00:00 __david__ joined
masak jnthn: "abuse of a proto"/"someone will probably find a way to [abuse it more]". you must be new here. :P 00:01
masak cackles evilly
colomon "This week's challenge is to figure out how to abuse proto...." 00:02
masak jnthn++ # nice blug post 00:03
00:03 whiteknight joined
dalek odel: e7e4edb | diakopter++ | / (3 files):
[dotnet] implement push, pop, unshift, shift in NQPArray
colomon jnthn++ # what masak said, plus sounds like good hacking!
jnthn 'twas a fun weekend :-) 00:05
lue jnthn++ # everything said above + reminding me to check on the P6 community more often 00:07
00:10 risou left
sorear woah, diakopter is hacking 6model now? 00:14
masak diakopter++ 00:16
jnthn I only wrote the "stuff to hack on" file yesterday and diakopter++ has already demolished over half the tasks. :-) 00:24
masak what a delightful destructive activity! 00:26
jnthn :-)
jnthn goes to get some rest - $dayjob tomorrow 00:28
night o/
sorear I should make one of those!
masak night, jnthn. 00:30
lichtkind good night everyone 00:35
perigrin jnthn: that just suggests you need to write a larger TODO list next time 00:38
or pick harder things
masak lol, I bloggethed: strangelyconsistent.org/blog/yet-an...f-november 00:42
lue masak++ # I feel like I haven't done Actual Real Coding™ as well 00:46
masak adds the '™' to the post 00:47
lue Web.pm is the thingy that (theoretically at least) would let me evilly code my site in P6 instead of, say, PHP, right? 00:49
masak right.
there's even been talk about making a Rakudo distribution optimized for that. 00:50
I think that would be a very nice goal.
flussence from what I understand, it'd be really easy to make a setting library to make rakudo php-like. 00:52
(auto http headers and stuff like that, not the functions :)
lue I tried building meself a site (no place online yet though), and I was wearing my P6-colored glasses while coding javascript, and couldn't bear it :D 00:53
diakopter perigrin: yes, harder things :)
(or just lots more things)
well-expressed/defined (aka easy) things are good too. 00:54
masak lue: have you seen the "Good Parts" talk by Doug Crockford? www.youtube.com/watch?v=hQVTIJBZook
perigrin diakopter++ 00:56
lue No, and I can't.
[ A ~10 year old laptop does not a video player make :/ ] 00:57
masak oh. sorry. :/
anyway, it's a good talk.
lue It's fine. I tried watching Patent Absurdity once by downloading it first. 00:59
It was the day I learned this computer cannot play any sort of 'video' or other visual wonder.
(I just tried compiling LibreOffice. Letting it run 24/7, it took days (as in several) and still didn't finish. *sob*) 01:00
01:01 leprevost joined
masak possibly a sign that it's time to upgrade? 01:01
lue I got this laptop for free though. :) 01:03
Compared to LibreOffice, Rakudo takes no time (I heard LibreOffice takes 2 hours on a quad-core!) 01:04
flussence I can sympathise with that, chromium is a nightmare to compile too :) 01:06
lue Then Chromium would be a Steve Moffat story for me ;) 01:07
flussence on my quad-core desktop, it takes longer than rakudo on my netbook...
I think that's due to it bundling half an OS worth of duplicate libs, though. Hopefully that'll be gone by version 10ish 01:10
colomon the last november for masak?!?!!! 01:12
masak well, I'll still exist afterwards :) 01:13
I hope.
lue as long as you don't have a legally binding contract with any deity that tied you to those novembers, you'll be fine. 01:14
colomon and you're making a secret announcement in december, eh?
masak yep.
sorear lue: do you also have free electricity/
colomon (well, the announcement isn't secret)
masak colomon: clue. you'll like it.
colomon \o/ 01:15
sorear lue: if not, a reasonable upgrade will pay for itself
power conservation tech improved a *lot* from 2000-2007, and 3 year old computers are still pretty cheap
lue My computer could be wasting energy and I don't even know it!? WOW! [not sarcastic there] 01:18
01:20 lichtkind left, masonkramer left
flussence I've actually checked mine a few times... 01:21
barebones P3 server - ~70W, P4 desktop - 200W, PhenomII X4 - 120W
colomon masak: I'm having a hard time imagining what you'd know for certain now and yet want to keep a secret.
masak colomon: good. :) 01:22
colomon but I like good surprises!
sorear the P4 is one of the worst processors in history as far as power consumption 01:27
they had to switch to the Core line cause P5 prototypes kept melting
masak that processor was hot! 01:28
lue the PPC G5 suffered like that, why it's not in any Apple laptops (unless I missed something and P5=G5) 01:33
sorear no, the P4 was an Intel chip, the G5 came from IBM 01:39
01:42 rgrau_ left
masak sleeps 01:43
01:43 masak left 01:48 Chillance left 01:55 pythonian4000 is now known as pythonian4000afk
diakopter afraid; I did stuff not in LHF.txt or ROADMAP 01:57
dalek odel: ad6fd43 | diakopter++ | / (3 files):
[dotnet] finish implementing pasttype while; implement until, repeat_while, repeat_until; add 14-while.t (all pass); add 05-comments.t b/c it passes.
diakopter actually I guess those are in ROADMAP under "Other types of loop" 01:59
sorear needs to make up a LHF and a ROADMAP 02:00
and finish up the docs
diakopter :)
sorear: jnthn's overall compilation approach is _incredibly_ similar to [what I thought I was inventing, novelly, in] Perlesque's 02:02
diakopter starts to guess it's kindof a common pattern 02:03
02:12 whiteknight left
sorear I'm going to need to implement real continuations at some point, hmm 02:16
02:19 lestrrat left 02:20 rjbs left 02:21 starcoder left, mdxi left
diakopter sorear: :P 02:26
02:29 rjbs joined 02:30 starcoder joined, starcoder left, lestrrat joined 02:33 starcoder joined, mdxi joined 02:42 starcoder left 02:46 starcoder joined 03:04 plainhao left 03:05 dnl joined
dnl moinsen 03:05
sorear hi 03:06
03:13 galf joined, galf left
diakopter sigh. stupid merge branch master, again. 03:15
[CursorBase] special thanks to NYTProf; apply some inlining and CSE for a 10-15% speedup when parsing grammar-heavy input. github.com/perl6/std/commit/c3fb774...e1a683181a 03:16
03:17 f00li5h is now known as skeleton_feline, pythonian4000afk is now known as pythonian4000 03:25 skeleton_feline is now known as f00li5h 03:56 ash_ joined 04:23 nadim left, Alias left 04:24 Alias joined 04:25 Limbic_Region left 04:27 nymacro joined 04:40 dnl left 04:46 Chat5550 joined 04:50 Chat5550 left 04:57 leprevost left 05:04 vx64z left 05:06 iWork joined 05:15 molaf joined 05:30 rvrgs joined 05:37 molaf left 05:52 Elise joined 05:53 Elise left 05:54 rvrgs- joined 05:55 rvrgs left 06:00 justatheory left 06:06 sftp joined 06:07 rvrgs- left 06:33 meppl joined 06:34 Chat0660 joined
Chat0660 hi room 06:34
06:34 kaare joined 06:35 drbean left, kaare is now known as Guest60740
sorear hello 06:35
Chat0660 whatsup
sorear How long of a timespan do you care about? 06:36
06:37 drbean joined
Chat0660 what do u mean 06:37
06:40 Chat0660 left 06:47 nadim joined 06:55 HarryS left 06:56 HarryS joined 06:57 ash_ left 07:08 uniejo joined 07:24 cognominal left 07:39 wallberg joined 07:43 QinGW joined 08:04 Bzek_ joined 08:08 Bzek left 08:10 jfried joined, clkao left 08:18 agentzh joined 08:27 clkao joined 08:28 sftp left 08:29 starcoder left 08:31 starcoder joined 08:33 Axius joined 08:38 starcoder left
jnthn o/ #perl6 08:42
moritz_ \o 08:45
uniejo |o| 08:46
LoRe /o\ 08:47
08:53 uniejo left 08:55 Axius left, uniejo joined 08:56 nymacro left 08:58 starcoder joined 09:00 Axius joined 09:03 nymacro joined 09:05 Axius left 09:06 Axius joined 09:08 _iWork_ joined 09:12 iWork left 09:32 nymacro left 09:36 Axius left 09:39 Guest60740 left 09:41 timbunce joined 09:42 kaare joined, kaare is now known as Guest94482 09:43 mavrc joined 09:50 Guest94482 is now known as kaare_ 09:56 lamstyle left 10:00 lamstyle joined 10:01 mavrc left 10:02 mavrc joined 10:03 QinGW left 10:06 nymacro joined 10:29 _iWork_ left 10:40 tadzik joined
tadzik o/ 10:41
jnthn o/ tadzik
10:48 tadzik left 10:50 tadzik joined 11:00 icwiener joined 11:05 masak joined
masak oh hai, #perl6! 11:05
moritz_ looks forward to masak's next November blog post 11:08
masak aka "masak's first November blog post"... :)
tadzik oh hai masak
so, is pls going to use Module::Tools maybe? 11:09
masak possibly. 11:10
will investigate. 11:11
but now, lunch &
11:14 rgrau_ joined 11:30 plobsing left 11:31 PacoLinux left 11:32 wamba joined 11:43 pythonian4000 is now known as pythonian4000afk 11:54 Trashlord left
jnthn diakopter: Just read over ad6fd437 - it looks good. :-) Yes, that was what the relevant item in the ROADMAP meant. 12:04
jnthn digs back into $dayjob after nice noms
12:07 kensanata joined 12:13 meppl left 12:14 iWork joined 12:15 agentzh left 12:17 plainhao joined, iWork left 12:18 iWork joined 12:21 iWork left 12:22 iWork joined 12:23 iWork left 12:25 satyavvd joined 12:31 Trashlord joined 12:32 am0c joined 12:33 tadzik left 12:34 am0c left
masak they what? twitter.com/pjf/status/29265124349 12:45
jnthn hehe 12:46
moritz_ giggles
masak now it feels like I'm missing out... :) 12:47
Perl 7.9. fancy that.
jnthn I want to think it's a typo for 5.9, apart from that woulda been a development version... 12:48
masak yeah. it's odd.
jnthn Literally *and* figuratively. ;-) 12:49
takadonet morning
sjn 5.7.9? :)
Perl6, version 7.9
moritz_ sjn: would be weird, since that's an (old) development version 12:51
sjn yep
doesn't seem like a typo either 12:52
sbp they're using a fractional base, not decimal
sjn 7 and 5 aren't really next to eachother on the keyboard
perhaps they mean the "september 2007" release? :-P 12:53
12:59 satyavvd left
moritz_ rakudo: sub factors($n) { ($n X/ 1..$n).grep: { .Int == $_ } }; say factors(15).join(', ') 13:00
p6eval rakudo 5f5bae: OUTPUT«15, 5, 3, 1␤»
masak rakudo: sub factors($n) { grep { $_ %% $n } 1..$n }; say factors(15).fmt(", ") 13:09
p6eval rakudo 5f5bae: OUTPUT«===SORRY!===␤Unable to parse blockoid, couldn't find final '}' at line 22␤»
masak rakudo: sub factors($n) { grep { $_ %% $n }, 1..$n }; say factors(15).fmt(", ")
p6eval rakudo 5f5bae: OUTPUT«, ␤»
masak huh. 13:10
rakudo: sub factors($n) { grep { $_ %% $n }, 1..$n }; say factors(15).fmt
p6eval rakudo 5f5bae: OUTPUT«15␤»
masak rakudo: sub factors($n) { grep { $n %% $_ }, 1..$n }; say factors(15).fmt
p6eval rakudo 5f5bae: OUTPUT«1 3 5 15␤»
PerlJam heh
masak rakudo: sub factors($n) { grep { $_ R%% $n }, 1..$n }; say factors(15).fmt # :) 13:11
p6eval rakudo 5f5bae: OUTPUT«1 3 5 15␤»
moritz_ tries to find nice non-junction versions of Damian's examples 13:13
masak moritz_: I had missed you in that thread. kudos for entering it.
jnthn I really want to read that thread and get to grips with it. 13:14
But multi-dispatch has been nomming all my brain tuits.
masak I find nowadays that I'm not too invested in junctions. they may be a central part of Perl 6, but they're not a central part of my Perl 6 coding. 13:16
they're occasionally nice, that's all.
sbp moritz_++ # entering the thread
jnthn I find them like hyper-operators. I won't use them in every bit of code I write, but when they're useful they're really useful.
PerlJam so ... you guys use || more than | in regex? :) 13:17
moritz_ jnthn, masak: I fully agree
masak PerlJam: I tend to use ||, because (I'm not aware that) | works as spec'd yet. 13:18
urgh. parens fail.
| doesn't work as spec'd in regexes. at least I'm not aware that they do.
13:18 MayDaniel joined
masak hence, I tend to use ||, for future compatibility. 13:18
I tend to avoid writing bitrottable code when I can. :) 13:19
13:20 dnl joined 13:21 Patterner left
jnthn The link with | in regexes is also more fluffy than deep implementation level too, I suspect. 13:22
masak moritz_++ # making a lot of sense in the thread
but urgh, sigil-less constants. do not want. :)
Util Having created a class ReallyLongClassName, I want to make a shorter typename R, so that I can say 13:23
multi sub infix:<+> ( R $a, R $b ) {...}
instead of
masak those sigils are there for a reason! why would I, after choosing a sigil-full language as Perl, choose to deprive the reader of the information that UNACCEPTABLE is an array?
Util multi sub infix:<+> ( ReallyLongClassName $a, ReallyLongClassName $b ) {...}
What is the Perl 6 syntax for such an alias, and does Rakudo support it yet?
masak Util: constant R := ReallyLongClassName
moritz_ Util: in a perfect world, my ::R := ReallLongClassName
masak Util: or subtype R of ReallyLongClassName, but that's assymetric in that ReallyLongClassName !~~ R, I think. 13:24
moritz_ masak: that will result in parse errors when you try to use the constant in place of a type name
masak moritz_: oh?
rakudo: class A {}; subtype B of A; say A ~~ B; say B ~~ A
p6eval rakudo 5f5bae: OUTPUT«Could not find sub &of␤ in main program body at line 22:/tmp/gMeharChHr␤»
Util Single colons instead of double colons on the := ?
masak rakudo: class A {}; subset B of A; say A ~~ B; say B ~~ A
p6eval rakudo 5f5bae: OUTPUT«Bool::True␤1␤»
masak interesting answer :) 13:25
moritz_ std: constant R = Int; sub f(R $x) { }
masak submits rakudobug
p6eval std a194beb: OUTPUT«Potential difficulties:␤ $x is declared but not used at /tmp/iCojxN1Gyb line 1:␤------> constant R = Int; sub f(R ⏏$x) { }␤ok 00:01 123m␤»
moritz_ std: constant R = Int; sub f(R $x) { $x }
p6eval std a194beb: OUTPUT«ok 00:01 123m␤»
moritz_ wonders why that works
masak Util: double colons are probably better.
jnthn rakudo: ::I = Int; my I $x; say $x.WHAT
p6eval rakudo 5f5bae: OUTPUT«===SORRY!===␤In "my" declaration, typename I must be predeclared (or marked as declarative with :: prefix) at line 22, near " $x; say $"␤»
jnthn rakudo: my ::I = Int; my I $x; say $x.WHAT
p6eval rakudo 5f5bae: OUTPUT«===SORRY!===␤Malformed my at line 22, near "::I = Int;"␤»
masak moritz_: why wouldn't it work?
jnthn :/
masak moritz_: oh, I see now. 13:26
moritz_ masak: because only type names are allowed before parameters
masak right, and the parser doesn't know that the constant represents a type name.
13:26 MayDaniel left
moritz_ www.perlmonks.org/?node_id=868728 13:28
13:28 estrabd joined
moritz_ thinks he now knows how to describe meta object protocols to the curious perl programmer 13:29
it just takes a lot of work to write down
13:29 MayDaniel joined, dnl left
masak the meta object protocol is just an API to OO stuff. 13:30
an OO API to OO stuff, even.
moritz_ right
jnthn moritz_: Dunno if any of my slides or blog posts make good things to link to.
moritz_ jnthn: they do, but I'm thinking of a somewhat different approach 13:31
jnthn I have explained this stuff, though it's often in the context of other things.
13:31 araujo left
masak likes jnthn's "innocent questions" approach :) 13:31
moritz_ let's start with a simple class, where each instance stores a callback
if it makes this callback invocable through a method call, it's a very simple MOP 13:32
if all instances share the same callbacks, it acts like a class implementation
masak sounds like the "let's build it ourselves" approach. 13:33
moritz_ indeed
13:33 Psyche^ joined, Psyche^ is now known as Patterner
masak I like that, too. 13:33
moritz_ starting from something very common (callbacks)
masak aye.
Util Thanks, all. Since none of these are supported in Rakudo yet, I will workaround, and make a note to revisit after one of those solutions is fully implemented. 13:34
jnthn Yes, showing how to build a simple one up is nice. 13:35
masak Util: huh? the subset thing seemed to work.
jnthn Worth noting that the core primitive in 6model is actually really, really simple too.
13:35 Trashlord left
jnthn Though its factoring is tied up in representation poly, which is less fun if you're still trying to get your head around meta-models. 13:35
13:36 lidden joined
jnthn But all it manages to do is dispatch methods and hold a list of attributes. 13:36
Util masak: creating the subset works, but using the subset in a param-list declaration is failing. Hmm. Let me look closer.
masak rakudo: class A {}; subset R of A; sub foo(R $x) { say $x }; foo(A.new) 13:37
p6eval rakudo 5f5bae: OUTPUT«A()<0x5db87f0>␤»
13:37 Trashlord joined
masak seems to work here. 13:37
of course, among the different approaches, this is the Wrong one, conceptually. 13:38
but it works.
13:40 MayDaniel left
jnthn e.g. you could accidentally a candidate narrower like this and affect multi-dispatch. 13:40
Since R is narrower than A here. 13:41
13:42 nymacro left
moritz_ rakudo: my ::A 13:42
p6eval rakudo 5f5bae: OUTPUT«===SORRY!===␤Malformed my at line 22, near "::A"␤»
jnthn std: my ::A;
p6eval std a194beb: OUTPUT«===SORRY!===␤Illegal redeclaration of symbol 'A' (see line 1) at /tmp/2_oKyMhSZP line 1:␤------> my ::A⏏;␤Multiple prefix constraints not yet supported at /tmp/2_oKyMhSZP line 1:␤------> my ::A⏏;␤Malformed my at
moritz_ that's a LTA error
jnthn std: my ::AAAAAAAAAAAAAAAAA; 13:43
p6eval std a194beb: OUTPUT«===SORRY!===␤Illegal redeclaration of symbol 'AAAAAAAAAAAAAAAAA' (see line 1) at /tmp/KnwUjRMmkd line 1:␤------> my ::AAAAAAAAAAAAAAAAA⏏;␤Multiple prefix constraints not yet supported at /tmp/KnwUjRMmkd line 1:␤------> my
perigrin masak: I've been telling people that MOPs are Object Models of how to build Object Models.
masak submits rakudobug
perigrin: yes, exactly. they're an OOD for OO.
perigrin yep
jnthn I tend to see a meta-object as just an object that describes how another object behaves and/or provides introspection for it. 13:44
13:44 araujo joined
moritz_ perigrin: do people actually understand that, if they haven't worked with a MOP before? 13:44
perigrin moritz_: sometimes 13:45
it depends on their exposure to OO
a lot of people write very procedural code in OO
13:46 dnl joined
masak sometimes that's what you want. 13:47
Util Aha! Subset *does* work. I was using Q instead of R, which fails (I think) because single caps are reserved for meta-ops.
Using Qu works: subset Qu of Quaternion;
masak Util: quoting. 13:48
Util: Q fails because it's used for quoting.
Util Oh , of course; R is a meta-op anyway.
masak it's a known, and quite annoying, bug.
Util Thanks!
13:49 perlygatekeeper joined
moritz_ std: subset Q of Int; my Q $x; 13:50
p6eval std a194beb: OUTPUT«ok 00:01 119m␤»
13:56 mkramer joined 14:02 PacoLinux joined 14:04 hanekomu_ joined 14:06 meppl joined 14:08 wamba left 14:30 orafu joined 14:42 tadzik joined 14:43 alester left
tadzik hello 14:48
14:49 cognominal joined, broquain1 is now known as broquaint
masak tadzik: hi! 14:51
"perl 1.0 patch 8: perl needed an eval operator and a symbolic debugger" -- wow. github.com/mirrors/perl/commit/a559...01be0bac10
Perl was small at that time.
tadzik oh wow
14:52 PacoLinux left
jnthn ...is that debugger really a source filter that inserts debug hooks? :-) 14:53
14:53 spq1 joined 14:54 kaare_ left
jnthn "What has been seen..." :-) 14:54
tadzik :)
seen smash 14:55
aloha smash was last seen in #perl6 2 days 23 hours ago saying "pmichaud: mornin'".
tadzik I'm curious about gil.di.uminho.pt/users/smash/rakudo-bench.html
15:01 ash_ joined 15:02 kanishka joined 15:10 dukeleto left 15:11 dukeleto joined, bluescreen joined, bluescreen is now known as Guest49255 15:13 uniejo left 15:16 nymacro joined 15:22 MayDaniel joined 15:24 slavik1 left
masak so far Perl 1 and Perl 6 are tied in moritz_++' poll. www.perlmonks.org/?node_id=868728 15:24
I wonder if people know that they're voting for Perl 1, and not for (say) vanilla Perl 5.
either that, or they're simply punie humans. :P 15:25
tadzik voted and changed the situation :) 15:28
15:29 _kaare joined, slavik1 joined 15:34 miso_soup joined
masak (Buddha Buck)++ # introducing *real* quantum theory on p6l, not the pop-culture trivialized versions of which junctions are but one example 15:38
colomon pmichaud++ # listening to his interview with merlyn now... 15:39
15:40 timbunce left
masak but, unfortunately, due to junctions only being a distant relative to real quantum superpositions, I doubt that "asking for the eigenstates of a quantum superposition is asking the wrong object for the property" translates to anything sensible in Perl 6. 15:40
of course, that's just one more argument for .eigenstates being the wrong name for the method.
15:41 Trashlord left 15:43 Trashlord joined, Trashlord is now known as Guest94158
pmichaud hello, all 15:44
masak hi pmichaud o/
tadzik hello, pmichaud
masak pmichaud++ # interview
tadzik indeed
15:44 nymacro left
pmichaud I'm stuck offline most of the day -- we're getting new broadband service at the house today 15:44
colomon pmichaud: listening to you talk right now...
masak today is Yapsi release day, by the way. 15:45
pmichaud yapsi++
colomon masak: \o/
jnthn o/ pmichaud
tadzik (Official Perl 6)++
jnthn pmichaud: Will you make #phasers tomorrow?
pmichaud colomon: how does the talk sound? I haven't heard it yet
colomon sounds great
pmichaud jnthn: definitely, unless something unexpected comes up
jnthn pmichaud: OK. I'll haz a question or two. :-) 15:46
masak question is, should I spend time adding some feature to Yapsi today, or should I just release it as-is? hm...
pmichaud masak: whatever facilitates the release
jnthn pmichaud: The main one is: if you were doing pasttype for, and the underlying implementation looked however you fancied, what would you do? Something like Parrot iterators, or... 15:47
masak pmichaud: good point.
tadzik masak: it didn't change much since the last release, did it?
masak no, not much.
but there was a bugfix that went in.
I know!
pmichaud jnthn: hmmm, I'll have to think about that one (1 day)
masak I'll restore Tardis to a working condition.
that's not much work.
tadzik "I'll implement threading!"
pmichaud I don't have a solid answer 15:48
jnthn pmichaud: It's fine, I wasn't planning to do it tonight. :-)
pmichaud: But would really like to have for loops soon. For one because not being able to iterate hashes is gonna block me soon. :-)
I figure we don't want the full-blown Perl 6 iterator model at NQP level though.
pmichaud I'm wondering if perhaps we do.
15:49 PacoLinux joined
tadzik jnthn: are you basically rewriting the whole NQP? 15:49
jnthn tadzik: Not really. It's just that pasttype<for> on Parrot compiles down to Parrot's iterator model. I don't have one of those yet for the 6model on .Net implementation though. 15:50
tadzik I see 15:51
15:52 justatheory joined
masak bottom icon at modules.perl6.org/fame-and-profit has bitrotted. :/ 15:53
pmichaud afk for a while
16:04 hanekomu_ left, iWork joined
masak wow, Tardis is not a very big project :) 16:04
this is promising. :)
flussence it's bigger on the inside! 16:05
masak I see now that I really had that one coming. :) 16:06
16:06 timbunce joined
masak good news! Tardis, while written for alpha and since unmodified, workd fine on master! \o/ 16:10
moritz_ masak: wow, that's impressive
masak maybe because it's so small.
colomon \o/
masak it's basically a subclassing of the Yapsi runtime.
hm. it runs, but not 100% flawlessly. 16:11
time to tinker a bit.
colomon time to add more tests! 16:12
masak indeed. 16:13
I haven't run the tests yet.
16:15 silug left 16:16 ggoebel joined 16:17 MayDaniel left
jnthn home & 16:19
16:19 stkowski joined
moritz_ masak: that would have been my first thing to do - running tests :-) 16:20
that's what you write them for, after all
masak mumbles something about backseat coders... :)
I'm going to run the tests. of which, I'm sure, there are several... :) 16:21
moritz_ "several" 16:22
16:22 miso_soup left
moritz_ mumbles something about backseat testers 16:22
masak :D
moritz_ :-) 16:23
masak I'm not sure we agree on the term "backseat X-er" here. but nvm.
16:24 plobsing joined
moritz_ I know it wasn't the best term to use, but worth the retort anyway :-) 16:24
masak nod 16:25
16:30 cdarroch joined, cdarroch left, cdarroch joined, sftp joined 16:31 mberends joined, simon1 joined, Ross^ joined 16:32 wamba joined
masak the tests tell me the same thing that running the program manually did. ordinarily, I would have run the tests first as well. but I'm a pessimist, and I actually thought I'd get a compiler error :) 16:32
moritz_ so you think your programs are half broken :-) 16:33
masak that's not a matter of debate :P 16:34
16:34 simon1 left
masak rakudo: my $a = "fool!"; $a ~~ / (foo) /; say $0 16:35
p6eval rakudo 5f5bae: OUTPUT«foo␤»
masak rakudo: my $a = "fool!"; $a !~~ / (foo) /; say $0
p6eval rakudo 5f5bae: OUTPUT«foo␤»
masak ooh!
masak resolves rt.perl.org/rt3/Ticket/Display.html?id=76892
16:39 patspam left
masak also interesting: last time I wrote this, I had the runtime do self.*tick rather than self.?tick -- my reasoning was that if there are several subclasses, each subclass will want to run their .tick method. 16:41
but this time it seems that I'm going with self.?tick again.
reasoning as follows: it's always the responsibility of the .tick method in the most derived class to determine how the rest of the dispatch is to be performed. 16:42
TimToady it can control it either way
masak is the result the exact same, though? 16:43
TimToady but one way you use lastcall, and the other way you use nextsame
masak ah, right.
TimToady all you're changing is the default
masak nod.
TimToady the main thing is that they all have to agree
masak anyway, I think I want the default to be .? and nextsame.
16:44 jhuni left
masak just as I want the default in my email client to be "reply" and not "reply to all" :) 16:44
by the way, I don't remember if cacm.acm.org/magazines/2010/11/1006...3/fulltext has been linked here yet. 16:47
colomon I haven't seen it linked here.
I think he's a bit daft. 16:48
masak please elaborate.
(I haven't read the article.)
colomon but I'm shocked at how many people out there responding to the article want nothing to do with Unicode.
He disses Perl, operator overloading, and heaven forbid the notion you might define your own operators. 16:50