Time Nick Message
23:52 dalek tablets: review: https://github.com/perl6/tablets/commit/ccd894ae3b
23:52 dalek tablets: minor text color tweak
23:52 dalek tablets: ccd894a | (Herbert Breunung)++ | docs/styles.css:
23:45 colomon fail('foo')
23:45 colomon (where AnEx is Exception)
23:45 timotimo yikes
23:45 colomon fail AnEx.new
23:45 colomon fail if !@lines;
23:44 colomon return fail()
23:44 colomon fail()
23:44 colomon my $exception = fail 42;
23:44 colomon for instance:
23:44 colomon or at least, a different understanding that Damian and I have.
23:43 colomon Actually, it looks exactly like it was written by different people with a different understanding of how fail is supposed to work. :)
23:42 colomon sorear: actually, looking at fail.t is giving me a headache.
23:40 dalek tablets: review: https://github.com/perl6/tablets/commit/3c8fc7632c
23:40 dalek tablets: added match object methods
23:40 dalek tablets: 3c8fc76 | (Herbert Breunung)++ | docs/appendix- (2 files):
23:34 colomon sorear: I'm back
23:18 dalek tablets: review: https://github.com/perl6/tablets/commit/a6d3f7dacf
23:18 dalek tablets: add backlinks
23:18 dalek tablets: a6d3f7d | (Herbert Breunung)++ | docs/appendix- (3 files):
22:55 colomon sorear: gotta run, bedtime for the little guy
22:54 colomon probably the smartest thing to do would be for me to ask the European gang how they are using fail in the morning while you're asleep.
22:53 colomon that does sound... disturbing.
22:52 sorear I'm extremely uneasy about the "void context eagerizes lists" thing
22:52 sorear what about in void context in lists?
22:52 sorear does it have to die in void context, for instance?
22:51 [Coke] https://gist.github.com/b44a0719b290f5a8ec16
22:51 [Coke] rakudo star spectest failures:
22:51 colomon but die "blah" can by caught by CATCH, right?
22:51 sorear which subset of the contradictory things do perl 6 programmers rely on?
22:51 colomon sorear: I don't doubt it.
22:50 sorear colomon: failure is a very magical type which does a number of mutually contradictory things
22:49 colomon return Failure "blah"
22:49 colomon sorear: naive understanding of it is it is basically shorthand for
22:48 sorear I think the idea of fail is that it's supposed to allow parallel operations to continue as if nothing happened, so that your hyper-maps can run to competion and return meaningless results
22:48 [Coke] jnthn: done.
22:47 colomon (that's Damian's code)
22:47 colomon fail if !@lines;
22:47 colomon things like
22:46 dalek Perlito: review: https://github.com/fglock/Perlito/commit/26ef33eef4
22:46 dalek Perlito: fix grammar-o
22:46 dalek Perlito: 26ef33e | coke++ | README:
22:46 colomon give me a second to dig up source
22:46 colomon sorear: honestly, I'm completely unclear on what it does differently from die, in practical terms.
22:45 dalek tablets: review: https://github.com/perl6/tablets/commit/5841485930
22:45 dalek tablets: clearify tilde op
22:45 dalek tablets: 5841485 | (Herbert Breunung)++ | docs/appendix-a-index.txt:
22:45 sorear what do people expect fail to do
22:45 colomon what's up?
22:44 colomon yes
22:44 sorear colomon: around?
22:44 [Coke] jnthn: star doesn't accept issues. rakudobug it?
22:41 * jnthn -> sleep, given I got none at all last night...
22:41 jnthn [Coke]: Please file an issue for the buglet, if you have chance.
22:36 [Coke] jnthn: I am trying "shock top lemon shandy".
22:35 [Coke] rakudo star buglet - "make rakudo-spectest" tries to run git pull on the exported t/spec
22:34 lichtkind jnthn: i had just a mouth full and thats anough, im currently und wheat grass juice - very healthy
22:34 jnthn lichtkind: Try more beer.
22:34 jnthn ;-)
22:33 lichtkind jnthn: had the page, the content is the problem :)
22:32 jnthn I wonder if we should consider \r\n as a single grapheme under NFG...
22:31 jnthn lichtkind: http://unicode.org/reports/tr18/tr18-14.html#Line_Boundaries
22:30 lichtkind just added confusion :)
22:30 lichtkind jnthn: this im currently reading too
22:30 jnthn See TR18 section 1.6 for a list of logical newlines.
22:30 jnthn C<\n> now matches a logical (platform independent) newline, not just C<\x0a>.
22:30 jnthn yeah, here
22:30 lichtkind jnthn: im currently reading s05 but its a bit confusing
22:29 jnthn I guess S05 has some language on exactly what \n does in regexes.
22:29 jnthn Whereas \n in a regex may match various options.
22:29 jnthn lichtkind: I suspect the difference may be that \n in a string has to commit to what it emits (platform wise)
22:23 lichtkind no my fault
22:23 lichtkind \n
22:23 lichtkind ma chat client ate that i suppose
22:22 jnthn oh :)
22:22 lichtkind TimToady: sorry i mant backsalsh n
22:22 jnthn lichtkind: You mean, do they use the same escape sequences?
22:22 dalek tablets: review: https://github.com/perl6/tablets/commit/9082169c3f
22:22 dalek tablets: hope to make more sense of escape seqences
22:22 dalek tablets: 9082169 | (Herbert Breunung)++ | docs/appendix-a-index.txt:
22:21 lichtkind TimToady: is \in "" and regex the same?
21:45 [Coke] nqp-2012.04.1 parrot-4.3.0 rakudo-2012.04.1
21:45 [Coke] has all of:
21:44 [Coke] looks like rakudo-2012.04.1/ (star) contains a copy of parrot.
21:41 lichtkind masak last time i checked spec had \a
21:41 lichtkind good night
21:38 masak 'night, #perl6
21:37 masak lichtkind: if the spec doesn't mention "\a", its silence could be taken as "do like Perl 5 does", i.e. support "\a".
21:37 timotimo i don't think it does
21:37 masak that rings a bell...
21:36 lichtkind rakudo doesnt know "\a"
21:35 timotimo don't you need it to check out parrot and nqp or something?
21:35 [Coke] (star, git) git is only required if you're creating a star release, not if you're a user, no?
21:25 lichtkind hahah
21:25 uvtc Gotsta go. o/
21:25 lichtkind uvtc: it looks good
21:25 jnthn moritz: Becuase if you don't, you get capital punishment?
21:25 lichtkind jnthn: the type we spelled of course uc it was just about the term
21:24 uvtc Ok, hope I did this correctly. Added a pull-request to rakudo/star.
21:23 moritz but when you meet mr. Regex, be sure to capitalize him properly
21:23 masak yes. we do the same for 'Str' and 'string', 'Int' and 'integer', 'Block' and 'block'...
21:21 jnthn If talking about them in general, "regex"
21:21 jnthn If you specifically mean that type name, Regex is probably right.
21:21 jnthn There's a type called Regex.
21:20 lichtkind uvtc: but is regex a name of a thing?
21:16 uvtc lichtkind, only proper nouns (like particular names of people, places, and things) get capitalized.
21:14 lichtkind question is what we count as name
21:14 lichtkind because officially in english names are uppercase
21:14 lichtkind one of the few thins im not decided
21:14 lichtkind uvtc: its fine i can fix it i just ask how we set rule
21:13 dalek tablets: review: https://github.com/perl6/tablets/commit/c18b5636bf
21:13 dalek tablets: whoops - switch to lowercase
21:13 dalek tablets: c18b563 | (John Gabriele)++ | docs/appendix-g-glossary.txt:
21:12 uvtc lichtkind, Oh shoot. Sorry, no. I saw "Regex" capitalized, and thought that was your convention. No, I prefer lowercase. Will fix both.
21:11 lichtkind uvtc: good you want to titlecase terms?
21:08 dalek tablets: review: https://github.com/perl6/tablets/commit/415a004929
21:08 dalek tablets: added "reify" to glossary
21:08 dalek tablets: 415a004 | (John Gabriele)++ | docs/appendix-g-glossary.txt:
21:04 lichtkind uvtc: so will you add it?
21:03 uvtc lichtkind, I know. Should've added a `;)`.
21:02 lichtkind uvtc: his name is larry
21:02 uvtc lichtkind, nice. :)
21:02 lichtkind uvtc: more important that lot of commits i have now clear picture what i want to have where and how to format things also made friends with markdown
21:02 uvtc If only this channel had a resident linguist...
21:01 lichtkind uvtc: and to reify i read that often but dont have slightest clue
21:01 lichtkind uvtc: but i filled in just yesterday 5, were getting there
21:00 lichtkind uvtc: thats what i ment with not answered
20:58 uvtc lichtkind, Oh, I see. You've got some terms in there without a definition.
20:57 uvtc lichtkind, "reify"
20:57 lichtkind uvtc: i added also lots where now by roughly 80, not all answered
20:57 lichtkind uvtc: why no just insert it rightaway, which is it?
20:55 uvtc lichtkind, Yes, lots of commits. I haven't been reading the Tablets recently though. Oh, I wanted to ask you: if I have a "term" I'd like added to the glossary, should I create an issue for that?
20:54 lichtkind uvtc: seen recent changes?
20:53 uvtc lichtkind, hi
20:53 lichtkind uvtc: cheers
20:46 uvtc (which recommends creating a separate branch)
20:46 uvtc Whoops, sorry, nm, found https://github.com/rakudo/rakudo/wiki/Frew%27s-Recommended-Workflow
20:46 moritz uvtc: a branch is generally safer
20:44 uvtc If I want to make a minor tweak to github.com/rakudo/star and then submit a pull-request for it, should I create a branch for that, or just make my changes to master?
20:42 moritz uvtc: arnsholt++ does all the heavy work for calling C functions, I just throw examples at him that don't work :-)
20:41 uvtc s/page/pace/
20:41 uvtc moritz, Cool. Thanks. I didn't want to ask you about it, because I can see things have been busy with the R* release, and also I think I saw that you're working at a breakneck page to get sqlite working. :)
20:40 moritz uvtc: I haven't forgotton about your proposal to reduce the output from the build; just haven't got aound to it
20:38 uvtc Ah. Yes. I brought up something related 17 days ago. It used to say subversion was a requirement. tadzik++ removed it, but "git" was not added to take its place.
20:35 uvtc For some reason the R* README doesn't mention that you need git installed to build R*. (BTW, on older Debian-based distros it was the "git-core" package. Now the package is just named "git".)
20:33 jnthn So it was a win both for pre-compilation performance and startup performance.
20:33 jnthn fwiw, it actually cut down compile time of CORE.setting as well as improving startup time.
20:33 jnthn We still have room for improvement.
20:32 jnthn Startup time improvements came from the "bs" work, for sure.
20:32 timotimo "bounded serialisation", it moved lots of startup time to the compile time of the interpreter itself by implementing serialisation of some bounded sort
20:30 uvtc Not familiar with the "bs" branch.
20:29 timotimo is that the startup bonus from the bs branch or even more awesome optimisations?
20:28 uvtc Just built R* 2012.04. Runs scripts faster than before. Quick startup.
20:24 timotimo understandably
20:23 moritz anyway, I love roles.
20:19 masak that's implicitly about code size, I guess.
20:19 masak well, the perennial question is "why don't you, like, split the setting into multiple parts?"
20:19 moritz well, I am :-)
20:17 masak well, in here people seem concerned about not duplicating code paths, but not about code size per se.
20:14 fglock special cases make the code big - not good for the browser
20:13 masak inneresting.
20:13 fglock this makes it easier to group problems together and helps minimize the special cases
20:09 fglock yes, very low-tech
20:08 masak how... manual. :)
20:07 fglock masak: I usually add the bugs to TODO-perlito5
20:03 fglock I thought it was more polite to ask first :P
20:03 jnthn fglock: Thanks :)
20:03 fglock jnthn: done
20:03 masak fglock: I was kidding. but you gave all the rest of us commit bits without asking... :P
20:02 fglock masak: because you already have one commit bit (I'm only giving one commit bit, if you want more you have to do it yourself)
20:01 jnthn fglock: Can't promise I'll find particularly many tuits to contribute to it, but sure, can't hurt to have one.
20:00 masak fglock: so, assuming I find one or two bugs in Perlito... where do I send my spam^Wreports?
19:59 masak granted.
19:58 [Coke] masak: please, you know you just file bug reports, anyway.
19:58 masak what, you ask him but not the rest of us? :P
19:57 fglock jnthn: do you want a commit bit for Perlito?
19:55 masak \o/
19:55 dalek DBIish: review: https://github.com/perl6/DBIish/commit/1499f8d030
19:55 dalek DBIish: they are now called fetchall-array and fetchall-hash
19:55 dalek DBIish:
19:55 dalek DBIish: rename methods, masak++
19:55 dalek DBIish: 1499f8d | moritz++ | / (2 files):
19:47 masak a hash.
19:45 moritz yes, but what would .fetchall-as-hash return?
19:45 masak I just think HoA looks unsightly.
19:45 masak .fetchall-as-hash?
19:44 moritz an array of hashes? or a hash of arrays?
19:44 jnthn .oO( .all-the-hashes )
19:44 moritz what would 'fetchall-hashes' return?
19:38 masak why not just fetchall-arrays and fetchall-hashes?
19:38 masak also, there's gotta be more memorable names than that :P
19:37 masak names in commit comment don't match names in patch.
19:36 dalek DBIish: review: https://github.com/perl6/DBIish/commit/c64574ab0a
19:36 dalek DBIish: Add fetchrow-AoH, fetchrow-HoA and mock-test them
19:36 dalek DBIish: c64574a | moritz++ | / (4 files):
19:27 dalek Perlito: review: https://github.com/fglock/Perlito/commit/f642820cbd
19:27 dalek Perlito: Perlito5 - perl5: fix 2 tests
19:27 dalek Perlito: f642820 | (Flavio S. Glock)++ | / (5 files):
19:13 fglock done :)
19:12 PerlJam fglock: perlpilot
19:10 fglock PerlJam: I can't see you in github, what is your name there?
19:08 fglock I think there are some cool things that could be done, I don't have time to do everything I'd like
19:08 PerlJam fglock: you could give me one right now, but I don't think I'd use it much (or at all)
19:07 PerlJam fglock: when I'm ready to actually contribute, sure.
19:07 fglock sorear: because it's fun
19:07 fglock PerlJam: want a commit bit?
19:06 dalek Perlito: review: https://github.com/fglock/Perlito/commit/7a5d100236
19:06 dalek Perlito: Perlito5 - perl5: document emitter bug (significant parenthesis)
19:06 dalek Perlito: 7a5d100 | (Flavio S. Glock)++ | TODO-perlito5:
19:05 PerlJam sorear: because you're *Awesome*. :)
19:05 moritz fishing for contributors :-)
19:05 sorear fglock: hmm? why>
18:59 fglock sorear: I've added you to Perlito in github
18:57 fglock o/
18:57 moritz \o sorear
18:56 sorear good * #perl6
18:55 dalek DBIish: review: https://github.com/perl6/DBIish/commit/37d1e6e0da
18:55 dalek DBIish: update credits
18:55 dalek DBIish: 37d1e6e | moritz++ | / (2 files):
18:51 moritz compared to what mberends++ and friends had to write some years ago
18:51 moritz this commit illustrates how much more mature and usable rakudo is today
18:51 dalek DBIish: review: https://github.com/perl6/DBIish/commit/ecdc343fb2
18:51 dalek DBIish: [Pg] make fetching of hashes a "bit" simpler
18:51 dalek DBIish: ecdc343 | moritz++ | lib/DBDish (2 files):
18:47 fglock all tuits are round
18:45 masak not sure I'll have lots of time to hack on it, but good to have a commitbit, I guess.
18:45 masak fglock: hey, thanks! :)
18:44 fglock yay
18:44 dalek Perlito: review: https://github.com/fglock/Perlito/commit/1e2d817323
18:44 dalek Perlito: Perlito5 - move node.js README instructions up
18:44 dalek Perlito: 1e2d817 | (Flavio S. Glock)++ | README-perlito5:
18:44 dalek Perlito: review: https://github.com/fglock/Perlito/commit/e518f94236
18:44 dalek Perlito: Perlito5 - fix README instructions; rebuild
18:44 dalek Perlito: e518f94 | (Flavio S. Glock)++ | / (2 files):
18:44 dalek Perlito: review: https://github.com/fglock/Perlito/commit/059f9aff02
18:44 dalek Perlito: Perlito5 - fix README instructions
18:44 dalek Perlito: 059f9af | (Flavio S. Glock)++ | README-perlito5:
18:42 fglock masak: I've added you to Perlito in github
18:41 dalek DBIish: review: https://github.com/perl6/DBIish/commit/26983c083c
18:41 dalek DBIish: [sqlite] implement disconnect method
18:41 dalek DBIish: 26983c0 | moritz++ | lib/DBDish/SQLite.pm6:
18:41 moritz fglock: thanks
18:41 fglock moritz: added
18:41 [Coke] fglock: Danke.
18:40 moritz (github username moritz)
18:40 moritz fglock: yes, can't hurt :-)
18:40 fglock moritz: want a commit bit? I seem to be in a commit bit mood today
18:39 dalek rakudo/nom: review: https://github.com/rakudo/rakudo/commit/dda5b09563
18:39 dalek rakudo/nom: hide some dispatchers from backtraces
18:39 dalek rakudo/nom: dda5b09 | moritz++ | src/core/Mu.pm:
18:38 fglock ok - I'll add it, you can remove later if it is too noisy
18:38 moritz I wouldn't mind
18:38 fglock you know it can be a bit off topic sometimes - perlito5 is about perl5, not perl6
18:37 moritz fglock: if so, follow the instructions in https://github.com/perl6/mu/blob/master/misc/dalek-push.txt
18:37 moritz fglock: do you want to have perlito commits announced in here?
18:35 fglock (added coke to Perlito)
18:32 fglock [Coke]: do you want a commit bit? what is your github name?
18:29 [Coke] whyforno perlito dalek updates?
18:29 * masak read that as 'disturbingly good karma' :)
18:25 fglock masak++ # distributing good karma
18:24 fglock README-perlito5 should be correct now
18:21 masak moritz can keep his karma point. he's probably done something I haven't ++'ed him for ;)
18:21 masak diakopter++
18:19 diakopter originally, and the first fix iteration, at least.
18:18 diakopter that was me
18:17 masak I think moritz++ implemented that.
18:06 timotimo oooh, you can give gist urls, that's sweet!
18:03 * gfldex files rakudobug
18:02 gfldex niecza++
18:02 p6eval niecza v16-23-gaa61ed5: OUTPUT«Potential difficulties:� $self is declared but not used at /tmp/EndlV58oNV line 10:�------> [32m method postcircumfix:<( )>([33m�[31m$self: |c) {[0m��i haz a sub�»
18:02 gfldex n: https://gist.github.com/2511366
18:00 p6eval ..pugs: OUTPUT«*** Cannot cast from VObject (MkObject {objType = (mkType "Foo"), objAttrs = <Hash:0xf6fdb119>, objOpaque = Nothing, objId = MkObjectId {unObjectId = 3}}) to VCode (VCode) at /tmp/F3IP3ojSqA line 19, column 1-7»
18:00 p6eval ..niecza v16-23-gaa61ed5: OUTPUT«Potential difficulties:� Unsupported use of | with sigil; nowadays please use | without sigil at /tmp/LGI4MhkOWQ line 10:�------> [32m method postcircumfix:<( )>($self: |$c[33m�[31m) {[0m� $self is declared but not used at /tmp/LGI4MhkOWQ line 10:�…
18:00 p6eval rakudo 256e1d: OUTPUT«Too many positional parameters passed; got 1 but expected 0 in sub somesub at /tmp/HwMYepvlnl:3 in sub METAOP_HYPER_CALL at src/gen/CORE.setting:10701 in method postcircumfix:<( )> at /tmp/HwMYepvlnl:12 in <anon> at src/gen/BOOTSTRAP.pm:815 in any <anon> …
18:00 [Coke] p6: https://gist.github.com/2511334
17:59 gfldex am i doing something silly or is this a bug? https://gist.github.com/2511334
17:58 fglock but dev is more fun
17:58 fglock the "release" version should pass all tests, this is "dev"
17:57 fglock yes, perl5-to-5 needs some fixing, the node.js backend works better atm
17:56 [Coke] fglock: https://gist.github.com/2511283 test failures, not sure if these are known.
17:55 [Coke] fglock: ah, much nicer, thanks!
17:55 fglock (updated)
17:54 [Coke] fglock: I'm just cut and pasting the first line into the shell. the readme makes it sound like that runs tests.
17:53 fglock I'll update the README
17:53 fglock note this is perl5-to-perl5; for perl5-to-js you need to compile perlito5.js
17:52 fglock [Coke]: this seems to work: $ prove -r -e 'perl -Ilib5 perlito5.pl -I./src5/lib ' t5
17:51 [Coke] fglock: ok, now if I run that, I get a lot of output that looks like perl5, not TAP.
17:45 * masak home
17:44 fglock the README is wrong, fixing
17:43 fglock looking
17:43 dalek tablets: review: https://github.com/perl6/tablets/commit/f79d992665
17:43 dalek tablets: undo silly dounle and triple rule for link names === is #equal-equal-equal again
17:43 dalek tablets: f79d992 | (Herbert Breunung)++ | docs/appendix- (3 files):
17:42 * [Coke] will play more later. danke.
17:41 [Coke] Can't locate object method "exp_stmts" via package "Perlito5::Grammar" (perhaps you forgot to load "Perlito5::Grammar"?) at x line 6.
17:41 [Coke] *** t5/01-perlito/24-strict.t
17:41 [Coke] the line in the readme has some failures, ala:
17:41 fglock build perlito5.js with this command: perl -Ilib5 perlito5.pl -I./src5/lib -Cjs src5/util/perlito5.pl > perlito5.js
17:40 fglock this works if you build perlito5.js: prove -r -e 'node perlito5.js -I./src5/lib -Bjs' t5
17:40 fglock the tests are now in t5/*/*.t (because there are more tests)
17:38 [Coke] do you mean `find t5 -name "*.t"` ?
17:38 fglock [Coke]: the Makefile was added to test perl5-to-perl6 compilation, that doesn't work yet - see the instructions in "README-perlito5" instead
17:36 fglock looking
17:35 [Coke] find: `t5/*.t': No such file or directory
17:35 [Coke] fglock: "make" in the top level fails:
16:54 moritz \o/
16:51 fglock perl is now in the "list of languages that compile to JavaScript": https://github.com/jashkenas/coffee-script/wiki/List-of-languages-that-compile-to-JS
16:51 moritz https://github.com/perl6/DBIish
16:22 timotimo i remember when i was told to just get a new clone of upstream to get rid of commits i made to the wrong branch
16:22 timotimo be glad you're not on mercurial, doing such recoveries there is probably much, much messier
16:21 moritz it's just what I usually try to avoid
16:21 moritz that's what I'm doing now
16:20 timotimo no worries, you should have a reflog of it and can push -f
16:19 moritz I pushed to the wrong repo
16:19 moritz oh dammit
16:16 moritz I like such patches :-)
16:16 moritz 4 files changed, 19 insertions(+), 101 deletions(-)
16:07 moritz roles++
15:35 p6eval rakudo 256e1d: OUTPUT«42»
15:35 moritz r: role R { has $.x }; class A does R { }; say A.new(x => 42).x
15:31 moritz \o sergot
15:30 sergot hi o/
15:23 * masak decommutes
15:22 PerlJam moritz: where's the google translate button when you need it?!?
15:21 dalek roast: review: https://github.com/perl6/roast/commit/5be6e89e6a
15:21 dalek roast: Fudge new test for Niecza.
15:21 dalek roast: 5be6e89 | (Solomon Foster)++ | S05-modifier/ignorecase.t:
15:20 p6eval niecza v16-23-gaa61ed5: OUTPUT«#<match from(0) to(2) text(aa) pos([].list) named({}.hash)>»
15:20 colomon n: my $a = 'aa'; say 'aa' ~~ /$a/
15:20 p6eval niecza v16-23-gaa61ed5: OUTPUT«Match()»
15:20 colomon n: my $a = 'aA'; say 'aa' ~~ /:i $a/
15:19 p6eval niecza v16-23-gaa61ed5: OUTPUT«#<match from(0) to(2) text(aa) pos([].list) named({}.hash)>»
15:19 colomon n: say 'aa' ~~ /:i aA/
15:17 moritz http://www.youtube.com/watch?v=hG6RyQCggdI # German only, I'm afraid
15:17 masak l'autopun: http://twitter.com/kazarnowicz/status/195880144899022848
15:16 jnthn :)
15:16 moritz jnthn: and she has to learn the lyrics, and at some point she mixes up some words, and the teach says "no, that would be second future at dawn" :-)
15:15 moritz jnthn: there's this nice sketch from Loriot about a woman who learns for her yodel diploma
15:15 PerlJam My trailing smiley was meant as a "deliberate nonunderstanding" marker
15:15 jnthn Maybe the morning conjugations are simpler, due to lack of coffee consumption, or something.
15:15 jnthn It would be neat if there was a language where you had to conjugate differently in the morning though.
15:13 masak PerlJam: the sun is up when the speaker tells the listener about the act. the positions of the sun and the moon then give clues as to whether the act described takes part in the past, present, or future.
15:12 moritz morning = past, midday = present, evening = future
15:12 moritz PerlJam: it's supposed to show the tenses
15:11 PerlJam masak: I don't understand what the time of day has to do with conjugation in those pictures :)
15:10 masak "Professional language redesigner on a closed spacetime loop. Do not attempt."
15:10 moritz says the language redesigner :-)
15:09 TimToady Do not redesign.
15:06 [Coke] on a closed spacetime loop.
15:04 PerlJam .oO( How many other people are resisting the Wizard of Oz reflex? )
15:02 PerlJam moritz: DBIish has a better substrate in Perl 6 than DBIx::Class does in Perl 5. I mean, we get handy things like Parcels and Captures and Cursors from the language already :)
15:02 * masak is impressed by the non-verbal teaching of verb forms over at http://en.wikipedia.org/wiki/File:Conjugaci%C3%B3n_de_correr.png
15:01 masak :P
15:00 TimToady or maybe "Professional language designer on closed course. Do not attempt." would be more appropriate :)
15:00 PerlJam moritz: indeed.
14:59 * TimToady wants a shirt that says: "Professional driver on closed course. Do not attempt."
14:59 moritz rather be a bit more explicit, than implicitly magical and backfiring
14:59 moritz PerlJam: I'd rather not repeat that kind of mistake
14:59 moritz PerlJam: DBIx::Class has this mechanism that a ->search call returns an iterator in item context, and returns all elements in list context. They consider it a complete nightmare
14:57 TimToady though it's arguably a good 2-level-ish huffman coding
14:57 PerlJam I was just thinking about DBIish roles for result sets and how they would be applied or used if it were my @results = $stmt.execute(42)
14:57 TimToady the Latin alphabet can't even represent all the sounds of English, let alone other languages, hence all the digraphs in both consonant and vowel space
14:56 jnthn It's not that it coerces it, so much as it STOREs the RHS in the array container.
14:56 moritz otherwise my @foo = 1, 2, 3; would put a Parcel into @foo, which would then be immutable
14:56 PerlJam that's what I thought.
14:56 moritz you'd need binding for any other type
14:55 moritz PerlJam: my @foo = ...; always coerces the LHS to an Array
14:55 PerlJam does "my @foo = mumble" constrain the methods available on @foo? i.e., could mumble return some object thingy that happens to do Positional, but also some other roles and would the methods of those other roles be preserved in @foo or would one need to use binding to keep them?
14:50 moritz which is usually interesting for things like INSERT, UPDATE or DELETE statements
14:50 moritz btw MiniDBI does have .rows, which returns the count of affected rows by the last statement
14:49 TimToady it's not something you derive by introspection
14:49 tadzik well, there's .do()
14:49 TimToady but we know this because you get different kinds of aphasias depending on which part of the brain you damage
14:48 moritz PerlJam: which, of course, leads to the question if we can't unify it with prepare in a nice way, and get rid of the statement handle abstraction entirely
14:47 huf but it's fairly situational, we verb nouns and noun verbs all the time
14:47 moritz PerlJam: hm, good question
14:47 * jnthn tends to use noun-y names for methods for "I have this right to hand" and verby ones for "I'll have to do some work..."
14:46 TimToady I believe so
14:46 PerlJam with laziness, why not my @results = $stmt.execute(42); # ?
14:46 moritz do all natural languages have the distinction between verbs and nouns?
14:45 TimToady the brain has at least two well-known but separate namespaces, one for nouns, and a different one for verbs
14:43 [Coke] so, if it's coming from the announcement, we can just give them the full url, not the /downloads one.
14:42 jnthn Agree "fetch" in the name feels like a kinda false minima though when it coems to naming.
14:41 tadzik good point
14:41 jnthn .elems, .chars, .bytes etc all return *counts*.
14:41 tadzik aye
14:40 jnthn Be careful with .rows
14:40 moritz tadzik: so you need to fetch all columns from a row inside the driver just for safety
14:40 tadzik I see
14:40 grondilu Yes I think it was in the announcement.
14:40 moritz tadzik: and once you've progressed to the next row, you can't access columns from the previous one
14:39 grondilu PerlJam: sorry I don't remember
14:39 moritz tadzik: because most C APIs progress row by row
14:39 PerlJam grondilu: and where did you get that URL from? the announcement?
14:39 moritz tadzik: that's dangerous
14:39 grondilu PerlJam: I just went to http://github.com/rakudo/star/downloads with my web browser and clicked to the button
14:38 tadzik moritz: how about rows being a lazy list or such?
14:38 tadzik moritz++
14:38 tadzik ooooh, coool
14:36 PerlJam [Coke]: I wonder if we shouldn't make a "project page" on github for R* and advertise that.
14:36 masak it's like feeling the need to call get methods .getSomething
14:36 masak +1 on dropping 'fetch_'
14:36 PerlJam grondilu: how did you get to the download button for rakudo-star? What led you to the github repo in the first place?
14:34 Juerd moritz: With DBIx::Simple I just named the methods after whatever they return.
14:33 Juerd moritz: That's nice but please consider having a look at DBIx::Simple's method names. A few things: "all" isn't necessarily true, as it only returns the *remaining* ones. And "fetch" is lame; duh, of course we're fetching; what else would we do? :)
14:31 masak also, "[I've had enough of this 137 crap]" made me smile. :)
14:29 masak (where we use `for` in Perl 6)
14:29 masak moritz: the situation feels somewhat analogous to file IO.
14:29 masak oh, it's a song about a phone number. I see.
14:28 moritz masak: so if you abort earlier, you might like the while-version better
14:28 PerlJam masak: 8675309
14:28 masak moritz: ah, point.
14:28 masak [Coke]: I read that xkcd comic, but I have no idea what Jenny's constant is.
14:27 moritz masak: if you don't exhaust the iterator, you might need to call .finish on the statement handle yourself
14:27 colomon never mind
14:27 colomon oh, wait, that was a typo on my part
14:27 colomon Insecure $ENV{PATH} while running with -t switch at t/fudgeandrun line 18.
14:27 colomon wow, Niecza complains loudly about t/spec/S02-magicals/env.t
14:26 p6eval rakudo 256e1d: OUTPUT«867.530902863768»
14:26 [Coke] r: say (7**(e-1/e)-9)*pi**2 # jenny's constant xkcd++
14:26 masak hm, when would I want to use the 'while' thingy rather than the 'for' thingy? :)
14:25 masak \o/ # .fetchrow
14:24 moritz Juerd: https://gist.github.com/2509680
14:22 grondilu Ahhhh ignore that ^
14:21 grondilu I've just compiled rakudo and now I'm running 'make install' and it takes much longer than it used to. Seems like it is retrieving some modules. That's new.
14:16 Juerd moritz: Could you give an example of p6ized DBI calling?
14:13 [Coke] (or admin options to hide those buttons)
14:13 [Coke] Probably worth opening github ticket to see if we (or they) can add some explanatory text.
14:11 [Coke] yup, that's a github thing.
14:11 grondilu I just clicked the big button 'download as tar.gz' on github
14:10 [Coke] that doesn't help.
14:10 [Coke] I do note that rakudo-star-2012.04.tar.gz doesn't have a description.
14:09 [Coke] yah, rakudo-star-c92df22.tar.gz isn't a tagged copy, either. it's "whatever was master at the time."
14:09 masak ah. [Coke]++
14:09 grondilu yes
14:09 [Coke] he probably hit "download as tar.gz" - that button is always there.
14:08 [Coke] how did grondilu find that tag?
14:08 [Coke] PerlJam: I disagree.
14:07 masak ...by Syrio Forel! :D
14:04 * colomon is watching his three-year-old get a tennis "lesson"
14:03 PerlJam grondilu's difficulty is good evidence that star releases should *NOT* be tagged I think
14:02 masak today's autopun: https://www.google.com/search?q=zerg+rush
14:01 colomon OOOOO, I can haz internetz!
14:01 tadzik so apparently tagging did confuse users
14:01 jnthn Yeah, the repo contains the things to build the relesae.
14:01 tadzik oh, right
14:01 jnthn Oh...as in, the contents of the star repo?
14:01 grondilu instead of 'rakudo-star-2012.04.tar.gz'
14:01 grondilu ah ok. What I did download was 'rakudo-star-c92df22.tar.gz'
14:00 masak oh phew
13:59 jnthn Yeah, the release cand I built yesterday has a modules directory with a bunch of modules
13:59 grondilu Well, maybe I just missed something
13:59 masak I thought the modules literally were packaged with Rakudo Star.
13:58 masak I'm a bit surprised at this too.
13:58 tadzik I don't think so
13:58 grondilu unless it is specific to 2012.04.1?
13:57 grondilu doesn't seem so.
13:57 jnthn Huh, I thought the modules were included in the rakudo-star?
13:57 * grondilu is using a public internet access
13:56 tadzik the worst thing about rakudo-star imho is the fact that you can't really make DESTDIR=foo install to package it nicely
13:56 masak and you're talking to us... how? :P
13:55 grondilu also, isn't the distribution of rakudo-star kind of weird? I mean, it's supposed to be a .tar.gz file, and yet once we untar it and do a "make", it still downloads the modules. I don't have a regular internet access so this is not much convenient for me :(
13:52 PerlJam DBI sort of assumes an SQL abstraction. Perl 6 has a good chance of making that very pluggable.
13:49 masak people need to question the fundamentals sometimes.
13:49 masak which is fine.
13:49 PerlJam maybe round 3, but my memory isn't that long
13:49 PerlJam I think were in 2 round of history rpeating itself wrt databases
13:49 masak right, noSQL is nothing new, really. just the name is new and hot.
13:49 PerlJam arnsholt: key-value stores were around in the 1970s too, they were just much slower :)
13:49 masak arnsholt: basically, Codd's idea was revolutionary and carried us all the way from the 70s until now.
13:48 tadzik grondilu: you can add it to Most Wanted list, or just write it yourself :)
13:48 tadzik grondilu: I don't think so
13:48 masak grondilu: maybe there should be.
13:47 arnsholt I mean, the rise of key-value stores is pretty recent, and relational has been dominant since the fall of structured DBs sometime in the 70s
13:47 arnsholt PerlJam: I've been assuming relational when people have talked about databases at least
13:46 arnsholt I think there's a Berkeley module in core, been there forever IIRC
13:46 grondilu BTW is there anything for BerkeleyDB in Perl6?
13:45 PerlJam fair assumption where?
13:45 arnsholt To be fair, relational has been a pretty fair assumption until recently
13:44 PerlJam In a way, even the Perl5 DBI is misnamed. It's not about interfacing to just /any/ database, but primarily relational databases.
13:43 tadzik BDI: BataDase Interface
13:43 JimmyZ or DBI::Link
13:42 JimmyZ any way, we needs DBI
13:42 masak yeah, DBI is fine too.
13:42 * moritz will think about it on his way home
13:42 JimmyZ +1 to PerlJam too
13:42 tadzik PerlJam++
13:41 PerlJam moritz: Just call it DBI and let the future sort it out
13:41 tadzik IdioDBI? :P
13:41 moritz it's not about simpler, it's about more idiomatic
13:41 moritz tadzik is right
13:41 tadzik afaiu it won't neceserilly be simplier :)
13:41 moritz PerlJam: well, at least I don't want to limit myself to staying smaller than MiniDBI
13:40 tadzik I don't think it's adequate
13:40 JimmyZ Simple version of DBI
13:40 JimmyZ I'm +1 to DBI::Simple
13:40 PerlJam moritz: how is it bigger?
13:40 moritz but Tiny < Mini
13:40 moritz now, I'm building something bigger than MiniDBI
13:39 JimmyZ DBI::Tiny ?
13:39 p6eval star 2012.04: OUTPUT«Actions Grammar»
13:39 moritz star: use JSON::Tiny; say JSON::Tiny.WHO.keys
13:39 moritz str: use JSON::Tiny; say JSON::Tiny.WHO.keys
13:38 p6eval rakudo 256e1d: OUTPUT«===SORRY!===Could not find JSON::Tiny in any of: /home/p6eval/.perl6/lib, /home/p6eval/nom-inst1/lib/parrot/4.3.0-devel/languages/perl6/lib, .»
13:38 grondilu r: use JSON::Tiny; say JSON::Tiny.WHO.keys
13:38 moritz grondilu: I don't think JimmyZ answered you
13:38 moritz JSON::Tiny is still as minimalistic as possible
13:38 grondilu I don't get it
13:38 p6eval rakudo 256e1d: OUTPUT«===SORRY!===Could not find DBI::Simple in any of: /home/p6eval/.perl6/lib, /home/p6eval/nom-inst1/lib/parrot/4.3.0-devel/languages/perl6/lib, .»
13:38 grondilu r: use DBI::Simple; say ::Simple.WHO.keys
13:38 JimmyZ or DBI::easy :)
13:38 tadzik and they often outgrow their names :)
13:38 tadzik we have a tendency for naming our modules ::Simple, ::Tiny, Mini* and so
13:37 p6eval rakudo 256e1d: OUTPUT«===SORRY!===Could not find DBI::Simple in any of: /home/p6eval/.perl6/lib, /home/p6eval/nom-inst1/lib/parrot/4.3.0-devel/languages/perl6/lib, .»
13:37 grondilu r: use DBI::Simple; say DBI::Simple.WHO.keys
13:37 JimmyZ DBI::Simple
13:37 grondilu Oh ok using it like this will do
13:36 grondilu Isn't there a dummy module I could use to show you something about using modules?
13:36 p6eval rakudo 256e1d: OUTPUT«&b»
13:36 moritz r: module A { our sub b() { } }; say A.WHO.keys
13:36 p6eval rakudo 256e1d: OUTPUT«&b»
13:36 moritz r: module A { our sub b() { } }; say A::.keys
13:35 p6eval rakudo 256e1d: OUTPUT«»
13:35 grondilu r: use Test; say Test.WHO.keys
13:35 p6eval rakudo 256e1d: OUTPUT«().hash»
13:35 grondilu r: use Test; say Test.WHO
13:35 PerlJam s/fomr/from/ # weird
13:35 masak grondilu: but if you're fine with that, just access its stash with, hm, .WHO
13:35 PerlJam well, size isn't the important consideration anyway, it's the shift fomr perl5think to perl6think
13:34 masak grondilu: a module only owns its "our"-scoped routines.
13:34 grondilu How do I get a list of routines in a module?
13:34 masak CentiDBI -- but I suggested that already, and I kinda like DBIish :)
13:34 moritz and I didn't like CentiDBI or DeciDBI :-)
13:34 moritz mu/micro and nano are also smaller than mini
13:34 Su-Shee nanoDBI ;)
13:34 PerlJam mu?
13:33 moritz I also considered 'MesoDBI', but "meso" is usually smaller than mini
13:33 arnsholt DBJ and DBE? =)
13:32 masak +1
13:32 moritz erm, yes
13:32 tadzik maybe DBIish, to be consistent?
13:32 PerlJam moritz: good idea ;)
13:31 * moritz goes with DBIsh and DBDish for now
13:31 Su-Shee DBItti :)
13:31 tadzik but then DBD is hard
13:31 * flussence ducks
13:31 flussence DBS, for "DBI Bikeshed"
13:31 tadzik erm, IBD
13:31 PerlJam DBIx ;->
13:31 tadzik IDB?
13:31 tadzik :)
13:31 moritz tadzik: I like the name, but I don't want to type it repeatedly :-)
13:31 PerlJam too long
13:30 tadzik how about DBIesque?
13:29 Su-Shee .oO(DBItje ;)
13:28 PerlJam moritz: a very worthy goal
13:27 moritz *more
13:27 moritz arnsholt: I don't want that level of engineering (yet?). I want something as simple as DBI, but mor 6ish
13:26 moritz arnsholt: and he uses ODBC as the internal API between DBI and DBDs
13:26 arnsholt Tim Bunce, right
13:26 moritz arnsholt: tim bunce worked a bit on DBDI, yes
13:26 arnsholt Inspired by ODBC or something IIRC
13:26 arnsholt moritz: Isn't whoever made DBI (his name escapes me ATM) also doing something along those lines?
13:25 Su-Shee or DBIje to honor mberends. ;)
13:24 moritz Su-Shee: :-)
13:24 Su-Shee moritz: DBIchen ;)
13:24 moritz PerlJam: I don't think we're ready for those big names yet
13:24 PerlJam or DBIvi DBDvi
13:23 PerlJam moritz: how about DBI6 and DBD6 ;)
13:23 moritz well, I have a bit of code already in a MiniDBI branch
13:23 * mikec is interested
13:23 moritz just planning
13:23 moritz well, it's not yet awesome. It's not at all yet
13:23 mikec cool
13:23 PerlJam awesome
13:22 moritz right
13:22 PerlJam and DBIish is a perl6 reimagining?
13:22 moritz and I don't want to deviate from that without first discussing it with mberends
13:22 moritz MiniDBI tries to be p5 DBI compatible as much as possible
13:21 moritz :-)
13:21 moritz well, it's like, DBI-ish
13:21 PerlJam what's "DBIish" mean?
13:21 moritz DBDish?
13:21 moritz when I create my 'DBIish' fork of MiniDBI, what should I name the DBD modules?
13:02 PerlJam food
13:00 [Coke] ah, that's better.
12:59 moritz rude?
12:58 [Coke] ... i don't like where this is going.
12:58 [Coke] nude!
12:56 moritz dude!
12:52 mikec_ dune!
12:52 [Coke] your made up word reminded me of dune's "kwisatz haderach".
12:51 moritz Juerd: fwiw I'm pondering a p6ized replacement for MiniDBI
12:48 tadzik what
12:48 [Coke] (catching up)
12:48 [Coke] prerekwizyt haderach!
12:32 tadzik yeah, it works now, for fork's sake :)
12:14 colomon fork++
12:14 moritz yes, because we've changed HTTP::Easy to my fork, which has the advantage of working :-)
12:13 colomon ooo, I see Bailador has also gone all green sometime in the last 24 hours. ;)
12:10 * jnthn is now mostly repaired and should be able to look at them soonish
12:08 colomon moritz++
12:07 moritz https://rt.perl.org/rt3/Ticket/Display.html?id=112624 and https://rt.perl.org/rt3/Ticket/Display.html?id=112626
12:07 tadzik good :)
12:06 moritz ABC requires one or two fixes from jnthn++ I'm afraid
12:06 colomon ;)
12:06 moritz I wanted to fix it, and found it already fixed :-)
12:06 moritz no, Testing
12:05 tadzik is ABC fixed?
12:05 tadzik colomon: I will after $work
12:05 colomon tadzik: you need to refresh the Emmentaler page. I fixed one project last night and removed another one from the ecosystem. ;)
12:03 tadzik err, wait...
12:02 tadzik colomon: nice idea for a sequel
12:02 colomon tadzik: while a sequel preventer would be nice, a remake preventer would be much more useful, IMO. ;)
12:02 Su-Shee http://en.wikipedia.org/wiki/Joss_Whedon large table of actors in the whedonverse ;)
12:02 Su-Shee haha. of course..
12:02 fglock o/
12:01 tadzik at least officially
11:59 tadzik moritz: shh, that's a secret project!
11:59 Su-Shee eliza dushku I think.
11:58 Su-Shee masak: yes!
11:58 Su-Shee eliza dushu of course and caiptn firefly who played the evil in buffy.
11:58 masak Su-Shee: Wash
11:58 Su-Shee and she who played in angel and in dollhouse whose name is on the tip of my tongue..
11:56 Su-Shee moritz: alan tudik (alpha and I forgot his name in firefly)
11:56 moritz unless tadzik++'s sequel preventer kills the characters, of course :-)
11:56 masak range? .oO( Jayne .. River )
11:55 Su-Shee whedon works with the same range of actors anyways.
11:55 Su-Shee they could.
11:54 masak can't they, theoretically, un-cancel Firefly?
11:54 Su-Shee god I wish they hadn't cancled it, too :)
11:54 colomon eh, at least it got two seasons and a decent chance to wrap up its stories. (I'm still scared by Firefly.)
11:51 Su-Shee god I wish they hadn't cancled it..
11:50 moritz \o
11:49 colomon o/
11:49 colomon loved that show
11:49 jnthn afternoon, #perl6
11:49 masak Dollhouse. watch it.
11:49 masak I also like the first scene of the pilot of "Dollhouse", where this line appears: "Have you ever tried to clean an actual slate?"
11:48 masak :)
11:47 Su-Shee anyways. it's very funny. :)
11:47 Su-Shee -is I added that accidently.
11:46 moritz s
11:46 moritz clean slate would be an idiomatic translation, ye
11:46 tadzik hah
11:46 Su-Shee tadzik: "start new, over" etc.
11:46 moritz "tabula rasa" means, literally, "table empty" (empty table)
11:46 Su-Shee tadzik: "making tabula rasa" -> clean slate, empty the table
11:45 tadzik what does that mean?
11:45 Su-Shee LOL masak: "this tabula is intentionally left rasa" is amazing. :) I think I need this a) as a shirt, b) as a signature and c) had to retweet it :)
10:13 moritz arnsholt: it doesn't really work, and I'm not sure the approach is right at all, but it's a fun attempt
10:13 moritz arnsholt: I've just pushed a new zavolaj branch, 'check-signature'
10:06 p6eval rakudo 256e1d: OUTPUT«0»
10:06 moritz r: sub f($x) { }; say &f.candidates_matching(1, 2).elems
10:06 p6eval rakudo 256e1d: OUTPUT«1»
10:06 moritz r: sub f($x) { }; say &f.candidates_matching(1).elems
10:06 p6eval rakudo 256e1d: OUTPUT«Method 'candidates_match' not found for invocant of class 'Sub' in block <anon> at /tmp/whcx9rFHd1:1»
10:06 moritz r: sub f($x) { }; say &f.candidates_match(1).elems
10:06 p6eval rakudo 256e1d: OUTPUT«===SORRY!===CHECK FAILED:Calling 'f' will never work with no arguments (line 1) Expected: :($x)»
10:06 moritz r: sub f($x) { }; say f.candidates_match(1).elems
09:58 arnsholt And I'll say "NativeCall makes calling out to C disgustingly simple" =D
09:56 moritz and in the end I'll say "look, I've written an SQLite backend for MiniDBI" :-)
09:56 moritz arnsholt: I like our modus operandi: I let you handle all the hard parts :-)
09:55 moritz arnsholt: no hurry
09:55 arnsholt Don't have time to debug it right now, but I'll look into it
09:53 arnsholt moritz: Well, that was annoying. But yeah, probably something fishy with zavolaj's pointers
09:40 moritz arnsholt: and https://gist.github.com/2507847 is the script that shows different behavior under valgrind and without it
09:39 moritz arnsholt: that one exhibts the "Malformed UTF-8" or "Unaligned end in UTF-8 string"
09:39 moritz arnsholt: make && PERL6LIB=lib perl6 t/40-sqlite.t
09:38 moritz just let me push the latest fixes
09:38 moritz arnsholt: yes, in MiniDBI, branch no_dsn
09:36 arnsholt moritz: Do you have your sqlite code on github or something?
09:36 masak "prereqs ok -- NOT" sounds like something Borat would say.
09:35 tadzik some some Configure.pm support may be necessary
09:35 tadzik so that would count as "prereqs ok -- NOT" in emmentaler, right?
09:34 tadzik I see
09:34 moritz tadzik: and the CPAN tester reports show that as "N/A" or "UNKNOWN" or so instead of PASS or FAIL
09:33 moritz tadzik: p5 modules run a configure script that detects dependencies, and dies if they are not fulfilled
09:33 moritz tadzik: no, p5 does it differently
09:32 tadzik discussion.init: http://tjs.azalayah.net/new.html shows failures in all the NativeCall-using modules now, because there are no libraries available and the tests fail. Any ideas how to resolve this? I suppose module tests should skip() if the library is not available, I think that's how it works in Perl 5
09:30 frettled That's plausible.
09:29 moritz I think it was krunen, but I'm not entirely sure
09:28 arnsholt Right. One of the Norwegians I think
09:28 frettled That wasn't me, that was someone else, but I might have said it.
09:28 arnsholt I think it was frettled who mentioned that the standard scholarly pronounciation of Egyptian is in no way based on actual reconstructions of the language
09:27 arnsholt Yeah, hieroglyphs are consonants-only
09:26 tadzik regarding Egyptian names, Imhotep, Amenhotep etc
09:26 tadzik istr talking about those on the hackathon
09:25 arnsholt Of course, that's not entirely true for the modern varieties, but that's the gist of it
09:25 arnsholt Alphabets that only have signs for consonants, not vowels
09:25 tadzik aha
09:25 arnsholt tadzik: Arabic and hebrew are the primary examples
09:24 phenny tadzik: "dinner" (pl to en, translate.google.com)
09:24 tadzik phenny: "obiad"?
09:24 tadzik abjads?
09:24 arnsholt Abuguidas fall in between alphabetic and syllabic
09:24 arnsholt But abjads are essentially alphabetic, except for the decision that you don't really care about vowels
09:21 masak kanji and hanzi are "the same", modulo the PRC's simplifications, and Japan's simplifications. but even after that, meanings have often diverged to different degrees.
09:21 arnsholt That a good enough answer? =)
09:21 arnsholt Abjad, yes-ish, abuguida, kinda
09:20 sorear for the purposes of this discussion, are abugidas and abjads alphabets?
09:20 masak neither of Japan's three writing systems is an alphabet. romaji is, though :P
09:20 arnsholt moritz: Ah, that's annoying. Especially different results when debugging
09:20 moritz that's the reason for the big JCK unification in Unicode
09:19 bonsaikitten frettled: but yeah, kanji is close enough to chinese that you can read a newspaper and get the general idea from the other language
09:19 frettled bonsaikitten: I am aware of that.
09:19 bonsaikitten frettled: japanese has three independent alphabets / writing systems
09:18 moritz and valgrind still reports invalid reads
09:18 frettled ISBN 0-684-85932-7
09:18 moritz arnsholt: I get different output when under valgrind :/
09:18 moritz arnsholt: there's still something wrong with the sqlite stuff
09:18 frettled bonsaikitten: I may be misunderstanding, but US intelligence sure didn't :)
09:18 bonsaikitten frettled: slight misunderstanding of written japanese
09:18 masak I think the two Japanese sets of kana count as syllabaries, not alphabets. </nitpick>
09:17 frettled Pretty cool cryptanalysis, actually.
09:17 frettled This meant that the US intelligence section intercepting the encrypted messages could recognize warship names from other text, simply by the change in alphabet
09:16 frettled The following is based on my recollection of a book I read on code-breaking during WWII: the Japanese navy had an "unbreakable" cipher known as Purple. Japanese uses to some extent the Chinese symbols for written language, and not an alphabet. But when communicating about naval warfare, they ended up spelling things out in their own alphabet to ensure that there were no misunderstandings.
09:16 brrt in which case, it is arguably good coding, in a way
09:16 masak frettled: sure. modern Mandarin is generally bisyllabic, compared to the monosyllabic Classical Chinese.
09:16 brrt i guess that human visual pattern recognition is actually good enough to deal with hanzi
09:15 brrt hmm
09:15 masak it's just a different factoring of things, that's all. gain some, lose some.
09:15 * brrt still would like to learn chinese
09:15 masak I think I used to believe that a bit more before learning Chinese. :)
09:15 brrt it is
09:14 masak it's a fairly strong statement to make that hanzi are somehow more of a mess than alphabets.
09:14 brrt alphabets are usefull especially because they prevent that mess
09:13 frettled masak: well, sometimes you need additional symbols to differentiate meaning
09:13 * moritz
09:12 masak by the way, who says you need an alphabet at all? 中国人做得很好，尽管只有汉字。
09:12 moritz https://de.wikipedia.org/wiki/Datei:Roman_Empire_125_de.svg
09:12 moritz https://en.wikipedia.org/wiki/File:Limes_Germanicus_2nd_c.png
09:12 brrt maybe english is the exception
09:12 frettled The funny thing is that the alphabet imposes changes on how people speak :)
09:11 brrt ah, that is very true
09:11 brrt not arguing that latin is the end-all alphabet
09:11 moritz brrt: actually the south of Germany has been under Roman ruling
09:10 brrt true, but germany never has, and it fits comfortably - except for umlauts - into latin
09:10 masak the Russian alphabet is a really good fit for Russian sounds. if you look at that and then at the stock of other Slavic languages, the ones with adapted Latin alphabets look like they made compromises of different kinds.
09:10 moritz like, being conquered by them
09:10 tadzik <obligatory brzęczyszczykiewicz video>
09:10 brrt many western european languages have 'fitted' latin to their own speech
09:10 moritz brrt: well, most of western europe had been under Roman influence
09:10 brrt very little, in fact :-)
09:09 tadzik clearly, you haven't heard enough Polish :P
09:09 tadzik "there are only so many sounds a human can make"
09:09 brrt but what most western european languages have done
09:09 brrt true
09:09 moritz and latin only used a tiny subset
09:09 brrt so, pretty much all languages can be captured by alphabet
09:09 brrt there are only so many sounds a human can make
09:08 brrt good question...
09:08 moritz the larger the difference in language is, the less likely it is that it maps well to the alphabet
09:08 moritz brrt: asked differently, why should they fit?
09:06 brrt most i know, at least
09:06 masak moritz: it has happened that I forget words in Swedish but remember them in English. but it's rare.
09:06 tadzik if not all of them
09:06 tadzik most of them, I think
09:05 brrt as an unrelated sidenote, why do many eastern european languages not fit into the latin alphabet?
09:04 tadzik yeah, no wonder I say "prerekwizyt"
09:04 phenny tadzik: "warunek wstępny" (en to pl, translate.google.com)
09:04 tadzik phenny: en pl "prerequisite"?
09:03 phenny tadzik: "prop" (pl to en, translate.google.com)
09:03 tadzik phenny: "rekwizyt"?
09:03 brrt the english word is frequently itself imported from france
09:03 brrt where you get a dutch version of an english word
09:03 tadzik yeah, it's apparently illegal Polish, according to two online dictionaries
09:02 brrt in the netherlands, a lot of people do that
09:02 moritz cute :-)
09:02 phenny moritz: "prerekwizyt" (pl to en, translate.google.com)
09:02 moritz phenny: pl "prerekwizyt"?
09:02 tadzik yeah. It's supposed to mimic "prerequisite"
09:01 phenny tadzik: "prerekwizyt" (en to en, translate.google.com)
09:01 tadzik phenny: "prerekwizyt"?
09:01 tadzik sometimes it results in abominations like...
09:01 tadzik I sometimes stick English words into Polish sentences, for I cannot recall the Polish word for it
08:59 moritz what about Swedish? :-)
08:57 masak also, somewhat annoying, sending off a lookup for the word for a concept in English, and $brain comes back with correct translations in French, Russian, Esperanto, and Mandarin, but no English.
08:56 masak I like the feeling of some part of my brain doing an instantaneous lookup of some word in some language I haven't used in a long time, all the while the rest of the brain goes "oh, is that so? how do I know that?" :P
08:54 brrt something i'd like to study in the future :-)
08:53 moritz if there are namespaces, I'm sure there's a lookup facility across them
08:52 masak I think jnthn's brain has a single namespace for words, and that's why he makes cross-language puns all the time :P
08:52 brrt sorear: fun, i read a lot of neurobiology, so i guess that is in many ways the 'other side' of approach
08:51 masak moritz: if you like this conversation, I strongly recommend Hofstadter's "I am a strange loop".
08:51 masak moritz: of course the interesting bits reside in the emergent phenomena from those signals.
08:51 moritz masak: ok, that usage I agree with :-)
08:50 masak moritz: well, I used it only as a rhethorical point against "either 'knowledge' or 'belief' is more defined than the other".
08:50 sorear (hi, I just got done reading four books on psycholinguistics)
08:50 brrt well, neither do i - used to when i was a child though
08:50 moritz but of course that could be context too
08:50 brrt that could be due to different 'construction routines'
08:50 moritz I for one don't have any trouble with some words that have the quite different meanings in English and German
08:49 sorear My mental dictionary absolutely never confuses English/sake with Japanese/sake
08:49 brrt in popular media, that is; i havent studied in any detail yet
08:49 brrt i recall reading it
08:49 moritz brrt: is that true that you have just one 'namespace'?
08:48 sorear brrt: I think it is wrong to say "one 'namespace'"
08:48 moritz masak: I kinda object to that sort of statement. It very much limits the concept of existence to some organizational layers
08:47 brrt or would it just be one, with different 'routines' per language
08:47 brrt does the brain have multipe 'compilers' for languages too?
08:47 brrt given that the brain has just one 'namespace' for words, regardless of how many languages you speak
08:46 brrt i was recently wondering
08:44 brrt very true
08:44 masak it's just electrochemical signals in wetware.
08:44 masak well, neither thing exists if you look closely enough.
08:43 brrt both stamements state the opposite thing as the 'defined' thing
08:43 * masak .oO( at least, that what I recall reading )
08:43 moritz and if that confidence is high, we call it "knowledge". The threshold isn't fixed.
08:43 moritz well, we humans tend to have some estimates about how confident we are about our memories
08:42 masak "black is just a very dark gray" -- "or gray is just a bright black :-)"
08:42 masak huh? our utterances are equivalent.
08:42 brrt although 'certain' is rather poorly defined
08:41 brrt i'd go with masak here
08:40 moritz or believe is just uncertain knowledge :-)
08:38 masak what is knowledge? knowledge is just sufficiently certain belief.
08:34 brrt and others don't :-p
08:34 brrt especially when you know it
08:34 brrt being wrong can be so much fun
08:33 masak all other things equal, I prefer to be right :P
08:20 tadzik and again, I'd like to be wrong
08:19 tadzik sorear: Heh, now that you say it I find it quite possible :)
08:18 sorear tadzik: you find it hardto believe that 6model in parrot will be after summer? :>
08:11 masak possibly.
08:09 moritz as weird as it sounds, I think that replacing the object model in parrot is less work than, say, cleaning up the call conventions
08:06 masak Perl 6: the Long Christmas.
08:04 brrt short summer, maybe
08:01 masak and on the summer.
08:01 masak depends on the Christmas.
07:58 brrt that'd be quick
07:58 frettled After summer and before Christmas.
07:56 tadzik It'd be lovely, but I find it quite hard to believe
07:56 tadzik whiteknight was saying something about "after summer" yesterday
07:55 brrt it seems like a huge project actually
07:55 masak it seems it'll require quite a big change in Parrot.
07:54 moritz 6model in parrot is still in the future, and it's not clear in which time frame we can expect it to land
07:54 moritz and that means PMCs
07:54 moritz for your project, you must use the status quo
07:54 brrt then i could wish for nothing more
07:53 brrt awesome
07:53 moritz sure
07:53 brrt somewhat relevant for my project, can i assume that whenever 6model is integrated, i can still use pmc's to connect the apache api to?
07:52 brrt i see
07:51 moritz all the "normal" objects are collected by parrot's GC
07:51 moritz mostly for things where order of destruction matters
07:51 moritz note that 6model uses refcounting only for certain things, not for the "normal" objects you use in the programming language
07:50 brrt :-)
07:50 brrt if so, will an integrated 6model-into-parrot use parrots garbage collecting, or the current refcounting?
07:50 moritz that's one of the few subsystems we haven't been tempted to reimplement in rakudo :-)
07:49 moritz yes
07:49 brrt but, does parrot use garbage collecting?
07:49 brrt this is more of a parrot discussion
07:49 brrt :-)
07:48 moritz you aren't the first to be confused by it :-)
07:48 brrt ok, i get it
07:48 moritz so, to reimplement the current PMCs on top of 6model, you also need a small MOP that provides enough functionality
07:47 moritz and PMCs use inheritance, for example
07:47 moritz that's something you build on top of 6model, as part of the MOP
07:47 brrt PMC does?
07:47 moritz 6model itself doesn't have a notion of classes, inheritance, roles etc.
07:46 brrt but, i still don't really understand the - planned - relation between PMC and 6model :-)
07:45 moritz PIR is not a small language
07:45 moritz aye
07:45 brrt so, big project actually
07:45 brrt i see
07:44 moritz well, it's a step towards replacing PIR
07:44 brrt so, he is going to implement PACT, which will replace PIR
07:43 moritz right
07:43 brrt the PACT thing
07:43 moritz he submitted two proposals, one was 6model, one was the compiler/compiler tools thing
07:43 brrt seems reasonable
07:43 brrt :-)
07:43 moritz at least in the scope of gsoc, benabik will write an abstract-syntax-tree-to-PBC compiler
07:42 * brrt is confused
07:42 moritz no
07:42 brrt benabik is going to integrate 6model into core and build pmc on top of 6model
07:42 brrt ... so, let me get this straight
07:41 moritz currently only nqp and rakudo use 6model
07:41 moritz it only uses parrot
07:41 moritz no
07:41 brrt winxed uses 6model?
07:41 brrt ok, i can live with winxed
07:41 moritz winxed, yes
07:40 brrt i was actually learning PIR a bit :-)
07:40 brrt if you get rid of PIR, what will be the 'smallest' language for parrot? winxed?
07:40 masak moritz: ok.
07:40 brrt also
07:39 moritz brrt: yes, 6model is in C
07:39 moritz masak: in my understanding, that's not quite accurate. It has an API for a MOP, but not a MOP itself
07:39 brrt pretty awesome, actually
07:39 brrt 6model is implemented in C?
07:39 brrt i'm reading the 6model overview pod right now
07:38 masak moritz: up until now, I've considered 6model a small, bootstrapping, MOP-agnostic MOP.
07:38 frettled In my experience, avoiding makefiles mostly means that you will end up reinventing them. Badly. I'm not sure why anyone would want to put themselves through that kind of pain.
07:38 tadzik frettled: cross-platform issues I suppose
07:38 masak the distinction between objects and meta-objects is deliberately blurred. meta-objects are just objects.
07:38 moritz brrt: 6model is not a meta-object model. It's a storage theme which makes it easy-ish to write your own meta-object model
07:37 moritz brrt: oh, and we want to get rid of PIR even before that. The plan is to compile from PAST or similar to PBC directly
07:37 brrt 6model is a meta-object model?
07:37 brrt so, pmc is an object model
07:37 moritz brrt: so there would need to be a minimalistic object model on top of 6model, in which PMCs would be implemented
07:36 moritz brrt: 6model is mostly a way to store objects and their meta objects, but doesn't come with any object model
07:36 brrt what i'm wondering is, what abstraction level are they, then?
07:36 brrt and if so, what will happen to compiling-to-pir?
07:36 moritz brrt: 6model and pmc aren't quite the same abstraction level
07:35 brrt if so, what will happen to pir?
07:35 frettled but these days, "cross-platform" usually means "works on RedHat AND Debian/Ubuntu"
07:35 moritz the other one is the "Malformed UTF-8" when getting the second row
07:35 * masak .oO( cross is incredibly hard to do )
07:35 brrt will it completely replace pmc?
07:35 brrt hey, i was wondering, if 6model is integrated to parrot core
07:35 frettled moritz: cross-platform is incredibly hard to do :)
07:35 moritz one is because 1.90 returns as 1.9, which IMHO isn't wrong
07:34 masak then again, I've just been targeting two platforms.
07:34 masak IME, it just means you can't use all the nifty syntax of GNU Makefile.
07:34 moritz sqlite status: 2 failing tests
07:33 moritz cross-platform makefiles are incredibly hard to write
07:33 masak and what makes you think you can avoid it? :)
07:33 frettled tadzik: why don't you want to use makefiles?
07:32 tadzik well, can work this way, yes
07:32 tadzik things like %*ENV<FOO>; shell 'echo $FOO'; are not windows-friendly
07:32 moritz %*ENV<foo> = 'bar'; is_run 'say %*ENV<foo>', "bar\n"
07:31 tadzik cross-platform testing may be tough
07:31 moritz testing shouldn't be too hard
07:31 tadzik and zavolaj eventually works without its help anyway :)
07:30 moritz \o/
07:30 tadzik no tests
07:30 tadzik moritz: worksforme
07:30 moritz tadzik: what's the state of the setenv branch?
07:29 tadzik I really don't want to fall back to using makefiles
07:29 tadzik imagine a Configure.pm, which gets require()d by Panda code and something gets run inside. This something sets LD_LIBRARY_PATH, so it stays for subprocesses. Thus when we later do perl6 --target=pir or run tests, the variable stays
07:29 moritz do makefiles under windows allow ENVVAR=value programtorun syntax?
07:28 tadzik hmm, I think setting LD_LIBRARY_PATH can be fixed outside of zavolaj after all
07:28 masak oh, point.
07:27 moritz note that much of the networking code is also in IO/Socket.pm
07:26 frettled Good late morning :D
07:26 frettled masak o/
07:03 masak probably the easiest way to get into things.
07:03 masak melatinini: or you could read the source code: https://github.com/rakudo/rakudo/blob/nom/src/core/IO/Socket/INET.pm
06:59 masak huh. raiph++
06:59 * masak backlogs
06:59 phenny masak: 01:50Z <raiph> tell masak http://irclog.perlgeek.de/perl6/2012-04-27#i_5503487
06:59 masak good actual morning, #perl6
05:01 sorear o/ moritz
05:01 moritz o/
05:00 melatinini sorear: thanks!
04:59 melatinini haha, ok, fair. i see that the rakudo star that came out today has some new www features :)
04:59 sorear melatinini: welcome to #perl6
04:59 sorear melatinini: trial and error? asking people who have used it?
04:54 melatinini hi, how can i explore the rakudo sockets interface? can't find documentation on io::socket::inet and i don't think it's the same as perl5's...
03:04 dalek ecosystem: review: https://github.com/perl6/ecosystem/commit/31f6037205
03:04 dalek ecosystem: Math::BigInt was made completely pointless once Rakudo got native big ints, so I'm removing it.
03:04 dalek ecosystem: 31f6037 | colomon++ | META.list:
01:50 phenny raiph: I'll pass that on when masak is around.
01:50 raiph phenny, tell masak http://irclog.perlgeek.de/perl6/2012-04-27#i_5503487
01:49 raiph phenny, let masak know about http://irclog.perlgeek.de/perl6/2012-04-27#i_5503487
01:47 p6eval rakudo 256e1d: OUTPUT«(signal SEGV)»
01:47 raiph r: macro f { quasi { state $a = 0; $a++ } }; f # but not this
01:47 p6eval rakudo 256e1d: ( no output )
01:47 raiph r: macro f { quasi { state $a = 0 } }; f # innerestin alternate, this works
01:47 p6eval rakudo 256e1d: OUTPUT«Cannot assign into a PMCNULL container in <anon> at /tmp/UisqLGWwy_:1 in block <anon> at /tmp/UisqLGWwy_:1»
01:47 raiph r: macro f { quasi { my $a = 0 } }; f # golf'd 4 masak
01:10 whiteknight Okay, pull request sent. Thanks [Coke]
01:06 whiteknight yes
01:05 [Coke] is this for nqp-latestS?
01:05 [Coke] pull request makes it easy.
00:52 whiteknight What's the current best way to submit changes to NQP, patch or pull request?
00:51 lichtkind good night
00:50 dalek tablets: review: https://github.com/perl6/tablets/commit/20b8c5158b
00:50 dalek tablets: simplify header and polish
00:50 dalek tablets: 20b8c51 | (Herbert Breunung)++ | docs/appendix-a-index.txt:
00:30 dalek tablets: review: https://github.com/perl6/tablets/commit/6c086df053
00:30 dalek tablets: polish links and format
00:30 dalek tablets: 6c086df | (Herbert Breunung)++ | docs/appendix- (3 files):
00:14 japhb 3. Automate the process of checking out and building particular versions, and do a run of each, compare across those.
00:13 japhb 2. Cache old results; allow comparisons of current run against an old result run.
00:12 japhb 1. Run all tests against a given version, dump timings. Do your own comparisons.
00:12 japhb Yeah, that was one. I can see three ways to do that:
00:09 sorear japhb: making changes to niecza and wondering how muhc slower it makes them
00:08 japhb sorear, BTW, what was the actual use case you were looking to perl6-bench for? (I have a couple use cases of my own that won't be served by the current 'bench' tool, so I'm curious if there is overlap in our needs).
00:01 japhb o/
00:00 sorear good * #perl6