Perl 6 - the future is here, just unevenly distributed

IRC log for #perl6, 2006-03-21

Perl 6 | Reference Documentation | Rakudo

| Channels | #perl6 index | Today | | Search | Google Search | Plain-Text | summary

All times shown according to UTC.

Time Nick Message
00:00 ayrnieu left perl6
00:00 ayrnieu joined perl6
00:00 ayrnieu left perl6
00:01 ayrnieu joined perl6
00:01 * ayrnieu grumbles.
00:08 pdcawley joined perl6
00:09 pdcawley_ joined perl6
00:20 Quell joined perl6
00:21 wolverian Juerd, hey, nice colours on juerd.nl :)
00:22 Juerd wolverian: It's random, so cache them while you can ;)
00:23 wolverian yeah, I noticed :)
00:25 Juerd It's some perl module that generates them. Most of the time, it's nice.
00:25 Juerd I haven't figured out how to calculate which one of <white black> will contrast most, though.
00:25 jisom convert to grayscale, measure the difference, if not enough, recalculate?
00:26 Juerd jisom: I don't know how to convert to grayscale properly.
00:26 Juerd The mean of R, G and B, certainly isn't useful.
00:27 jisom I know swing has one in it's source....but it might be sufficient to try
00:27 wolverian Juerd, hold a sec, I'll get you the coefficients
00:32 FurnaceBoy Juerd, you can also add 0x80,0x80,0x80 to get a highly contrasting colour.
00:32 FurnaceBoy add and modulo
00:33 TimToady won't that turn gray into gray?
00:33 FurnaceBoy no, it would turn 0x80,0x80,0x80 into white :-)
00:33 FurnaceBoy sorry, into black
00:33 TimToady ah, gotcha.
00:34 FurnaceBoy but 0x7f,0x7f,0x7f into white
00:34 * particle_ sees a red door
00:34 FurnaceBoy it's sorta handy for overlaying stuff.
00:34 TimToady don't eat the red door.
00:34 FurnaceBoy where you have a continuously changing background like an image
00:34 FurnaceBoy so cannot pick a single colour to contrast
00:35 Juerd FurnaceBoy: No, adding 0x80 xx 3 doesn't help on bright colours...
00:35 TimToady Seems like xor with 0x80,0x80,0x80 does the same thing...
00:35 Juerd FurnaceBoy: Oh, modulo.. Hmmm...
00:35 FurnaceBoy TimToady, no
00:35 FurnaceBoy hmm
00:35 FurnaceBoy yes
00:35 FurnaceBoy :-)
00:36 Juerd FurnaceBoy: I guess it would create something contrasting, but I'd like to stick to black and white.
00:36 FurnaceBoy yes
00:36 FurnaceBoy it's really only useful for image-like backgrounds
00:36 jisom a % 0x80 => a & 0x7f
00:36 wolverian Juerd, Color::Scheme has a convenient looking ->distance() method, btw. anyway, here are the coefficients for the channels if you want to calculate the contrast for a human eye: 0.30 0.59 0.11 (RGB). then calculate the distance as if the channels were dimensions, normalising if you want.
00:37 FurnaceBoy that's not actually based on the eye; those are the traditional NTSC coefficients
00:37 wolverian oh, sorry.
00:37 FurnaceBoy well, they're based on the eye to the extent the 195x NTSC coeffs are :-)
00:37 FurnaceBoy mostly still used
00:37 wolverian do you have better ones?
00:37 FurnaceBoy no, was just nitpicking. ;_)
00:37 wolverian ok :)
00:37 Juerd wolverian: I lack math skills to understand that, though. What's a coefficient, and everything from "calculate the distance" I understand in English, yet I don't know what it actually means :)
00:37 FurnaceBoy the NTSC are still commonly used
00:38 FurnaceBoy Istr a paper on perceptual neutral density ... hmm.
00:38 wolverian Juerd, each value there is multiplied against the two colours' difference for that channel. i.e. 0.3 * (col1.r - col2.r) etc.
00:38 Juerd wolverian: But I only have one colour, the background colour.
00:39 wolverian oh, then I misunderstood what you wanted. :)
00:39 FurnaceBoy well,
00:39 FurnaceBoy those coeffs still give you a useful grey conversion
00:39 FurnaceBoy g = .30r + .59g + .11b
00:39 Juerd Given a certain background colour, I want to know if white or black is the better foreground colour.
00:39 FurnaceBoy note (.30+.59+.11) = 1.0
00:40 wolverian Juerd, then you _do_ have two colours, or three, to be exact.
00:40 FurnaceBoy yes, you can grey-ise as above
00:40 FurnaceBoy then check whether it's >.5 or <.5
00:40 Juerd wolverian: Really?
00:40 wolverian Juerd, white, black, and your bg colour.
00:40 Juerd FurnaceBoy: Ahh. I'll try that.
00:40 Juerd wolverian: Hmm.
00:40 wolverian FurnaceBoy's method should work fine here.
00:41 FurnaceBoy fwiw, I believe Photoshop's RGB->Grey Scale conversion still uses those exact coefficients.
00:41 FurnaceBoy absent colour management settings.
00:41 wolverian you don't need to calculate the real distance, only compare distances, and that works fine with manhattan distance
00:41 wolverian (too many 'distance's there..)
00:42 wolverian FurnaceBoy, er, isn't NTSC YUV?
00:42 wolverian FurnaceBoy, or do you mean those coefficients are what it comes to after translating to RGB? sorry, slow tonight..
00:43 Arathorn NTSC's only YUV if you happen to digitize it into chrominance & luminance
00:43 * Arathorn promptly stops talking, given he can't remember how the colourburst is encoded in PAL, let alone Never The Same Color twice
00:44 FurnaceBoy "The coefficients 0.299, 0.587 and 0.114 properly computed luminance for monitors having phosphors that were contemporary at the introduction of NTSC television in 1953, however, these coefficients do not accurately compute luminance for contemporary monitors."
00:44 wolverian FurnaceBoy, thanks.
00:44 FurnaceBoy I would concur with that last remark, those coeffs aren't quite right for usual screens
00:44 wolverian well, screens differ hugely in colour reproduction anyway
00:44 FurnaceBoy yep.
00:44 FurnaceBoy and Juerd only needs a "close enough for govt work" answer :-)
00:44 wolverian LCDs have really bad curves
00:45 * wolverian hates his laptop's "black", which is more like gray
00:45 FurnaceBoy this guy is a colour theory master http://www.poynton.com/
00:45 wolverian FurnaceBoy++
00:45 Juerd FurnaceBoy: This works quite well. I used a threshold of 0x90 instead of 0x80, though.
00:46 FurnaceBoy yeah, that's probably exactly the result of the gamma shift since 1953 :-)
00:46 Juerd wolverian: Almost all TFT black is gray, because TFT's natural colour is white.
00:46 wolverian yeah. OLED please. :)
00:46 FurnaceBoy yeah. subtractive
00:46 Juerd wolverian: The pixels don't manage to fully block that.
00:46 FurnaceBoy but this Samsung is mightly black...
00:46 Juerd FurnaceBoy: And probably glossy too, then.
00:46 FurnaceBoy except it still suffers a lot at the dark end of the range
00:46 FurnaceBoy playing MPEG for instance
00:47 wolverian Juerd, hm, is juerd.nl using the new code now? I'm getting somewhat non-contrasty results for the date title on the frontpage
00:47 Juerd I hate glossy screens.
00:47 FurnaceBoy no, it's fairly matte
00:47 FurnaceBoy just a good 1:700 or more
00:47 Juerd wolverian: Yes, it is.
00:47 ayrnieu black looks quite good on this Apple Cinema 20" -- but yes, glossy.
00:47 Juerd wolverian: Did you get white?
00:47 wolverian I got white text on a very light blue
00:47 wolverian let me check the colours
00:47 Juerd I'll get it back to 0x80 then
00:48 Juerd See /css.plp, the comments at the top
00:48 Juerd I need to base things differently for the code blocks
00:48 Arathorn is now known as Aragone
00:49 wolverian I got (128, 255, 247) for the bg of the <h2> and white as the text colour
00:49 wolverian (that's 80fff7)
00:51 Juerd I was using the wrong colours for the background :)
00:51 Juerd The calculation was based on the pure colour, not the lighter tint
00:51 wolverian ah. :)
00:52 wolverian ow, got a bad one again
00:53 Juerd Yea, I'm lost in my single letter variables, and meaningless array indices.
00:53 wolverian Color::Scheme looks really nice. I'll try this on my css files.. :)
00:54 Juerd It should be okay now
00:54 Juerd wolverian: Beware... you'll like it, your visitors will like it, and it will cost you more time than you now think it does.
00:54 wolverian seems to be
00:54 wolverian Juerd, oh, I don't have any visitors.
00:55 Juerd http://tnx.nl/4603WGFV  # horrible code
00:55 wolverian eww :)
00:56 Juerd Worse is that it leads to things like
00:56 Juerd    background: <:= $c[2][0] :>;
00:56 Juerd    color: <:= $c[2][4] :>;
00:56 wolverian egh, my $b = ...$b[0]...$b[1]...; # wow that is confusing :)
00:57 Juerd Yea.
00:57 Juerd And don't make the mistake of thinking 'c' stands for colour. That's purely coincidental.
00:59 Juerd I've written more than enough crappy code in my life to know what is and what is not maintainable. Still, for private site things, I *like* messing around with crappy code.
00:59 wolverian procrastinating..
00:59 Juerd What's that?
00:59 wolverian putting off something by distracting yourself with something else
01:00 Juerd Ah, I did that all day without knowing there was a word for it :)
01:00 wolverian :)
01:00 Juerd Spent half a day rewiring parts of my car. Should have done taxes.
01:00 ayrnieu procrastination also refers to a kind of clever task management.
01:00 Juerd I'm going to bed now
01:01 Juerd Thanks for helping me with the colour issue
01:01 Juerd And thanks for delaying my sleep with that ;)
01:01 * Juerd --zzz &  # .zZ
01:01 wolverian night :)
01:02 jisom finally I got it working....in a mostly correct way but nothing more than a "floating point" error
01:04 mako132_ joined perl6
01:12 ajani joined perl6
01:14 ajani is now known as ajani_
01:15 ajani_ is now known as ajani
01:16 Juerd I'm still awake. Damn you, internet! :)
01:16 * Juerd afk now
01:16 * wolverian pushes Juerd to the bed
01:17 justatheory joined perl6
02:05 p5evalbot joined perl6
02:05 GabrielVieira joined perl6
02:52 buu joined perl6
02:59 xern joined perl6
03:12 amnesiac joined perl6
03:21 buu joined perl6
03:45 Khisanth joined perl6
04:03 KingDiamond joined perl6
04:40 stevan_ joined perl6
05:08 elmex_ joined perl6
05:11 stevan joined perl6
05:31 vel joined perl6
05:33 svnbot6 r9671 | Darren_Duncan++ |  r3250@darren-duncans-power-mac-g4:  darrenduncan | 2006-03-20 12:58:49 -0800
05:33 svnbot6 r9671 | Darren_Duncan++ |  ext/Rosetta/ : updated the class complement of Rosetta.pm
05:33 svnbot6 r9672 | Darren_Duncan++ |  r3253@darren-duncans-power-mac-g4:  darrenduncan | 2006-03-20 15:18:27 -0800
05:33 svnbot6 r9672 | Darren_Duncan++ |  ext/Rosetta/ : completed initial Rosetta Shell, which executes
05:33 svnbot6 r9673 | Darren_Duncan++ |  r3256@darren-duncans-power-mac-g4:  darrenduncan | 2006-03-20 16:36:33 -0800
05:33 svnbot6 r9673 | Darren_Duncan++ |  ext/Rosetta/ : final smaller updates synchronizing with perl 5 version CPAN release v0.722.0
05:33 dduncan well, the server woke up, thankfully
05:36 vel joined perl6
05:40 mj41__ joined perl6
05:46 Kattana joined perl6
05:49 nirgle joined perl6
05:56 Quell joined perl6
06:15 svnbot6 r9674 | Darren_Duncan++ |  r3262@darren-duncans-power-mac-g4:  darrenduncan | 2006-03-20 22:12:30 -0800
06:15 svnbot6 r9674 | Darren_Duncan++ |  ext/Rosetta/ : following today's release on CPAN of the Perl 5 flavor, incremented all major version numbers ; also, a few pod fixes
06:16 amnesiac joined perl6
06:35 GeJ joined perl6
07:01 xern joined perl6
07:04 FurnaceBoy_ joined perl6
07:10 iblechbot joined perl6
07:10 FurnaceBoy_ left perl6
07:24 dduncan left perl6
07:24 drrho joined perl6
07:27 stevan joined perl6
07:32 Quell joined perl6
07:36 Quell joined perl6
07:37 ghenry joined perl6
07:41 lisppaste3 joined perl6
07:43 nothingmuch joined perl6
07:47 nothingmuch morning
07:50 Supaplex zzZzzz
07:58 nothingmuch ingy: ping
07:59 beppu_ joined perl6
08:06 marmic joined perl6
08:31 gaal ingy: ping
08:34 Supaplex ZZZzzz....ong...*snort*
08:40 wilx joined perl6
08:42 kane_ joined perl6
08:50 Chiicco24 joined perl6
08:50 stennie joined perl6
08:50 Chiicco24 left perl6
09:15 nnunley joined perl6
09:50 nnunley_ joined perl6
09:51 KingDiamond joined perl6
10:02 penk joined perl6
10:07 kattana_ joined perl6
10:20 stevan joined perl6
10:22 nnunley_ is now known as nnunley
10:44 xinming joined perl6
10:57 pjcj joined perl6
10:58 nivrrex joined perl6
10:59 nivrrex left perl6
11:00 szbalint joined perl6
11:05 chris2 joined perl6
11:28 ingy gaal: yes?
11:28 nothingmuch ingy ping too
11:28 nothingmuch can i have that script you use to make a symlinked @INC entry to a dist like layout?
11:28 ingy nothingmuch: elk?
11:29 nothingmuch ingy: møøse
11:30 ingy nothingmuch: sure, looking for it...
11:31 ingy nothingmuch: http://svn.kwiki.org/kwiki/bin/make-lib-tree.pl
11:32 ingy nothingmuch: put it in your path and then just tell it about various sources
11:32 ingy it creates ./lib/
11:32 nothingmuch thanks
11:32 ingy and remembers all old sources in ./lib/.sources/
11:32 ingy and remembers all old sources in ./lib/.sources
11:32 ingy rather
11:33 nothingmuch møøse
11:33 ingy so when you run it with no arguments, it refreshes
11:33 ingy elk
11:33 nothingmuch yay
11:33 nothingmuch btw, what's your current email?
11:33 ingy it doesn't recurse though
11:34 ingy make-lib-tree.pl ~/src/ingy/* ~/src/pugs/perl5/*
11:34 ingy [email@hidden.address]
11:34 nothingmuch why not [email@hidden.address]
11:34 ingy laziness
11:35 nothingmuch i'll give you ingy@møø.se if i manage to buy that domain
11:35 nothingmuch it's not allowed
11:35 nothingmuch but i think i sort of managed
11:36 ingy ?
11:36 ingy not allowed?
11:37 ingy besides it would be [email@hidden.address] :p
11:37 renato joined perl6
11:37 nothingmuch ingy: .se don't like ø
11:38 nothingmuch make [email@hidden.address] an alias please
11:38 nothingmuch i will never manage to remember anything else
11:39 osfameron nothingmuch: have you got moo.se ?
11:40 nothingmuch osfameron: no
11:40 nothingmuch it's just parked =(
11:42 ingy how about mooooo.se?
11:42 nothingmuch no, that's lame
11:42 ingy or m00.se
11:42 nothingmuch i wouldn't mind that if i don't get møø.se
11:43 ingy ø_0
11:45 gaal ingy: hi. quick yaml question
11:45 gaal I want heredoc-style quoting, like |
11:45 gaal but want to avoid the need for indentation in that quoted node
11:45 gaal can I?
11:46 ruoso joined perl6
11:46 ingy gaal: no
11:46 gaal too bad :/
11:46 gaal I understand why it wouldn't be
11:46 ingy how would you know the end?
11:47 gaal |<<"END" :-P
11:47 ingy would require heredoc
11:47 ingy yep
11:47 ingy what's the use case?
11:48 gaal I wanted to make yaml the encoding for this: http://search.cpan.org/~gaal/Te​mplate-Patch-0.01/bin/metapatch
11:48 gaal the templates need to be trivially human-editable
11:49 ingy yaml is trivially human-readable, but not always trivially human edit-able
11:49 gaal yes indeedy
11:49 gaal I appreciate the tradeoff, it just leans the wrong way for my current application. oh well :)
11:50 nothingmuch couldn't '-|' be extended into '-
11:50 nothingmuch '-| __marker__'
11:50 nothingmuch <non indented text here
11:51 nothingmuch __marker__
11:51 nothingmuch ?
11:51 ingy the cool thing about the verbatim block, is that you can trivally embed any yaml in yaml
11:51 ingy yaml is .il technology :p
11:51 nothingmuch oops, didn't notice backlog
11:51 gaal bad nothingmuch
11:51 nothingmuch sorry, i skipped just that line
11:51 nothingmuch i swear
11:52 gaal nothingmuch: i wonder, is oo.se maybe free? :)
11:52 gaal isn't.
11:53 nothingmuch checked =)
11:53 nothingmuch i wouldn't have minded m.oo.se
11:53 nothingmuch it's ver del.icio.us
11:53 nothingmuch i mean ver.y
11:53 nothingmuch or v.er.y
11:54 gaal e.lk (sri lanka)
11:54 ingy :D
11:54 ingy i.lk
11:55 * ingy goes back to Chinese lesson
11:57 mako132_ joined perl6
11:59 svnbot6 r9675 | broquaint++ | * Added PadWalker 1.0 to Makefile.PL.
12:07 * ruoso is porting lrep-compiler to PCR
12:10 Juerd $foo[7] = "bah";
12:10 Juerd Wrong channel.
12:10 nothingmuch ruoso: PCR?
12:10 nothingmuch Pugs::Compile::Rules?
12:14 gaal Juerd: yes, for it to be right for this channel you'd have to have said @foo[7] = "bah";  :-)
12:15 Juerd gaal: Oh? $foo[7] = "bah"; works well in Perl 6.
12:15 Juerd It doesn't touch @foo, but it does work.
12:16 Juerd In fact, it was a PHP example, making $foo[7] the direct translation, in Perl 6.
12:16 Juerd As PHP has only array (hash, really) *references*
12:16 ruoso nothingmuch, yes
12:16 Aragone $foo[7] in perl6 makes $foo an arrayref, then?
12:17 ruoso Aragone, yes
12:19 gaal takes $foo to be an arrayref, rather
12:19 gaal ?eval my $foo = "scalar"; $foo[7] = "oops"; $foo.perl
12:21 gaal anyway, that's "*** Can't modify constant item: VUndef"
12:21 gaal which is probably not the best error message possible here :/
12:21 * gaal woos evalbot back
12:22 * ruoso is recompiling the new Perl6.p6 that already uses PCR
12:24 nirgle left perl6
12:27 * ruoso wonders if fglock renamed some capture nodes
12:27 Juerd gaal: Well, $foo is nicely autovififithingied if it's undef :)
12:27 gaal Juerd: ACK
12:28 Juerd That single feature that's keeping me away from Ruby.
12:58 pmurias joined perl6
12:58 pmurias hi roso
12:58 svnbot6 r9676 | pmurias++ | the hybrid wird not-realy working port to the PCR
12:58 ruoso pmurias, I was wainting for you
12:59 pmurias nice of you
12:59 ruoso pmurias, :) so... is this commit *the* commit?
13:00 pmurias yes
13:01 ruoso does it uses PCR from CPAN or from pX?
13:01 pmurias tested only on the pX one, you have to send PERL5LIB to run it
13:02 ruoso ah... ok...
13:02 ruoso I'll change the Makefile to reflect that
13:05 fglock joined perl6
13:05 ruoso pmurias, hmmm.... not working...
13:06 ruoso pmurias, oh... ok... saw the commit log now
13:06 pmurias hi fglock
13:07 pmurias the part which I could get right yesterday was the old rules and PCR rules interop
13:07 pmurias s/could/couldn't/
13:07 ruoso pmurias, that's the hard part
13:07 ruoso pmurias, :)
13:07 Limbic_Region joined perl6
13:08 pmurias ruoso: it's mostly a matter of determining how the rule API changed
13:09 ruoso pmurias, that's exactly the question I was about to make to fglock
13:10 ruoso fglock, ?
13:11 ruoso Not an ARRAY reference at ...Pugs/Runtime/Rule.pm line 94
13:11 iblechbot joined perl6
13:11 ruoso this was the error I was getting in the try I was making before your commit
13:11 fglock hi! - I'm trying to commit the $0,$1 implementation - but got to merge first :(
13:12 ruoso fglock, is the "Not an ARRAY" error related to that?
13:13 fglock ruoso: looking...
13:14 fglock ruoso: that's a problem with the match{state}, which is an internal variable - it should never happen
13:15 ruoso fglock, only when you mix old with new code... :)
13:15 ruoso fglock, how the state should looks like...
13:15 ruoso fglock, so I can wrap it...
13:16 fglock ruoso: state is an array with [ number, state ]
13:18 Aragone (btw, you do know that the concat function in Pugs::Runtime::Rule2 in the released PCR 0.01 on cpan is mangled?)
13:19 svnbot6 r9677 | fglock++ | pX/../PCR - $0, $1
13:20 fglock Aragone: looking...
13:21 fglock Aragone: looks ok - http://search.cpan.org/src/FGLOCK/Pugs-Com​piler-Rule-0.01/lib/Pugs/Runtime/Rule2.pm
13:21 fglock but this file is gone in the new version
13:21 fglock work &
13:21 Aragone it tries to recurse to Pugs::Emitter::Rule::Perl5::concat()
13:21 Aragone though
13:21 Aragone (it's fixed in the new version)
13:22 Aragone rather than itself
13:22 kanru joined perl6
13:23 ruoso fglock, I see that 'push @list, rule' now pushes a string instead of a sub...
13:23 ruoso in PCR
13:25 ruoso fglock, I'm a little bit confused..
13:28 fglock ruoso: the list is processed into method calls in the end of Pugs::Grammar::Rule
13:28 ruoso fglock, ah ok
13:29 ruoso fglock, this should be changed, shouldn't it?
13:30 * ruoso getting back to Perl6::Bible to see how the syntax should looks like
13:31 svnbot6 r9679 | pmurias++ | $grammar_namespace now passed to the PCR emitter
13:32 fglock ruoso: there are 2 kinds of rules in PCR: methods and objects. @list expects a list of objects - so we need to curry the methods into rule-objects. I don't know if this is correct
13:33 ruoso fglock, so... the emitted code should already wrap it...
13:33 ruoso fglock, and push @list, \&rule... would just work
13:34 fglock ruoso: yes, but Pugs::Grammar::Rule is a bit odd (it's not emitted automatically)
13:35 ruoso fglock, but it should represent the same thing, shouldn't it?
13:36 fglock ruoso: [OT] where does gnome mounts a SMB share?
13:36 ruoso fglock, it doesn't
13:37 ruoso fglock, that's the great shit of gnome vfs
13:37 ruoso fglock, and that's why I pray for hurd every day on my like
13:37 ruoso s/like/life/
13:37 fglock ruoso: it represents the same thing, at high level - put p5 can't have method==object
13:38 pmurias ruoso: ever heard of FUSE?
13:39 elmex joined perl6
13:39 ruoso pmurias, yes... Yeat Another Linux Kernel Dirty Hack
13:40 ruoso fglock, but... are rule methods accessed any other way than as rule objects?
13:42 fglock ruoso: $grammar->rule() vs. $rule()
13:44 ruoso fglock, ah ok... I got confused again...
13:46 ruoso so... push @list, rule should have a different syntax, shouldn't it?
13:47 ruoso or one could tie @list to something that makes it work
13:48 fglock ruoso: not sure - I think it should just be emitted in a different way
13:49 kisu joined perl6
13:49 ruoso fglock, tie in the emitted code
13:50 ruoso turn "my Rule @list;" into the tie code...
13:50 ruoso fglock, what do you think?
13:52 * ruoso thinking again
13:57 ruoso fglock, in "my $rule = rule { ... }"... which value would have $rule?
14:02 fglock ruoso: $rule is a Pugs::Compiler::Rule object
14:03 ruoso fglock, so... doesn't $grammar->rule->() makes sense?
14:06 ruoso pmurias, named captures already are written as $<name>
14:06 ruoso pmurias, this is one of the problems
14:09 fglock ruoso: what if I mix objects and methods in @array?
14:11 ruoso fglock, good question...
14:16 ruoso fglock, I added a 'return '' unless defined $n;' in P::E::R::P5::emit_rule
14:17 ruoso before the die "unknown node"
14:18 fglock ruoso: but that is an error in the AST - you should emit a warning
14:18 pmurias i'll take a break from the computer now, if someone check why does the make program AST leak into the PCR while compiling Rules.p6 I'll be gratefull
14:18 ruoso fglock, hmmm... I noticed it now...
14:18 ruoso fglock, the rules simply were not there... :)
14:20 b_jonas joined perl6
14:21 Qiang joined perl6
14:22 pmurias s/make/main/
14:27 ruoso fglock, how should literals appear in a rule AST?
14:27 ruoso fglock,  'term' => '\\)'
14:29 ruoso unknown node: $VAR1 = '\\(';
14:30 fglock ruoso: escaped chars are not defined yet in PCR - see the 'XXX' in the grammar
14:30 ruoso fglock, oh... ouch...
14:30 ruoso fglock, any workaround?
14:31 fglock ruoso: I can implement it - one sec
14:39 chris2 joined perl6
14:41 fglock ruoso: done - I'll write some tests now
14:41 svnbot6 r9680 | fglock++ | pX/../PCR - escaped chars (untested)
14:42 ruoso constant( '\(' )
14:43 ruoso fglock, shouldn't it be constant('(') ?
14:45 fglock ruoso: testing...
14:47 fglock ruoso: fixed - committing
14:48 ruoso fglock, nice... thanks...
14:48 svnbot6 r9681 | fglock++ | pX/../PCR - fixed escaped chars; tests
14:55 svnbot6 r9682 | fglock++ | pX/../PCR - '\n' escaped char
14:57 obra seen nothingmuch?
15:02 bsb joined perl6
15:02 svnbot6 r9683 | fglock++ | pX/../PCR - updated comments, TODO
15:06 ruoso fglock, 'Not a SCALAR reference' at 'sub { ${ $grammar->p6ws( @_ ) } }'
15:08 b_jonas joined perl6
15:08 vel joined perl6
15:08 nothingmuch joined perl6
15:14 svnbot6 r9684 | ruoso++ | Huge refactoring... importing work from lrep-pcr (pmurias++) to here... going further... It is broken by now... I still didn't understand why...
15:16 ruoso fglock, why does it try to dereference to scalar if it is a code?
15:17 iblechbot joined perl6
15:20 svnbot6 r9685 | ruoso++ | updating TODO
15:20 ruoso fglock, lrep-compiler has the most up-to-date in svn... there is this 'not a scalar reference' problem... please take a look if you have some time
15:21 ruoso pmurias, I took the work you were doing in lrep-pcr back to lrep-compiler... there is this problem I explained to fglock... but it's almost there...
15:25 * ruoso have to go for lunch...
15:29 justatheory joined perl6
15:31 particle_ left perl6
15:56 nothingmuch joined perl6
16:01 frobnitz joined perl6
16:02 FurnaceBoy joined perl6
16:19 b_jonas joined perl6
16:21 kisu joined perl6
16:29 stennie joined perl6
16:41 Ara4n joined perl6
16:41 Ara4n audreyt: ping?
16:43 KingDiamond joined perl6
16:46 b_jonas joined perl6
17:06 b_jonas joined perl6
17:10 ruoso joined perl6
17:12 willie0220 joined perl6
17:14 svnbot6 r9686 | fglock++ | Pugs::Grammar::Base update
17:16 DesreveR joined perl6
17:17 gilby joined perl6
17:19 ruoso fglock, ?
17:19 fglock ruoso: hi
17:20 ruoso fglock, what does ``sub { ${ $grammar->p6ws( @_ ) } }'' means?
17:20 willie0220 left perl6
17:20 ruoso fglock, it generated this code... and it causes error...
17:21 fglock p6ws returns a match object - $$match returns the data structure
17:21 ruoso fglock, hmmm...
17:21 ruoso fglock, it says 'Not a SCALAR reference'
17:21 SamB joined perl6
17:21 nothingmuch joined perl6
17:22 fglock ruoso: don't reuse the p6ws method in the rule grammar - it was generated by the 'broken' PCR/lrep
17:22 fglock did you compile p6ws yourself?
17:22 ruoso fglock, no
17:23 ruoso fglock, I still need to recompile the lrep grammar
17:23 ruoso fglock, if p6ws shouldn't be used, what should be?
17:23 ruoso fglock, or should I define a new p6ws method?
17:24 fglock ruoso: we'll need to write a grammar compiler, in order to fix PCR::Rule
17:24 fglock ruoso: yes, defining the method yourself will be fine
17:24 ruoso fglock, lrep is a grammar compiler...
17:24 ruoso fglock, ok
17:25 fglock ruoso: but it doesn't use PCR, so PCR thinks it's broken
17:25 * ruoso have a meeting now
17:30 gilby joined perl6
17:44 justatheory joined perl6
17:49 mj41_ joined perl6
18:19 GabrielVieira joined perl6
18:19 GabrielVieira ho there
18:19 GabrielVieira *hi
18:24 nothingmuch joined perl6
18:26 amnesiac joined perl6
18:28 bernhard joined perl6
18:44 stevan joined perl6
18:45 GnuVince joined perl6
18:47 GnuVince left perl6
19:11 KingDiamond joined perl6
19:11 justatheory joined perl6
19:14 KingDiamond joined perl6
19:15 fglock ugh - "Unknown error / Compilation failed in require. / BEGIN failed--compilation aborted."
19:17 GeJ joined perl6
19:40 DesreveR is now known as r0nny
19:53 larsen joined perl6
20:03 chris2 joined perl6
20:03 Limbic_Region joined perl6
20:18 ruoso fglock, now, after fixing p6ws (at least I think I fixed it), it just doesn't match anything...
20:20 fglock ruoso: I have compiled a grammar with lrep+PCR - now I'm trying to integrate the compiled grammar back into PCR
20:21 fglock ruoso: I'll paste the p6ws I've got here
20:21 ruoso ok
20:21 pasteling "fglock" at 200.17.89.80 pasted "p6ws (p5)" (12 lines, 341B) at http://sial.org/pbot/16448
20:22 fglock this was compiled with my current PCR/lrep
20:22 fglock (which doesn't work)
20:25 m joined perl6
20:29 chris2_ joined perl6
20:31 chris2_ is now known as chris2
20:32 ruoso fglock, still doesn't work...
20:32 * ruoso have to work now...
20:32 ruoso :(
20:44 ruoso it really does not match anything
20:44 pjcj joined perl6
20:47 hexmode joined perl6
20:48 fglock ruoso: my version is passing some internal tests already, but it is still too broken for committing
20:49 ruoso fglock, ok... I'm a little bit busy now...
20:50 ruoso fglock, but it seems it's all PCR fault now :)
20:51 fglock ruoso: can you paste your p6ws?
20:51 ruoso fglock, please leave a BUGS file with some help if I come back and you're not here...
20:51 ruoso fglock, the same as yours
20:51 ruoso fglock, it was like that already
20:52 ruoso fglock, but I have to work now... I'll be back later... and I'll probably work on it...
20:53 ruoso fglock, if you like, I can help you in PCR when I come back
20:53 fglock ruoso: sure
20:53 ruoso fglock, just leave some pointers somewhere...
20:54 ruoso so I can know what to do...
20:54 fglock ruoso: I'll try to make a PCR-dev dir
20:54 ruoso ok
20:55 fglock ruoso: I spent a few hour to find out what 'Unknown error' was about :(
20:56 fglock hours
20:56 ruoso fglock, :) that's the fun of debugging .... heh
20:56 ruoso fglock, when you hit your head on the table
20:57 ruoso fglock, hmmm... looking better... I think it still can be my fault...
20:57 ruoso :)
20:57 fglock ruoso: does it pass all tests?
20:58 ruoso fglock, PCR? yes
20:58 ruoso lrep I can't even compile...
21:00 pmurias C
21:00 pmurias C
21:01 pmurias C
21:01 pmurias C
21:08 ruoso fglock, hmmm... I can't find where I'm wrong
21:09 fglock ruoso: write a test file and run it with PCR alone
21:10 ruoso hmmm... I think <rule>* is matching nothing and just returning
21:15 * ruoso have to work.... really... :)
21:15 Juerd debian--  # your netinst images SUCK
21:16 Juerd Busybox is NOT ENOUGH.
21:16 Juerd And debian-installer's partitioning/raid/lvm thing is way too tedious.
21:16 mj41__ joined perl6
21:17 svnbot6 r9687 | ruoso++ | trying to figure out whats wrong...
21:17 wolverian submit bugs?
21:18 ruoso fglock, if I change * to + the match returns undef...
21:19 Juerd wolverian: Was that meant for me?
21:19 wolverian yes.
21:20 Juerd wolverian: Won't help. If they agreed, they wouldn't have made these sucky cd's.
21:20 * Juerd goes to find out how to debootstrap a different arch from the running arch
21:20 wolverian Juerd, someone else might read the bug report and get ideas. who knows.
21:20 Juerd I don't believe in that.
21:21 FurnaceBoy_ joined perl6
21:21 ruoso Juerd, debootstrap already does it...
21:22 Juerd Good.
21:22 Juerd Back to manual installations, again.
21:22 Juerd At least then you get something that:
21:22 Juerd a. works
21:22 Juerd b. works in time
21:22 Juerd c. understands perl
21:22 Juerd d. can be automated easily
21:23 Juerd And then I can install it remotely, which means I don't have to listen to the extreme noise this 1U box makes
21:24 fglock ruoso: compiling a language using itself was not supposed to be easy, right? :(
21:25 fglock ruoso: circularity means the bugs can be anywhere
21:25 fglock ruoso: they are actually everywhere
21:26 Juerd Oh, and vi
21:26 Juerd vim, even
21:26 Juerd nano sucks
21:28 fglock Juerd: I use nano a lot (and vim, gedit, scite)
21:28 ghenry joined perl6
21:29 Juerd fglock: Then you suck :)
21:31 fglock "you are what you eat"
21:43 m joined perl6
21:43 svnbot6 r9688 | fglock++ | pX/.../PCR/PCR-dev.tar.gz - compiles it's own grammar; bugs! - some
21:43 svnbot6 r9688 | fglock++ | matches return undef
21:50 ruz joined perl6
21:52 fglock ruoso: now I know what are most of the errors - I'll restart the grammar migration stepwise
21:53 fglock ruoso: I can't replicate the errors you found (p6ws, +/*)
21:54 fglock later &
21:54 fglock left perl6
22:07 svnbot6 r9689 | ruoso++ | this way it simply doesnt work
22:10 GabrielVieira joined perl6
22:16 justatheory joined perl6
22:21 Limbic_Region joined perl6
22:35 FurnaceBoy_ is now known as FB|afk
22:44 Blues-Man joined perl6
22:46 Blues-Man hi
22:47 Blues-Man left perl6
23:03 siosiosios joined perl6
23:03 diotalevi_ joined perl6
23:04 hlen joined perl6
23:10 jisom joined perl6
23:15 Ara4n Juerd: you playing with my ghastly xul/flash takahashi hack?
23:15 * Ara4n is keeping half an eye on his apache logs
23:15 Ara4n (when his DSL isn't dead, that i)
23:16 Ara4n is)
23:16 Juerd Ara4n: No
23:18 Ara4n fairenough
23:18 * Ara4n just interested to know if it's gone and segfaulted any more innocent GREs
23:34 rgs joined perl6
23:36 FB|afk is now known as FurnaceBoy
23:46 whiteg joined perl6

| Channels | #perl6 index | Today | | Search | Google Search | Plain-Text | summary

Perl 6 | Reference Documentation | Rakudo