAR1 = {
'install_dir' => 0,
gaal looking at it... 16:05
btw "perl -Iinc util/build_pugs.pl -h"
Aankhen`` Ah.
Thanks. =)
Oh, PugsBuild::MiniYAML needs to handle escaped characters. 16:09
gaal fixed it, committing.
Aankhen`` Merci. :-D
gaal that wasn't the problem
Aankhen`` OK.
gaal the problem was that my true/false handling wasn't anchored
Aankhen`` Oh. 16:10
gaal so your string contained an "n"
Aankhen`` Yeh.
gaal and embarassingly, the mini parser normalized that as a "false".
Aankhen`` Committed?
gaal src
sec 16:11
Aankhen`` waits impatiently.
gaal up 16:12
i hope there are no more bugs soon. i have to prepare dough for the pizza :) 16:13
Aankhen`` I shall try to refrain from finding them. ;-)
OK, I managed to hook it up. 16:14
Now let me see if I can override it using PREFIX= on the command line.
gaal here, take this BugSquasher2000. may it serve you well.
Aankhen`` Yep.
gaal hold on
why PREFIX=?
Aankhen`` What do you mean?
gaal we're trying to unify the configuration 16:15
so expoty PUGS_BUILD_OPTS=PREFIX=c:\whatever:
oh crap
that won't work
Aankhen`` Yeh.
gaal because i split that var on :
tsk, tsk.
Aankhen`` Oh, thataway.
I was thinking of it in terms of how you have no `export FOO=bar` on Windows.
svnbot6 r5754, gaal++ | be more careful normalizing true/false values 16:16
gaal anyway, do what you think's right :)
i'll be back :)
Aankhen`` A'right.
I think I'll be off for dinner too.
See ya later.
gaal :) & 16:18
svnbot6 r5755, Aankhen++ | * added `install_dir` key to config.yml -- path to install Pugs in. 16:25
r5755, Aankhen++ | * modified Makefile.PL and PugsBuild::Config to handle the new key.
gustins hey :)
here is someone from germany? 16:26
iblech gustins: Yep
gustins from which city?
iblech Augsburg, near Munich 16:27
gustins it's great... it's in bavaria?
iblech Yep
gustins :)
iblech autrijus: FYI, parsing the output of -CPIL of Prelude/JS.pm: ~~30s here. The actual compile: ~~0s :)) 16:30
kev009 is perl6 anywhere near release state? 16:31
PerlJam kev009: define "near"
you give me an epsilon and I'll tell you if we're within it.
kev009 say, 6 months? 16:32
PerlJam no.
kev009 how about a year 6?
PerlJam oh ... you'll probably have to define "release" too as my idea of a release is proably different than yours
kev009 beta and or RC will be fine. 16:33
PerlJam 1.5 years? I think we *might* have a beta by then (but I'm optimistic) 16:34
kev009 PerlJam: ok. any good source to read what will be new/changed? 16:35
I've heard a lot of hype but haven't seen much material
PerlJam kev009: dev.perl.org
And you can use pugs now to try things out.
kev009 ok 16:36
PerlJam pugs is a full-featured implementation of perl6. The main downside is that it's slow.
Limbic_Region s/is a/will be/ 16:37
s/it's slow/it's currently/ 16:38
PerlJam right, right.
Limbic_Region err currently slow even
PerlJam pugs rocks though.
Limbic_Region pugs will likely end up being more full featured then the official p6 compiler
PerlJam I still get goose bumps when I write perl6 and pugs actually runs it. 16:39
Limbic_Region at least WRT backends being targeted anyway
osfameron Limbic_Region: is it definite that pugs won't be the/a official p6 compiler ?
PerlJam once we get to some form of AST, we can get anywhere :) 16:40
Limbic_Region osfameron - nothing is definate
but it certainly looks that way for a couple of reasons
PerlJam osfameron: It's the official bootstrap compiler.
iblech PerlJam: As proven by -CPIR, PIL-Run and PIL2JS :)
PerlJam yeah, PIL2JS still makes me boggle a bit :-) 16:41
Limbic_Region PerlJam - "anywhere" is a bit of a stretch - I assume you mean given any turing complete language 16:42
iblech Oh, and WRT speed: The only really slow part of current Pugs is its evaluator. The parser and -CPIL are fast :)
Limbic_Region and even then - there are some things that are just way too hard to realistically do even with turing complete languages
PerlJam Limbic_Region: how about "anywhere someone would want to go" then? 16:43
Limbic_Region I would have just said most anywhere 16:44
*shrug* - what do I know - I am just a lurker these days
PerlJam "anywhere" is close enough for me then.
measure with a micrometer, mark with chalk and cut with an axe
Khisanth iblech: but WRT to speed aren't most people concerned about the evaluator? :) 16:45
Limbic_Region chuckles at the idea of BASIC with coroutines and continuations
Khisanth - well, when PIL-PIR is complete it shouldn't be much of an issue 16:46
iblech Khisanth: Well :)
Limbic_Region: Right.
PerlJam iblech: well, the issue just gets moved from pugs-space to parrot-space really. 16:47
er, Limbic_Region
Limbic_Region there are still a lot of optimizations to be made in Parrot itself and there are certainly a large number of optimizations to be made in the PIL-PIR process
IIRC - parrot already meets or exceeds p5/python/ruby (and even C) in many benchmarks 16:48
well - I do recall correctly - the problem is my recollection is from some time ago and things change
PerlJam Limbic_Region: your recollection is pre-chip. The jury is still out on the changes that chip and leo are making. 16:49
Limbic_Region agreed
admittedly I have only been following p6.i through the summaries since late november of last year 16:50
I would be very much interested in mapping how much of the p6 language has been spec'd against what hasn't been spec'd against how much has been implemented in Pugs against how much is currently supported in PIL-PIR 16:52
ingy masak: pong 16:59
svnbot6 r5756, gaal++ | cleanups for default config.yml. Note that we need an empty makefile target 17:14
r5756, gaal++ | so that when the user edits their config.yml pugs gets recompiled. But the
r5756, gaal++ | initialization can't be in the rule because of the circular dependency that
r5756, gaal++ | would entail.
eric256_ curses his web host....3 days of odd errors with no explanations, then today it suddenly starts working agian 17:19
dudley Does PREFIX=mumble work now? 17:33
last I heard, it only partially worked. Some things defaulted to your p5 prefix regardless 17:34
gaal A checked it in; i didn't try it
Aankhen`` Uh, I didn't check in PREFIX=. 17:37
I just added an `install_dir` key in config-template.yml, and support for it in Makefile.PL and PugsBuild::Config. `install_dir` functions as an alias of PREFIX. 17:38
s/alias of/alias for/
(I got tired of always using `perl Makefile.PL PREFIX=foo` :-P)
gaal ah.
ok.
dudley Aankhen``: does that install things where you want them? 17:39
Aankhen`` On my system, yes, pretty much.
Everything except Perl6::Pugs is installed into the specified directory.
dudley Is your PREFIX different from your p5 prefix?
Aankhen`` Yes. 17:40
I'm on Win32, if that makes a difference to you.
eric256_ wonders why you renamed it install_dir instead of PREFIX 17:42
although i think yours makes more sense. lol
dudley No, there's just a warning in INSTALL about how PREFIX is broken. If it's not, I'll take that out.
Aankhen`` Ahh, I tried to make it a little more user-friendly.
As you say, `install_dir` makes more sense to me than `PREFIX`.
dudley: As I recall, I never did have trouble with PREFIX. It was the Debian (or perhaps Linux in general) crowd that had a lot of problems. 17:44
Hola nothingmuch. 17:46
eric256_ i just nevere make install....no reason to chance it ;)
Aankhen`` Heh. 17:47
nothingmuch oops 17:48
hola
Aankhen`` What's shakin'? 17:49
nothingmuch not much
eric256_ anyone here familar with the current state of pugs scopes?
nothingmuch is depressed
hand won't heal
I can't play the bass
Aankhen`` Awww. 17:50
eric256_ S06 says that global subroutines can be prefixed with * but don't need to be if it is unambiguos.... or am i reading that rwong?
nothingmuch i think that makes sense.
my sub foo { }
foo() will beat any global foo
but if there's no 'my sub foo' you don't need to say 'our sub foo' or something to use it 17:51
eric256_ but do global subs need to be declared with *? is it only optional when calling them?
while i'm at it .... wouldn't "as global" be better than a * for subs? ;) 17:53
nothingmuch i don't think so
think of 'sub name block' as a sort of macro
that compiles down to
BEGIN { &name := sub block }
eric256_ but we have "as primitive" and "as builtin", etc. 17:54
nothingmuch ....? 17:55
eric256_ i'm not sure where attributes fit in that macro of yours...are they droped into the actual block ? sub test as builtin { ... } is &test := sub {is builten; ... }
svnbot6 r5757, iblech++ | * PIL2JS:
r5757, iblech++ | * my @a = "hi" is now the same as my @a = ("hi",) (i.e., it works now).
r5757, iblech++ | * my %h = (a => 1, b => 2) and my %h = (a => 1) work now, too.
r5757, iblech++ | * Test.pm doesn't die anymore when reporting a failed tested (because it
r5757, iblech++ | accessed $?CALLER::CALLER::POSITION, which is now stubbed).
r5757, iblech++ | * Various parameters ($_, $^a) were seen as pass-by-name only because of me
r5757, iblech++ | misunderstanding isNamed. Fixed.
r5757, iblech++ | * Many new primitives and operators: &nothing, &split (Str only), &not,
r5757, iblech++ | &true, &join, &map, &grep, &sum, &infix:<..>, &infix:<^..>, &infix:<..^>,
r5757, iblech++ | &infix:<^..^>, &infix:<**>, &infix:«<=>».
r5757, iblech++ | * Updated README to reflect PIL2JS's progress.
r5757, iblech++ | * unEVALed some tests in t/operators/ so PIL2JS can run them.
eric256_ he just likes all them ++'s
nothingmuch it's not really a macro
eric256_ i understand, but then where do attributes fit in such that is global doesn't make more sense 17:56
PerlJam eric256_: Why does it have to be either/or? Can't it be "and"? 17:57
(personally, I like the use of hungarian notation in this instance though)
eric256_ oh i don't care if its both PerlJam ... i just have never realy gotten used to this punctuation... and since these cases 17:58
wont happen often as global seemed like a nice fit
but them i'm still not realy secure in my understanding of attributes, hence the questions ;) in seek of knowledge 17:59
Aankhen`` stares at iblech.
iblech Aankhen``: ? :D 18:00
Aankhen`` iblech++
Oh, and gaal++; gaal++; # for rushing to the rescue earlier =)
iblech I'm uploading the complete testsuite, compiled to JS, currently. Many tests pass :) 18:01
(Of course, even more tests fail) 18:02
eric256_ hmmm....no response...must mean its time to head off for lunch 18:11
iblech m19s28.vlinux.de/iblech/stuff/not_p...js-demo/t/ -- Our testsuite compiled to JavaScript :) 18:22
eric256_ thats kind of sick 18:23
lol
masak has finally understood the difference between "svk up" and "svk pull" 18:26
and i think i like it 18:27
iblech FYI, many tests in t/builtins/ mostly pass. t/oo/, by contrast, dies all the time. Part of the errors are because &eval is unimplemented, though 18:29
Aankhen`` Hey iblech, you're running dual AMD Athlons? 18:30
iblech Yep :) Very useful for compiling thousands of tests :) 18:31
Aankhen`` Heh.
gaal all smokers: if you have lots of ram, you can speed up smokes. in your config.yml, uncomment ghc_heap_size and precompile_modules
Aankhen`` Are those regular Athlon MPs or Athlon 64s?
gaal this makes your pugs slower to build but faster to smoke
Aankhen`` I uncommented the precompile_modules line while building. `nmake test` was smokin' fast. ;-) 18:32
Aankhen`` bangs his head on the wall.
s/ fast//
gaal you can also inline additional arbitrary modules into pugs but you need lots of ram for every one.
iblech Aankhen``: Regular Athlon MPs 2000 MP+
Aankhen`` iblech: Ah, I see. 18:33
eric256_ after chaning config.yml do you need ot perl Makefile.PL agian? 18:34
Aankhen`` eric256_: Yes.
gaal ideally no :) 18:35
Aankhen`` Well, you definitely need it for install_dir.
Dunno about the others.
BBIAB.
eric256_ i often spell unoptimized wrong. so that is a nice option to have in the config file 18:36
gaal you have it there
and the old targets still exist 18:37
or am i misunderstanting you?
eric256_ i like the option in config.yml so i don't have to always type nmake unoptimized....since i tend to spell it wrong the first time ;)
gaal aaahhh :)
oaky
yes, well it's not all bad this thing 18:38
eric256_ do modules have to be compiled at make time? or is there a while to precompile them for use later?
s/while/way/ 18:39
gaal no, we don't have dynamic loading yet 18:40
eric256_ i keep hoping...dunno why, nothing i need to precompile...it will just be fun. ;) 18:42
includeing Test.pm in the compiling is taking forever.
masak hm, wizard.p6 is still b0rken
i got stuck in an infinite loop after killing my first army of frogs
eric256_ thats new. 18:43
masak i'll see if i can find the bug
anyway, i get a lot of "army of frogs is attacking you. what will you do?" followed by "the army of frogs is dead!" all the time 18:44
Ctrl+C doesn't help, so i had to open a new terminal and kill pugs
can anyone confirm this behaviour? 18:45
svnbot6 r5758, fglock++ | * fixed tests in Set::Infinite::Functional
dudley masak: Ctrl+C worked for me, but otherwise, yes, it happens here 18:46
masak i haven't found the bug in the source, but it is either a simple error in wizard.p6 or a pugsbug 18:47
update on the wizards.p6 bug: it seems that when the frog army is shifted from the array, the array doesn't get shorter 19:05
gaal frog traces! 19:06
masak at first sight, there seems to be one shift to many, since both line 59 and line 176 do a shift
gaal: :)
s/to many/too many/
pasteling "masak" at 130.238.83.176 pasted "I can't explain this error in a one-liner, could someone help me?" (5 lines, 153B) at sial.org/pbot/11922 19:22
gaal @.b
masak d'oh! thx 19:23
gaal the error location is confus[ing|ed]
Khisanth hmm
gaal but that's a well-known issue. 19:24
pasteling "masak" at 130.238.83.176 pasted "ok, next one :/ i get a "can't modify constant item" when i shift from an rw member -- why?" (4 lines, 260B) at sial.org/pbot/11923 19:29
Aankhen`` ?eval class Test { has Int @.b is rw; method b { @.b.shift() }; my %h; %h<test> = Test.new(:b("testing", "testing")); my $b = %h<test>.b 19:34
evalbot6 Error: unexpected "{" expecting trait, ";" or end of input
svnbot6 r5759, chromatic++ | Added tests for global destruction: call DESTROYALL() on all active objects.
r5759, chromatic++ | (This is the last feature Test::Builder needs to work completely.)
r5760, chromatic++ | Allow the option of using svk, if it doesn't look like an svn checkout.
r5760, chromatic++ | Hopefully somewhat portable to other shells.
Aankhen`` ?eval class Test { has Int @.b is rw; method b { @.b.shift() } }; my %h; %h<test> = Test.new(:b("testing", "testing")); my $b = %h<test>.b
evalbot6 Error: Can't modify constant item: VUndef
Aankhen`` ?eval class Test { has Int @.b is rw; }; my %h; %h<test> = Test.new(:b("testing", "testing")); %h<test>.b
evalbot6 ['testing', 'testing']
Aankhen`` ?eval class Test { has Int @.b is rw; }; my %h; %h<test> = Test.new(:b("testing", "testing")); %h<test>.b.shift
evalbot6 Error: Can't modify constant item: VUndef
gaal nothingmuch, ping 19:37
nothingmuch pong 19:38
gaal eilara pointed out that in HTT, it's kinda silly to color the percentage box yellow on failures
nothingmuch HTT?
gaal it oughta be red. maybe differently tinted red, but red
TTH
nothingmuch oh
gaal i am low on wakies 19:39
nothingmuch i like it moving from yellow to red
depending on percentage
gaal but yellow has another meaning
nothingmuch "unexpected success"?
autrijus greetings. I'm about to sleep now; this is just a heads-up to say I'm not dead. I have merely been caught up in $work for a couple days; after that I've stumbled upon this horrible thing called the Sims. :-/
gaal so maybe have it move from pink to blood red
hello autrijus :)
nothingmuch autrijus: sims are fun to kill 19:40
the "party fire" scenario is by far my favourite
make a house
gaal autrijus: may you never meet GTA: SA
autrijus nothingmuch: indeed, isn't it. but that kept me in Win32, and that means no Pugs work ;)
nothingmuch invite all your friends
autrijus gaal: good, I don't even want to know about it
nothingmuch make 3-4 fireplaces
close off all thie doors
and put lots of furniture next to the fireplaces
sometimes if you build a house over that place, it will be haunted
gaal nothingmuch: smoker to the core.
autrijus anyway, I think I'll take drastic measure to eradicate the game and boot into the productive freebsd tomorrow 19:41
but I need to sleep now. g'nite :) *wave*
nothingmuch gaal: the i don't want to change the yellow, it works well for me
i disagree with the difference of meaning
gaal maybe bonus tests should be blue then? 19:42
anyway the colors shouldn't be the only cue
i gave the TTH spiel to quite a few people by now.
some are colorblind
:-/ 19:43
nothingmuch if they are colorblind they should just use the regular test harness output
or find a wayh to denote stuff with symbols, or shades of grey
red and green blindness is by far the most common, if my facts are right
so this is a big problem
gaal i think the scaled summary is cute but not really useful 19:44
svnbot6 r5761, Stevan++ | Perl6::MetaModel - (Perl 5)
r5761, Stevan++ | * ::Class & ::Role no longer call $class_name->meta which helps in breaking circularity
r5761, Stevan++ | * ::Object is being broken down in an effort to break circularity as well
r5761, Stevan++ | * ::MetaModel same thing here,... tweaks and hacks to break circularity
r5761, Stevan++ | * ::MetaClass, just removing a comment which is no longer relevant
r5762, Stevan++ | org.perl6.metamodel - (yes, thats right, it's Java)
r5762, Stevan++ | * first stab at the MetaModel in Java, very primative still
r5762, Stevan++ | ** please don't hate me ;) **
nothingmuch scaled summary?
gaal without it we can supply alternate CSS
whoa stevan :)
stevan :) 19:45
nothingmuch gaal: TTH is all about cuteness
it's a way to get on top of lots of data
and it does it by being visually attractive
stevan gaal: I waas going to do C#, but my Mono install is broken
gaal candy ne effective
not necessarily at least 19:46
C#?
oh
stevan it is an experiment in the solidity of the metamodel 19:47
gaal nothingmuch: i should get you that book by the graph guy
stevan Java is not as forgiving as P5, JS
nothingmuch graph guy?
gaal first i have to remember its name
and his name
there's a really good book about visual presentation
stevan Tufte?
gaal yeah, i think so 19:48
stevan Edward Tufte
gaal yes
nothingmuch is frustrated
every time someone new sees TTH i suddenly have to get educated 19:49
and this results in 10 people editing CSS
and then I have to merge all that
and eventually the change isn't even noticable
eric256_ hehe /me comes back from lunch to find that the build failed because of exhausted heap. oops
gaal isn't the solution to that the commit bit? :-P 19:50
nothingmuch nope, because there are conceptual conflicts
gaal eric, uncomment ghc_heap_size
nothingmuch people who are unhappy with what the color-czar says
which I think was stevan, theorbtwo and I a while back
eric256_ or recomment precompile_modules: [ ext/Test/lib/Test.pm ]
nothingmuch should just replace what they want to replace.
by editing the CSS or subclassing or whatever
gaal eric, if you do precompile test your smoke will fly 19:51
nothingmuch gaal: how do I precompile test?
eric256_ i'll give it another go then... just in the name of science
gaal nm: vi config.yml
brb
nothingmuch restarts smoke loop
gaal (nm: rename TTH as Bikeshed. then people will literally be able to color it.) 19:52
nothingmuch =) 19:53
Khisanth ?eval '«'..'»'
evalbot6 Error: unexpected "\\" expecting term postfix, operator, postfix conditional, postfix loop, postfix iteration, ";" or end of input
eric256_ and you would expect?
i'm guessing that looks different with unicode support. lol 19:54
Khisanth heh
it actually works, just not with evalbot
eric256_ i wonder if evalbot is unicode friendly
iblech ?eval "äöü"
evalbot6 *** unexpected "(" expecting term postfix, operator, postfix conditional, postfix loop, postfix iteration, ";" or end of input at -e line 8, column 4
nothingmuch nope 19:55
that sucks
Khisanth where did the ( come from?
nothingmuch IRC is byte leveel
haskell is unicode friendly
pugs uses haskell strings
iblech Khisanth: Probably from examples/network/evalbot/evalhelper.p5
nothingmuch so my conclusion is that haskell expects utf-16 or something
Khisanth ['«'..'»'] didn't seem to work in ghci 19:57
what is the $ operator called in Haskell? 19:59
iblech Function application 20:00
f $ x = f x
f (long expression, parens needed) = f $ long expression, parens can be dropped thanks to the $ operator 20:01
Khisanth hrm so . is like using $ twice? 20:07
iblech (f . g) x = f (g x) = f $ g x = f $ g $ x, so yes 20:09
Khisanth thanks 20:10
iblech np
Aankhen`` G'night. 20:17
svnbot6 r5763, fglock++ | * several improvements and more docs in Set::Infinite 20:40
r5764, fglock++ | * more tests in Set::Infinite
sili are those commit comments? 20:41
:/ 20:42
vague++
Khisanth yes, you didn't know?
sili rhetorical
Khisanth hrm there is no test of :by 20:43
iblech There is 20:44
Sec
t/operators/adverbial_modifiers.t 20:45
Khisanth what else can it be used with besides the range op? 20:46
svnbot6 r5765, putter++ | Another cm of progress on modules/Grammars/rx_grammar.pl.
iblech No idea 20:47
Khisanth how about working with chars and strings? 20:48
putter f:x(1)(2,3) => f(x=>1,2,3) or something like that. doesnt work yet. (which makes emulating rx a pain;)
eric256 perlbot nopaste 20:58
perlbot Paste your code here and #<channel> will be able to view it: sial.org/pbot/<channel>
putter & 21:02
pasteling "eric256" at 66.102.136.66 pasted "Just me agian" (14 lines, 1.1K) at sial.org/pbot/11927 21:13
eric256 i'm very talented at getting errors during make today
what a proud skill that is ;). The paste is an error i got while nmake unoptimized with Prelude and Test pre compiled 21:14
iblech eric256: Ooh! You found a bug in GHC! 21:16
(I think.)
svnbot6 r5766, iblech++ | * More unEVALs and minor test fixes in t/subroutines/. 21:18
r5766, iblech++ | * Usual svn properties.
r5766, iblech++ | * PIL2JS:
r5766, iblech++ | * ?(?0) no longer returns true (in JavaScript, all objects are true -- even
r5766, iblech++ | new Boolean(false)).
r5766, iblech++ | * Native JavaScript functions may safely return arrays now.
r5766, iblech++ | * Initial support of &?SUB and &?BLOCK.
r5766, iblech++ | * $*OS = "browser".
r5766, iblech++ | * Arity reporting is sometimes wrong (this is an (already tested against)
r5766, iblech++ | pugsbug), so &statement_control:<for> and &map no longer refuse to work
r5766, iblech++ | with 0-ary functions, but instead simply call them and hope they actually
r5766, iblech++ | accept at least one param.
r5766, iblech++ | * Fixed &substr and &split to autocoerce their arguments into Strs.
r5766, iblech++ | * Fixed pil2js.pl to accept files ending in .p6 as Perl 6 sources.
eric256 i'm not sure that excites me 21:21
lol
svnbot6 r5767, iblech++ | Test.pm -- Skip checking using junctive &infix:<|> when running under PIL2JS.
r5767, iblech++ | This is so tests don't die if one of their subtests fail.
eric256 trying it without Test.pm precompiled to see if the trouble is there or with the heap size.. 21:22
yep. works fine without Test.pm included 21:24
that must suck up the memory
iblech Seems so
eric256 @.monsters.perl.say; my $x = shift @.monsters; @.monsters.perl.say; 21:31
any ideas why that prints the same thing beofre AND after the shift? 21:32
hmmm.... @.monster is a property of an object, does shift not work on the proxy object that is being used?
?eval class Foo { has @.test }; my $x = new Foo; $x.test = (1,2,3); $x.foo.say; 21:33
evalbot6 Can't exec "pugs": Text file busy at examples/network/evalbot//evalhelper.p5 line 42.
eric256 ?eval class Foo { has @.test }; my $x = new Foo; $x.test = (1,2,3); $x.foo; 21:34
evalbot6 Error: No compatible subroutine found: "&foo"
eric256 ?eval class Foo { has @.test }; my $x = new Foo; $x.test = (1,2,3); $x.test;
evalbot6 [1, 2, 3]
eric256 ?eval class Foo { has @.test }; my $x = new Foo; $x.test = (1,2,3); shift; $x.test; $x.test;
evalbot6 Error: unexpected ";" expecting term
eric256 ?eval class Foo { has @.test }; my $x = new Foo; $x.test = (1,2,3); shift $x.test; $x.test;
evalbot6 [2, 3]
eric256 or not
svnbot6 r5768, Stevan++ | org.perl6.metamodel - (Java version) 21:37
r5768, Stevan++ | * MetaClass now supports superclasses and can calculate MRO
r5768, Stevan++ | - added test for MRO calcs
eric256 hmmm. definitly works in tests... must be missing something 21:38
i think i'm going crazy now. 21:40
svnbot6 r5769, eric256++ | wizard.p6 - debug info added,bug shifting off of an Objects property 21:43
eric256 can someone run the updated wizard.p6 for me? and confirm that the debug output is the same all three times... i've gotta be missing something obvious but this is all right there in the middle of a method.... 21:45
?eval class Foo { has @.test }; my $x = new Foo; $x.test = [1,2,3]; shift $x.test; $x.test; 21:46
evalbot6 [2, 3]
eric256 hmm thoguth maybe some () versus [ ] thing.....i feel all alone here. ;)
pasteling "iblech" at 80.81.17.170 pasted "eric256: Debug output" (15 lines, 1.1K) at sial.org/pbot/11929 21:47
eric256 well its not just me
see in that output how @.monsters stayed the same even after shift @.monsters ?
i still havn't managed to recreate it with smaller code though, so something is going on 21:48
thanks for running that BTW ;) 21:50
svnbot6 r5770, iblech++ | * Usual svn props.
r5770, iblech++ | * unEVALs in t/subroutines/ and t/var/type.t.
r5770, iblech++ | * PIL2JS -- Initial support for always-existant $_.
eric256 hmm. turns out its an issue with the constructor 22:03
?eval class Foo { has @.test }; my $x = Foo.new( :test(1)); shift $x.test; $x.test; 22:07
evalbot6 \1
eric256 ?eval class Foo { has @.test }; my $x = Foo.new( ); push $x.test, 1; shift $x.test; $x.test;
evalbot6 []
svnbot6 r5771, iblech++ | Many more unEVALs in statements/ and syntax/. 22:24
r5772, eric256++ | * added tests for bug with OO contructor + list attributes
r5773, eric256++ | * fixed test count in attribute_list.t"
r5774, iblech++ | * Usual svn props. 22:30
r5774, iblech++ | * One more unEVAL: t/general/atoms.t
r5774, iblech++ | * PIL2JS: &last, &next, and &redo work in while and until loops now, too.
r5774, iblech++ | * PIL2JS: Implemented &next (but not yet working for &for, only for the other
r5774, iblech++ | loop constructs).
eric256 eval is evil 22:31
iblech Not generally, but many tests don't have a chance of passing in PIL2JS as PIL2JS doesn't have &eval 22:32
And eval(...)s slow down execution
eric256 are those tests that are fixed now? or do you just have a better way around evals? 22:33
iblech Right, most tests have long been fixed. 22:35
(And the unEVAL was simply forgotten)
eric256 ahh 22:38
wolverian there's a Java backend now too? 22:39
iblech wolverian: No, only the beginning of a port of Perl6-MetaModel to Java 22:41
Hm... PIL2JS passes more tests I thought it would :) 22:42
wolverian well, that's still something!
iblech uploads the testsuite compiled to JS again
m19s28.vlinux.de/iblech/stuff/not_p...js-demo/t/ 22:50
Ok, need to sleep now :) 22:51
Night all :)
QtPlatypus Just to confirm, At the moment there is no way to get at a code's PIL from withing Pugs. And there is no function corrasponding to "show" from src/AST/Internals.hs 23:13
If that understanding is incorrect, please correct. 23:14
Excalibor greetings 23:27
stevan wolverian: the Java backend is an experiment 23:40
in porting a highly flexible object model to a highly restrictive runtime 23:41
it might also come in handy with the JS backend since JS and Java can (in some cases) talk to one another easily
QtPlatypus: I think you can do print as PIL 23:42
QtPlatypus: IIRC, that is how the backends work currently, but calling Pugs to parse the P6 and export PIL, which they then parse and run in the runtime 23:43
23:46 Maddingue__ is now known as Maddingue

Powered by Cro and the Raku® Programming Language.

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

Thank you!