SVK users: search.cpan.org/dist/SVN-Mirror/ | paste: sial.org/pbot/perl6 | pugs.blogs.com | pugscode.org | pugs.kwiki.org | www.treehugger.com/files/th_images/paradigm.jpg
Set by audreyt on 17 August 2006.
00:07 frankg joined
frankg read about perl6 inards the other day but I'm curious about the outer stuff now -- is the syntax completely diff from perl5? 00:08
as far as the OOP stuff goes?
zgh different enough, they're not going to backwards compatibility for the syntax
the object model is the same I think, at least the perl6 object model is available in perl5 now with Moose 00:09
markstos frankg: syntax for Perl6 objects is described here: dev.perl.org/perl6/doc/design/syn/S12.html
lambdabot Title: Synopsis 12: Objects -
frankg cool - that's what I need
markstos frankg: I think you'll be happy that writing "my $self = shift" frequent is no longer necessary. :) 00:10
You can skip that and just write self.other_method if you need to.
frankg yes, nice change -- I really want to see how it compares to python and ruby
zgh favorably
:-)
frankg good to know
00:25 nekokak joined 00:26 xdg joined 01:24 dolmans joined 01:32 mauke_ joined 01:35 justatheory joined 01:38 zgh left 01:42 mauke_ is now known as mauke 02:07 Khisanth joined 02:11 evalbot_12583 joined 02:26 agentzh joined
agentzh hi, there~ 02:26
audreyt: i feel like writing about Judy on win32 and smartlinks.pl on your blog site. :) 02:27
but first of all, i need to setup the auto-updating mechanism on feather. ;-) 02:28
Juerd++
TreyHarris: those two bugs in smartlinks.pl were fixed last night. feel free to complain to me if you encounter another bugs. ;-) 02:31
*bug
02:35 mako132_ joined
markstos More Perl6 blogging: use.perl.org/~markjugg/journal/30714 02:35
lambdabot Title: Journal of markjugg (792)
TreyHarris markstos: "new() is not needed for simple objects"? you mean defining a new sub is not required? 02:37
markstos TreyHarris: right. See dev.perl.org/perl6/doc/design/syn/S12.html and search on the page for... 02:40
lambdabot Title: Synopsis 12: Objects -
markstos All classes inherit a default new constructor from Object.
Perl5 has no default new(), and Perl6 does.
Although, I'm sure one of the method-making or object-making modules in Perl5 has a default like this. 02:41
wolverian markstos, "ww// has been replaced with < &gt"
er, s/ww/qw/
markstos, you're probably missing a trailing ; on the entity
markstos fixing... 02:42
TreyHarris markstos: sorry, yes, I knew that defining a new was not required. but that bullet point makes it sound like *calling* new() isn't required 02:44
02:44 agentzh joined
markstos Ah. I'll see if I can clarify. 02:46
TreyHarris (if you'll recall, I'm the one who informed you of that ;-)
markstos Oops. Everyone looks sort of the same on IRC. 02:48
TreyHarris heh, true :-)
markstos If only we had floating photos next to our names.
02:49 kanru joined
markstos Except those can be annoying sometimes... 02:49
TreyHarris: I've updated the post now to give you a credit mention at the bottom. Thanks! 02:50
wolverian: you too!
wolverian thanks :) 02:51
markstos++ # spreading the word
02:52 putter joined
TreyHarris I need to get people to start giving karma to trey++. I'm only TreyHarris on freenode, because somebody else grabbed the nick from nickserv first 02:55
but thanks :-) markstos++
02:59 agentzh joined
agentzh oh, dear...the terminal of feather is pretty slow here... 03:00
03:05 xinming_ joined 03:09 hikozaemon joined
markstos Ok. I think I found a bug. I'll explain it here, if it sounds like a bug to you, too, I'll submit a formal test for it. 03:12
putter audreyt, etal: Did a paragraph ever get written sketching what is interesting about p6 from a CS perspective? Ie, the type system, STM, parallel stuff, rules, syntactic flexibility, etc. 03:13
I'm mainly thinking of a concise jargony characterization. Though a longer sketch might be interesting too.
Moving on, (stevan - the tradition continues;) has there been any thought on some standard sub/macro names for low-level elements of p6 behavior? Like method calls or multi dispatch. Basically the prelude can define a lexically scoped MOP, which the normal "make prelude go fast" tools can usually compile away.
So aggressive flexibility for both syntax _and semantics_. :) 03:14
markstos If declare a simple parent and child class in the same file, things are fine. However, if I split it up into a parent class file, a child class file, and script to invoke the child, the bug appears. The child file can't find the parent unless I add "use Parent" to the child file. Shouldn't "class Child is Parent" be enough ?
svnbot6 r12584 | agentz++ | [util/smartlinks.pl] 03:15
r12584 | agentz++ | - commented out "use YAML::Syck;", since we no longer need it here.
agentzh ?eval 1+2
evalbot_12583 3
markstos Adding "use Parent" seems wrong, because it's traditionally a "has-a" relationship. 03:16
agentzh ?eval 1+2 03:17
evalbot_12583 3
agentzh oh, too bad...feather's apache uses the ISO-8859-1 charset...how can i change that to UTF-8? 03:22
Juerd? audreyt? 03:23
03:25 weinig is now known as weinig|zZz
agentzh has found an answer from Google. :D 03:27
TreyHarris didn't Damian say something at YAPC about writing up a "why Perl 6 should be interesting to academics" monograph? 03:28
markstos: not sure if that is wrong or not. it follows Perl 5, certainly--becoming a subclass of module A requires that you use module A before you express that relation. TimToady was saying last night that 'does' may need to have implicit requiring sometimes. perhaps 'is' would be covered by the same thing 03:29
(C<but> too) 03:30
s/have implicit/cause implicit/
markstos TreyHarris: Thanks for the response. I just asked on p6l for the Synopsis (12) to be clarified regarding this matter... it doesn't say one way or the other right now. 03:31
03:35 justatheory joined 03:36 traecer joined
markstos What's the equivalent for (scalar @a) ? 03:37
( Is there a handy p5 -> p6 conversion chart ? )
PerlJam markstos: scalar @a still works. You just have to use it in the right context to get the right value out :) 03:38
markstos Ah. 03:39
I also found Perl6/Perl5/Differences.pod, which I'm reviewing.
I think the answer might be @array.elems 03:40
TreyHarris markstos: yes, that's what you want
PerlJam: AFAIK, scalar @a doesn't work. pugs doesn't like it
?eval my @a = 1..5; scalar(@a)
03:40 evalbot_12583 is now known as evalbot_12584
evalbot_12584 Error: No compatible subroutine found: "&scalar" 03:40
markstos Right. That's what I got. 03:41
TreyHarris ?eval my @a = 1..5; +(@a)
evalbot_12584 5
TreyHarris that works too
but .elems is clearer
?eval my @a = 0..5; +(@a)
evalbot_12584 6
PerlJam could have sworn scalar was still in the perl lexicon 03:42
TreyHarris no. it's $(), @(), %() now. but autoboxing means the scalar value of @a is not what you want: 03:43
?eval my @a = 0..3; $(@a)
evalbot_12584 [0, 1, 2, 3]
TreyHarris scalar @a is just an array as a reference (what we'd call an arrayref in Perl 5)
er, s/as a reference/as a scalar/
03:46 justatheory joined
TreyHarris +(), ?(), and !() are sub contexts of scalar for numeric, boolean, and boolean-negating 03:50
markstos I'm trying to set a default value for a class trait. It looks like a specc'ed but unimplemented feature: 03:51
TreyHarris ?eval my @a = 0..3; my @b; [+(@a), ?(@a), ?(@b)]
evalbot_12584 [4, Bool::True, Bool::False]
markstos ?eval class Foo { has $.prop is rw = 1; }
evalbot_12584 Error: unexpected "=" expecting trait, "handles", comment, ";" or "}"
TreyHarris markstos: yes
markstos TreyHarris: Is there a recommended workaround, short of declaring a method for that?
I suppose I would just write $.prop = 1; at the top of the file... 03:52
TreyHarris ?eval class Foo { has $.prop is rw; START { $.prop = 1 } }
evalbot_12584 Error: Cannot cast into Hash: VRef <Sub:0x894c330>
markstos TreyHarris: Is the START block necessary to set a class default ? 03:53
TreyHarris ?eval class Foo { has $.prop; method BUILDALL { $.prop = 1 } } 03:54
evalbot_12584 undef
TreyHarris markstos: no, was looking for a workaround
markstos What about just:
TreyHarris ?eval class Foo { has $.prop; method BUILDALL { $.prop = 1 } }; my Foo $bar .= new; $bar.prop
evalbot_12584 \undef
markstos ?eval class Foo { has $.prop; $.prop = 1; } 03:55
TreyHarris ?eval class Foo { has $.prop; method BUILD { $.prop = 1 } }; my Foo $bar .= new; $bar.prop
evalbot_12584 Error: Cannot cast into Hash: VRef <Sub:0x894c330>
\1
TreyHarris ?eval class Foo { has $.prop; method BUILD { $.prop = 1 } }; my Foo $bar .= new; $bar.prop
evalbot_12584 \1
TreyHarris that's what you want
well, no, not what you want. but it's a workaround that works :-)
markstos ?eval class Foo { has $.prop; $.prop = 1; }
evalbot_12584 Error: Cannot cast into Hash: VRef <Sub:0x894c330>
markstos TreyHarris: Thanks! 03:56
TreyHarris generally you don't want executable lines inside a class entry. 03:57
s/entry/definition/
must go for awhile
&
markstos That makes sense.
TimToady that should be a submethod BUILD. 04:26
markstos TimToady. Thanks. I'll change my code. 04:27
TimToady and if it works, it should probably be seeing $!prop since it's trying to set the internal storage, not the virtual. 04:31
*setting
but the OO implementation is all getting reworked soon... 04:32
04:44 markstos left 04:58 avarab joined 05:14 agentzh joined
TreyHarris < TimToady> but the OO implementation is all getting reworked soon... 05:15
talk about quietly dropping a bombshell :-) 05:16
05:17 lambdabot joined
wolverian I think he means pugs's internals ... 05:18
05:28 Khisanth joined
TreyHarris is there any workaround to overload an object's stringification, since neither *infix:<as> nor prefix:{'~'} seem to currently work? 05:34
05:40 jferrero joined 05:49 xinming joined
agentzh feather.perl6.nl/~agentzh/syn/ 06:09
lambdabot Title: Index of /~agentzh/syn
wolverian nice.. what's up with the huge dividers? 06:11
06:14 marmic joined
agentzh dividers? 06:14
wolverian between the sections 06:16
e.g. immediately after the table of contents, there are two of them, with "back to top" in between
the dividers seem oddly big, vertically
agentzh okay, i'll remove the dividers soon. ;-) 06:17
i'm also feeling uncomfortable with dev.perl.org's CSS. 06:18
wolverian: what's your favourite CSS for pod2html?
wolverian I don't have one.. they all seem rather unsexy
agentzh is sad...
wolverian the CPAN CSS is ok 06:19
agentzh i'll look into ActiveState's CSS and also JSAN's.
agentzh thinks CPAN's CSS is terrible.
wolverian I'd like to see one where the table of contents is a sidebar
and the titles aren't in ALLCAPS
:)
(doable-ish with CSS) 06:20
agentzh hmm, maybe we should get someone to write a new CSS file. 06:21
06:21 Aankhen`` joined
wolverian hm, the index <ul> isn't marked with a specific class or id 06:22
agentzh btw, feather.perl6.nl/~agentzh/syn will be automatically updated every one hour. ;-) 06:23
lambdabot Title: Index of /~agentzh/syn
agentzh wolverian: the index is generated by Pod::Html. :)
wolverian yup 06:26
it doesn't validate, btw :-) 06:27
missing one alt attribute
er, that was a random CPAN page, sorry
well, validator.w3.org/check?uri=http%3A%...2FS02.html 06:28
ouch
lambdabot Title: Result for feather.perl6.nl/~agentzh/syn/S02.html - W3C Markup Validator
agentzh wolverian: fixing 06:30
wolverian agentzh++
I'm fearfully low on tuits, sorry
agentzh np
06:33 baest joined
agentzh wolverian: sigh. most of the HTML errors are caused by Pod::Html... 06:56
lambdabot Im going to be a movie star. I just applied online for this new movie with chris farley. It only cost me $50!
wolverian agentzh, really? I wonder why it claims to be xhtml 1.0 strict then..
lambdabot i'm 19
wolverian and what does search.cpan.org use then? (it has only sporadic errors)
agentzh i'll upgrade my local Pod::Html then. 06:57
wolverian there's Pod::Xhtml too, from BBC :)
agentzh looking 06:58
wolverian oh.. search.cpan.org is html4
agentzh i'm wondering which module search.cpan.org is using. 07:01
svnbot6 r12585 | agentz++ | [util/smartlinks.pl]
r12585 | agentz++ | - first attempt to fix HTML errors, wolverian++
07:03 ludan joined
agentzh is pondering switching to Pod::Xhtml. 07:03
xinming hmm, anyone here who uses emacs for editing perl 6 scripts? 07:06
TreyHarris xinming: I do 07:17
07:23 neoesque joined, Lorn joined
xinming TreyHarris: hmm, Ok, since in perl 6, we can't use `sub { } sub { }`, so, we need `sub { }; sub { }` but the problem is, if you use indent in cperl-mode, It will make a newline for the ';' How can we avoid this? 07:25
TreyHarris: I mean the C-M-\ 07:26
07:26 iblechbot joined
xinming TreyHarris: try a small example with C-x h C-M-\ 07:26
07:35 avar joined
TreyHarris xinming: i'm not seeing that behavior 07:41
can you paste a small script so i can try what you're trying?
xinming lisppaste3: url 07:42
lisppaste3 To use the lisppaste bot, visit paste.lisp.org/new/perl6 and enter your paste.
xinming TreyHarris: Ok.
lisppaste3 xinming pasted "indent problem for perl6" at paste.lisp.org/display/24531 07:45
07:45 SSelva joined
xinming TreyHarris: I think you've customized your cperl-mode, Mine is not configured, I'm lazy to do customization if it doesn't hurt. :-) 07:46
TreyHarris xinming: no, i'm seeing the same behavior
but why do you have semicolons there?
agentzh wolverian: the HTML code generated by Pod::Xhtml contains almost 0 error. :)
wolverian agentzh, almost? :( 07:47
agentzh there's 1 warning according to my Firefox.
xinming TreyHarris: Because, in perl 6, 'sub { } sub { }' is not allowed. 07:48
TreyHarris that's true
but you're not on the same line
a closing brace as the last thing on a line always terminates a statement
so the semicolon is not required
xinming is ashamed for a long time not follow the synopsis... 07:49
TreyHarris: thanks
TreyHarris "sub foo { ... }; sub bar { ... }", but "sub foo { ... }\nsub bar { ... }"
thought i think there's an exception for "if $foo { ... }\nelse...." 07:50
agentzh wolverian: sadly the HTML pages from Pod::Xhtml are ugly... :/
xinming I understand now, for two sub { } in a line without semicolon would raise an error.
TreyHarris xinming: exactly
xinming but if they are not in a line, then, they are legal.
wolverian agentzh, surely you can change the CSS?
xinming thinks, It's like a small trick in perl 6. :-) 07:51
agentzh wolverian: well, it's not a CSS issue. Pod::Xhtml inserts too many "<br /><br />" into the page.
wolverian oh
time for a subclass
Pod::Xhtml::Cute
agentzh hehe 07:52
wolverian: subclassing now... 07:53
wolverian :-) 07:54
agentzh++
agentzh i wonder if i should send Pod::Xhtml::Cute to CPAN. :)
TreyHarris xinming: it's very like the trick about everything being a closure now, or $foo = @bar not being a reference but still making $foo have the content of @bar. it's all very subtle stuff that TimToady++ has put a lot of thought into so that you don't have to think about it most of the time :-) 07:55
07:56 kane-xs joined
TreyHarris audreyt: it's midafternoon, where have you been all day? we miss you. :-) 07:59
08:05 nothingmuch joined
TreyHarris moose! 08:07
08:10 Lorn_ joined
agentzh wolverian: i think postprocessing the HTML source is the simplest way. Pod::Xhtml is not well designed and subclassing it is particularly hard. 08:10
wolverian agentzh, oh. well then. 08:11
HTML::TreeBuilder to the rescue!
(or if you can find a valid SGML POD outputter, XML::LibXML would work too)
agentzh wolverian: why not plain regexes? 08:12
i think plain regexes are fine.
only a few lines of code...
wolverian because it's reinventing the wheel 08:13
the various ways that HTML can be broken are so mindbogglingly, well, various that you don't want to touch it yourself
agentzh hehe, i don't just want another dependency and extra cost. 08:14
s/don't just/just don't/
everything you said is quite true. :)
wolverian mm, well discuss it with audrey and others if you think you don't want to make the decision 08:15
agentzh hmm, the HTML source is generated from Pod::Xhtml, so i think it's fine as long as Pod::Xhtml doesn't change.
wolverian well if you have a dependency on Pod::Xhtml already.. adding another isn't much 08:16
and HTML::TreeBuilder is rather common
08:17 elmex joined
agentzh hehe, okay, i'll leave that to someone else as an exercise. 08:17
wolverian sure
agentzh in the meantime, i'll use the regexes. 08:18
:)
wolverian XML::Liberal is nice too 08:19
agentzh looking now 08:20
oh dear...Pod::Xhtml doesn't support "=hegin html"... 08:25
svnbot6 r12586 | masak++ | * realized, through Conrad Schneiker's mail to perl6-users, that there's a search interface for #perl6 archives at colabti.de/irclogger/irclogger_log_search/perl6
r12586 | masak++ | * added 14 new references to READTOO that turned up by searching archives for "paper"
r12586 | masak++ | * (still need to backlog just to be sure, though)
r12586 | masak++ | * removed section headings "Current" and "Older" in READTOO, since they were not used, and no clear motive exists for moving things to Older right now
agentzh *begin
lambdabot Title: #perl6 irc log search
08:28 traecer left 08:44 buetow joined
dolmans say there is a class Point with two attributes x and y, are these legal to constructor an object: new Point(x=>2, y=>3) and Point(x=>2, y=>3)? 08:52
08:53 Khisanth joined 08:54 agentzh joined
agentzh i think i should stick with Pod::Html...Pod::Xhtml is not satisfactory at all... 08:57
09:02 chris2 joined
xinming ?eval class A { method update { $.b = "abcdefg" }}; A.new.update; 09:09
09:09 evalbot_12584 is now known as evalbot_12586
evalbot_12586 \"abcdefg" 09:09
xinming hmm, should this raise an error instead of returning the value in $.b?
because $.b hadn't been declared yet. 09:10
nothingmuch ingy: congratulations! 09:17
09:32 MacVince joined 09:34 zgh joined 09:40 avar joined 09:42 integral joined 09:50 drrho joined 09:58 dakkar_ joined 10:04 ruoso joined 10:28 visq joined 10:40 MacVince left 10:47 agentzh joined 10:53 Aankhen`` joined, Southen joined 10:55 alinbsp joined 10:59 frederico joined
svnbot6 r12587 | agentz++ | [util/smartlinks.pl] 11:02
r12587 | agentz++ | - stripped extra newlines
r12587 | agentz++ | - got rid of "=begin html"/"=end html"
r12588 | agentz++ | [util/smartlinks.pl] 11:26
r12588 | agentz++ | - added more docs to usage outputs.
r12588 | agentz++ | - added the --fast option with which smartlinks.pl won't
r12588 | agentz++ | update the synopses from the web.[util/smartlinks.pl]
r12588 | agentz++ | - confirmed that option --syn-dir works
r12588 | agentz++ | - added docs for the --syn-dir option to usage outputs.
r12589 | agentz++ | [t/02-test-pm/5-todo.t] 11:32
r12589 | agentz++ | - removed too advanced tests where use "[,] ..."
agentzh fglock: it seems to me the latest v6.pm still can't compile the new Test.pm with &todo defined. are you using an older version of Test.pm? 11:34
fglock: sorry, my mistake. it compiles now. ;-) 11:38
11:42 iblechbot joined 11:56 elmex joined 11:57 nothingmuch_ joined 12:15 weinig|zZz is now known as weinig 12:18 Limbic_Region joined
Limbic_Region TimToady - thanks for fixing all_parse.t and some of the examples that weren't passing due to syntax errors 12:26
Limbic_Region was afraid he was going to screw something up
12:28 baest joined 12:29 takanori joined
agentzh oh, it's very quiet here today... 12:30
12:32 weinig joined
Limbic_Region I can see 12:38
there are certain people that seem to be a catalyst for in channel participation - a synergy gets started and more and more join in 12:39
when they are absent - people tend to work on stuff on their own
both are good things IMO
Juerd 05:23 < agentzh> Juerd? audreyt? 12:41
agentzh?
agentzh Juerd: hi
Juerd: cron is really helpful. :) 12:42
Juerd agentzh: Yes, it is. Is this your first experience with it? 12:43
agentzh Juerd: and the .htaccess file also solves the charset problem of Apache.
Juerd: yes. i seldom or never use Linux.
feather++ 12:44
Juerd agentzh: If you have some spare time to write these things in FAQ-form for the website, I'd be delighted
(The site is in the pugs repository, under docs/feather)
agentzh Juerd: i'd love to!
Juerd Thanks
agentzh my pleasure. :) 12:45
currently i'm writing an article on smartlinks.pl for audreyt's blog. :)
will do that in an other day. :) 12:46
Juerd What are smartlinks?
agentzh please look at the outputs of smartlinks.pl and then you'll see: 12:47
feather.perl6.nl/~agentzh/syn/S04.html
lambdabot Title: S04
agentzh :) 12:48
Juerd Ooh... Visual coverage :)
agentzh yeah :)
with this tool, i'll be more motivated to write tests for pugs. :) 12:49
Juerd It's great that people create selfmotivators :)
agentzh ...util every paragragh of the synopses is covered.
hehe
*until
Limbic_Region fglock ping 13:02
13:03 fglock joined
Limbic_Region fglock ping 13:03
oh, he is using CGI:IRC - it takes a second for things to load 13:04
fglock ping
13:04 vytautas joined
svnbot6 r12590 | fglock++ | v6 - updated command line options, docs (requires a Module::Compile patch) 13:06
13:09 fglock joined
Limbic_Region tries again 13:10
13:10 fglock joined
Limbic_Region doesn't try again 13:10
fglock ping
fglock Limbic_Region: pong - sorry, my browser was frozen
Limbic_Region no worries - here's the deal 13:11
for the last two days I have tried to update PCR via CPAN.pm - it keeps telling me it is up to date 13:12
pasteling "agentzh" at 210.22.200.67 pasted "The first sketch of my post on smartlinks.pl to Audrey's blog. Please comment. :)" (107 lines, 4.1K) at sial.org/pbot/19212
Limbic_Region so I decide to go and download and install it manually
I click the download link from cpan and get
fx.saintjoe.edu/pub/CPAN/authors/id....16.tar.gz - 404 not found
any idea what's going on? 13:13
integral (that could be indicative that your mirror is behind; I can get it from a mirror near to me)
agentzh Limbic_Region: change your CPAN mirror and try again? 13:14
Limbic_Region agentzh - I tried several but perhaps they are all not good
agentzh really? i'm impressed.
Limbic_Region tries another one at random
fglock Limbic_Region: try cpan.pair.com
Lorn_ fglock: hi :) do you read my reply? 13:15
Limbic_Region got it
agentzh oh, the latest one on my mirror is 0.12. sigh. 13:16
fglock Lorn_: I'll read it now
[particle] agentzh: good blog post 13:18
agentzh particle: thanks! 13:19
[particle] i'm just working on adding some unique ids to the pge test suite
i think i may change those to smart links
...that format, anyway.
agentzh particle: nice
[particle] right now, they disallow ws chars
i could instead wrap them in <>, with optional first char
this would be a superset of pod links 13:20
agentzh is there any sample? url?
[particle] today
right now it exists only in coke's working copy of parrot
agentzh ah, okay. no hurry. :)
fglock Lorn_: create a local copy with: ' svn co svn.openfoundry.org/pugs pugs ' 13:23
lambdabot Title: Revision 12590: /
13:28 cjeris joined
audreyt agentzh: great writeup! I have a feeling that, like TestMatrix, this will propagate to the CPAN community in due time 13:28
I'd like to annotate my module's tests against POD too :)
agentzh hehe
thank you.
audreyt it's interesting how Pugs has become a net exporter of software engineering technologies into CPAN... 13:29
agentzh yes, it's fun.
-Ofun
audreyt glad to know that the optimization worked 13:30
audreyt spent a whole day with GHC on AIX 4.3... happily everything worked out for $job
Limbic_Region audreyt - a kludgey version of the example parseability tester is in the repo now
audreyt Limbic_Region: yes, I noticed, I noted TimToady's commits too :)
Limbic_Region++
Limbic_Region TimToady added a try block which fixed a problem he was having locally plus a half a dozen fixes to the examples
oh good
audreyt TimToady++
Limbic_Region was afraid $work had completely consumed you
yes TimToady and audreyt plus plus 13:31
audreyt the client site has no internet
so... yeah.
(and it's in 8th floor, so citywide wifif didn't work there eitheR)
Limbic_Region :-(
is there any reason for me to test the Win32 whacky behavior or have you not had a chance to look locally yet? 13:32
agentzh audreyt: is there a delay between saving posts and showing posts on your site?
audreyt I havn't got a chance. fortunately tomorrow is $job-free day
agentzh: no there is not
agentzh is trying to find a button to click... 13:33
Limbic_Region well, don't spend a lot of time on it - your time is better used elsewhere - there are kludgey work arounds and "working" is what's important
audreyt true but, a failing test in t/ will make me feel better :) 13:34
or a failing test in ext/File-Find/t, come to think about tthat.
now the perf work is done, I plan to spend tomorrow looking at regressions
before I swap in the new MOP
so that the Judy/Interning regressions doesn't get mixed with MOP regressions 13:35
so if a failing test is there tomorrow I'll look at it... probably first
as ext/ sorts higher than t/
Limbic_Region is going to run a new smoke this morning
audreyt agentzh: you put the state in "Draft"
agentzh oh... 13:36
Limbic_Region is there a reason that parrot has to be part of the path to run rules tests or can PCR handle those tests?
agentzh what can i do now then?
audreyt nothing... it's posted
agentzh++
agentzh okay, finally see the "publish" option...
audreyt ...which I pressed for you. it's on air now :) 13:37
agentzh++
agentzh :)
oh, it looks nice. 13:38
13:39 mako132 joined
agentzh except the titles... 13:39
i mean subtitles...
13:40 cmarcelo joined
cmarcelo hello 13:40
Limbic_Region repeats his earlier question 13:41
agentzh cmarcelo: hi
Limbic_Region is there any reason Parrot needs to be accessible to run the full test suite - can't PCR fill the role of rules?
cmarcelo agentzh++ # smartlinks is cool =) 13:44
agentzh Hmm, good question. :)
cmarcelo: thanks!
audreyt Limbic_Region: yes it can. 13:45
want to help writing a bridge?
also, which one should be fallback?
modifier? pragma?
use rx 'PCR';? 13:46
currently Pugs expects to send and receive rule match objects in a particular format 13:47
Limbic_Region audreyt - sorry, was on PerlMonks
audreyt PGE_Match 2 5 [PGE_Array [PGE_Match 2 5 [] []]] []
a dumper form PGE into that format is in parrot/runtime/parrot/library/PGE/Hs.pir 13:48
Limbic_Region not sure as far as writing a bridge - /me thinks fglock much better qualified - but here are some thoughts
audreyt if PCR match object can dump to that format
then we can trivially call out to PCR
13:48 mdiep joined
audreyt the production rules are: 13:48
Limbic_Region 1. We should have a single module or pragma that does the fallback for us - that way test scripts are abstracted and any rules plugin can be added later
audreyt data MatchPGE
= PGE_Match Int Int [MatchPGE] [(String, MatchPGE)]
| PGE_Array [MatchPGE]
| PGE_String String
| PGE_Fail
13:49 hexmode joined
audreyt the "Int" part and "String" part are normal numbers and double-quoted strings; the "MatchPGE" part is recursive 13:49
i.e. it can contain submatches. (either positional in the [MatchPGE] array, or named with the ("name", match) syntax)
Limbic_Region 2. That module/pragma should also be our universal translator - we should not expect PGE and PCR to maintain uniformity 13:50
audreyt a "bridge" is a perl5 module, or a new method to PCR's Match, that causes this format be dumped
Limbic_Region 3. This should all be done by someone who knows what they are doing ;-)
audreyt sure. the translator for PGE, for example, ins in src/pge/run_pge.pir
which is outside PGE proper 13:51
Lorn_ fglock: the problem with Module::Compile it is a v6.pm problem? or Module::Compile problem?
13:51 Lorn_ is now known as Lorn
fglock audreyt: this command line syntax now works, requires a Module::Compile patch to ignore the '-e' switch: echo 42.say | perl -e 'use v6-alpha' 13:51
Lorn_: it's a perl5 command line problem
audreyt fglock: oh wow
fglock: patch url?
fglock audreyt: one sec
audreyt Limbic_Region: do you think calling it re::engine makes sense? 13:52
or re::override? or just "re"?
Lorn fglock: hmm, ok, i finished the instalation of svn, and i'm sync now. 13:53
audreyt or %ENV<PUGS_P6REGEX_ENGINE> ?
13:54 Alias_ joined
Alias_ seen audreyt? 13:54
audreyt Alias_: hi! ltns
Alias_ indeed, trips to other countries and dealing with personal/family issues and eve online (darned Schwern) 13:55
audreyt oy. but I see you survived and are back to the modlist
Alias_ Yeah, I'm gradually starting to spool up normal life again 13:56
audreyt vanilla at $job has proved to be invaluable. thanks for getting xdg on the boat :)
Alias_ WRT P6 sparse arrays, am I right in thinking that for all current cases where we use hashes with positive integer keys, it's now going to be faster to use arrays?
audreyt why, yes.
Alias_ And less memory too? 13:57
Alias_ thinks maybe, but not much less
audreyt sure.
well, one pointer less for each key
Alias_ righto
audreyt so not much.
Alias_ That's a lot though, when you're data is just a boolean though 13:58
s/though//
audreyt er, yes.
Judy has another Set type
which can be used natively to implement Set/Junctions some day I hope
optimized for the case where your data is just a boolean
Alias_ What is $work at the moment, that Vanilla is used for
audreyt it's an jifty application running on AIX and Win32. 13:59
Alias_ Because Jifty is one of my test cases for Win32, and currently is fails horribly
Lorn fglock: i got some error, on perl -Ilib -e 'use v6-alpha' ' "hello world".say '
Alias_ s/is/it/
Lorn fglock: pastebin.com/773231
buubot The paste 773231 has been copied to erxz.com/pb/2846
audreyt fails, as in not passing tests for its deps? well I use "notest install"...
Alias_ eep
ok
Just FYI also, if you disable the strange and sort of broken privilege escalation protection, Vanilla works on the latest Vista too 14:00
audreyt I mean, andreas implemented notest in CPAN shell, I gotta use it for something... 14:01
k, noted :)
fglock Lorn: you need to run perl5/Pugs::Compiler::Rule 'make install'
Khisanth well there is a but in Cache::Simple::TimedExpiry in the latest version so that particular dep isn't going to pass either way ...
Lorn fglock: ok
14:01 rashakil_ joined
Alias_ Also, kane is finally going to add a CPANPLUS_IS_RUNNING-like environment variable that actually does what it says 14:01
So I might also soon get enough of Module::Install::With able to be finished to start porting bits of M:I over to it 14:02
audreyt good. we need to probe that in Module::AutoInstall
exactly the same way it probes for the .cpan lock
Alias_ nods
See M:I:With code for my centralisation of all the "who am I playing with" helper functions
Once all that works, I'd like to go through the rest of the code and switch everything over to it 14:03
audreyt now that 5.10 is looming with the inevitable CPANPLUS borgification
Alias_ Which still freaks me out
Module::Build isn't ready yet, CPANPLUS arguably isn't ready yet
audreyt it's very important that the, ehem, dual core" situation does not cause breakage
Alias_ Any eta on 5.10
audreyt Module::Build is already in
Alias_ It's in, but it doesn't work
As long as the big problems are fixed before 5.10, I'm happy though 14:04
audreyt rgs makes the call on eta, not me :)
Alias_ true
audreyt as long as there are big problems, I'm pretty sure there won't be a 5.10 :)
Alias_ Dunno if rgs is aware of them all though
Maybe I should email him
(I really don't have time for this stuff though)
audreyt s/him/p5p/ 14:05
Alias_ Well yes, but him first
Khisanth email him your IRC logs :)
audreyt so I can ge tthe mail too and chime in (yes I'm back to p5p but havn't had much time except lurking
Alias_ I got a ticking off from Bill/Richard at YAPC for not contacting people quitely first enough before I, as they put it, open up all guns blazing :)
14:05 bpphillips joined
Alias_ quietly 14:05
e.g. the rt.cpan anon web thing 14:06
audreyt *nod*
fglock audreyt: mm - now it seems to work even without the M::C patch - maybe I was doing something wrong
audreyt fglock: cool then, as I think M::C wouldn't interfere
that's fantastic :) 14:07
Alias_ BTW, how soon should I have before PPI use v6 support starts to become needed
audreyt as soon as the first CPAN module that uses v6 becomes a dependency for a non-v6 module
which... I don't know how soon.
definitely before this christmas.
Alias_ ok
hrm
sigh
Limbic_Region audreyt - $work called me away - will back log and let you know what if any input I can give - TTFN
Limbic_Region & # $work
Alias_ I guess I need to look into generalising the implementation of hidden content I created for heredocs 14:08
audreyt but you can work reactively, i.e. punt until someone screams at you :))
Alias_ Oh I am
Lorn fglock: the delay for print 'hello world' its normal?
Alias_ The PPI bug queue is already embarrasingly long, and "No patches accepted" is starting to yield fruit
audreyt alias++ 14:09
Alias_ audreyt++ # Everything I ever learned about encouraging contributors
audreyt :D
Alias_ (... I stole ruthlessly from you)
sigh, ok, back to it then 14:10
audreyt my pleasure. *curtsies*
Alias_ I just saw ingy's grant acceptance
14:10 zgh joined
Alias_ Which means I probably need to get YAML::Tiny finished too, because YAML.pm gets yoinked out and replaced :/ 14:10
fglock Lorn: how much delay?
Alias_ ETOOMUCHTODO 14:11
On the plus side, Mango++
[particle] ETOOLITTLETIME :)
Alias_ indeed
Khisanth $Alias.clone; :)
fglock Lorn: it needs some time to precompile the compiler, but that only happens once
Alias_ Khisanth: I'm a Gemini... so often I've wished I was _actually_ twins
Lorn fglock: the time is user 0m6.510s 14:12
Alias_ Khisanth: But then I wonder, would I be the cool twin, or the shy twin...
Lorn fglock: and the cpu go to 100% while compile...
Khisanth well if they are clones then all of them is you
Lorn ops
Alias_ audreyt: oh, also on the WTF queue, WTF is with YAML::Syck putting a blank line after --- 14:13
Lorn interprets ( confuse now )
Alias_ audreyt: YAML::Syck is less compatible with YAML than YAML::Tiny is, in my tests
(and I only test the JSON-subset of data) 14:14
audreyt Alias_: ...a blank line?
Alias_ ---\n\ncontent
audreyt $ perl -MYAML::Syck -e 'print Dump "content"'
--- content
Alias_ Reported by (I forget) 14:15
hang on
audreyt Lorn: if you save your file instead feeding it via STDIN to perl -e, then the compilation is cached 14:16
Alias_ audreyt: checkout svn.phase-n.com/svn/cpan/trunk/YAML-Tiny 14:17
edit t/lib/MyTests.pm to enable $COMPARE_SYCK, then make test 14:18
The only tests for YAML::Tiny are to compare it's behaviour against YAML
(and now Syck)
its
So either 1. YAML.pm needs to change, 2. YAML::Syck needs to change 3. erm.. dunno 14:19
14:22 prefiks joined
audreyt Alias_: YAML::Syck::Dump( @$object ) doesn't workl; Dump only handles one arg at the moment 14:22
agentzh g'night & 14:23
audreyt I don't see the blank line
14:23 agentzh left
audreyt agentzh: g'nite! 14:23
Alias_ ok, will skip tests for multi-document tests
audreyt I changed it to
local $@;
my $syck_out = eval { YAML::Syck::Dump( $object->[0] ) };
which test reveals the blank line thing? 14:24
Lorn audreyt: i not use -e, i create a file "use v6-alpha; "hello, world".say;" 14:25
audreyt ah. then second time around it should be faster, yes 14:26
ingy hola
fglock Lorn: did you have Cache::Cache installed?
s/did/do/
audreyt ingy: congrats re grant 14:27
Lorn fglock: Cache::Cache is up to date.
audreyt I see there's a C-based libyaml now too :) 14:28
Alias_ (but it isn't finished...?) 14:29
See Perl Foundation blog
OK, changed tests to skip if not @$object == 1 and blank line bug goes away 14:30
hmm... and other bugs found :)
svn update and repeat make test 14:31
syck doesn't strip comments? 14:32
fglock Lorn: it takes about 6s (windows XP, crusoe 1GHz) and 1s (ubuntu, pentium 4, ???GHz) 14:33
14:33 vel joined
ingy audreyt: thanks 14:34
I found out about that here first ;) 14:35
audreyt x: foo
#bar
fglock Lorn: and it takes about half the time when running the file a second time
audreyt ingy: the snippet above, syck considers it a line continuation and YAML.pm considers it a comment
fglock biab &
ingy audreyt: syck is correct I believe...
audreyt Alias_: syck is correct I believe... 14:36
ingy haha
Alias_ hrm... will YAML.pm be fixed?
audreyt now ingy gets his grant, sure :)
ingy I'm not entirely sure actually
Alias_ Because the tests it fails are from real world sample files I got from Plagger and Vanilla Perl
So any change has real world impact
ingy Alias_: I'll ask on #yaml 14:37
Alias_ ok, thanks
ingy: Checkout and make test on svn.phase-n.com/svn/cpan/trunk/YAML-Tiny to see the cases where YAML and YAML::Syck disagree on comments 14:39
Lorn fglock: hmm, i had do something wrong, because in perl -e or file.pl the time is, about 7s 14:41
ingy 06:44 < xitology> seems to be a comment :)
06:45 < xitology> ' #' cannot appear in a plain scalar.
audreyt (though it's isomorphic to the foo/#bar one above)
Alias_ (right) 14:42
ingy xitology is the best spec guru we have
Alias_ So I should assume a syck bug?
ingy looks so 14:43
Alias_ ok
ingy looks forward to integrating libyaml
14:43 iblechbot joined
ingy audreyt: is there a url for my grant acceptance? 14:43
audreyt news.perlfoundation.org/2006/08/gra..._to_1.html 14:44
14:44 ivas joined
audreyt use.perl.org/article.pl?sid=06/08/22/1016231 14:44
lambdabot Title: use Perl | 2006 3rd Quarter Grant Votes
audreyt "The PyYAML port is to be done by Ingy (Brian Ingerson) and was approved because frankly, YAML support in Perl ain't all that great and there are few better qualified that Ingy to get this done right."
ingy I hope that is true... 14:46
I think it is
But I'll be doing it pugstyle anyway
Alias_ Ditto. If there are any, they are most certainly "few"
audreyt I can't think of one... 14:47
nothingmuch_ Alias_ || audreyt: can you please pull magic to give IVORW comaint on Tie::FTP? 14:50
it doesn't show up in my comaint giving page
Alias_ oh right, yes
14:50 nothingmuch_ is now known as nothingmuch
audreyt I'll let alias do it :) 14:52
Alias_ nothingmuch: You can't because GIFF holds primary
Alias_ didn't even know audreyt had perms :)
nothingmuch GIFF?
audreyt I've had it for years now :)
Alias_ Tie::FTP GIFF first-come 14:53
nothingmuch wtf
it's not in his author dir
Alias_ audreyt: on Tie::FTP, or as PAUSE admin?
nothingmuch hmm
i may have passed it on
but not primary
audreyt Alias_: pause admin
nothingmuch i never do that =/
odd
audreyt I think I'm still pause committer too
not sure
Alias_ You can't pass on first-come 14:55
only primary
"Added IVORW to co-maintainers of Tie::FTP." 14:56
IVORW will probably need to upload again
If they have already...
audreyt ...or reindex...
Alias_ or that
audreyt (background: openfoundry.org was running on a forked copy of PAUSE, and I hacked on PAUSE.pm a bit to make it uploadable to CPAN) 14:57
but then we switched everything to svk and that plan was dopped
clkao you mean chromatic switched everything to svk? 14:58
audreyt not sure what that has to do with chromatic 14:59
14:59 ivas left
audreyt we as in the openfoundry-version-control-plan-as-drafted-by-clkao :) 14:59
Alias_ clkao: So... SVK::GUI... :) 15:00
clkao audreyt: everything2 15:01
audreyt lol
15:03 plisk joined
svnbot6 r12591 | fglock++ | * v6 - export V6NOTIDY=1 - and Test.pm compiles in less than 4s 15:05
15:07 ivas joined
audreyt wow 15:08
my macbook just fell from 2m height to floor
LCD case cracked open and everything froze
Limbic_Region does the smoke harness abandon if a test is taking "too long"?
obra poor audrey :/ 15:09
audreyt (it fell LCD-side down first)
...and then it rebooted itself
PerlJam fell? Or was it *pushed*? /me suspects foul play.
fglock audreyt: ugh
audreyt then everything continued to work just fine.
the plastic case is rather easy to push back in, too.
Alias_ Just like your intestines when you suffer a gut wound
(doesn't mean everything is ok though) 15:10
Limbic_Region t/examples/all_parse.t works fine for me outside of the harness but the smoke report showed 100% failure???
audreyt PerlJam: fell. I was in the shower, the computer is placed on the shelf, which I guess became a bit slippery.
I'll probably refrain from IRC'ing from the shower for a few days.
Alias_ PerlJam: Questions about the use of a computer in the shower will not be addressed
PerlJam Alias_: I have no such questions. I take a laptop with me to the strangest places too :) 15:11
Alias_ audreyt: I think you really need a ToughBook specifically for that case
audreyt I'm surprised at how tough this macbook has proven itself to be.
[particle] suggests dry cleaning
Alias_ Speaking of macs, I should fire up my new mac
audreyt esp. the LCD part, which if it was the old Asus, would have broken ~4 times already
Alias_ wonders what sort of screen it takes
Time to buy it a monitor methinks 15:12
audreyt but it's clearly a sign I should stop abusing it :)
g'nite
15:13 bpphillips joined
Alias_ What. The. Hell 15:14
Mac OS 9!!!
Hm.. background picture of someone else's girlfriend... 15:15
I think this mac wasn't cleaned before it was sold
Alias_ prepares for some Voyeurism
PerlJam Alias_: I take it when you said "new" you meant "new to you"?
Alias_ I hate Macs, but I'm starting to need one for things like File::HomeDir testing
And other platform issues
So I bought a G4 tower for $100 on eBay 15:16
hmm.. and apparently it's hung already 15:17
fglock clkao: try running v6 make test with 'export V6NOTIDY=1'
Alias_ I guess now I need to get a OS X license 15:18
clkao fglock: uhm?
Alias_ oh man, cooperative multiprocessing really sucks 15:19
fglock clkao: it disables the call to Perl::Tidy, goes 25% faster here
clkao oh ok. i don't need v6 to be any faster
otherwise i will actually start using them
too dagerous ;)
15:19 marmic joined
fglock clkao: :) 15:20
clkao fglock: do the tests i included pass now?
i saw lots of failure yesterday
fglock clkao: I don't know yet - I'll try again in a few hours 15:21
15:21 aholanda joined
Alias_ hrm, if a program hangs in OS 9, there wasn't much you could do was there 15:23
clkao ok 15:24
fglock clkao: I'll look for a zip() implementation - this should fix hash.t 15:25
lunch &
15:28 kanru joined 15:30 FurnaceBoy joined
nothingmuch Alias_: am I maint right now? 15:54
Alias_ yes
nothingmuch okay
thanks
nothingmuch wasn't sure from the statement 15:55
[particle] where can i find the pod spec for perl6?
audreyt pugs/docs/Perl6/Spec/Documentation.pod
[particle] thx!
audreyt damian says he'll commit a new rev in a few days 15:56
15:56 Aankhen`` joined
audreyt it's still in flux... 15:56
[particle] will the smartlink format fit into std pod?
audreyt clkao: to support zip() and each(), Data::Bind need to support the idea of pipes
Alias_ What is changing?
audreyt namely, the callconv would be
func($obj, \@pos1, \%nam1, \@pos2, \%nam2, ...) 15:57
at perl5 land
in zip(1,2,3;4,5,6), the (1,2,3) are in pos1, aand 4,5,5 are in pos2 15:58
clkao what what? why second pos and named?
oh ok
audreyt feeds
clkao and how do we use them?
uhm wait
isn't that [\[1,2,3],\[4,5,6]] ? 15:59
hm, i don't know ; much yet
audreyt sub zip (*@@slices) { } 16:00
clkao @.@
lambdabot Not enough arguments to @.
audreyt then @slices[0][0] is 1
@slices[1][0] is 4
lambdabot Unknown command, try @list
audreyt @slices[0]<foo> is not defined
lambdabot Unknown command, try @list
audreyt but if you had said zip(1,2,3, foo=>4 ; 5,6,7)
then it would be 4.
yes, very complicated. 16:01
clkao "hide until" next week
audreyt (and pugs's new ast migration is partly to address this)
clkao sorry, i have svn::mirror craziness to fix
audreyt which simply had no counterpart in old ast
clkao *nod*
16:01 Alias_ joined
audreyt oh. "svk pull" for pugs now always do a mergeback and report checksum mismatch 16:01
clkao what's the status compiling the ast to classes?
audreyt on my laptop
clkao audreyt: a mergeback?/? how come? 16:02
audreyt I don't know. I just ^C for now
truly weird. might be the new pull logic
clkao nopaste the output?
hmm
audreyt and just for pugs not other checkouts
comp to mooseclass is done but the ast itself is being integrated
maybe in another couple weeks when it's stable and working 16:03
great. openfoundry down again so I can't test :/
till tomorrow then 16:04
obra morning, audrey 16:08
16:09 xinming joined 16:14 alinbsp joined 16:24 xerox joined
xinming audreyt: ping 16:27
16:27 zgh joined
Alias_ obra: Someone responded to my post about web anon-rt.cpan pointing to a CPAN module for interfacing with a blog comment anti-spam thing... it may or may not be useful 16:29
16:29 Psyche^ joined
Alias_ Your thoughts on that avenue? It means pushing the bug reports over the network to the service before accepting... 16:29
obra Alias_: if you can sell robrt on it, I'll do it. but I'm tired of battling the commentspam. (note that many of our best spammers are humans. the human tests Do Not Helo) 16:30
Help, even.
Alias_ Yeah, they said
I am stunned at the idea of human spammers being profitable
What the hell are they trying to do
Just Google points? 16:31
Because we could always not auto-link URIs
16:31 fglock joined
obra Alias_: there's a lot of cheap labor out there. 16:31
Alias_: it doesn't help
I've tried.
If you can solve the spam problem, lots of folks would be interested. 16:32
Alias_ And yet we allow anonymous emails
Well, that's what I find the most confusing
Email is ok, but web is not
Juerd email is mostly to-one, not to-many.
The to-many's have moderation anyway
Alias_ Juerd: If you email to rt, it's to-many 16:33
Juerd Less impact per message. Less incentive to clean up.
obra targetted email that runs through spamassassin lets in relatively little spam. email spammers aren't interested in targetting a specific address
Juerd Alias_: "mostly" :)
obra blog spammers are targetting a specific venue.
Alias_ But rt isn't a blog, which means they are RT spammer?
Juerd I get more and more targeted spam
obra Alias_: for their purposes, it is
Alias_: I'm committed to following perl.org's lead on this one. Robrt has put much more energy into this than I have. 16:34
Alias_ And we put an email address to email on the front page... which means humans will just email that
I'm just really confused I guess
I don't understand how they won't just adapt
And register a new bitcard account every time or something
obra Because bitcard does address verification. 16:35
And spammers are historically very not cool with having their addresses verified.
Alias_ I asked about doing address verification for the public web bug reports, and got told no
We already require an email address
16:36 zgh joined, shachaf joined
Alias_ You report via the web, but email the contact email to verify before you let it leave the holding area 16:36
obra Right. I'm not about to build ANOTHER address verification system. that's part of what we use bitcard for
Alias_ bitcard is an account though, it needs a password and so on 16:37
obra Yes.
If you solve the world's spam problems, I'll turn back on anonymous comments. 16:38
But. it was making rt.cpan unusable for everyone.
Alias_ I just did... you verify addresses, without scaring people off by making them subscribe to something
Juerd I'm okay with having my address verified. 16:39
I'm absolutely not okay with having my address stored in a central database.
obra Alias_: go sell this to robert 16:40
Alias_ ok
Juerd (snail mail address, that is)
obra Juerd: we're talking about email only
16:40 Psyche^ is now known as Patterner
Juerd Oh, okay. Language thing, then. I just now realised that "address" can also mean "e-mail address". 16:41
The spam problem is solved by changing e-mail.
Radically. 16:42
16:43 zgh joined
Juerd Meanwhile, (tubular bells)++ 16:43
Alias_ I figure we just need to create a proper immune system for the intarweb 16:44
xinming How to type the smile face in utf-8 please? 16:45
I've searched pugs slides from audreyt, But I can't find it... :-/
oops. sorry, found it, It's too small in my screen. :-/
Juerd For smileys in unicode, I always go to search.cpan.org/~autrijus/ 16:48
For copy/pasting :)
kolibrie Juerd: I plan to buy a plain, boring, nice keyboard. Do you have any brand suggestions? 16:50
Juerd Key Tronic 16:51
The ErgoForce models, KT2001pro and such.
kolibrie goes to look them up (thanks Juerd) 16:52
Juerd I'm really starting to hate my kinesis keyboard, by the way. 16:54
xinming Juerd: hmm, If I write some test, Is it ok to use some chinese unicode characters? Since I really don't know where to find the utf8 characters. :-/ 16:55
Juerd It's better for my hands, but not comfortable. It's annoying.
xinming: Why wouldn't it be?
As long as it's not profane :)
cjeris Juerd: which kinesis? i tried a kinesis maxim and it was okay but too stiff
Juerd cjeris: Contoured. 16:56
I loved it. I'm starting to hate it.
cjeris oh, ok. can't use those, wrist pronation is part of what i can't take
Juerd What is pronation?
cjeris turning your palm flat instead of vertical/inward
Juerd With this keyboard, my wrists are almost perfectly straight
cjeris i have a Comfort because it lets you get higher tilt
Juerd I want IBM to create a split model of their UltraNav keyboards. 16:57
xinming Juerd: Well, Why I thougt of this is because, some people might not understand what I means, hmm, but now, I'll write the test first. :-)
Juerd Not fixed split, but two separate parts.
obra I'd even support a fixed split ultranav 16:58
Juerd I'd support it 16:59
But when they do something new, they have the option to do it GOOD.
cjeris Comfort is big, heavy, clunky, expensive, and a bit too tall, but the infinite adjustability is lovely
Juerd well.
Big is no option, because big means the mouse is too far away.
The contoured keyboard is really great for typing text, but it absolutely sucks for code. 17:00
cjeris in that case you remove the numeric-keypad module from the right and reattach it on the left :)
obra Juerd: at the moment, the only keyboard I can find that I can type on for long periods and travel with is an MS Natural Elite. What else should I look at?
xinming Juerd: hmm, what does quote protection mean? hmm, is "qq" in quote protection, or not?
17:01 justatheory joined
xinming hmm, BTW, I know what is the difference between q and qq, But just don't know, what the quote protection will protect, protect variable, or protect the string itself. :-P 17:01
Juerd xinming: I don't know what quote protection is. 17:08
obra: Depends on your specific needs. I've used dozens of keyboards intensely, and have had the complete range of physical problems.
obra: And if there's one thing I learned, it's that every problem demands its own keyboard. 17:09
The other thing is that the ideal keyboard doesn't exist.
xinming Juerd: hmm, can you explain about the q:w and q:ww?
Juerd (It would be a dual ultranav)
xinming: No, sorry.
obra *nod* I've got cubital tunnel syndrome. I really only type comfortably with a split keyboard in my lap. The Kinesis is too bulky to travel with. the datahand is too space alien. the fingerworks..I need some feedback. the ultranav is, while sexy, still too cramped. 17:10
Juerd xinming: I'd have to look it up in the synopses. You can do that yourself. I find all this far from intuitive, and way too complex and customizable.
I want to create a desk chair one time. 17:11
That lets you have your arms vertical. It'll look silly.
obra heh. i spend too much time in cafes
xinming Juerd: the problem currently is... I've tried that, and pugs treat them the same.. :-/
?eval my $a = 123; my @b = q:w /$a/; 17:12
17:12 evalbot_12586 is now known as evalbot_12591
evalbot_12591 ["\$a",] 17:12
xinming ?eval my $a = 123; my @b = q:ww /$a/;
evalbot_12591 ["\$a",]
Juerd I often want things that I can perfectly describe, but they don't exist.
They're physically possible, and there'd probably be a market for them, but for some reason they don't exist.
Alias_ nods
Juerd I don't have the means to create them, or to talk to people who could.
This is very frustrating. 17:13
cjeris obra: if you can tolerate a little key stiffness, look at a kinesis maxim for traveling? the stiffness was my only problem with mine
Alias_ The trouble is, having them custom made is expensive
The only time I did that was with my skating gloves
obra I've thought about the maxim. stiffness: more or less than an IBM clicky?
cjeris never used an ibm clicky. it's stiffer than the $2 made-in-taiwan you get with a new pc. it loosens with wear.
Juerd Right now, I want a portable case with adjustable compartments for computer parts, that I can put in my car so that I always have the right parts with me.
A camera bag comes close. 17:14
Alias_ I have a great pair of inline skating leather gloves, but the leather on the palms was shit, so they wear out. Since nobody makes the gloves any more, I went to a ladies purse repair man and had him upgrade the gloves
17:14 bernhard joined
Juerd But isn't large enough, and is too portably designed. 17:14
Alias_ They rock now, but they are VERY expensive gloves :)
I guess there's just no market for a hundred dollar skating gloves :/
Juerd One of the biggest problems with the kinesis contoured is the stiffness, and that it uses cherry switches 17:15
obra Alias_ I bet that's not true
Juerd Who the hell, in ergonomics land, uses cherry switches! Maltron and Kinesis. But these switches SUCK! NO FEEDBACK!
obra oh. the switches on a kinesis don't bother me.
Juerd Yeah, a clicky speaker thingy.
Alias_ obra: Let me check
cjeris Juerd: feedback?
Juerd cjeris: feedback
obra Juerd: I've been very tempted to spend time with a HHK Pro
Juerd obra: Happy H 17:16
?
cjeris i don't understand what you mean by feedback, i guess.
obra yeah.
Juerd cjeris: A subtle feelable click.
obra: I find them annoying, feedback wise.
Alias_ obra: They go up to about $60, but they are really bulky and heavy aggressive skating gloves
Juerd I don't remember exactly why.
obra Juerd: the Pro or the regular one?
Juerd I don't remember
obra (The Pro has entirely different switches)
Juerd Oh, that's weird.
cjeris oh. there's never been a click on any keyboard i've ever used :) 17:17
Juerd Hm - let's try the UltraX keyboard here
cjeris except a VT100 i typed on once a long time ago
17:17 c6rbon joined
Juerd cjeris: I don't mean audible click. 17:17
obra www.pfu.fujitsu.com/en/hhkeyboard/l...yspec.html
cjeris no, i know what you mean. like the click at the bottom of an ibm selectric buckling-spring type?
Juerd I'm now using the UltraX keyboard 17:21
My desk is too low :)
17:21 glasser left
Juerd Or my chair to high. 17:21
Hm
This is better.
I have no idea if I could get used to this.
17:21 jferrero joined
Juerd I'll give it a try for some time. 17:21
The kinesis is now in the storage closet again. 17:22
Hmm. I can mod this one to dvorak, probably
Let's see.
integral juerd++ # reminding me to change my desk height 17:23
Juerd nssssssssssssssvttttttttttttlt==tr==l//ssss[000000000000000000][eeeeeeeeeeeeeeeeeeeeeeeeeeqyeeqqqqqqqqqqqa,,,,,,,,,,,,,,,,,,,,,,,,,,,qqqaaaaaaaaaaaaaa--------------------- 17:24
z//--/////////////////////////////////-vwwwwzzzzzzzz
[particle] thinks Juerd is writing perl 7
cjeris i can't ever get my keyboard low enough to have 90 degree elbows ... my arms are too long compared to my torso :(
Juerd ok il unfocus
TimToady openfoundry's http server appears to have pined for the fjords. 17:31
[particle] yes, audreyt noticed that earlier
the fjords are calling me, too...
17:35 nothingmuch joined 17:40 chris2 joined 17:44 visq joined
Alias_ obra: It appears I've had my robert's mixed up for a while then? Which robert is that robert, his link on noc.perl is useless 17:47
obra Robert Spier 17:48
Juerd This was by far the hardest keyboard to physically remap :) 17:50
I've even killed one key 17:51
(But sacrificed the windows key to fix it)
cjeris you mean you just rearranged all your keycaps? 17:52
Juerd Yes, but it's a laptop-ish keyboard
So that's hard :)
cjeris: That's what desks and chairs are for. You can't change your body, but you can change your furniture 17:53
Hm, with a lagged ssh, trying out a new keyboard is weird.
Alias_ ingy? 17:59
nothingmuch Alias_: thanks, btw.. i was a bit distracted and forgot that part ;-) 18:03
Alias_ which part? 18:04
The getting primary in the first place part? :)
Juerd THIS KEYBOAD UCK 18:10
Juerd thow the keyboad aay
18:11 zgh joined
Juerd Ahhh, trusty keytronic. 18:13
All I need now is a wrist rest. But I can do without. 18:14
Oh, right. The right shift was broken on this one. 18:15
Juerd starts to cry :)
xinming TimToady: what's doesn't the quote protection mean please? 18:18
18:22 hexmode joined
TimToady I saw the references in the backlog, but have no idea what you're referring to. 18:26
Limbic_Region xinming - you have some context for your question - URL or email msg-id or something - I have no clue either 18:27
TimToady okay, I see 'quote protection' in S02 18:28
that's Unix-shell-style quote protection.
<a 'b c' d> produces ("a", "'b", "c'", "d") 18:29
<<a 'b c' d>> produces ("a", "b c", "d")
it also means that <<$a '$b'>> interpolates $a but not $b 18:31
<<$a "$b">> interpolates both, but $a will be split into separate words while $b won't.
it's still to be decided exactly how "shellish" we want to get here. 18:32
since various shells have various idiosyncracies when it comes to interpolation and such 18:33
so <...> is basically q:w/.../ while <<..>> is qq:ww 18:34
18:35 _bernhard joined
zgh tries in pugs 18:40
aw
eval: <<a b "c d" e>>
buubot zgh: Error: Can't find string terminator " a" anywhere before EOF at eval line 1.
zgh eval: <a b "c d" e>
buubot zgh: Error: Can't locate Text/ParseWords.pm: Too many open files at /usr/lib/perl/5.8/File/Glob.pm line 152.
zgh well it works better than that in pugs
TimToady ?eval <<a b "c d" e>>
evalbot_12591 ("a", "b", "\"c", "d\"", "e")
zgh thanks 18:41
TimToady not impl
zgh how do you use qq:ww?
?eval qq:ww/a b c/
evalbot_12591 ("a", "b", "c")
zgh ah
neat
zgh goes back to playing with pugs
TimToady <<...>> is merely a pretty version of qq:ww/.../ 18:42
just as <...> is a pretty version of qw/.../
zgh qq:ww doesn't work in pugs yet either with regards to shell quoting
oh well
xinming TimToady: thanks for your explanation, then, I'll leave this in the test first.
18:49 bernhard_ joined 18:52 Eimi joined 18:54 weinig is now known as weinig|away
Limbic_Region is the repo down? 18:56
[particle] yes
18:58 prefiks joined
Limbic_Region thanks 18:58
19:08 _bernhard joined
ingy last ingy 19:15
19:21 bernhard_ joined 19:38 theorbtwo joined 19:50 jferrero joined 20:00 xerox joined
Limbic_Region fglock ping 20:02
fglock pong 20:03
Limbic_Region I just tried to install v6 from CPAN 20:04
many many many failures in the test suite - seem similar to the ones reported by CPAN testers on linux despite myself being on Win32 20:05
are you aware of these or would you like a test output sent to you?
fglock Limbic_Region: I've seen the failures - I'll check with clkao to see how he controls the MANIFEST (only passing tests should enter the MANIFEST) 20:07
zgh do you need a specific perl to use v6?
Limbic_Region that's what I suspected - or else the failing ones should be flagged as todo
zgh i thought mine was too old
fglock Limbic_Region: I think it is part of the smoking process
Limbic_Region think what is part of the smoking process? 20:08
fglock zgh: not sure - how old?
Limbic_Region: the tests that pass the smoke enter the MANIFEST list and get published
zgh 5.8.7, on x86_64 20:09
it's the default for ubuntu I think
fglock zgh: that should work
zgh fglock: thanks I'll try again later
Limbic_Region ahh - well, force install then? 20:10
fglock Limbic_Region: yes 20:12
Limbic_Region k 20:13
20:49 eden_c joined 21:02 fglock joined 21:25 neoesque joined 21:27 hexmode joined 21:32 beppu joined
TreyHarris is there a problem with svn? 21:34
beppu ?eval %ENV<USER>
evalbot_12591 Error: Can't modify constant item: VUndef
TreyHarris svn.openfoundry.org/
[particle] it's down
beppu ?eval say %ENV<USER>
evalbot_12591 Error: Can't modify constant item: VUndef
TreyHarris ?eval $+USER
evalbot_12591 Error: Can't modify constant item: VUndef
beppu tf?
21:34 markstos joined
TreyHarris ?eval say $+USER 21:35
evalbot_12591 Error: Can't modify constant item: VUndef
beppu ?eval say 'wtf?'
evalbot_12591 OUTPUT[wtf? ] Bool::True
TreyHarris ?eval say $?PUGS_VERSION
evalbot_12591 OUTPUT[Perl6 User's Golfing System, version 6.2.12, June 26, 2006 (r12591) ] Bool::True
markstos ?eval class Super { method foo { $?CLASS; } }; class Kid is Super {}; say Kid.new.foo;
evalbot_12591 OUTPUT[Super ] Bool::True
beppu ?eval %ENV.keys.sort.map:{ "$_=%ENV<$_>\n" }.say;
evalbot_12591 OUTPUT[ ] Bool::True
TreyHarris beppu: looks like environment variables are hosed
beppu I was actually wondering why my map statement wasn't doing what I thought it should be doing. 21:36
TreyHarris beppu: this is new, my r12537 is fine
markstos I want to get the current class back. But $?CLASS doesn't do that in my example-- it returns the parent class, where $?CLASS is defined. Is there a way to get the class that's currently being executed, versus one were $?CLASS is defined.
beppu TreyHarris: on your pugs, do you get a bunch of "key=value" pairs printed when you run the last ?eval I did? 21:37
TimToady ?eval %ENV.keys.sort.map:{ say "$_=%ENV<$_>" }
beppu I'm getting "key=" with no value.
evalbot_12591 ()
markstos beppu, I have r12583 and I get keys with no values. 21:38
beppu Am I doing something wrong or is it pugs?
TreyHarris yes. "CLICOLOR=\nCOLUMNS=\n". but no values, just keys
beppu advice? 21:39
markstos ?eval say %ENV.perl;
evalbot_12591 OUTPUT[\undef ] Bool::True
TimToady ?eval say %*ENV.perl
evalbot_12591 OUTPUT[\undef ] Bool::True
TreyHarris ?eval %ENV.sort.map:{ "$_.key()=$_.value()}\n" }.say;
markstos beppu: try say %ENV.perl on your machine. That works for me.
evalbot_12591 Error: No compatible subroutine found: "&key"
TreyHarris beppu: the above works on my pugs 21:40
markstos It shows the values.
TimToady might be not considered "safe" by evalbot
TreyHarris sort on a hash sorts by key
beppu %ENV.perl doesn't work for me. :( 21:41
TimToady works here
beppu time to recompile pugs, perhaps 21:42
TimToady though theoretically it should have to be %*ENV
unless you've imported it into your namespace
markstos I now the meta model work is still in process, but is there currently a way to get the current class? I'm looking for something like "say self" would return, but without the extra <Obj:foo> decoration.
TreyHarris beppu: for sort %*ENV.kv -> $var, $val { say "$var=$val"}
TimToady which do you mean by "current class"? 21:43
lexically, or self's actual?
markstos uh, self's actual ?
beppu the for loop works. thanks Trey
markstos Even if code is stored in the parent, I want the child's name to be returned.
TreyHarris beppu: ummm... not so sure actually.... 21:44
TimToady TreyHarris: that can't possible be correct
*possibly
beppu oh wait... ;)
TimToady it would sort the keys independently of the values
TreyHarris yes, it does 21:45
beppu I took out the sort, but it looks like it comes back to me in sorted order, anyway. 21:46
(and the keys and values match up this time . ;)
TimToady .kv will come out in the right pairings but as separate list items 21:47
PerlJam beppu: don't rely on implementation accidents :)
beppu ;)
Is the .perl method like a built-in Data::Dumper ? 21:48
TreyHarris it's odd: %*ENV.sort:{$^b cmp $^a} gives me a good reverse-ordered list of lists (not pairs), but %*ENV.pairs.sort:{$^b.key cmp $^a.key} gives me "No compatible subroutine found: "&key""
TimToady markstos: that'll be whatever we rename .ref to, I suspect 21:49
beppu %ENV.perl was working for me... silly me forgot to actually `say` it.
markstos beppu: Yes, like D::D.
TreyHarris it looks like '.key' is not defined for ::Pair::HashSlice 21:50
which seems quite odd.
TimToady I've just been dithering about which interrogative pronouns to attach to which metainfo.
markstos TimToday: "self.ref" works today-- I'm happy. Thanks. I can find/replace later. :) 21:51
beppu I'm curious about why: %ENV.keys.sort.map:{ "$_=%ENV<$_>\n" }.say; doesn't work
TimToady if it's something checked into pugs, I'll fix it myself when I make up my mind.
markstos I would have guessed something like "self.string", or "self.name". 21:52
TimToady: CGI::App p6 isn't in the pugs SVN repo yet. I'm attached to using darcs at the moment, but I may import it one day yet, and live with svn or svk. 21:53
TreyHarris beppu: something funky is going on, because %*ENV.keys.sort.map:{ "::$^a::" } is not interpolating $^a. or $_.
i'll write a test as soon as openfoundry is back up
TimToady it'll probably be self.what or self.WHAT, or maybe who. But who might be the new name for .SKID
beppu TreyHarris: thanks, man.
TimToady would make a great skit about a baseball team...
markstos :) 21:54
self.meta.class_name seems straight forward, but maybe too long.
TimToady independent of that is whether these are all submethods of .META, or if .META should be renamed to .HOW
markstos meta does evoke introspection for me, which seems appropriate. 21:55
HOW doesn't do that for me. It evokes a Native American greeting...
TimToady and it's likely that a .what or .who would really mean .META.what or .MEAT.who anyway.
TreyHarris TimToady: tracking this down... is it possible that superclass dispatch does not happen in classes implemented in Haskell? 21:56
TimToady s/AT/TA/
Alias_ TimToady: Can you play Real movies?
markstos ah, just shorthands.
TimToady just as .does is short for .META.does
or .^does
Alias_: I could probably figure out how to play one. 21:57
Alias_ abc.net.au/science/broadband/cataly...ice_hi.ram
Some interesting stuff for software design types on how many More Ways To Do It there should be :)
Although only at a basic level, not very hardcore 21:58
TimToady okay, thanks, but my wife is napping right now so I'd better wait.
Alias_ More of a stepping off point
sure, not urgent, it's back of the brain stuff
TimToady thanks
obra *snicker* 21:59
TreyHarris yes, key and val are not implimented for Pain::HashSlice. is there any reason I shouldn't just implement it? i'm wondering if there's some deep reason it's unimplemented that i'm just not getting, or if it's just an oversight. 22:02
s/Pain/Pair
22:05 rashakil_ joined 22:06 Khisanth joined 22:08 markstos left 22:13 rashakil_ joined 22:17 diakopter joined 22:18 zgh joined 22:23 cjeris left
Juerd Hmmm 22:28
Can we still change the internal representation of text strings?
22:33 mdiep joined 22:43 eden_c left 22:53 visq joined 23:07 avar joined 23:15 yhsuan joined, lollan joined 23:18 neoesque joined 23:19 mdiep joined 23:28 mdiep joined 23:35 chromo joined, chromo left 23:46 glasser joined 23:48 JarJarBinks joined 23:50 frederico joined
TreyHarris ?eval my %a = %(zip('a'..'z';1..26)); my @p = %a.pairs; @p[0].key 23:51
evalbot_12591 "a"
TreyHarris ?eval my %a = %(zip('a'..'z';1..26)); my @p = %a.pairs; @p[0].value 23:52
evalbot_12591 \1