Camelia, the Perl 6 bug

IRC log for #perl6, 2012-08-25

Perl 6 | Reference Documentation | Rakudo | Niecza | Specs

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

All times shown according to UTC.

Time Nick Message
00:07 colomon FROGGS++: Awesome!  I was actually trying and failing to get MP3 playback working in Niecza a month or two ago.  (Not Niecza's fault as far as I know, I think it was Mono issues.)
00:10 gongyiliao joined #perl6
00:15 dayangkun joined #perl6
00:16 nebuchadnezzar joined #perl6
00:16 Circlepuller joined #perl6
00:20 s1n joined #perl6
00:25 jaldhar joined #perl6
00:29 Circlepuller_ joined #perl6
00:31 fgomez joined #perl6
00:47 dayangkun joined #perl6
01:01 tokuhiro_ joined #perl6
01:05 gongyiliao joined #perl6
01:18 popl joined #perl6
01:28 lumi_ joined #perl6
01:48 bonifatio joined #perl6
01:49 dayangkun joined #perl6
02:00 skids joined #perl6
02:52 thou joined #perl6
02:54 bonifatio joined #perl6
02:55 dayangkun joined #perl6
03:27 japhb Very strange time anomaly in Rakudo 3188aee:  time perl6 -e '' => .477s,  time perl6 -e '0' => 0.934s,  time perl6 -e 'use JSON::Tiny;' => 0.594s
03:28 japhb jnthn, pmichaud, moritz: Any ideas ^^ ?
04:21 dalek joined #perl6
04:21 PerlJam joined #perl6
04:27 tadzik joined #perl6
04:27 hugme joined #perl6
04:28 pmichaud joined #perl6
04:29 Juerd joined #perl6
04:29 masak joined #perl6
04:30 topo joined #perl6
04:30 [Coke] joined #perl6
04:31 p6eval joined #perl6
04:31 Util joined #perl6
04:37 PerlJam joined #perl6
05:01 desertm4x joined #perl6
05:34 orphu joined #perl6
05:44 erkan joined #perl6
05:44 erkan joined #perl6
05:56 kaare_ joined #perl6
06:04 Telgalizer joined #perl6
06:22 araujo joined #perl6
06:22 araujo joined #perl6
06:44 FROGGS joined #perl6
06:55 thou rn: ^9.roll(30);
06:55 p6eval rakudo 5ceefd, niecza v19-37-g0567c61:  ( no output )
06:55 thou rn: ^9.roll(30).perl.say;
06:55 p6eval rakudo 5ceefd, niecza v19-37-g0567c61: OUTPUT«(9, 9, 9, 9, 9, 9, 9, 9, 9, 9, 9, 9, 9, 9, 9, 9, 9, 9, 9, 9, 9, 9, 9, 9, 9, 9, 9, 9, 9, 9).list␤»
06:57 thou rn: (^9.roll(30)).perl.say;
06:57 p6eval rakudo 5ceefd, niecza v19-37-g0567c61: OUTPUT«0..^30␤»
07:06 GlitchMr joined #perl6
07:17 masak good morning, #perl6
07:20 felher o/ masak
07:21 masak oh, nice. now I see a face when I read the nick "felher". felher++
07:23 felher yeah. Same here with you and the other guys i met :)
07:26 * felher wishes everyone attending the hackathon a happy and productive time is off for $job
07:45 moritz \o
07:46 FROGGS joined #perl6
07:56 FROGGS joined #perl6
08:05 thou seen dalek
08:05 aloha dalek was last seen in #perl6 3 hours 44 mins ago joining the channel.
08:18 Psyche^ joined #perl6
08:20 leont joined #perl6
08:23 SamuraiJack joined #perl6
08:44 eiro hello all
08:45 diakopter ~hi
08:46 diakopter what's up?
08:48 eiro seen tons of new people at yapc. impressed by the young people and women: it seems we're not missing diversity anymore
08:48 aloha tons of new people at yapc. impressed by the young people and women: it seems we're not missing diversity anymore was last seen in  15577 days 8 hours ago .
08:50 diakopter well, that is an extremely tiny sample
08:51 masak we may have some diversity, but I'm pretty sure we need more of it ;)
08:51 tadzik good morning #perl6
08:54 betterworld joined #perl6
08:56 masak thou: the thing you need to understand about prefix:<^> and other prefix operators is that they're looser than terms. for some reasons, we don't always expect that with things like ^9.roll(30). but that's the way it falls out.
09:00 moritz \o
09:06 immortal joined #perl6
09:06 immortal joined #perl6
09:15 betterworld joined #perl6
09:31 FROGGS joined #perl6
09:32 masak from http://repository.readscheme.org/f​tp/papers/ai-lab-pubs/AIM-349.pdf : 'Note well that we did _not_ bring forth a clean implementation in one brilliant flash of understanding; we used an experimental and highly empirical approach to bootstrap our knowledge.'
09:33 masak this is from the first of the "Lambda Papers". a good read if one (like me) is interested in how closures work and the early discoveries about them.
09:36 sftp joined #perl6
09:46 erkan joined #perl6
09:46 erkan joined #perl6
09:49 erkan joined #perl6
09:49 erkan joined #perl6
09:54 felher masak++ # for the link. I sure want to read this if i happend to have some free time next week
09:58 erkan joined #perl6
09:58 erkan joined #perl6
10:23 erkan joined #perl6
10:23 erkan joined #perl6
10:29 moritz is Tim Smith here on IRC?
10:31 moritz seems not; I've commented on the commit instead
10:41 spider-mario joined #perl6
10:49 whiteknight joined #perl6
10:55 tadzik .u »
10:55 phenny U+00BB RIGHT-POINTING DOUBLE ANGLE QUOTATION MARK (»)
10:55 tadzik does anyone have an xkbmap with this?
10:56 tadzik I used to have that on alt+[], but I seem to have lost that one
10:56 ObseLeTe joined #perl6
11:02 erkan joined #perl6
11:02 erkan joined #perl6
11:05 erkan joined #perl6
11:15 erkan joined #perl6
11:15 erkan joined #perl6
11:19 desertm4x joined #perl6
11:21 crab2313 joined #perl6
11:29 am0c joined #perl6
11:31 crab2313 r: say 1+3 ?? $_ !! 123;
11:31 p6eval rakudo 5ceefd: OUTPUT«Any()␤»
11:31 crab2313 n: say 1+3 ?? $_ !! 123;
11:32 p6eval niecza v19-37-g0567c61: OUTPUT«Any()␤»
11:37 wk_ joined #perl6
11:42 birdwindupbird joined #perl6
12:02 cognominal joined #perl6
12:08 telex joined #perl6
12:09 JimmyZ joined #perl6
12:24 tadzik re woman at yapc, that was at least one good reason to work at booking :)
12:24 tadzik there seemed to be many of them working there
12:35 leont Try Cluj.pm. At their meetup two weeks ago there were about as many women as men, and they're programmers not recruiters. They're doing something right there.
12:35 tadzik indeed
12:37 jtpalmer joined #perl6
12:41 desertm4x left #perl6
12:41 GlitchMr joined #perl6
12:42 marmay joined #perl6
13:05 FROGGS joined #perl6
13:06 FROGGS hi there
13:06 tadzik hi FROGGS
13:09 awwaiid tadzik, I have » as compose+>+>
13:09 phenny joined #perl6
13:09 tadzik awwaiid: yeah, I prefer to having «» close
13:09 tadzik I put them on altgr+[]
13:09 awwaiid ah
13:09 tadzik like “”, ‘’ and so
13:10 mst I ... suspect I would need a new editor for unicode.
13:10 awwaiid I know «» can be used in perl6, what are “”, ‘’ for?
13:10 rindolf joined #perl6
13:10 awwaiid er... s/can/are/
13:11 awwaiid you knowwhatimean
13:11 rindolf Hi all.
13:11 awwaiid hi rindolf
13:13 tadzik awwaiid: nothing. It's just that I don't always quote, but when I do, I do it with proper characters
13:13 tadzik also, Polish for example uses „” rather than “”
13:13 tadzik and ,,'' looks stupid
13:13 awwaiid ah
13:16 am0c_ joined #perl6
13:33 skids joined #perl6
13:45 moritz what are modules that (possibly indirectly) use sockets and have tests?
14:23 flexibeast joined #perl6
14:35 __sri is there a reason .spurt and .spew have different names in rakudo and nqp?
14:37 tokuhiro_ joined #perl6
14:38 mst __sri: nqp is a baby perl6 so it doesn't have the post-puberty capacity to spurt yet and can only spew?
14:38 * mst hides
14:38 __sri :D
14:41 skids Is the ".clear" method mentioned in http://perlcabal.org/syn/S06.html#Feed_operators for feed taps still likely to be implemented, and if so, when and how was it to be called?  Each time the feed tap occurs in a new feed?
14:54 JimmyZ hmm, I can't bootstrap panda
14:55 JimmyZ https://gist.github.com/3466747
15:08 JimmyZ looks like 'mkdir' and 'open' doesn't support no-ascii GBK strings.
15:32 thou joined #perl6
15:39 JimmyZ joined #perl6
15:39 leont Non-ascii on Windows is a bit of a bitch
15:40 moritz non-ascii on windows is also a bit of a bit :-)
15:40 moritz (but less so, I think)
15:40 thou moritz: hi, you asked about Tim Smith. it is i
15:40 leont Not sure what parrot uses inside, if they use the non-wide interfaces, like perl 5 does, it is indeed problematic
15:43 tokuhiro_ joined #perl6
16:07 _jaldhar joined #perl6
16:09 hello_wol joined #perl6
16:13 masak good morning, #perl6
16:18 * moritz too late
16:18 moritz good morning masak
16:22 masak so. let's merge macros-d2, shall we? :)
16:23 moritz does it work? :-)
16:23 masak I have one known/outstanding bug, but I don't consider it a blocker.
16:23 masak the rest works, to my knowledge. I'd rather have what I have in maste^Wnom.
16:23 masak I'll be happy to write a todo test for the known bug.
16:25 moritz so, go ahead and merge :-)
16:25 * masak .oO( Tim Smith -- is it thou? )
16:25 * masak .oO( yes, it is I )
16:25 jaldhar_ joined #perl6
16:26 masak well, when I say "merge", I really mean something more like "rebase, write spectests, and then merge" ;)
16:26 masak but I have some time to do that tonight, so that's what I'll do.
16:26 moritz why rebase instead of merging?
16:27 ObseLeTe joined #perl6
16:27 moritz if you don't expect any fallout from the merge, you can merge locally, spectest and then push
16:28 masak hm, I meant rebase interactively, to make the commits make me appear less stupid ;)
16:28 moritz don't worry about that
16:28 moritz we pretty much know stupid (or not) you are :-)
16:28 japhb Developers (me included) have an irrational fear of looking stupid.  Don't bother with that, you've already proven yourself here.  :-)
16:29 masak *sigh* -- I must've phrased myself badly.
16:29 masak :)
16:29 masak I actually have very little fear of appearing stupid. if I did, I would be much less chatty on this channel.
16:30 japhb Long ago when I worked at a place that had Quake LAN games, my nick was "Dumbass" so no one could claim I was being uncharacteristically stupid. ;-)
16:30 masak the last five years have been one big journey of discovery for me. I've learned loads, much thanks to daring to look stupid when that was the easiest way to learn things.
16:31 masak the "make the commits make me appear less stupid" is more of a meme than anything else. I'm actually not afraid of leaving commits behind that make me look stupid.
16:31 masak but I do like leaving commits behind that are nicely compartmentalized and "do one thing well".
16:32 japhb Ah, that is a worthy goal.
16:32 masak actually, I don't seem to need an interactive rebase this time. they look fairly nice as it is.
16:33 japhb \o/
16:33 masak there's a bit of "no, let's do it like this" in there, but in a sense, that's good historical documentation too.
16:33 japhb True
16:33 masak for D1, I rebased too much, I think, making the work look like one big commit :)
16:35 japhb Some of that "historical documentation" should probably become condensed into code comments or Pod.  It's nice to read a current rev, wonder why one of a few possible designs was chosen, and see a nice comment that says "I couldn't do A because foo, or B because bar, so I went with C, which seems to be working"
16:38 masak on a scale between "leave informative comments" and "leave informative commits", my preference seems to fall more towards the former than most programmers'
16:42 japhb Works for me ... going all the way to the latter seems to only work when most of the team does it, or the repo is curated (in the way that the Linux kernel is)
16:45 masak agree.
16:56 masak moritz: as to the more general way to interpret the question "why rebase instead of merging?": I vastly prefer commits in a straight line to branches and merge commits.
16:58 leont You can do both
16:59 leont That way you can have merge commits (useful) without having non-linear history (annoying)
16:59 moritz masak: if there are commits that break features, I much prefer a merge commit
16:59 masak leont: I don't understand how that can be.
17:00 leont You first rebase your branch on master, then merge it into master using --no-ff
17:00 masak leont: ah. nice.
17:00 masak moritz: you mean breakages that are discovered later? why are merge commits nicer in that case?
17:01 leont Trick I learned on p5p, actually. Think it was demerphq's.
17:01 moritz masak: no, I mean if a branch breaks the build, or some features that previously worked
17:06 MayDaniel joined #perl6
17:07 masak why would I want to merge work from a branch that breaks the build? I don't understand the use case.
17:09 moritz well, eventually you fix it in the branch
17:09 moritz adn then you merge the branch
17:09 moritz instead of rebasing all the commits on top of master/nom
17:10 moritz so that from the perspective of the master branch, there is one commit (a merge commit) that didn't break anything
17:10 masak ok, now I understand your scenario.
17:11 masak why is a merge commit better there?
17:11 moritz because a git bisect won't stumbler over the commits that broke the build, for example
17:11 moritz *stumble
17:12 moritz and it's easier to see which commits were associated with a certain feature
17:17 masak seems to me you're exclusively talking about the use of git-rebase --interactive that squashes commits.
17:18 masak I kinda like the way the topology is flattened. I'm not necessarily advocating squashing commits in the process.
17:19 rindolf joined #perl6
17:23 moritz masak: well, squashing commits in one thing, but I wouldn't squash a whole non-trivial branch
17:25 mst right, for that purpose I prefer to rebase onto master
17:25 mst and then force git to produce a merge commit anyway
17:25 mst that tends to look really nice in the relevant browsers etc.
17:26 moritz what's the benefit over straight merging?
17:34 masak the usual benefit of rebasing.
17:35 masak you get conflicts in commit order, not all-at-once.
17:55 mst straight merging is a crutch for people who don't have git.
17:55 moritz mst: that's not an answer to my question -)
17:55 mst it's one of those "try it with the rebasing for a bit, then go back to without, then scream, then understand" type things
17:56 mst if you don't immediately see the value in having a clean commit history, you're probably not going to be swayed by any argument I can think of, so TIAS is probably an easier answer
17:57 moritz mst: I do value a clean commit history, I just don't see why a rebase + merge is supposed to (significantly) cleaner
17:57 * [Coke] waves from the airport.
17:57 moritz if many of the commits have conflicts, then I can see the value. If not, not
17:58 mst so ... you're saying "there are two cases - one where the rebase is superior, and one where the rebase is neutral"
17:58 mst at this point, I would argue that the argument for defaulting to doing the rebase has just been made :)
17:58 cjbot joined #perl6
17:59 moritz ok, I agree now :-)
17:59 mst oh, the other thing is it means that after the rebase, I can do a straight 'git diff master' to get a clean diff set to review
18:00 * [Coke] is very happy he has a chance to git pull everythig before the flight
18:00 mst which during the reviewathon I just did for dbix-class was really quite handy
18:01 masak mst++ # swaying moritz ;)
18:04 [Coke] cjbot:[D help
18:04 [Coke] cjbot: help
18:04 cjbot Run by Coke, I relay tweets about Perl 6.
18:05 [Coke] the sleeps are probably the wrong way to stop us from killing the CPU long term.
18:05 [Coke] glad to see he hasn't fallen asleep, though.
18:06 moritz [Coke]: yes, it's not a long term solution, just a workaround
18:07 mst well, I'll be doing a bit of async I/O stuff
18:07 mst I guess you can bolt a timer API on the top in the process of rewriting my code to not be completely horrible
18:07 mst (since I'm sure it'll require that)
18:23 cjbot joined #perl6
18:27 sorear good * #perl6.
18:27 diakopter wb
18:27 rindolf sorear: hi.
18:28 sorear I am safely home, pmichaud, masak, anyone else who wanted to kno0w
18:28 moritz glad to hear that
18:29 rindolf sorear: that's great.
18:30 sorear hello rindolf.  Are you just saying hi or can I help you?
18:30 rindolf sorear: no, no help needed in particular.
18:30 rindolf sorear: just saying hi.
18:34 masak sorear! \o/
18:34 masak glad to hear you've arrived safely.
18:36 GlitchMr perl6: print ucfirst 'the world'
18:36 p6eval rakudo 5ceefd, niecza v19-37-g0567c61: OUTPUT«The world»
18:36 GlitchMr But... Perl 6 doesn't have ucfirst... does it?
18:36 masak GlitchMr: correct.
18:36 moritz GlitchMr: it's deprecated in rakudo, but it won't be replaced until we have tc as a repalcement
18:37 moritz aloha: don't fall asleep
18:37 masak GlitchMr: if you're worried about the name being taken when you wanted such a sub in your program, the good news is that your sub will lexically shadow the setting's :)
18:38 moritz nr: say Nil
18:38 p6eval rakudo 5ceefd, niecza v19-37-g0567c61: OUTPUT«Nil␤»
18:38 GlitchMr Sometimes I would like to replace log with something more... useless
18:38 moritz why doesn't that say Nil() like other type objects would?
18:39 moritz GlitchMr: and you're free to do that
18:39 GlitchMr But that would probably be probably too confusing
18:39 moritz isn't Perl 6 wonderful?
18:39 GlitchMr Using log() for logging instead of logarithm
18:40 GlitchMr JavaScript does that. But in JavaScript we have Math.log() and console.log()
18:40 GlitchMr (also, lack of consistency on first letter ;))
18:42 moritz nr: say Nil.new === Nil
18:42 p6eval niecza v19-37-g0567c61: OUTPUT«Unhandled exception: No value for parameter '$l' in 'infix:<===>'␤  at /home/p6eval/niecza/lib/CORE.setting line 0 (infix:<===> @ 1) ␤  at /tmp/0UPX3nNvy8 line 1 (mainline @ 4) ␤  at /home/p6eval/niecza/lib/CORE.setting line 4137 (ANON @ 3) ␤  at /home/p6e…
18:42 p6eval ..rakudo 5ceefd: OUTPUT«True␤»
18:42 GlitchMr Another Niecza bug?
18:42 moritz "No value for parameter"?
18:43 GlitchMr Isn't === comparable to "is" (mutable types) and "==" (immutable types) in Python?
18:45 sorear dunno, I don't know python
18:45 sivoais joined #perl6
18:46 sorear moritz: niecza's signature binder treats Nil as an omitted argument
18:48 sorear although this seems to not (longer?) be by spec
18:48 moritz I'm pretty sure you should be able to === to Nil
18:51 sorear I am not so sure where I stand on this particular consistency/DWIM tradeoff
18:52 sorear since we've established that === autothreads, it's not a low-level binding
18:53 moritz if the OKNess proposal goes anywhere, that will probably change anyway
18:54 * TimToady_ is happy to report that our living area in Belgium has a decent internet connection
18:54 diakopter ahoy
18:54 TimToady joined #perl6
18:56 TimToady also that viv can now parse nearly 800 lines of viv
18:56 phenny TimToady: 23 Aug 13:12Z <sorear> tell TimToady see colomon 06:12 and 05:13 PDT, I think we are looking at a stdbug
18:56 sorear TimToady: awesome!
18:57 leont viv == perl5 on perl6?
18:57 TimToady yep
18:57 moritz erm, isn't that the other way round?
18:57 moritz it's the p6-to-p5 compiler that runs STD.pm6
18:57 TimToady well, it's a perl5 parser written in perl6
18:58 TimToady anyway, got heredocs working earlier today, at least for the '' and "" forms
18:58 ingy an someone give me a url for a grammar of perl6?
18:58 ingy c
18:58 sorear the way STD is currently set up, there is 6000 lines of Perl 6 which implements a grammar (std/STD.pm6), and 4000 lines of Perl 5 which implements the Perl 5 emitter (std/viv)
18:59 sorear github.com/perl6/std, STD.pm6
18:59 sorear https://github.com/perl6/std/blob/master/STD.pm6
19:00 TimToady well, viv itself is 4300 lines, so I don't know where you're gonna find the -300 lines :P
19:00 sorear TimToady recently restarted a Perl 6 gramar for Perl 5 (std/STD_P5.pm6)
19:01 sorear I think "viv can parse 400 lines of viv" means "std/viv+std/STD.pm6+std/STD_P5.pm6 can parse 400 lines of viv"
19:01 sorear STD_P5 is not independent of STD
19:01 TimToady correct
19:05 * masak had not realized this
19:05 TimToady it uses the same EXPR opp parser, for instance
19:08 eiro hello all
19:08 masak heiro!
19:08 masak TimToady: oh, that makes sense.
19:09 TimToady tomorrow we'll be going to see The Pageant of the Golden Tree in Bruges
19:10 moritz aloha has stopped reporting commits
19:10 leont Bruges is fairly high on my to-visit list, I it rather beautiful
19:10 eiro just read about viv: exciting project!
19:11 * masak .oO( vive la viv! )
19:11 * moritz found http://www.imdb.com/title/tt0780536/ quite amusing (re Bruges :-)
19:13 sorear hello eiro.
19:14 leont moritz: watched that movie on my first date with $GF, we're romantic like that :-p
19:15 * eiro .oO( viva! )
19:19 sorear so, I'm listed in the credits for Rakudo 2012.08
19:19 sorear why?
19:20 __sri viva la viv!
19:20 moritz sorear: somebody ++'ed you in a commit message, I think
19:21 masak moritz++ did :)
19:21 masak e5e9e79e1212c3c5bc26d2524bed3e5ebbc2639c
19:21 masak https://github.com/rakudo/rakudo/commit/​e5e9e79e1212c3c5bc26d2524bed3e5ebbc2639c
19:22 moritz sorear: you diagnosed the proble with Num.new not being well-suited for subclassing
19:32 GlitchMr Is there a difference between &log and \&log?
19:33 moritz nr: say (\&log).WHAT
19:33 p6eval rakudo 5ceefd, niecza v19-37-g0567c61: OUTPUT«Capture()␤»
19:33 moritz nr: say (&log).WHAT
19:33 p6eval niecza v19-37-g0567c61: OUTPUT«Code()␤»
19:33 p6eval ..rakudo 5ceefd: OUTPUT«Sub()␤»
19:34 GlitchMr oh, I see...
19:34 GlitchMr \&log simply doesn't work
19:35 masak indeed.
19:35 GlitchMr Without any warnings... what is Capture anyways?
19:35 masak a Capture is the thing you get from doing \
19:35 masak it captures containers rather than values.
19:36 masak containers, also known as "value cells".
19:37 GlitchMr my $arr = (3; 4)
19:38 GlitchMr Is it modern comma operator?
19:39 moritz GlitchMr: http://doc.perl6.org/op/infix/:
19:42 masak moritz: er, that's a colon.
19:42 moritz oh, I misread
19:45 sorear oh, something I need to do now that I have network again
19:45 moritz sorear: fix aloha? :-)
19:45 sorear don't we use dalek here?
19:45 moritz erm, right
19:46 moritz fix dalek, then :-)
19:46 sorear why should I care if aloha is broken
19:46 moritz never mind aloha
19:46 sorear well, https://github.com/sorear/niecza​/compare/0567c617ab...55e895a832
19:47 masak oh, that's why my commits don't show up on the channel... :)
19:47 sorear bailador runs unmodified + niecza no longer depends on write access to the install directory + parrallel make debreakage
19:47 moritz masak: did you merge?
19:47 sorear the berlin layover was amazingly productive
19:48 tadzik wow, nice!
19:48 tadzik sorear++
19:48 masak moritz: no, not yet. I'm just about to rebase onto nom locally. spectesting before and after.
19:48 tadzik I wonder how many more than 20req/s can you squeeze there :)
19:59 eiro no __END__ and __DATA__ in perl6 ?
19:59 masak eiro: not with that syntax, no.
20:00 masak eiro: http://perlcabal.org/syn/S26.html#The+%3dEND+block
20:00 masak huh, I thought we renamed that one =finish
20:00 masak is S26 on the web out-of-date?
20:03 moritz might be
20:03 moritz if it's just --doc=HTML output, I can easily refresh it
20:04 eiro thanks :)
20:04 masak rebase successful. rebuilding and respectesting.
20:10 moritz tadzik: I meant pod-gist to be a replacement for Pod::Block.gist, not .Str :-)
20:10 moritz Pod::To::HTML doesn't seem to like the change :/
20:13 sorear (o/ everyone)
20:14 moritz and t/spec/S26-documentation/09-configuration.t doesn't like it either :(
20:14 masak post-rebase spectest run succeeded. pushing.
20:14 colomon sorear++
20:15 masak r: say "what SHA is p6eval on now?"
20:15 p6eval rakudo 5ceefd: OUTPUT«what SHA is p6eval on now?␤»
20:15 masak it'll have b7e189 in a little while :)
20:17 * masak is excited at this
20:17 masak I haven't added tests yet, but I will during the weekend.
20:18 moritz masak: does that mean I should wait with pushing my local patch until you're done?
20:18 masak no, it means I'm done.
20:18 moritz (I just want to add a test file)
20:18 masak I've already pushed.
20:18 moritz \o/
20:18 masak r: say "how about now?"
20:18 p6eval rakudo 5ceefd: OUTPUT«how about now?␤»
20:19 * masak is lazy and doesn't look at the cron table
20:20 * moritz triggers a rebuild manually
20:22 masak moritz++
20:22 moritz oops, https://github.com/rakudo/rakudo/commit/​1965724f02e0b9afd0fa52e40648ab316dbc1ddd contained stuff I wanted to put into a separate commit
20:22 moritz moritz-- # still using git commit -a
20:23 masak whoa. I never used -a by virtue of it being obviously a bad idea from the start.
20:23 masak I *want* to add files (or parts of them) manually.
20:23 masak (in other words, I like the staging area)
20:27 moritz r: say 42
20:27 p6eval rakudo 196572: OUTPUT«42␤»
20:27 moritz that's the new rakudo
20:27 jnthn evening o/
20:27 moritz good UGT morning, jnthn
20:28 jnthn no...it's so not morning :P
20:29 moritz in UGT it's always morning when you join a conversation :-)
20:30 masak jnthn: good morning!
20:31 masak r: macro demo($ast) { quasi { "This is the " ~ {{{$ast}}} ~ " unquote splicing on the channel" } }; say demo "first"
20:31 p6eval rakudo 196572: OUTPUT«===SORRY!===␤Method 'position' not found for invocant of class 'QAST::Unquote'␤»
20:31 masak :/
20:32 masak r: macro demo($ast) { quasi { {{{$ast}}} } }; say demo "first"
20:32 p6eval rakudo 196572: OUTPUT«===SORRY!===␤Method 'position' not found for invocant of class 'QAST::Unquote'␤»
20:32 masak seems to be a systematic error. investigating.
20:32 masak oh!
20:32 masak hm, I have such a method here...
20:33 jnthn NQP revision bump fail?
20:33 masak possibly.
20:35 moritz 2012.08-2-gc223607 added it
20:35 masak yes, looks kosher.
20:35 moritz and it works locally
20:35 masak I have that as the current NQP_REVISION
20:36 masak moritz: manual rebuild didn't update nqp?
20:36 moritz masak: I have a theory
20:36 moritz (it should, because I ran the same command as cron does)
20:37 moritz maybe it did a 'git pull' during an existing rebuild run
20:37 masak :/
20:37 moritz $ p6eval@feather:~/nom/nqp$ ./nqp --version
20:37 moritz This is nqp version 2012.08-4-g8fac254 built on parrot 4.4.0 revision RELEASE_4_4_0
20:38 moritz that is very odd
20:39 * moritz cleans out and tries again
20:39 sorear jnthn!
20:40 moritz in particular it's odd that it has a newer NQP revision than was ever in tools/buil/NQP_REVISION
20:41 moritz and now NQP rebuild segfaulted
20:41 moritz wtf.
20:42 masak o.O
20:49 moritz r: macro demo($ast) { quasi { {{{$ast}}} } }; say demo "first"
20:49 p6eval rakudo 196572: OUTPUT«===SORRY!===␤Method 'position' not found for invocant of class 'QAST::Unquote'␤»
20:50 * masak .oO( never do live demo! )
21:09 moritz r: macro demo($ast) { quasi { {{{$ast}}} } }; say demo "first"
21:09 p6eval rakudo 196572: OUTPUT«first␤»
21:10 masak {{{\o/}}}
21:10 masak r: macro demo($ast) { quasi { "This is the " ~ {{{$ast}}} ~ " unquote splicing on the channel" } }; say demo "second"
21:10 p6eval rakudo 196572: OUTPUT«This is the second unquote splicing on the channel␤»
21:11 moritz r: macro call($a) { quasi { {{{$a}}}() } }; call { say "OH HAI" }
21:11 p6eval rakudo 196572: OUTPUT«OH HAI␤»
21:13 masak r: macro call($a) { my $x = "how very"; quasi { say $x; {{{$a}}}() } }; my $x = "lexical"; call { say $x }
21:13 p6eval rakudo 196572: OUTPUT«how very␤how very␤»
21:13 masak oh!
21:13 masak bug.
21:13 moritz they are not hygienic?
21:13 masak this would happen if the unquotes don't get their context set properly, I guess.
21:13 masak they should.
21:14 masak they're meant to be hygienic.
21:14 * moritz could say lots of things about writing tests early in the process, but decides not to
21:15 moritz instead, I'll gravitate towards the bed
21:15 moritz 'night &
21:17 masak 'night :)
21:17 masak I usually tend towards writing tests early in the process, too.
21:18 masak anyway, this is an excellent test to put in a macros-d2.t file. moritz++
21:18 masak hm, if the above happens, then that means...
21:18 masak r: macro call($a) { quasi { {{{$a}}}() } }; my $x = "lexical"; call { say $x }
21:18 p6eval rakudo 196572: OUTPUT«Any()␤»
21:18 masak yeah.
21:23 bubaflub joined #perl6
21:38 Moukeddar joined #perl6
21:42 * [Coke] spends a flight trying to convert partcl to partcl-nqp, is stymied, and then remembers that pmichaud already did most of this work.
21:42 [Coke] sorear: how was the trip?
21:44 [Coke] UGT: universal grumpy time.
21:45 * [Coke] notes that "class foo is bar; <class def>" fails horribly in nqp, but only if you target==pir.
21:48 cjbot joined #perl6
21:49 silug joined #perl6
21:49 [Coke] is it still necessary to "use NQPP6Regein nqp?
21:49 [Coke] is it still necessary to "use NQPP6Regex" in nqp?
21:50 [Coke] (I hope not, since it's failing. ;)
21:53 [Coke] can someone help me with a nqp-rx -> nqp regex issue?
21:54 diakopter probably not, but I'm curious what the problem is
21:54 diakopter (probably not *I)
21:54 [Coke] https://gist.github.com/3471303
21:55 diakopter I think ** went away
21:55 * [Coke] remembers to checkout a copy of the specs this time..
21:56 diakopter try % instead?
21:57 [Coke] see, you were helpful. /(
21:58 [Coke] ;)
21:58 masak diakopter: ** went away for separators. not for numbered/range quantifiers.
21:58 diakopter ah.
21:58 diakopter (good thing this case was a separator)
21:59 [Coke] if I change it to %, it now dies with:
21:59 [Coke] Unable to parse blockoid, couldn't find final '}' at line 7, near "token meta"
22:00 diakopter :S
22:02 [Coke] updated gist
22:02 diakopter maybe it was %%
22:02 diakopter (sorry!)
22:02 * diakopter no clue
22:04 [Coke] that would impact what the regex did, but not stop it from compiling
22:04 sorear [Coke]: trip was great
22:09 cjbot joined #perl6
22:11 dalek joined #perl6
22:11 sorear 1. ssh into sorear@feather3.perl6.nl
22:11 sorear 2. sudo -u p6eval tmux attach
22:12 sorear 3. notice that the dalek frontend is NOT RUNNING
22:12 [Coke] glad to hear it. I am travelling for work to a programming conference, but it's to learn about "agile". :P
22:12 sorear 4. attempt to fix
22:12 * [Coke] stares at regexes.
22:13 cognominal joined #perl6
22:16 sorear o/ cognominal
22:16 [Coke] so, ** -> % or %% seems to be the right answer, but doing a simple dropin replacement doesn't work.
22:20 dalek niecza: 3826c2a | sorear++ | Makefile:
22:20 dalek niecza: Add an extra dependency to make parallel make happy
22:20 dalek niecza: review: https://github.com/sorear/niecza/commit/3826c2af21
22:20 dalek niecza: 9fad9f7 | sorear++ | / (5 files):
22:20 dalek niecza: First attempt to use a per-user directory for the module cache
22:20 dalek niecza:
22:20 dalek niecza: Data is stored in Local Application Data\NieczaModuleCache, which Mono will
22:20 dalek niecza: map to $HOME/.local/share/NieczaModuleCache.  The --obj-dir object actually
22:20 dalek niecza: does something now (how long has it been broken?) and can be used to isolate
22:20 dalek niecza: multiple instances of Niecza, and is used internally to make bootstrapping
22:20 dalek niecza: work.
22:20 dalek niecza:
22:20 dalek niecza: One known wart is that $*PERL<version> may become out of date because it is
22:20 dalek niecza: only updated when the setting needs to be recompiled, and the build system
22:20 dalek niecza: no longer always forces a recompilation of the user-wide setting.
22:20 dalek niecza: review: https://github.com/sorear/niecza/commit/9fad9f7d24
22:20 dalek niecza: 55e895a | sorear++ | Makefile:
22:20 dalek niecza: Update mkpackage for new run/Run.Kernel.dll file setup
22:20 dalek niecza: review: https://github.com/sorear/niecza/commit/55e895a832
22:20 sorear whee, dalek is working
22:20 sorear [Coke]: ** -> +%
22:21 masak [Coke]: what sorear said. % isn't a quantifier, it's a quantifier *modifier*.
22:21 masak 'night, #perl6
22:26 sorear bye masak
22:26 popl joined #perl6
22:28 tadzik oops, so I broke rakudo :(
22:29 sorear o/ tadzik
22:29 tadzik sorear: o/
22:43 jeffreykegler joined #perl6
22:45 cognominal \o: sorear
23:15 lateau joined #perl6

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

Perl 6 | Reference Documentation | Rakudo | Niecza | Specs