The Return of the Journal : pugs.blogs.com/ | pugscode.org | pugs.kwiki.org | paste: paste.lisp.org/new/perl6 or sial.org/pbot/perl6
Set by GammaRay on 31 December 2005.
00:25 Southen joined 00:29 scook0 joined
PJF Somewhat off-topicly, for but the antipodeans in the room, perl.net.au/ is coming off beta soon, and discussions are held in #PerlNet . 00:31
01:13 wilx joined 01:27 frederico joined 01:45 DaGo joined 02:04 elmex joined 02:09 sapper joined 02:26 stennie left 02:27 stennie joined 02:49 SamB joined 02:53 verve_ joined 03:02 vel left 04:00 zuulvin_ joined 04:46 Amnesiac joined 04:56 Khisanth joined
audreyt nnunley: say again? 05:06
nnunley: what are you trying to do with parrows? 05:07
rafl: YAML 0.49_70 is on CPAN; I think 0.50 will be released in this hackathon
05:08 whiteg joined 05:11 whiteg joined 05:12 buu joined 05:48 Khisanth joined 05:49 DaGo joined 05:53 Aankhen`` joined 06:10 Khisanth joined
audreyt goes to zhubei for hackathon & 06:15
gaal have fun! 06:17
audreyt :)) 06:18
gaal: how's YAML::Syck? will you (or did you) have cycles to work on that for the next ~4hrs
if not I'll give it a go
gaal I have something that should work but doesn't :( 06:19
on two counts it doesn't:
1. the perlapi isn't returning the right blesstype (obviously i'm using it wrongly) 06:20
2. the tag isn't emitted, though I'm doing the same thing as in .Hs!
I
'll send you the patch; I have to go to $work now
06:20 nothingmuch joined
gaal btw, the implicit "e" arg in the .hs source was very useful for me for learning, but I think of maybe putting it back to an explicit arg because it isn't doing something special there and someone who didn't follow your lesson might find it confusing :) 06:22
sent 06:26
06:28 khisanth_ joined
gaal also, I was wondering if there was a golfier way to patmatch against records? the last change meant some of the code looks a lot clumsier :( 06:30
audreyt ok, please mail me the patch 06:31
like, in 5mins, if possible :)
yeah, the implicit ?e should go
oh, I didn't parse the "sent"
patmatch against records is verbose since hs doesn't have punning; but you can match it asif it's positional. 06:32
i.e. for data Foo = MkFoo { foo :: Int, bar :: Int }
you can patmatch it as
f (MkFoo x y) = ...
totally ignoring the label bit
also, YAML's emitting of undef as completely null 06:33
is valid and legal
gaal really?
audreyt yup
checked with ingy
gaal interesting. but I like ~ better (visual cue)
please tell me if you got it, sometimes my outgoing mail sucks. 06:34
audreyt it's a matter of taste
I got it
ok, gotta run now
ciao!
gaal ok (anyway the tick with null is parsing it, too :-)
enjoy
positional patmatches feel like tighter coupling than is warranted.. is there a meta-_ character that means "I don't care about the rest of the positional fields?" 06:36
06:55 mjl69 joined 07:20 nothingmuch joined 07:30 lisppaste3 joined 07:37 loftcity joined
nnunley audreyt: I'm trying to get the -g option working under PIL. 08:20
gaal (not that named field changes don't change the ABI... but at least that only requires recompilation, not source code changes) 08:35
08:39 nothingmuch joined 09:14 kakos joined 09:33 bernhard joined 09:40 iblechbot joined 09:45 whiteg joined 10:16 chris2 joined 10:20 kane_ joined 10:22 GeJ joined 10:40 whiteg joined 10:46 kanru joined 10:57 kanru joined
gaal audreyt: so, emitting the tag doesn't work because we (oughtta) find out the blessing at the reference level, but that's where we hit 'case SVt_RV' and immediately recurse into perl_syck_emitter_handler again. too bad there's no way to pass an implicit arg here too :) 10:58
as for why sv_reftype returns 'REF' and not the blessed class... I'm still pondering that 10:59
ugh, an ugly hack is to take advantage of the fact we have a void pointer and pass a struct with the class tag in the case of a blessed reference, and a real SV in the other cases. the handler can look for a magic cookie to determine what input it received. I don't believe I just suggested this. 11:01
11:01 xinming joined 11:02 kanru joined
gaal ...memory management of this is simple at least; it can all be done with an automatic variable 11:16
audreyt gaal: yo 11:45
I've fixed the tag problem
via extra emitter bonus field 11:46
clkao waits for miyagawa
audreyt clkao: @ airport? 11:47
clkao hai 11:48
nnunley Konnichi wa clkao-san 11:59
gaal ah, the bonus field is certainly the right way to do it :)
but did you find out why the sv_reftype was being weird?
clkao nnunley: nihao
gaal wow, I actually understood what you said. 12:00
nnunley About the most Japanese that I can write. :/
audreyt gaal: yes, sv_retype should test the SvRV
gaal nihao's in Chinese, and I got that too :)
audreyt not the sv itself
now, the hash style thing, I still don't have a clue. 12:01
clkao audreyt: i need gugod's address
gaal there's an enum, maybe we just need a different value?
or does that control inlining?
audreyt clkao: 勝利五路31號 12:02
clkao cant resad
audreyt gaal: I have no idea
clkao: they are html
clkao orz
gaal hmmm magic is stored in the hash, not the reference. ok, makes sense
so the bless keyword is highly magical 12:03
i mean of course it is
but more so than obvious
audreyt perl -e 'print "\x{52dd}\x{5229}\x{4e94}\x{8def}31\x{865f}"'
clkao: better?
clkao ya got it already
audreyt gaal: svn up svn.kwiki.org/audreyt/YAML-Syck/
gaal and address is five characters? so cool
wah!? 12:05
*** glibc detected *** double free or corruption (fasttop): 0x082cd2c0 ***
at the end of 'make'
audreyt yeah, there might be an double free going on 12:06
gaal urp, not at the end of make, at runtime.
clkao got him!!
audreyt attempted a fix. 12:07
clkao: oooh!
gaal Safefree == ?
audreyt gaal: Safefree = free
clkao audreyt: dinner yet?
audreyt see r67
clkao: I ate a lot of cookies
clkao: gugod went out to buy some food material
but not proper dinner yet
clkao oh ok. so we will see meet you directly
audreyt ingy's working on releasing YAML 0.50 -- we've got the callback API figured out and YAML::Syck is going to roundtrip well with taguri 12:08
yup
12:09 Alias_ joined
Alias_ Anyone know who set up the logging bot? 12:09
leo obra: ping 12:10
gaal audreyt: can't the safefree+NULL always be done in cleanup? 12:12
audreyt gaal: no, because HV/AV can't propaget their tags to elems
*propagate 12:13
gaal ....do they need to? 12:15
12:15 elmex joined
audreyt they must not 12:15
gaal just the syck_emit_seq / _map needs to be tagged
audreyt if you freeze at end of cleanup, they will
gaal oh, of course, we're reusing the bonus. duh. 12:16
but that's where the double free is happening then, no?
hmm, no. 12:17
Safefree == free + double free guard? then in cases where there *isn't* a tag it'll be freed twice. 12:19
should be "if (tag) free (tag)" 12:20
12:22 r0nny joined
gaal heh. running this in gdb, of course there is no double free :-( 12:23
Alias_ ping?
erm.. guys, that log bot? Anyone know what it is and where I can get one? 12:24
gaal Alias_: I don't know, sorry. maybe talk to the colabti admin? alternatively, iblech runs a private bot, ask him for a copy of the code? 12:28
Alias_ colabti?
gaal Alias_: colabti.de/irclogger/irclogger_log/perl6
ah, a hop and a half away: colabti.de/IrcLogger2.html 12:29
Alias_ aha, found it
yep
12:33 iblechbot joined
audreyt okay... enough YAML for now 12:38
ingy threatens to port PArrows to perl5 and have a working/robust/stable Rules implementation
on perl5
that's probably the next h'thon item
audreyt goes back implementing PILN repr types and the Code type
gaal maybe ingy wants to come to .il in feb?
audreyt I don't know; I'll ask 12:39
Alias_ So is Ingy doing the "travelling around the world" thing too? 12:40
audreyt apparently so
12:52 frodo72 joined 12:55 whiteg joined 13:08 azuroth joined
avar made a megapugs hack 13:12
well
not exactly
audreyt mm?
miyagawa arrived :) 13:13
avar avar.lir.dk/mw/HEAD/wiki/Perl_6 <- perl 6 pastebin
using the syntax highlight from util/perl6.vim
audreyt ooh
avar \o/
rafl audreyt: Great. 13:16
azuroth any idea how to get `.nsv` streams to play under linux?
apparently google does. 13:17
avar so, uh, cool?;)
nnunley avar: Pretty cool. 13:19
avar I'm still trying to find a way to make vim work using stdin / stdout 13:22
i.e. echo food | vim ... - > yummie
so I'm just using temporary files now 13:23
13:23 whiteg joined
gaal avar: I think you can't, but -f and --cmd might help. 13:28
yay l33t :) 13:30
avar shell_exec( "vim -u {$this->mVimrc} -e +'setf perl6' +'run! syntax/2html.vim' +':w {$this->mOutFile}' +':qa!' {$this->mInFile}" );
just using that right now
azuroth how annoying. xine reckons it can handle it, but it doesn't seem to be able to 13:35
13:38 whiteg_ joined 13:39 ilogger2 joined 13:40 ilogger2 joined 13:41 mjl69 joined 13:44 PJF left, stennie joined 13:46 alexe_home joined
gaal azuroth: maybe mplayer? 13:51
13:51 kane_ joined, G2 joined
obra leo: pong 13:59
leo obra: it was re META.yml - but too late for this release 14:08
obra oh no :/ 14:09
leo well, do you have a working meta.yml?
obra well, it sounds like andreas can fix it for us.
oh. does parrot have _no_ meta.yml?
leo exactly 14:10
obra ah.
nod. ok.
we can fix it next month
leo well - if there is a quick solution, I can still include it in 0.4.1
clkao obra! 14:11
obra isn't enough of a META.yml wizard to definitely get it right
clkao: hi!
leo I think we should discuss this on #parrot
obra and broken metadata is scarier than no metadata
and andreas can fix it in PAUSE
leo ok
obra we just need to get him a list of what parrot has that it shouldn't 14:12
leo yup - please send a TODO to p6i
gaal modulo not working, I have YAML round-trip working.
so audreyt, when you get a moment, I need some explanation of IVars :) 14:13
obra leo: done 14:17
leo thanks
gaal I modulorized not working. Roundtrip YAML works now, committing. 14:24
svnbot6 r8607 | gaal++ | * eval :lang<yaml> deserializes pugs objects. 14:26
gaal this probably breaks horribly with fancy classes :) 14:28
?eval class Foo; has $.x; my $x = Foo.new(x=>"xxa").yaml; say $x; my $y = eval $x, :lang<yaml>; say $y.yaml
evalbot_8606 OUTPUT[--- !pugs/object:Foo x: xxa pugs: unexpected node: <Scalar>
gaal here, '07...
?eval class Foo; has $.x; my $x = Foo.new(x=>"xxa").yaml; say $x; my $y = eval $x, :lang<yaml>; say $y.yaml 14:30
evalbot_8606 OUTPUT[--- !pugs/object:Foo x: xxa pugs: unexpected node: <Scalar>
azuroth ?eval say "sarah smells";
valbot_860 is now known as evalbot_8607
evalbot_8607 OUTPUT[sarah smells ] bool::true
gaal ?eval class Foo; has $.x; my $x = Foo.new(x=>"xxa").yaml; say $x; my $y = eval $x, :lang<yaml>; say $y.yaml 14:31
evalbot_8607 OUTPUT[--- !pugs/object:Foo x: xxa --- !pugs/object:Foo x: xxa ] bool::true
gaal tada!
azuroth wow, it must be true
Alias_ ?eval say "azuroth suxx0rs"; 14:32
evalbot_8607 OUTPUT[azuroth suxx0rs ] bool::true
Alias_ ;)
gaal ?eval say "everything you read"
evalbot_8607 OUTPUT[everything you read ] bool::true
azuroth you'll pay for this Alias_ 14:33
gaal praises Syck and why
azuroth sar 14:36
ul8o
14:37 azuroth left 14:57 Limbic_Region joined 15:02 bsb joined
audreyt gaal: pong 15:09
?eval my $x; $x = \$x; $x.yaml 15:13
evalbot_8607 "--- \n<ref>: ~\n"
15:15 nothingmuch joined
avar ?eval my Num $x = 5.5; $y = \$x; $y.yaml 15:19
evalbot_8607 Error: Undeclared variable: "$y"
avar ?eval my Num $x = 5.5; my $y = \$x; $y.yaml
evalbot_8607 "--- \n<ref>: 11/2\n"
nnunley 5.5 is being converted to a fraction? 15:22
audreyt yeah, bad idea 15:23
since I don't think yaml supports rationals natively
15:30 lisppaste3 joined
audreyt gaal: hm, where did you parse the tags stuff in syck load? 15:38
I can't seem to find the relevant line.
ahh, syckNodeTag 15:41
gaal rehi 15:50
hello. :)
I was looking at serializing rules previously, but was interrupted.
audreyt yo 15:51
want to take a stab in adding parsing tags to YAML::Syck? :)
gaal sure
audreyt k. I'll release 005 15:52
gaal btw please inspire the haskell' folks to add some kind of syntax making patmatches against sting prefixes less clunky :)
Alias_ 005? 15:53
audreyt ("foobar", rest) <- splitAt 6, string
s/,//
gaal: isn't that good enough?
Alias_ Is this a versioning system based on spy code numbers?
audreyt Alias_: no, it's YAML::Syck 0.05
now with object dumping and no memory leaks
Alias_ ah :)
gaal hey, but did you figure out the double free? I didn't 15:54
Alias_ Remind me again why YAML stopped being for outputting simple reports your secretary could read?
audreyt gaal: I added tag=NULL
Alias_ :)
audreyt Alias_: well, if your data type is simple, it still does that :)
Alias_ riiiight
audreyt like, if you only use the JSON subset of YAML.
Alias_ Did someone solve me "prove that a file only contains the JSON subset"
my
gaal one serious issue with yaml is that it serializes badly to dead trees 15:55
Alias_ For the purposes of sanity checking YAML data
->load( $filename, json_subset => 1 )
gaal (... think "page break") 15:56
audreyt: so can i svn up?
audreyt gaal: sure
Alias_ gaal: sure it does, within sufficiently simple subsets
audreyt Alias_: sure... it's called "dump with JSON.pm ;)" 15:57
Alias_ If you like UGLY
The thing that most attracted me to YAML was readability, not the hardcore serialization aspect
Speaking of which, have you solved the Storable problems yet? 15:58
gaal "I have this excellent compression algorithm, if you happen to have data that is equivalent to what I had in mind"
gaal needs to $work a bit... audreyt, can the tagparse stuff wait a few hours or did you want to get 005 out your-today? 15:59
s/equivalent/equal/ 16:00
re: splitAt: (a) I don't like having to count and (b) does the compiler generate the same code it did for an explicit 'a':'b':'c': patmatch? 16:01
anymoose: gotta go, bbiab :b<~2hr>
&
audreyt gaal: no, probably not. hm 16:02
Limbic_Region audreyt - were you only joking when you journaled that objects would be done this weekend? 16:06
or just that one aspect of objects?
16:07 marmic joined
audreyt Limbic_Region: no, objects as specced as S12 16:08
Limbic_Region: but I've been distracted into YAML hacking... will go back to objects in a few mins
Limbic_Region heh 16:09
s/distracted/conned/
?eval ([+] 1..100).yaml 16:13
evalbot_8607 "--- 5050\n"
Limbic_Region heh - cool
?eval { "this is $^weird" }.("cool").yaml 16:14
evalbot_8607 "--- this is cool\n"
Limbic_Region sweet
Juerd_ ?eval [].yaml 16:15
evalbot_8607 "--- []\n\n"
Juerd_ Sometimes things just make sense.
16:21 aufrank joined
aufrank ?seen schwern 16:21
gaal i am teh eevil audreyt distraxx0r lol
aufrank are they fun distractions? 16:22
16:22 Alias_ left
audreyt gaal: I implemented blessed obj loading ;) 16:22
gaal too fast for me, as usual :) 16:23
audreyt nah, because clkao pressured me into it
that's the thing about hackathons
clkao bwahaha
audreyt clkao: threatened to switch svk to Y::Syck
clkao now, what do you need from me
audreyt per-checkout .svk ;)
clkao mm 16:24
audreyt gaal: do you have any inkling about the hash map style?
the hs call Just Worked
I can't seem to figure out how it differes from the p5 one
clkao: YAML::Syck 0.07 out.
clkao k
audreyt 3 releases in an hour ;)
aufrank nice work, any parrot people in the room 16:29
audreyt leo++ 16:30
aufrank oh yeah, that's how we do it
leo++
gaal sorry, i keep getting abducted by cows. 16:32
leo thanks folks
gaal audreyt: is YAML::Syck's syck the same version as pugs'? 16:36
why mentioned he's realeasing a new version soon, incidentally.
the API exercise for emitting looks precisely the same to me, so only an implementation difference can explain it.
audreyt gaal: I took from svn snapshot 16:37
so probably a bit delta
let me look at cping from pugs
gaal either that, or he's playing tricks on us: if (caller eq 'audreyt') { play_tricks }
geoffb Have to keep audreyt amused somehow, she clearly doesn't have enough things to work on ... 16:38
audreyt gaal: yeah, reverting to the pugs version works
okay, releasing 0.09.
gaal whee :) 16:39
aufrank syck is a yaml parser, is that right?
gaal +emitter 16:40
yeah
aufrank k
audreyt 0.09 released.
gaal ?eval my $x = "hi aufrank".yaml; say $x; my $y = eval $x :lang<yaml>; say $y 16:41
evalbot_8607 OUTPUT[--- hi aufrank hi aufrank ] bool::true
gaal dances the round trip
audreyt gaal++
gaal audreyt++ for $releases_per_hour 16:42
aufrank nice!
gaal audreyt: also does objects, and simultaneously available at a CPAN near you for p5. ("after a pause", as leo++ so punningly put it) 16:44
audreyt very punny
-Opun
aufrank groans
audreyt *rimshot* 16:46
gaal oooh! rxRule is a String! I see rule serialization coming soon :)
aufrank we can pass around grammars in yaml 16:47
gaal i bet that'll change sooner or later though.
anyway, this looks like an excellent time to go home. see you later :) & 16:48
aufrank audreyt: what representation do you think rule objects will use? p6opaque?
bye gaal
audreyt aufrank: probably p6opaque, yes 16:49
aufrank does each representation define its own .yaml method? 16:50
or do some representations do a yamlizable role? 16:53
audreyt probably they need to be able to coerce into YamlNode 17:05
aufrank so is eval $x :lang<yaml> the same as eval $x.as<yaml> ? 17:07
Limbic_Region ?eval .25.yaml 17:12
evalbot_8607 "--- 1/4\n"
Limbic_Region ?eval 1.2.3.yaml
evalbot_8607 Error: unexpected "." expecting exponent, term postfix, operator, postfix conditional, postfix loop, postfix iteration, ";" or end of input
audreyt aufrank: in the future, yeah, I think so ($x as YAML::Node) 17:31
oh wait, you mean $x is a string?
audreyt is confused
miyagawa: svn.kwiki.org/audreyt 17:38
miyagawa: going to release JSON::Syck just for fun? :) 17:40
miyagawa maybe, yeah :9
audreyt that will contribute toward your missing 13 modules
miyagawa right, exactly
gaal rehi 17:47
ooooh automatic yaml coersion
or coercion. now we have to learn *two* languages.
audreyt XD 17:49
aufrank is cleaning house for the imminent return of my $girlfriend 17:52
gaal the ruby convention of letting user-defined objects live in a separate tag namespace appeals to me
aufrank oh, wait
that would make her local to this block, right? 17:53
gaal vs. p5's !perl/CGI
aufrank my $*girlfriend, I guess?
gaal you could say our $girlfriend, but that doesn't look too good.
aufrank she is classy, though
gaal has $.girlfriend 17:54
audreyt gaal: hm, how does ruby dump?
gaal !rub:object/CGI
*ruby
I made pugs do that too
audreyt that is correct 17:55
gaal let YAML::Syck do !perl
audreyt however, p5 doesn't really work that way
!perl/@CGI
gaal for compat with YAML.pm
audreyt means an array ref blessed into CGI package
there's nothing object-ish about that
aufrank ruby calls it marshalling, right?
audreyt since in perl5, every thing is not an object.
gaal well, do Dump(CGI->new)
lol # "since in perl5, every thing is not an object."
money is no object! 17:56
Juerd_ I object.
gaal romance languages, ready, set, go! 17:57
aufrank gaal: I think the pickaxe book has an example of marshaling and YAML
gaal I subject (you to trite puns)
17:57 justatheory joined
gaal I think ruby's syck can install symbols right into the namespace. ahh, joy, in p6 that'll be interesting again. 17:58
aufrank yeah, gaal it's page 414-417 in the pickaxe.
gaal I thought I was taking a break from the crazy exporter stuff and now it's chasing me!
aufrank: thanks, is that book online? 17:59
gaal is only picking up ruby now
aufrank gaal: no, not online unfortunately 18:01
but google seems to think that syck will eventually be included in ruby, if it's not already
www.rubygarden.org/ruby?YamlInRuby
gaal I believe it is. 18:02
chris2 syck is included since 1.8.0 18:03
aufrank gotcha
chris2 and it heavily fuds inside ruby's symbol table :|
gaal bbiab & 18:04
audreyt bbl too & 18:18
rafl is looking for testers for the parrot 0.4.1 Debian packages: files.perldition.org/debian/ 18:26
Juerd_ pakket-architectuur (i386) komt niet overeen met het systeem (powerpc) 18:29
;)
rafl Juerd_: I can upload the source package as well, if you like. 18:30
Juerd_ It's not a matter of me liking or wanting, it's a matter of you seeking testers :)
rafl Juerd_: Uploaded. 18:35
18:37 lisppaste3 joined
Juerd_ rafl: Compiles and installs 18:43
And runs
rafl: Didn't have libicu34-dev, as Ubuntu doesn't provide it. It's in Build-depends, though. 18:44
18:45 Limbic_Region joined
rafl Juerd_: That's fine. I want the Debian version to be linked against icu. 18:55
Juerd_ What is it? :)
rafl International Components for Unicode library 18:56
Juerd_ Sounds scary
rafl Did your camera already arrive, btw?
It is, of course.
Juerd_ No, not yet
I'm hoping it arrives the coming four days, because I want to use it Friday :) 18:57
Limbic_Region Parrot's continued dependency on ICU for unicode support is unfortunate
rafl Limbic_Region: Why? 18:58
leo Limbic_Region: perl Configure.pl --without-icu
Limbic_Region well, for one its requirement of c++
the original plan for Parrot was to be very basic ANSI C compliance - for portability sak3e
rafl Juerd_: I guess so. I told my housemate to send it three or four days ago because I wasn't at home. I guess it'll arrive soon. 18:59
Limbic_Region for another, the plan was that Parrot's Unicode interface would be very generic allowing many/any unicode library to be used
and AFAIK, that hasn't happened
leo Limbic_Region: we have reduced the need for icu vastly - if it isn't there all compiles and tests too
Juerd_ As proven by my system :)
Limbic_Region leo - and still retained unicode support?
Juerd_ Who needs unicode ;) 19:00
leo sure, if ICU is installed and used, you get enhanced functionality
Juerd_ leo: Isn't it better for a VM to provide a single stable interface?
Limbic_Region so then has Parrot been building in its own unicode library and not just an API to external libraries? 19:01
Limbic_Region guesses he should follow Parrot development more than just the summaries
Juerd_ That isn't so platform dependent that things differ even based on whether a certain lib is available...
leo Limbic_Region: no, w/o ICU there is no unicode support, that's it
Limbic_Region right - so my original sentiments stand 19:02
Juerd_ leo: But don't we *always* want unicode support, and not depend on something external for it?
Limbic_Region not that making Parrot compile without ICU isn't a step in the right direction
leo patches welcome for you own unicode lib ;-)
Juerd_ It's not as if this icu thing is omnipresent. 19:03
leo if we got a better lib we can use it - np
Limbic_Region right - someone made a massive patch which mostly worked so it got accepted
no one else stepped up to the plate to support a different library
Juerd_ leo: This should perhaps be on a somewhat higher level of priority than "patches welcome"
Perl 6 requires unicode support, and IIRC, Parrot was supposed to be used by Perl 6... 19:04
leo and icu provides unicode for paot - where is the problem actually? 19:05
rrr
Limbic_Region well, getting ICU to build did get easier Juerd_
Juerd_ leo: Eh, unavailability of binary packages for my system.
leo we don't build icu - we use it as a lib
Limbic_Region not the point
if you can't build the lib
then you can't use unicode
leo Juerd_: compiling icu one from src isn't a biggy 19:06
Limbic_Region and building ICU isn't trivial
Juerd_ leo: And, as L~R said, portability, although I'm just copying his argument now :)
Limbic_Region though it has been a year or more since I did it
c++ isn't that big of a deal but it does go against the original dependency plan for Parrot
leo Limbic_Region: - then please don't tell rants
Juerd_ leo: Oh, but there you are mistaken. It is. Compiling anything from source IS a biggy, because it interrupts my standard work flow, and makes me WAIT for software to install.
leo Juerd_: what system are you using w/o any icu lib? 19:07
Juerd_ When it comes to non-toy things, I either use available binary packages, or another tool altogether.
leo: Apparently Ubuntu
Limbic_Region leo - I think you are misunderstanding my rant 19:08
leo and ubuntu doesn't have any ICU binary package?
Juerd_ Oh, my previous apt-cache search had a spelling error. There is some icu available for this platform. An old version, but available.
Juerd_ was searching for uci
leo ok then
Juerd_ s/.*// :) 19:09
Limbic_Region I am not against ICU support for Parrot - I am against the idea that it is the only option for Unicode support
but as you said - patches and volunteers welcome
leo Limbic_Region: it isn't - please provide another package
it's currently the only one we have that's it 19:10
Juerd_ Limbic_Region: Does ICU have a bad reputation?
leo same with libgmp
Limbic_Region Juerd_ - no, not that I know of 19:11
leo Juerd_: it provides much more than needed for usual defs of needed
Limbic_Region the developers were quite helpful at trying to resolve the build problems I was having on Cygwin
but it is a monster
Juerd_ I see
For a libicu-lite ;) 19:12
s/For/Fork/
leo that's not a new idea - but we for sure don't do it 19:13
Limbic_Region actually, I tried
but that was still when ICU was part of the parrot trunk
leo reinventing just smaller wheels is still a bad idea
Limbic_Region well, I wasn't reinventing anything - just trimming the enormous amounts of fat 19:14
Juerd_ leo: Through histories, wheels in general have become available in much lighter forms.
Though history in plural doesn't quite make sense in English :)
leo that's still wellcome of course - trimming existing stuff - otoh it is a lot of work 19:15
Juerd_ leo: As it turns out, if the spokes are reliable, you can do with less :)
And if you have a nice steel rail to roll on, rubber can be left off the wheel too
The ultimate realisation is of course that the wheel itself was a bad idea :) 19:16
leo all ok - what parrot will need is one PMC (e.g. UnicodeString) that DTRT 19:17
19:26 hcarty joined
rafl Juerd_: Which version? 19:27
Juerd_ rafl: 28 19:32
rafl Juerd_: Maybe report a bug against the icu src package to include 34 as well. 19:33
Juerd_: Should not be much work as it's in Debian already.
Juerd_: Erm, packages.ubuntu.com/dapper/libs/libicu34 19:34
Juerd_ Ahh, dapper
But dapper is too scary on powerpc
It barely even runs on i386 sometimes :)
rafl I know a lot of people using dapper. Hrm.. 19:35
Juerd_ I've had quite some trouble with dapper on my laptop 19:36
19:36 Amnesiac joined
svnbot6 r8608 | gaal++ | YAML serialization - 19:41
r8608 | gaal++ | * emit PGE Rules (rx:g/a/.yaml works)
r8608 | gaal++ | * normalize emitted booleans
r8608 | gaal++ | * src/Data/Yaml/Syck.hsc cleanups
r8608 | gaal++ | - YamlNode structure shouldn't change much henceforth
r8608 | gaal++ | - remove implicit vars
19:50 Lartuv joined
clkao ?eval rx:g/a/.yaml 19:58
valbot_860 is now known as evalbot_8608
evalbot_8608 "--- !pugs/Rule \nrule: a\nglobal: true\nstringify: false\nadverbs: \n g: 1\n"
clkao subalashi! 19:59
gaal clkao: other direction coming up. 20:05
btw, what does "stringify" mean, anyway? :) 20:06
aufrank gaal: are you joking or in earnest? 20:13
stringify means "make into string" 20:14
gaal aufrank: I meant in the context of a PGE Rule record member :)
aufrank oh
nevermind then :)
gaal what does it mean for a Rule to say "no stringify"?
aufrank not a clue 20:15
gaal argh why does dist/ need manual cleaning after hsc hacking? :( 20:17
okay, I have Rule deserialization, but it segfaults when used. 20:24
probably I need to copy values.
just redumping out the deserialized Rule works, though. 20:25
so I'll checkin, if $you can fix please do :)
svnbot6 r8609 | gaal++ | Yaml serialization - 20:29
r8609 | gaal++ | * parsing rules works (but they segfault when applied, probably need
r8609 | gaal++ | a copy made?)
gaal ?eval my $r = rx:g/a/.yaml; (eval $r :lang<yaml>).yaml.say 20:31
evalbot_8608 pugs: can't deserialize: pugs/Rule
gaal oh yes you can!
call your younger brother.
aufrank goes to airport 20:32
bye all
gaal ?eval my $r = rx:g/a/.yaml; (eval $r :lang<yaml>).yaml.say
valbot_860 is now known as evalbot_8609
evalbot_8609 OUTPUT[--- !pugs/Rule rule: a global: true stringify: true adverbs: g: 1 ] bool::true
gaal bye aufrank
audreyt: I'm still getting double-frees with r70 20:37
is anyone still awake from the hackathon? 20:39
ah, great, scratch that: make clean fixed my problems :) 20:45
20:57 Southen_ joined 21:02 brentp joined 21:09 brentp joined 21:29 kanru2 joined 21:34 rantanplan_ joined
gaal audreyt: got perl -MYAML::Syck -le 'print Dump(\qr/foo/)' to work, but bizarrely, not Dump(qr/foo/). maybe tomorrow :) 21:35
for now, zzZ & 21:36
buu ?join #perl 21:41
evalbot_8609 you should only do ?join in a private message so other bots don't accidentally come.
buu arfgh
sigh
Now ?join gets the svnbot and not the evalbot
21:47 hcarty left 22:05 bsb left 22:18 DaGo joined
gaal realizes a SYB approach is actually warranted to transform all native types like Rules (and probably Roles etc) to yaml 22:28
gaal goes back to sleep
22:38 dduncan joined 22:57 penk joined
dduncan so apparently Parrot 0.4.1 was just released 23:04
since audreyt's proposal had Pugs 0.2.11 going out just after this, how much time do I have left to get in changes first? 23:05
... will check journal ...
svnbot6 r8610 | stevan++ | PIL/Native/Roles/* 23:12
r8610 | stevan++ | - small comment update for Num.pil
r8610 | stevan++ | - added new Seq role which implements an immutable sequence type with
r8610 | stevan++ | the canonical set of list operations (ala LISP/Scheme/etc), all it
r8610 | stevan++ | needs is to have .head and .tail defined
r8610 | stevan++ | PIL/Native/Classes/*
r8610 | stevan++ | - a new folder with a set of core PIL^N classes which will be used to
r8610 | stevan++ | build the code Perl6 classes (maybe,... uhm, I think, it's my current
r8610 | stevan++ | thought direction at the moment at least)
r8610 | stevan++ | - added immutable Sequence and LazySeqeunce clases, neither is really
r8610 | stevan++ | tested (waiting on the repr refactor from audreyt++), and LazySequence
r8610 | stevan++ | has some issues when a methods must return a new sequence (apply, filter,
r8610 | stevan++ | & reverse) since that sequence won't be lazy (I have to ponder this
r8610 | stevan++ | some more before I can decide what is the best approach for it, ideally
r8610 | stevan++ | they will return lazy sequences too).
23:27 vel joined
svnbot6 r8611 | stevan++ | docs/notes/PILN_core_classes.pod 23:27
r8611 | stevan++ | - start of a document describing the core PIL^N classes/roles,
r8611 | stevan++ | think of these as being like Parrot's core PMC types, but more
r8611 | stevan++ | powerful than the native PIL^N methods. More to come on this
r8611 | stevan++ | later.
23:30 lisppaste3 joined
dduncan regarding 6.2.11 release date, the last journal mention was on dec 24th, which said in early january just after Parrot 0.4.1, and both of those prerequisites now exist 23:57
aside from getting all the tests to pass or skip, are there any other todo items now? 23:58
23:58 mlh joined

Powered by Cro and the Raku® Programming Language.

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

Thank you!