»ö« Welcome to Perl 6! | perl6.org/ | evalbot usage: 'perl6: say 3;' or rakudo:, niecza:, std:, or /msg p6eval perl6: ... | irclog: irc.perl6.org/ | UTF-8 is our friend!
Set by sorear on 4 February 2011.
diakopter pmichaud: have you thought about ways to coalesce redundant/unneeded Cursors (for efficiency) 01:15
diakopter pmichaud: (or, perhaps I should ask: is there any possibility of coalescing Cursors for fastpath mode, then constructing the correct Cursor structure when the matching fate is chosen.. or something like that) 01:24
diakopter pmichaud: or some other way of using a fastpath, far less heavyweight mode for patterns that are detected to not need all the features of a Cursor and related 01:28
moritz \o 05:18
sorear o/ 05:19
japhb moritz, do you have graphviz on whatever machine you are doing htmlify for perl6/doc on? 06:45
jnthn morning, #perl6 06:46
wow...short backlog is short.
diakopter morning
dalek c: 1c5e9d8 | (Geoffrey Broadwell)++ | lib/Perl6/Type (2 files):
Cache the inversion of all type relationships in the type graph
 06:47
c: 11bb98c | (Geoffrey Broadwell)++ | lib/Perl6/TypeGraph/Viz.pm:
Add Perl6::TypeGraph::Viz, which uses GraphViz to visualize type graph relationships
c/generate-svg: 06c2000 | (Geoffrey Broadwell)++ | / (2 files):
Generate type graph visualizations as SVG files (and .gitignore the new html/svg/ directory)
 06:48
japhb phenny, tell moritz Feel free to merge the generate-svg branch for perl6/doc once you have graphviz installed on your htmlify system. 06:49
phenny japhb: I'll pass that on when moritz is around.
japhb afk, but will scroll back &
dalek kudo/nom: e1905aa | moritz++ | / (3 files):
Str.path, IO::Path.new(Str:D), tests

Also rename IO::Path.dir to .directory, as per spec Preserve backwards compatiblity for now		 07:23
moritz the shared module installations to ~/.perl6/lib is starting to annoy me 07:29
phenny moritz: 06:49Z <japhb> tell moritz Feel free to merge the generate-svg branch for perl6/doc once you have graphviz installed on your htmlify system.
moritz I want to keep a "production" installation of rakudo for running p6rd, htmlifying doc.perl6.org etc 07:30
moritz speaking of which, p6rd has segfaulted after several days of operation 07:37
sorear what does p6rd do? 07:38
moritz inform us about release dates
p6rd: upcoming
moritz eeks 07:39
moritz Net::IRC::Bot depends on Str.bytes 07:39
diakopter moritz: what other examples of long-running p6 programs are there 07:45
moritz diakopter: dunno, maybe Bailador apps 07:46
p6rd: upcoming
p6rd parrot 2012-08-21, rakudo 2012-08-23, niecza 2012-08-27
moritz p6rd: 2012-12 07:47
p6rd parrot 2012-12-18, rakudo 2012-12-20, niecza 2012-12-31
bonsaikitten is there a RSS feed for that? :) 07:50
GlitchMr robert.duckduckgo.com/?q=chomp+perl6 08:37
:)
moritz that looks nice 08:38
tadzik wow, cool!
GlitchMr Also, robert.duckduckgo.com/?q=ACCEPTS+perl6 (ignore "making some information available" for now) 08:40
Additionally, if you know name of class, you can skip Perl 6 part 08:41
robert.duckduckgo.com/?q=Str.chomp
This works too
GlitchMr ... I should implement inheritance now :P 08:42
So you could do stuff like Str.any too
moritz robert.duckduckgo.com/?q=X%3A%3AAdHoc+perl6 # doesn't work yet :/ 08:43
GlitchMr Hmmm... I know
I'm going to fix ita
sergot hi o/ 08:48
moritz \o sergot 08:49
GlitchMr++ # ddg stuff
jnthn GlitchMr++ 08:49
GlitchMr Actually, i lied about Str.chomp-like stuff always working 08:50
There is one edge case where it doesn't 08:51
robert.duckduckgo.com/?q=Any.any
moritz hpmor.com/notes/progress-12-08-01/ :( no new chapter 09:34
GlitchMr doc.perl6.org/type/Junction#See%20Also 10:18
Those things are making parsing harder than it needs to be... 10:19
But whatever :)
moritz GlitchMr: what's the problem with it? 11:10
methods are always below a level 1 heading called "Methods"
or "Methods supplied by..."
jnthn Mmm...curry lunch :) 11:13
Ewww...excel generation afternoon.
.oO( can't win 'em all... )
 11:14
moritz so, who will write a spreadsheet writer for Perl 6? :-) 12:45
moritz nr: say Capture ~~ Cool 12:49
p6eval rakudo c1bfbb, niecza v19-15-g051783d: OUTPUT«False␤» 12:49
moritz doc.perl6.org/svg/type-graph-Any.svg 12:51
that's... a lot of stuff inheriting from Any
we might not want that all in the chart, just the superclasses and roles
dalek c/generate-svg: 584bf60 | moritz++ | lib/AST.pod:
document AST

  (masak, please proof-read and augment)		 12:58
jnthn moritz: Did you mean for that to go in the branch? 12:59
moritz erm, no
but I meant to merge the branch anyway
jnthn :) 13:00
moritz japhb++ did some really cool work there
dalek c: fd36986 | moritz++ | lib/Block.pod:
document Block
 13:00
c: 06c2000 | (Geoffrey Broadwell)++ | / (2 files):
Generate type graph visualizations as SVG files (and .gitignore the new html/svg/ directory)
c: 5f4d151 | moritz++ | lib/ (7 files):
Merge branch 'master' of github.com/perl6/doc
c: 584bf60 | moritz++ | lib/AST.pod:
document AST

  (masak, please proof-read and augment)
c: db1a7dd | moritz++ | / (3 files):
Merge remote branch 'origin/generate-svg'
jnthn Sadly, here at $client I don't have anything that will render an SVG installed :'/
Will have to have a look when I get back home.
moritz not even firefox?
that's cruel :-) 13:02
jnthn I could install Firefox I guess.
jnthn was moved from machine to machine for weeks on end, so hasn't invested much time in making this one the way he likes it until it seems things settle down a bit. 13:03
Tip for anybody who does do a Perl 6 spreadsheet emitter thing: give the programmer a DOM for it that doesn't suck. :/ 13:05
daxim perl6: my $foo; say $foo.WHAT 13:14
p6eval rakudo c1bfbb, niecza v19-15-g051783d: OUTPUT«Any()␤»
moritz jnthn: how would a non-sucking DOM for a spreadsheet look like? 13:19
moritz I'd imagine there's not too much to it 13:20
like, a collection of worksheets
and each worksheet is a 2D grid of cells
and a cell knows if it holds a string, number, date or formula
I'd imagine there's not too much to it # I can imagine what Excel or Gnumeric core devs would think about that line :-) 13:21
jnthn Things like, if you can format stuff within a cell as well as at the level of the cell as a whole, try for making the API for doing the two consistent. :-)
[Coke] (spreadsheet writer) you mean like the wonderful p5 one? that's a big enough project that it seems better to spend cycles being able to run p5/convert existing p5 code.
moritz [Coke]: I've read that the one for XLSX is slightly smaller than the one for XLS files :-) 13:22
[Coke] I've spent the last six months working on a project in java to generate .xlsx files quickly from sql/stored procs, so it's kind of near to my brain.
if I were writing something, I'd just do .xlsx. it's just xml and zip and a few thousand pages of DTDs. 13:23
moritz [Coke]++ # few thousand pages of DTDs 13:24
it's not a project I'd do for fun
[Coke] moritz: me either! ;) 13:25
I'm cheating, of course, being a perl programmer, and ignoring the DTDs unless I have to. ;)
jnthn Nor me :)
moritz is shelling out to perl considered kosher in the Java world? :-) 13:26
flussence I'd give up and use CSV... 13:27
rjbs I didn't know Javaites kept kosher. 13:29
jnthn
.oO( Java code always looked a little porky to me... )
 13:31
Oh my, now it spat out a corrupt Excel document?! 13:33
dalek c: b1b87b7 | moritz++ | / (2 files):
[htmlify] emit PNG images along with SVG

also change the output directory		 13:43
moritz jnthn: doc.perl6.org/images/type-graph-Any.png for you :-)
doc.perl6.org/images/type-graph-Cool.png less convoluted 13:45
jnthn whoa!
moritz japhb: I think we should omit subclasses if there more than $n of them (where $n might be 10 or 20 or so) 13:46
jnthn doc.perl6.org/images/type-graph-Numeric.png is nice, though makes me wonder if showing the things under Real wouldn't also have been nice. 13:47
moritz it's always just one level for consumers 13:48
currently
jnthn *nod*
japhb++
These are great.
moritz maybe we should generally limit to $n consumers, rather than to one level
but it's something I'll happily leave to japhb++
jnthn Is it working of data from relection, ooc? 13:49
That is, can this be applied to any bunch of types?
Or is it fairly specialized to what's in CORE?
moritz currently it gets the data from type-graph.txt 13:50
jnthn k
moritz but it goes through an indirection layer (Perl6::TypeGraph, with Perl6::Type nodes)
shouldn't be hard to fill it via introspection instead
jnthn yay :) 13:51
moritz and still reuse the dot/svg/png generation
UncleFester6 r: pir::say "abc" 13:52
p6eval rakudo c1bfbb: OUTPUT«===SORRY!===␤pir::say missing a signature␤»
moritz r: nqp::say 'abc'
p6eval rakudo c1bfbb: OUTPUT«abc␤»
moritz r: pir::say__vS 'abc'
p6eval rakudo c1bfbb: OUTPUT«abc␤»
UncleFester6 some p6 modules depend on pir::load_bytecode and wound build for me with parrot anymore - same missing signature error 13:53
s/parrot/rakudo
moritz that's since the toqast branch merge
I've submitted a pull request for MIME::Base64 13:54
UncleFester6 other module I noticed would be Diget::MD5
Digest::MD5
moritz snarkyboojum: please apply it :-)
UncleFester6: I'll patch-and-pull-request that one when I come home 13:56
dalek c: cb90010 | moritz++ | htmlify.pl:
[htmlify] stop supporting custom output dirs

since static files directly live in html/, there is not much point in allowing different output directories, which would then be incomplete. Gets rid of some repetitive argument passing		 14:02
moritz r: say Routine.^mro 14:12
p6eval rakudo c1bfbb: OUTPUT«Routine() Block() Code() Any() Mu()␤»
dalek c: 982e21f | moritz++ | lib/Routine.pod:
document Routine
 14:26
GlitchMr I've updated my DuckDuckGo thing, so you can do stuff like "Str.say" or "perl6 Mu". It isn't yet merged, I'm now waiting for merge :). 14:27
moritz GlitchMr++
moritz ++crazedpsyc # anticipated merge :-) 14:27
GlitchMr "Str.say" shows entry for "Mu.say", but I don't think it matters much 14:28
moritz it's a very good start
jnthn Once its merged, probably good to blog about it too :) 14:30
GlitchMr I wonder why they don't use JSON or something like this, but whatever
jnthn finds it quite exciting :)
GlitchMr Also, github.com/GlitchMr/zeroclickinfo-...output.txt 14:31
JimmyZ wonders whether there is a ticket or not for github.com/perl6/doc/blob/master/h...fy.pl#L74, the needless colon 14:33
moritz JimmyZ: does rakudo handle it without the semicolon? 14:35
JimmyZ moritz: nope
hoelzro does anyone know what library Github uses to do its syntax highlighting? 14:36
moritz JimmyZ: you meant rakudobug? yes
JimmyZ moritz: yeah
moritz JimmyZ: there's even a ROADMAP item for it
crazedpsyc moritz: hah, robert is the fathead guy, I'm staying out :)
moritz ++robert then :-) 14:37
GlitchMr hoelzro: pygments
hoelzro GlitchMr: thanks
JimmyZ hoelzro: I think it's here : github.com/github/ 14:38
hoelzro JimmyZ: I looked at their markdown lib, but I couldn't find details on the syntax highlighting
hoelzro wants to add Perl 6 support to GH
GlitchMr Send your language to pygments and GitHub should update someday 14:39
...
JimmyZ hoelzro: github.com/github/linguist
GlitchMr Except, I think that there could be a problem... most of Perl 6 scripts use .pl extension 14:40
And .pm
moritz and .pod
JimmyZ hoelzro: they use this library at GitHub to detect blob languages, highlight code, ignore binary files
hoelzro JimmyZ: I just saw that, thanks =) 14:41
hoelzro uses .p6/.pm6
I'm guessing I'm allowed to dip into the content of the file a bit to look for 'use v6' or something 14:42
JimmyZ hoelzro: For disambiguating between files with common extensions, they use a bayesian classifier. github.com/github/linguist/blob/ma...ssifier.rb
hoelzro nice 14:43
[Coke] Oh. I still have a project for pmichaud to show test coverage by test by implementation, don't i.
moritz pull request for Digest::MD5 submitted
GlitchMr But, would GitHub classify Perl 5 and Perl 6 as different languages :P? 14:44
[Coke] I suppose I should start that by moving my gist into an actual project.
moritz cosimo++ is usually pretty quick in applying them
GlitchMr: if we ask them nicely, why not?
[Coke] moritz: how would they, if we're using .pl everywhere?
(same way we do?)
moritz [Coke]: what JimmyZ said 14:45
GlitchMr They classify Python 2 and Python 3 as this same languages... but differences are rather small, so whatever
moritz [Coke]: irclog.perlgeek.de/perl6/2012-08-02#i_5868680
GlitchMr Perl 5 and Perl 6 are definitely different 14:46
[Coke] you'd think I could see that far up the scroll. huh. ;)
moritz you'd have to move your eyes :-)
PerlJam They could just use file(1) ;)
moritz better than fail(1) :-) 14:47
GlitchMr glitchmr@strawberry ~/g/z/perl6_doc> man fail 14:49
No manual entry for fail
[Coke] . o O (Ugh. so blah)
moritz r: say fail(1)
p6eval rakudo c1bfbb: OUTPUT«1␤ in method gist at src/gen/CORE.setting:9624␤ in sub say at src/gen/CORE.setting:7024␤ in block at /tmp/pKPe2gswrV:1␤␤»
flussence
.oO( failure is not an option? )
GlitchMr I guess I don't have program which provides fail manual page
tadzik fail: command not found
that has a double meaning :)
GlitchMr But, I think the fail man page exists... 14:50
Except it's fail(9)
man.cx/fail
[Coke] github.com/coke/perl6-roast-data -- was originally in a gist. 15:13
moritz [Coke]++ 15:15
jnthn beer, decommute &
[Coke] now pmichaud can open an issue there. ;) 15:19
dalek c: 1ca287f | moritz++ | lib/X/Syntax.pod:
X::Syntax
 15:20
c: f576a26 | moritz++ | lib/X/Syntax/NegatedPair.pod:
X::Syntax::NegatedPair
 15:24
PerlJam wonders what his brain is doing sometimes 15:31
at first glance I read that as "X::Syntax::NegatedBear" 15:32
(weird, huh?)
moritz firing neurons?
while writing docs I just realized that we could include the key of the malformed pair in the error message
I hate that, getting stuck in code improvements while I actually wanted to write docs 15:36
it's a good case for writing docs earlier than I usually do 15:37
cosimo++ already merged the Digest::MD5 patch \0/ 15:38
cosimo moritz: thanks for that 15:39
is the current rakudo branch nom? 15:41
moritz yes 15:42
cosimo I'm a bit out of the loop :-| 15:43
dalek kudo/nom: d71ad18 | moritz++ | src/ (2 files):
include key of negated pair with value in the error message
ast: 425b786 | moritz++ | S32-exceptions/misc.t:
test for extra information in X::Syntax::NegatedPair
 15:45
moritz decommute&
gfldex r: role R { our $.r; }; class C does R { our $.c = LolImUndefined; } 15:55
p6eval rakudo c1bfbb: OUTPUT«(signal SEGV)»
gfldex i'm in your language make your compiler crash 15:56
[Coke] eek.
gfldex jnthn: ^^^ please note
arnsholt Only happens if it has a role? 15:57
gfldex yes 15:58
gfldex it's RT#114380 15:58
arnsholt r: role R { has $.r; }; class C does R { our $.c = LolImUndefined; } 16:00
p6eval rakudo c1bfbb: OUTPUT«===SORRY!===␤CHECK FAILED:␤Undefined routine '&LolImUndefined' called (line 1)␤»
arnsholt Right. So somehow related to the role having a package variable, it seems
gfldex r: role R { my $.r; }; class C does R { our $.c = LolImUndefined; } 16:08
p6eval rakudo c1bfbb: OUTPUT«(signal SEGV)»
moritz r: role R { my $.r } 16:18
p6eval rakudo c1bfbb: ( no output )
moritz r: role R { my $.r }; class C does R { our $.c } 16:19
p6eval rakudo c1bfbb: OUTPUT«(signal SEGV)»
moritz r: role R { my $.r }; class C does R { }
p6eval rakudo c1bfbb: OUTPUT«(signal SEGV)»
moritz golfed!
pmichaud moritz: ping
moritz pmichaud: pong
pmichaud I'm reading the "platforms problems" thread on parrot-dev, and I'm a little concerned. Parrot's thinking of abandoning the ability to run on *BSD and *Solaris? 16:20
(and I'm asking you because you've been somewhat aware/involved of the branches involved, I think :)
I guess my first question is: Do we (Rakudo) care about BSD and Solaris? My thinking is "yes we do". 16:23
moritz yes, we do 16:26
pmichaud will the io_cleanup and/or thread branch merges give Rakudo any features/benefits that it doesn't already have, ooc? 16:27
pmichaud I guess I should just write my concerns to parrot-dev. 16:30
that's going to be a delicate email to write. :-/
PerlJam pmichaud: "aaiiiigggghh! What are y'all doing?!?" :) 16:36
GlitchMr Just wondering, when Perl 6 search will be added to DuckDuckGo how can I blog about it? 16:48
GlitchMr I guess I will have to make a blog :P 16:51
felher Is there a special pass-through-ast-of-only-matching-subrule thing? I.d: if i have a grammar "... regex a() { <a> | <b> | <c> | <d> } ... ", is there a special whatever that does " method a($/) { if $/.exists("a") { make $/<a>.ast; } elsif $/.exists("b") .... } "? Im using " method a($/) { make $/.values[0].ast; } " for now, which is not bad. I'm just wondering if there is a special way to express this 16:56
which i'm not aware off.
Oh, btw: i mean "regex x() { <a> ..." and "method x($/) {...", not regex and method "a" :) 16:58
pmichaud email sent 17:00
felher: there's not a "pass through" mechanism for the .ast items (yet?) 17:05
some of us have toyed with the notion of having one but as yet we do not.
felher pmichaud: okay. Thanks :) 17:06
moritz pmichaud: sorry, have been called away
pmichaud I speculate we could have a method that looks through .caps and returns the first defined .ast that it encounters.
moritz pmichaud: and yes, we want some stuff that the io_cleanup branch gives us 17:07
(like, .encoding on sockets, the possiblity to add a recv_bytes method to socket)
pmichaud are these things that we can only obtain through the io_cleanup branch? 17:08
moritz that's something I can't answer
I understand the current state of the IO subsystem is a mess, and it would be really much more effort to add such things, though still possible 17:09
pmichaud well, we'll see what responses my email gets. :-) 17:11
moritz it sounds like "negate the ability to run Parrot on" is a bit of an exaggeration 17:13
pmichaud yeah, that's probably true.
moritz there are also TODOed windows tests, and yet you can still run parrot on windows
or maybe s/run/crawl/ :-) 17:14
pmichaud But "drop support for" sounds much more than "todo some tests"
and if I/O itself doesn't work on those platforms... that's a pretty big feature to lose. Thus we need more information. :) 17:15
moritz agreed 17:15
pmichaud okay, thanks for the info. 17:18
PerlJam pweb 17:35
oops, ww
[Coke] pew pew! 17:36
masak salut, #perl6 17:54
japhb o/ 17:56
moritz, jnthn: Thank you for taking a look at the type graph visualizations -- and the positive comments :-) 17:57
dalek rl6-roast-data: 67c8561 | coke++ | perl6_pass_rates:
today
 17:58
[Coke] tries to decide if that's useful.
japhb For the Any, Exception, and Metamodel cases, I have some code locally that I've been playing with to segment the complete typegraph, and then draw these in an easier-to-digest form. 17:59
[Coke] settles on yes, since you ahve a link to the latest results.
jnthn good evening o/
japhb o/
[Coke] ... where did August 1st go.. whoops. ;)
moritz down the Tiber, I'm sure
japhb thinks about the "limit consumers" heuristic suggested by jnthn++ and moritz++ 18:00
[Coke] would it be helpful to put in the latest copies of each of the spec test run files? 18:01
eh, sure.
oops, meeting.
sorear good * #perl6 18:42
diakopter sorear: howdy 18:50
(u do) 18:51
jnthn o/ sorear , diakopter 18:51
dalek rl6-roast-data: 3ec3e6e | coke++ | / (5 files):
Add the test run output files daily.

Start with the previous run's output.		 18:53
[Coke] there: now raw.github.com/coke/perl6-roast-da...ummary.out will have the latest automatically in case folks want to try to fix those, or catch breakage. 18:54
dalek rl6-roast-data: 550280d | coke++ | spec_ (3 files):
remove echos - they predate running via cron.

remove commented out code		 19:04
rl6-roast-data: 4aa1b06 | coke++ | LICENSE:
Add LICENSE
rl6-roast-data: 3526c40 | coke++ | README:
Add README
rl6-roast-data: 1f5526f | coke++ | spec_ (3 files):
reformat code
[Coke] annoyingly, having this in a public repo in a gist makes it more likely I'm going to hack on it. :) 19:05
diakopter :) 19:07
[Coke] *instead of a gist, I mean.
[Coke] deletes the gist.
[Coke] rakudo is seeing a lot of these: 19:15
use of uninitialized value of type Any in string context in sub is at lib/Test.pm:81
r: use Test: say is(Any, 3); 19:16
p6eval rakudo c1bfbb: OUTPUT«===SORRY!===␤Confused␤at /tmp/RMwHNaaZbC:1␤»
[Coke] r: use Test; say is(Any, 3);
p6eval rakudo c1bfbb: OUTPUT«use of uninitialized value of type Any in string context in sub is at lib/Test.pm:81␤␤not ok 1 - ␤use of uninitialized value of type Any in string context in sub is at lib/Test.pm:84␤␤# got: ''␤# expected: '3'␤False␤»
[Coke] Should that warn, or just return false? 19:17
moritz yes, it should warn 19:19
since &is does string comparison
[Coke] r: no warnings; 19:20
p6eval rakudo c1bfbb: OUTPUT«===SORRY!===␤CHECK FAILED:␤Undefined routine '&warnings' called (line 1)␤Undefined routine '&no' called (line 1)␤»
[Coke] r: no Test;
p6eval rakudo c1bfbb: OUTPUT«===SORRY!===␤CHECK FAILED:␤Undefined routine '&Test' called (line 1)␤Undefined routine '&no' called (line 1)␤»
dalek ecs: bbbdb78 | moritz++ | S32-setting-library/Exception.pod:
[S32::Exception] X::Syntax::NegatedPair should report the key
 19:21
moritz r: :!a(1) 19:24
p6eval rakudo c1bfbb: OUTPUT«===SORRY!===␤Argument not allowed on negated pair␤at /tmp/h6yAYGvHeC:1␤»
moritz idly wonders why rakudo didn't update on p6eval 19:26
dalek c: c837ee8 | (Geoffrey Broadwell)++ | lib/Perl6/TypeGraph/Viz.pm:
Improve algorithm for finding neighbors of requested node visualization; this makes the Numeric visualization way better, for instance
 19:30
c: 804072a | (Geoffrey Broadwell)++ | htmlify.pl:
Write specialized visualizations for Any, Exception, and the Metamodel
c: b0b42c9 | (Geoffrey Broadwell)++ | .gitignore:
Fix .gitignore to skip html/images/ instead of html/svg/
japhb iterates towards some form of sanity ... 19:32
dalek albot: 369b28b | moritz++ | sync.pl:
teach sync.pl about pugs
 19:36
albot: cf3cbb9 | moritz++ | evalbot.pl:
remove toqast target

it has been merged into nom by now
albot: 6d61bd8 | moritz++ | build-scripts/rebuild-nom. (2 files):
[nom rebuild] switch to much simpler rebuild script

since building and running happens on two different machines, rsync provides snapshotting semantics for us. No need to keep two separate install locations anymore
japhb moritz, jnthn: In case it wasn't obvious, the .svg versions of the visualizations are clickable -- each node is a link to its corresponding type documentation
moritz japhb: yes, I've noticed that :-) 19:37
though I'd prefer absolut paths in the SVGs
because I'm sure some user agents will screw up with relativ paths :-)
japhb OOC, why?
moritz or we, when we move the images around 19:38
japhb ah, hmm.
japhb wonders if there are any user agents that *can* render an SVG, but *can't* get the relative paths correct.
moritz I've generally made bad experience with relative paths and later changing URLs
or relative URLs and mod_rewrite insanity :-) 19:39
but in the end it's just a slight preference
japhb I've had the opposite experience: I commonly have tools move between "sites" or parts of the same site, and so only the relative urls work. 19:40
japhb shrugs ... it's even easy to override the $.url-base :-)
moritz \o/ 19:41
japhb++ # flexible design
japhb heh
moritz one more thing 19:42
japhb ?
moritz we probably want thumbnails of those images to be included in the HTML for the type
japhb imagines moritz in a black turtleneck in front of a 5m high screen
moritz because the images themselves are often too large for direct inclusion
japhb moritz, yeah, I was thinking something similar, just hadn't quite decided how I wanted to do the thumbnails. 19:43
I'd like to have the thumbs for something not terribly complex (a half-dozen nodes, say) to still be quite readable. 19:44
moritz in case we don't have a better idea, convert $source.png -resize {$x}x$y $dest.png
for small $source.png that will produce a readable $dest.png
japhb nodnod 19:45
moritz though it seems that you can pass site="x,y" to dot too
though of course not as a command line argument. It needs to be in the input file :( 19:46
r: :!a(1)
p6eval rakudo c1bfbb: OUTPUT«===SORRY!===␤Argument not allowed on negated pair␤at /tmp/H3fTFz0sgC:1␤»
moritz r: :!a(1)
p6eval rakudo d71ad1: OUTPUT«===SORRY!===␤Argument not allowed on negated pair with key 'a'␤at /tmp/b3lQGrZFFx:1␤»
japhb moritz, well, we're generating the dot data, so injecting it into the input is workable. :-) 19:47
japhb Since there can be only one link off the thumbs, how about we make the PNGs be the "thumbs", and have them link to the full-size SVGs? 19:48
moritz I'd love to simply link to the URL without extension
and let content negotiation figure out whether to deliver the .png or .svg version
not sure how well it works with contemporary browsers :-) 19:49
japhb Do browsers request SVG in preference to PNG?
moritz that's a very good question
dalek c: 7bdd313 | moritz++ | lib/X/Syntax/NegatedPair.pod:
[X::Syntax::NegatedPair] document .key, error message
 19:50
japhb Well, let's start with the way we know will work, see what that's like, and go from there.
flussence chromium doesn't seem to care what the response is with images, it sends "Accept: */*" 19:51
it sends a full list for html/css/js files, though 19:52
oh wait, not for js either
firefox sends for a .gif: "Accept: image/png,image/*;q=0.8,*/*;q=0.5". Looks like it wouldn't load svg. 19:53
sorear i'm pretty sure you can use css3 fallbacks to do this :p 19:55
dalek albot: a8b8afd | moritz++ | sync.pl:
[nom] stop syncing outdated dirs
 20:01
moritz japhb: I have an idea for injecting HTML into the output 20:06
japhb: we can teach Pod::To::HTML to copy '=FOR HTML' sections verbatim to the output
and then it'll be easy to add some HTML that embeds the image etc. 20:07
wait, S26 says that should be =begin Html ... =end Html 20:08
japhb There's no tag for having an image in the Pod directly? 20:09
moritz hm, there is 20:10
=Image www.perlfoundation.org/images/perl_...32x104.png 20:11
though not sure if you can make the image a link
I remember actively discouraging tadzik++ from parsing block elements inside of X<...> constructs
for sanity reasons :-)
japhb It seems kinda silly not being able to make an image link (though I agree with what you said to tadzik) 20:12
moritz hm, S26 is full of options 20:15
maybe we can make =begin para L<URL HERE> \n block contents \n =end para work that way
but not today& 20:19
dalek c: 76924b1 | (Geoffrey Broadwell)++ | / (2 files):
Support sized Viz output, and make PNG visualizations 8" x 3" max (rendered at 96 dpi) for use as "mostly readable thumbnails"
 20:22
c: 4bf7649 | (Geoffrey Broadwell)++ | lib/X/Syntax/NegatedPair.pod:
Merge branch 'master' of github.com:perl6/doc
japhb Time for lunch &
japhb Mmm, Goan food ... 21:13
jnthn I'd go an' have some of that!
japhb :-P
masak trust jnthn to find the puntential in any utterance... :) 21:16
masak .oO( he should change his nick to jnpun... )
jnthn But it'd contain a vowel and risk pronuncability :P 21:17
masak well, I considered jnphn, but that wasn't as visually evident. 21:17
and given how IRC is mainly visual (even though it's really a metaphor for spoken conversation)... 21:18
japhb Anyone around who can htmlify the perl6/doc and run the sync script? moritz seems to be asleep, and my public key access to [email@hidden.address] seems broken 21:21
(You'll need URI and Pod::To::HTML for Perl 6, and graphviz installed and in your PATH for generating the visualizations) 21:23
masak nr: role R { our $.r }; class C does R {}; say "alive" 21:41
p6eval rakudo d71ad1: OUTPUT«(signal SEGV)»
..niecza v19-15-g051783d: OUTPUT«alive␤»
masak adds this to rt.perl.org/rt3/Ticket/Display.html?id=114380 21:42
lue hello world o/ 22:48
gfldex r: role R { my $.r; }; class C does R { our $.c = LolImUndefined; } 23:13
p6eval rakudo d71ad1: OUTPUT«(signal SEGV)»

Powered by Cro and the Raku® Programming Language.

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

Thank you!