»ö« Welcome to Perl 6! | perl6.org/ | evalbot usage: 'perl6: say 3;' or rakudo:, niecza:, std:, or /msg camelia perl6: ... | irclog: irc.perl6.org | UTF-8 is our friend!
Set by sorear on 25 June 2013.
Mouq github.com/perl6/std/commit/fadaff1e 04:32
github.com/rakudo/rakudo/commit/1b6cbf6 04:50
std: s||| <-- STDBug 04:55
camelia std 082981d: OUTPUT«===SORRY!===␤Unrecognized regex metacharacter < (must be quoted to match literally) at /tmp/RHgx0p84GT line 1:␤------> s||| <⏏-- STDBug␤Can't call method "from" on unblessed reference at STD.pm line 54407.␤FAILED 00:…»
Mouq std: s&&& #? 04:58
camelia std 082981d: OUTPUT«===SORRY!===␤Regex not terminated at /tmp/E66xfkDwP7 line 1 (EOF):␤------> s&&& #?⏏<EOL>␤Parse failed␤FAILED 00:01 120m␤»
TimToady std: s|x|| 05:01
camelia std 082981d: OUTPUT«ok 00:01 122m␤»
TimToady it's the initial || or && thing
I guess the code is assuming you're smart enough not to expect a null regex to be meaningful there :) 05:03
so it's one of those bugs that well-meaning people will never see...
std: s|||a||b||c|| 05:04
camelia std 082981d: OUTPUT«===SORRY!===␤Unsupported use of suffix regex modifiers; in Perl 6 please use prefix adverbs at /tmp/OSuxjd9E8l line 1:␤------> s|||a||b⏏||c||␤Parse failed␤FAILED 00:01 121m␤»
TimToady heh
now that's probably a realer bug
or not, depending on how you think of it 05:05
it's definitely close to the DIHWIDT category
Mouq has no idea what s|||a||b||c|| is supposed to be
moritz \o
TimToady std: s/||a||b||c//
camelia std 082981d: OUTPUT«ok 00:01 122m␤» 05:06
moritz Mouq: s/// can have custom delimiter characters
TimToady but the parser does check for terminator before || where normal ops can be, just not the leading one, which is special-cased
moritz Mouq: seems it was an attempt to use || as delimiter
TimToady well, ||| since it was a subst :)
Mouq Ah. I wasn't reading the last || as |{''}| 05:08
TimToady the line in question is in nibbler, [ <.normspace>? < || | && & > ]?
that allows a null first case dis or conjunction 05:09
Mouq [ <.normspace>? <!stopper> < || | && & > ]?
TimToady it's eating the || and then expecting more cases
could fix that, if it really needs fixing :)
can't hurt, I suppose 05:10
probably closer to what the user was expecting, so I guess it's good
Mouq is just looking at this stuff to try to get s&a&b& etc. working on the stdsigspace branch
Mouq is almost up to spectest 05:11
TimToady Mouq++ # persistence 05:12
Mouq The big thing I don't know how to fix is that sigspace calls <.ws>, but not the locally defined <.ws> :( It appears to have done so before, which is weird since I haven't changed the ast generated much 05:13
TimToady why is ws locally defined? the std grammar doesn't redefine it... 05:15
if something more specialized eats the ws first, then any ws called from EXPR is a no-op 05:16
Mouq rn: grammar A { token ws { x }; rule foo {foo bar} }; 'fooxbar' ~~ /<A::foo>/ 05:17
camelia ( no output )
..niecza v24-98-g473bd20: OUTPUT«Unhandled exception: Cannot dispatch to a method on A because it is not inherited or done by Cursor␤ at /home/p6eval/niecza/lib/CORE.setting line 288 (Mu.dispatch:<::> @ 4) ␤ at /tmp/YKc8wUz4OW line 1 (ANON @ 4) ␤ at /home/p6eval/niecza/lib/CO…»
lue
.oO(When is s# # # going to work? I don't think I can go on without it.)
Mouq r: grammar A { token ws { x }; rule foo {foo bar} }; say 'fooxbar' ~~ /<A::foo>/
camelia rakudo 2ce544: OUTPUT«｢fooxbar｣␤ A::foo => ｢fooxbar｣␤␤»
TimToady std: s♯foo♯bar♯ 05:18
camelia std 082981d: OUTPUT«ok 00:01 122m␤»
TimToady there you go
Mouq Oh, wait. It does work now. What do ya know? 05:19
*on stdsigspace
lue std: m☃canhaz☄ 05:29
camelia std 082981d: OUTPUT«===SORRY!===␤Unrecognized regex metacharacter ☄ (must be quoted to match literally) at /tmp/VW1r5W4rMT line 1:␤------> m☃canhaz☄⏏<EOL>␤Couldn't find terminator ☃ at /tmp/VW1r5W4rMT line 1 (EOF):␤------> …»
TimToady all you have to do is persuade the Unicode consortium to define those as a matched pair 05:31
lue
.oO(Stupid STD, ☃ is only part of Ps, not Pe)
TimToady you can't fight The (snow)Man 05:32
lue I'll find the more artistically-minded members of the consortium and send them a letter "On Symbolism of the disjunction between Earth and Space, expressed through officially-recognized character pairings." 05:33
TimToady fortunately or unfortunately, they really believe in the "uni" part of what they're doing :) 05:34
what you really need to do is persuade them to add characters like COMBINING LEFT BRACKETNESS and COMBINING RIGHT BRACKETNESS 05:35
lue COMBINING PS CONTROL ADVERB :) [hey, they already got the LTR/RTL indicators...] 05:37
"Because sometimes mathematicians do crazy stuff like [0,∞[ and it'd be nice to let my text editor know that the second [ is supposed to be Pe, not Ps." 05:38
TimToady might be a tough sell, though, considering how all the other char traits tend to stick to the base char, not the combiners...
so what you need is just a LEFT BRACKETNESS character with a COMBINING IS REALLY THE FOLLOWING CHARACTER 05:39
TimToady to the first approximation, mathematicians should be ignored rather than shot 05:40
lue Curse you TimToady, for I now wish to attempt the most high-brow letter for such a proposal. But tomorrow o/ 05:43
diakopter TimToady: in that case, the first approximation has about 1 significant digit and a thousand degrees of freedom 05:48
TimToady To whomever eventually stoops so low as to canvas these logs for more cute sayings of mine to inflict on the quote files of the world, please feel free to capitalize my sentences and put punctuation on the end without extra annotations like those silly brackets you always see. Thank you. 05:49
Mouq n: my \term:<❄> = "☃☃".ords;say ->$a,$b{->$snow-flakes{eval "('$a'.comb$snow-flakes'$b'.comb)$snow-flakes❄"}(Q ☃ **☃ )+$a.comb*$b.comb*1000+❄}(|<sno w!!>).chr 05:52
camelia niecza v24-98-g473bd20: OUTPUT«☃␤»
Mouq n: my \term:<❄> = "☃☃".ords;say ->$left-hand,$right-hand{->$snow-flakes{eval "('$left-hand'.comb$snow-flakes'$right-hand'.comb)$snow-flakes❄"}(Q ☃ **☃ )+$left-hand.comb*$right-hand.comb*1000+❄}(|<sno w!!>).chr 05:54
camelia niecza v24-98-g473bd20: OUTPUT«☃␤»
Mouq ^^ Slightly more poetic
Mouq n: my \term:<❄> = "☃☃".ords; my \term:<❆> = ❄+❄+so True; say ->$left-hand,$right-hand{->$snow-flakes{eval "('$left-hand'.comb$snow-flakes'$right-hand'.comb)$snow-flakes❄"}(Q ☃ **☃ )+$left-hand.comb*$right-hand.comb*❆*❆*❄*❆*❄*❄+❄}(|<sno w!!>).chr 05:58
camelia niecza v24-98-g473bd20: OUTPUT«☃␤»
TimToady amazing what you can make with two hands and five fingers each 06:00
maybe you can use ✌ and ✋ 06:01
dalek p/stdsigspace: 3cfcbaa | Mouq++ | src/QRegex/P6Regex/ (2 files):
Fixes to work correctly on Rakudo
 06:13
p/stdsigspace: 9a1455c | (Tobias Leich)++ | src/vm/moar/ (2 files):
wire up and use nqp::backendconfig

This is just lacking a fresh stage0 to work.		 06:15
Mouq makes pull requests for nqp and rakudo 06:19
I also have updates to roast to get it up to spec that I'll push when (slash if) the PRs get accepted 06:25
Or should I push them now?
TimToady dunno, probably need to collect more opinions on that 06:26
.oO(push driven development)
long run, can't hurt, but maybe someone has a good reason to hold off short-term 06:27
otoh, if the new tests are closer to spec, who's to complain? 06:28
only fudgers, I guess 06:29
but hey, fudgers are people too
Mouq How do you throw unthrown exceptions in NQP? 06:35
moritz Mouq: you can push the roast changes in branches 06:44
Mouq: there are no NQP-level unthrown exceptions, afaict. a Perl 6 failure can be thrown simply calling .Str, for example 06:45
(if it's not handled yet)
moritz Mouq++ # stdsigspace 06:46
masak morning, #p6 06:49
moritz masak, masak
dalek ast/stdsigspace: 5e06d2c | Mouq++ | S05- (4 files):
Test for sigspace semantics correctly
 06:50
ast/stdsigspace: d20823e | Mouq++ | S05-mass/rx.t:
Fix faulty regex tests
masak moritz: www.masak-masak.blogspot.se/ 06:51
"Malaysian food adventures (and mishaps)"
moritz masak: :-) 06:52
masak seems "masak" is Malay for "cook" or "cooked" or "ripe".
moritz ripe, cooked masak! 06:53
dalek kudo/nom: e51cbdd | moritz++ | / (3 files):
[Configure] [JVM] move some generated files to gen/jvm/
 06:56
kudo/nom: 8b61933 | moritz++ | tools/build/Makefile-Parrot.in:
[Build] [parrot] generate some files in gen/parrot/
kudo/nom: 73344a6 | moritz++ | LHF.markdown:
remove LHF.markdown (out of date)

all the TODOs in there are already done, and nom has been the default branch for quite some time
hoelzro o/ #perl6 07:01
moritz \o hoelzro, * 07:02
moritz any objections to merging Mouq++'s stdsigspace branches? 07:17
Mouq moritz: me! 07:18
moritz Mouq: you object?
how very OO :-) 07:19
Mouq Heh. Give a minute to compile and I can fix the bug with not throwing null regex exceptions
moritz Mouq: ok. Don't hurry :-)
masak decides to play bad cop 07:20
Mouq: hurry! go go go go!
compile faster! 07:21
moritz masak: we should really try good cop/bad cop in real life at some point :-)
masak oh man. 07:22
yes.
FROGGS we should totally invent a board or card game where one is the null pmc access and the others need to catch (or rethrow) it
masak moritz: complete with throwing Mouq in an actual interrogation room and everything. :P
moritz masak: and the good cop handing out chocolate and the bad cop handing out super spicy chilli sweets or so :-) 07:23
masak djungelvrål :)
arnsholt Salmiak is fun for terrorizing non-Scandinavians =D 07:24
masak no, the bad cop goes "YOU THINK I'M GOING TO GIVE YOU CANDY, DO YOU? DO YOU!?"
FROGGS www.youtube.com/watch?v=OtV47HMjmY8
moritz masak: - "yes" - "AND YOU WOULD BE RIGHT! HERE YOU GO!"
FROGGS with a flashlight pointing to mouq 07:25
masak moritz: that's a rather agreeable bad cop ;) 07:26
arnsholt FROGGS: Have you seen the DailyWTF guy's review of Salmiak? Pretty hilarious =D
FROGGS arnsholt: umm, no :/ 07:27
moritz masak: I keep think of the Order of Chaos (HPMoR)
masak yeah, chaotic cop. 07:28
arnsholt FROGGS: www.salmiyuck.com/
Mouq r: my %objections = :Mouq<me>; %objections<Mouq>:delete; say +%objections 07:42
camelia rakudo 2ce544: OUTPUT«0␤»
Mouq moritz: If no one else objects, I'm okay with them being merged 07:45
moritz Mouq: I'll wait for jnthn to wake up before merging 07:45
Mouq Fine by me 07:46
Mouq bed &
moritz good night, and dream of correctly sig'ed spaces
masak not to mention correctly spaced sigs. 07:51
plane &
jnthn morning o/ 09:06
yoleaux 07:46Z <nwc10> jnthn: hide from moritz :-)
hoelzro ahoy jnthn
jnthn
.oO( how do you hide on IRC? :) )
 09:07
hoelzro /hide? 09:10
jnthn Was there not some unresolved issue with sigspace + nqp-m? 09:19
moritz jnthn: I hope that's resolved; the pull request contains a Moar bootstrap update 09:21
but I guess I should run tests with moar
jnthn I'm curious how the Moar bootstrap update was done 09:24
Given it has no make bootstrap-files target yet
Because we're not quite ready to commit to bytecode back-compat yet.
moritz but ins't that same problem with the existing bootstrap files? 09:28
erm, rephrase 09:29
don' the existing bootstrap files have the exact same problem?
jnthn No because for now we re-generate stage0 from the cross-compiler 09:32
So I'm curious if that was also updated
(And if so where the PR for it is.)
moritz jnthn: ah, so a possible approach would be to merge the code into MoarVM/nqp-cc too... and then, does it matter which compiler generated the bootstrap files for NQP? 09:37
jnthn moritz: Yeah, though maybe simpler for everyone is that I spend a couple of hours today looking over the bytecode stuff and figuring out what changes I'd like before we stabalize it. 09:40
moritz: I don't think there'll be much. 09:41
moritz jnthn: ok, then I'll simply wait for now
jnthn Yeah, I'll have time to do it in a little bit. Just got a couple of small $dayjob errands to take care of first. 09:45
moritz ok, no hurry at all
dalek p: db75765 | jnthn++ | / (19 files):
Add m-bootstrap-files target; close the loop.

We now work the same way with MoarVM's stage0 as for JVM and Parrot. Note this means that from here on, MoarVM must maintain bytecode back-compat at least long enough to keep stage0 building!		 12:01
jnthn .tell Mouq The bootstrap loop is closed on NQP on Moar now, and there's an m-bootstrap-files target. So it should be possible to do the sigspace updates for it also. :) 12:02
yoleaux jnthn: I'll pass your message to Mouq.
jnthn moritz: If you want to take the pull request and do that bit of the task, feel free also. :) 12:03
dalek rl6-roast-data: 9f2ba87 | coke++ | / (3 files):
today (automated commit)
 12:42
[Coke] niecza++ 12:44
jnthn [Coke]: Something needs updating due to recent build changes? 12:49
[Coke] also gets the .../rakudo.parrot/install/bin/parrot: not found error during a parrot only build.
jnthn: yes, working on it
jnthn k :) 12:50
dalek kudo/nom: 2fceb75 | coke++ | Configure.pl:
build - fix typo, throw entire error message
 12:54
jnthn ducks to avoid being hit by a flying entire error message 12:56
dalek kudo/nom: 493505c | coke++ | Configure.pl:
this is perl 5. Also, iterate correct hash.
moritz and here I thought die() took a list, just like print() 12:57
[Coke] here I thought I was editing perl 6. ;)
[Coke] ooh, you made --gen-nqp work with jvm? nice. 12:58
[Coke] moritz: is "make all" still going to work if I do a build with backend=jvm || backend=parrot ? 12:58
(will I still get a ./perl6 in both cases?)
moritz [Coke]: it totally should 13:00
[Coke] excellent. I'm too lazy to wait and find out now, so we'll see after lunchtime. ;)
dalek rl6-roast-data: 341cdae | coke++ | doit:
no more merging in pugs results by hand

Either it runs on this machine, or it doesn't.
rl6-roast-data: b8880ca | coke++ | bin/rakudo. (2 files):
track changes from new build setup.

  moritz++
[Coke] There. anyone gets sad when pugs goes to zero, I could use your help getting it running on the "new" test box. 13:01
hoelzro [Coke]: you're breathing new life into Pugs? 13:07
[Coke] hoelzro: oh hellz no 13:08
I'm just keeping it on life support.
but I'm tired of doing the pugs run on feather and manually merging in the results to the run on the new test box.
masak [Coke]++ # life support 13:09
[Coke] I did a bunch of work about... a year ago? to fudge tests for pugs so we'd get an idea of how much of it was still spec/working. (about 1/3)
I think we got one or two ods back into bugs as a result, but only reaaaaaaaaly low hanging fruit.
*mods
into *pugs
wow. 13:10
-> Dayjob
masak is there anything *other* than "give it a MOP" that could be done to Pugs that would be productive and interesting and -Ofun?
hoelzro ah, I see
[Coke] masak: not even that, if the last 2 years are any indication. 13:11
have fun, though. there's a perfectly good test suite to keep you honest in the meantime!
now really, dayjobbing.
masak :) 13:13
moritz masak: maybe bring the multi dispatcher to current spec? 13:14
masak ooh 13:18
colomon [Coke]: yeah, the (rakudo parrot based) smoke test completely failed overnight too. 13:29
dalek p/stdsigspace: db75765 | jnthn++ | / (19 files):
Add m-bootstrap-files target; close the loop.

We now work the same way with MoarVM's stage0 as for JVM and Parrot. Note this means that from here on, MoarVM must maintain bytecode back-compat at least long enough to keep stage0 building!		 13:32
p/stdsigspace: fa7ac95 | Mouq++ | / (19 files):
Merge branch 'master' of github.com/perl6/nqp into stdsigspace
p/stdsigspace: e96eed0 | Mouq++ | src/vm/moar/stage0/ (11 files):
Update MoarVM bootstrap
moritz tries again to merge and test the branches 13:34
LlamaRider o/ #perl6 13:51
Has any of you wonderful people successfully installed NativeCall for the JVM-based Rakudo?
diakopter hi LlamaRider 13:52
jnthn LlamaRider: NativeCall on Rakudo JVM is still a work in progress, mostly by arnsholt++. It supports basic calls so far, but is missing (last I checked) the array/struct/callback bits. 13:53
LlamaRider I see. Thanks jnthn! I guess it's back to Parrot for me then... only that Parrot stopped installing properly on Ubuntu 13.10 due to some missing Unicode support ... but hopefully it's just my local upgrade gone wrong somehow. 13:54
tadzik do you have libicu installed?
LlamaRider no and what's worse: E: Unable to locate package libicu 13:55
LlamaRider seems it got torn out of universe (or wherever it was) 13:55
moritz LlamaRider: libicu-dev? 13:56
diakopter LlamaRider: libicu-dev I think
masak LlamaRider! \o/ 13:57
LlamaRider libicu-dev is already the newest version. 13:58
well that's good news
diakopter LlamaRider: oh, the paths changed
moritz: it's the header location change
moritz LlamaRider: I created a symlink /usr/include/unicode to /usr/include/x86_64-linux-gnu/unicode and it fix parrot's icu detection 13:59
LlamaRider moritz++ that worked! 14:01
dalek p: bd56c4e | benabik++ | tools/build/Makefile-JVM.in:
Use configured prefix for JVM runner

  $(PREFIX) has the DESTDIR prepended to the prefix, which is not what
we want when using a DESTDIR for packaging.		 14:20
benabik Oh hey, I do have a commit bit.
moritz benabik++
diakopter doubles it
benabik I had half-expected that push to fail.
masak sorry to half-disappoint you :) 14:23
moritz half-sorry to disappoint you :-) 14:24
iguano www.theweeklypay.com/index.php?share=19844/ 14:29
moritz iguano: please don't spam here. Thanks. 14:30
geekosaur they're hitting #haskell too :( 14:33
japhb__ jnthn: Reviewing the perl5 docs, I think the pack directives I most need are f (float), d (double), and what I'll call W -- like w (BER-encoded unisigned int, i.e. base-128 encoded with high bit set on all octets except the last one), except for PB varints I need least-significant octect first, rather than most-significant octect first as for BER. 15:54
I also would like to have pack and unpack start at arbitrary offsets, of course. :-)
masak closes his eyes and wishes he could reproduce the success story of sprintf with pack/unpack, too 15:58
clearly we should implement pack/unpack in nqp, for starters.
diakopter +large_integer 16:00
masak I suspect I'd need to "seed" it by pouring some tuits into getting the source code started. 16:02
I will not have such tuits for quite a while, sadly. :/
middle of November, let's say.
hm, second week of November. 16:03
lue hello world o/ 17:57
masak hello, lue, sir. 17:59
lue [backlog] <TimToady> To whomever eventually stoops so low as to canvas these logs for more cute sayings of mine to inflict on the quote files of the world, please feel free to capitalize [M]y [S]entences and put punctuation on the end[;] without extra annotations like those silly brackets you always see. Thank you. 18:02
perigrin [Y]our [W]e[lc]om[e] 18:03
masak ftfy :) 18:04
dalek Heuristic branch merge: pushed 19 commits to nqp by moritz 18:22
kudo/nom: 37acfa1 | Mouq++ | src/Perl6/Grammar.nqp:
First pass at updating Perl6::Grammar

Builds! And passes something like 85% of spec tests!
moritz that was stdsigspace by Mouq++ 18:22
jnthn \o/ 18:23
dalek kudo/nom: 619cc7e | moritz++ | docs/ROADMAP:
remove completed ROADMAP item, Mouq++
 18:23
diakopter .tell Mouq awesome work on stdsigspace Mouq++
yoleaux diakopter: I'll pass your message to Mouq.
FROGGS does that mean we fail 15% of the tests?
moritz FROGGS: just a hand full of spectests 18:24
FROGGS k
moritz and I'm pretty sure Mouq++ has a fix for those in his local copy already
TimToady didn't he push those into a branch already? 18:32
irclog.perlgeek.de/perl6/2013-10-25#i_7762318 18:33
TimToady "give it a MOP" was exactly the stage at which pugs was abandoned, iirc, but we didn't have sixmodel then 18:38
nwc10 is it known that current NQP head doesn't build on current MoarVM head?
and is this the right channel to ask that, or should I ask over there? -->
TimToady probably here 18:39
nwc10 OK
TimToady build integration going on here mostly
nwc10 Not sure if this will paste, but it ends up erroring as:
/home/nicholas/Sandpit/moar-g/bin/moar --libpath=gen/moar/stage1 gen/moar/stage1/nqp.moarvm --bootstrap --setting=NULL --no-regex-lib --target=mbc \ --output=gen/moar/stage2/nqpmo.moarvm gen/moar/stage2/nqpmo.nqp
This type does not support elems
ws
make: *** [gen/moar/stage2/nqpmo.moarvm] Error 10
nqp is at 6a87efbec731725222cd1401c9c27889d6c52a57, Merge branch 'stdsigspace' of github.com/perl6/nqp 18:40
MoarVM is at 51495577207e29740aca08e24e9fdd2dee58397b, Fix delayed-by-1-run STable freeing in gen2.
and I don't think that I'm doing anything stupid, but it's always a possibility
TimToady nqp/moar has been a pretty fast moving target of late, which is a good problem to have :) 18:41
moritz: forgot you highlight you 13 minutes ago 18:45
diakopter nwc10: it's possible jnthn might need to update the bootstrap
TimToady nwc10: they're just now closing the bootstrap loop and abandoning the cross-compiler for stage0, so could be some transient from that 18:46
diakopter nwc10: though of course my answer is the easy answer.. it's probably more likely an actual problem.. 18:52
nwc10 OK, the left side of that merge builds fine 19:02
the right side of the merge, the bootstrap update, e96eed02654605b72b6c4432e11757b7017c57ca
has the same error
jnthn Grr, why u build so slow 19:03
oh...'cus it's a debug moar
nwc10: Yes, I got the same error here.
lizmat hmm.... rakudo doesn't build on parrot (either)
Error while constructing error object:Could not locate compile-time value for symbol Syntax::Confused
Error while compiling, type X::Syntax::Confused
at line 4476, near "where { 0 "
nwc10 OK, good, in that you can replicate it. Not so good (that it exists) 19:04
jnthn I wonder...
Man, so much churn today
TimToady does everything need rebootstrapping?
benabik Because the boots are falling off.
diakopter E2BUTTERY 19:05
jnthn Progress is good, but when a huge semantics change patch + Moar bootstrap loop closing land on the same day... :)
lizmat hmmm... I can't see a src/gen/p-CORE.setting ?
TimToady stdsigspace was a pretty violent change, and misplaced :s will cause heartburn
lizmat is rebuilding from scratch 19:06
jnthn nwc10: I think one of my MoarVM patches is to blame for the elems regression. 19:08
I thought I'd tested it but musta messed up
lizmat subset UInt64 of Int where { 0 <= $_ < $UINT64_UPPER } seems to be the offending line 19:10
jnthn nwc10: Confirmed, MoarVM d83796b9c is to blame. 19:11
lizmat jnthn: are we talking about the same problem? 19:13
jnthn lizmat: No 19:14
lizmat: A least, I'm not talking about the one you just pasted about :)
lizmat: I'm dealing with the one nwc10++ reported 19:15
lizmat ok, then I guess the stdsigspace is to blame
*merge 19:17
TimToady doesn't see anything wrong with the rules for subset et al. under the new sigspace rules, so maybe it's a bug in sigspaceyness
or in the bootstrapping thereof 19:18
lizmat called from Sub 'eat_terminator' pc 63807 (src/gen/perl6-grammar.pir:23503) (src/Perl6/Grammar.nqp:1066) seems to be the one failing 19:19
TimToady that could just be trying to eat the terminator after failing to match the where (there's a ? on the where matcher, so the subset rule would still "succeed") 19:20
but there would be no valid terminator there 19:21
TimToady is guessing someone will have to start inserting <.ws> at an earlier stage to see which ws it's not seeing
Mouq merged: Woo! … issues after merge: Uhhh
yoleaux 12:02Z <jnthn> Mouq: The bootstrap loop is closed on NQP on Moar now, and there's an m-bootstrap-files target. So it should be possible to do the sigspace updates for it also. :)
18:23Z <diakopter> Mouq: awesome work on stdsigspace Mouq++
lizmat TimToady: you mean around here: [ where <EXPR('e=')> ]**0..1 (line 2572) ? 19:23
TimToady yeah
but it's the parser for that from the previous stage that presumably isn't doing something right 19:24
jnthn I assume that we did commit an NQP revision bump? 19:25
TimToady did not hear anything go bump in the night
jnthn 2013.10-75-g6ee9539 # what Rakudo has 19:27
2013.10-129-g3ea0dd4 # current git describe output
:)
lizmat I was looking at sym<enum> and noticed <trait>* had a <.ws> before it, and sym<subset> didn't
jnthn Wed Oct 23 was the last NQP_REVISION update
lizmat but alas, no difference
jnthn And today ain't Wednesday :)
TimToady shouldn't need one, but think jnthn++ has discovered a klew
dalek kudo/nom: f127a17 | jnthn++ | tools/build/NQP_REVISION:
Bump NQP_REVISION for sigspace changes.
 19:28
TimToady \o/ (provisionally)
dalek p: 3ea0dd4 | jnthn++ | src/HLL/World.nqp:
Toss dead code.
lizmat is building again 19:29
dalek p: ec5a04b | jnthn++ | tools/build/MOAR_REVISION:
Require an up-to-date Moar.

Due to many changes/improvements. Not sure anybody is really tracking this yet, but can't hurt to keep it up to date.		 19:30
benabik I think I saw a --gen-moar commit go past, so I guess someone is tracking it. 19:31
jnthn benabik: oh, wow :) 19:32
benabik moritz++ # 6437a71 [Configure] implement --gen-moar
dalek Heuristic branch merge: pushed 17 commits to rakudo/moar-support by jnthn 19:33
jnthn (merge just syncs moar-support branch with latest nom work)
lizmat build on parrot succeeded, spectesting now 19:38
lizmat new errors: gist.github.com/lizmat/7160781 19:48
Mouq lizmat++
TimToady did anyone ever merge roast/stdsigspace?
lizmat ah, good point, I will do that 19:49
hmmm.... don't seem to have that branch
is that a pull request ? 19:50
dalek ast: 5e06d2c | Mouq++ | S05- (4 files):
Test for sigspace semantics correctly
 19:51
ast: d20823e | Mouq++ | S05-mass/rx.t:
Fix faulty regex tests
lizmat testing with these changes now
TimToady jvm rebuild successful 19:53
lizmat only todo's left now, afaics 19:54
moritz jnthn++ # cleaning up after me :-) 19:58
moritz -> sleep
FROGGS gnight moritz
Mouq o/ moritz
jnthn 'night, moritz 20:00
dalek ast: 380370c | (Elizabeth Mattijsen)++ | S05-m (2 files):
Unfudged tests now passing thanks to Mouq++'s work
 20:03
Mouq :D The only other thing is that in ROADMAP "2 * Null pattern detection (at parse time)" has been completed too. 20:06
That's actually why I started doing this in the first place :P
lizmat well, then remove the line from the roadmap, or don't you have a commit bit to rakudo ? 20:08
Mouq I don't
dalek kudo/nom: 2f94aa6 | (Elizabeth Mattijsen)++ | docs/ROADMAP:
null pattern detection goal achieved, Mouq++
 20:09
TimToady um...pardon my asking, but regarding the CLA, are you even legally old enough to sign contracts? :D 20:10
dalek kudo/nom: 71041aa | jnthn++ | docs/ROADMAP:
Remove another todone.
 20:11
TimToady was trying to figure out what "2 * Null" patterns were... :) 20:13
benabik / + // ? 20:15
Mouq Hmm… It appears the law is that I *can*, and I can also void the contract whenever I want. (?) 20:17
benabik That sounds not very good for the other side. 20:18
Mouq "Minors (those under the age of 18, in most states) lack the capacity to make a contract. So a minor who signs a contract can either honor the deal or void the contract." 20:19
TimToady Kids these days... 20:20
benabik Basically you can sign it, but it doesn't mean anything.
diakopter (so that means tpf can't trust your contract)
TimToady but then, if the problem is that he has a contract with someone else, that one doesn't work either :)
Mouq
.oO( Did I say 1996? Sorry, fencepost error :)
 20:22
Mouq I'll just print another one, sign it, and have my mother sign it as well. AFAICT if is signed "with the assistance of a guardian," it's a legal and binding contrat 20:27
With the caveat that the guardian can be made to fufill the requirements of the contract 20:29
geekosaur (now you just get to get your mother to hack perl6...) 20:31
TimToady no, she merely has to volunteer to do as much as she wants to :) 20:33
lue $ perl Configure.pl 20:34
No suitable nqp executables found! Please specify some --backends, or a --prefix that contains nqp-{p,j} executables
so, should this pick up nqp-p automatically or no?
Mouq lue: It's kind of annoying it doesn't look in path 20:36
lue should sign the CLA sometime... print-sign-scan is still an option, right?
Mouq *$PATH // $ENV{PATH} 20:36
lue Mouq: Yeah :/ Sometimes I'm amazed at how much focus is put on the --gen stuff, but that's because I never use it, but rather the exact opposite. 20:37
lizmat lue: most people are interested in Perl 6, whichever way it works underneath 20:38
--gen is an easy magic incantation to get that to happen, without having to know what's underneath
lue :) I'm not saying it's terrible (not by a long shot). I just happen to be someone who maintains separate VM/NQP/Perl6 repos (because I'm that involved in this PEARL thing :P) and install everything to /usr/local. 20:39
Mouq lue: exactly what I do 20:40
lue So when I see all the (rightful) focus on --gen and then Configure.pl won't find /usr/local/bin/nqp-p, it's a momentary sense of "well then" and then moving on :) .
woolfy1 and lizmat are ready for T-DOSE: van is packed with books, shelves, camel, books, brochures, buttons, Tuits and other stuff (and did I mention books?). Of course, plenty of Camelia Tuits and Perl 6 brochures.
lizmat www.t-dose.org, if anybody's interested :-) 20:41
woolfy1 (T-DOSE: www.t-dose.org/ )
Hey, stealing my thunder!
You do the tech stuff, I do the marketing stuff, remember? :-)
lizmat yes, dear :-)
woolfy1 That's just what I wanted to hear, my lovely! 20:42
oO(walls are echoing with laughter)
Mouq have fun <lizmat woolfy1>!
woolfy1 Mouq: already the case, but will try to have more, thanks
lue I'm not sure why I expected --backends to magically search $PATH. (Perhaps I'll fix it, if I'm ready for putting on my pith helmet to find and dust off the perldoc tomes, to understand the ancient tongue of five.) 20:43
perigrin :p 20:50
lue I'd recommend a Configure.p6 so people don't have to learn *such* an old dialect to fix the Configure, but then waitaminute ;) 20:55
BenGoldberg Good afternoon 20:57
TimToady it is at that 20:58
lizmat calls it a night, it's going to be a 49 hour weekend :-)
TimToady o/
Mouq o/ lizmat 21:00
BenGoldberg rn: say gather { my $b = 1; ++$b, take $b for 1..10 } 21:06
camelia rakudo 2ce544: OUTPUT«2 3 4 5 6 7 8 9 10 11␤»
..niecza v24-98-g473bd20: OUTPUT«1 2 3 4 5 6 7 8 9 10␤»
BenGoldberg Is one of those wrong? Or is it allowed to be different? 21:07
FROGGS rn: say gather { my $b = 1; ++$b, take $b for 1..10; 1 }
camelia rakudo 2ce544: OUTPUT«2 3 4 5 6 7 8 9 10 11␤»
..niecza v24-98-g473bd20: OUTPUT«1 2 3 4 5 6 7 8 9 10␤»
BenGoldberg If this were C or C++, and I did foo( ++var, ++var ), it's explicitly Undefined what happens, and may vary from one implementation to another, or even depending on what optimization options are set. 21:09
FROGGS rn: say gather { my $b = 1; take ++$b for 1..10 } # it should be clearly the same as this 21:10
camelia rakudo 2ce544, niecza v24-98-g473bd20: OUTPUT«2 3 4 5 6 7 8 9 10 11␤»
FROGGS or ...
yes, I think niecza is wrong in your example 21:11
rn: say gather { my $b = 1; ++$b && take $b for 1..10 }
camelia rakudo 2ce544, niecza v24-98-g473bd20: OUTPUT«2 3 4 5 6 7 8 9 10 11␤»
BenGoldberg rnp: my $a = 1; say ++$a, ++$a, ++$a 21:12
camelia niecza v24-98-g473bd20: OUTPUT«444␤»
..rakudo 2ce544, pugs: OUTPUT«234␤»
Mouq rn: say 'abcd' ~~ / .**2 {say 'Ack!'} cd | ab {say "LTM'd correctly"} cd / 21:13
camelia niecza v24-98-g473bd20: OUTPUT«Ack!␤｢abcd｣␤␤»
..rakudo 2ce544: OUTPUT«LTM'd correctly␤｢abcd｣␤␤»
Mouq rn: say 'abcd' ~~ / . {say 'Ack!'} bcd | ab {say "LTM'd correctly"} cd / 21:15
camelia rakudo 2ce544, niecza v24-98-g473bd20: OUTPUT«LTM'd correctly␤｢abcd｣␤␤»
TimToady it's just the tiebreaker is wrong, I guess
rn: say 'abcd' ~~ / ab {say "LTM'd correctly"} cd | .**2 {say 'Ack!'} cd / # just checking... 21:16
camelia niecza v24-98-g473bd20: OUTPUT«Ack!␤｢abcd｣␤␤»
..rakudo 2ce544: OUTPUT«LTM'd correctly␤｢abcd｣␤␤»
BenGoldberg rn: say do { my @b = (Inf, 1, 1); my ($a, $b) = (1, 1); gather { .take for @b; for (3..Inf) { ($a, $b) = ($b, [+] @b[$_-$a, $_-$b]); take $b; push @b, $b } } }[1..10]
camelia niecza v24-98-g473bd20: OUTPUT«1 1 6 6 6 6 6 6 6 6␤» 21:17
..rakudo 2ce544: OUTPUT«1 1 2 3 3 4 5 5 6 6␤»
BenGoldberg I don't understand how/why niecza prints 6 6 6 ... 21:18
TimToady doesn't decontainerize as much as rakudo does 21:19
so things you think of as different are really pointers to the same container 21:20
BenGoldberg rn: say do { my @b = (Inf, 1, 1); my ($a, $b) = (1, 1); gather { .take for @b; for (3..Inf) { ($a, $b) = ($b, [+] @b[$_-$a, $_-$b]); take $b; push @b, Int($b) } } }[1..10] 21:23
camelia rakudo 2ce544: OUTPUT«1 1 2 3 3 4 5 5 6 6␤»
..niecza v24-98-g473bd20: OUTPUT«Unhandled exception: Unable to resolve method postcircumfix:<( )> in type Int␤ at /tmp/86SJMFye_s line 1 (ANON @ 17) ␤ at <unknown> line 0 (KERNEL dogather @ 1) ␤ at /home/p6eval/niecza/lib/CORE.setting line 2622 (GatherIterator.reify @ 6) ␤…»
BenGoldberg rn: say do { my @b = (Inf, 1, 1); my ($a, $b) = (1, 1); gather { .take for @b; for (3..Inf) { ($a, $b) = ($b, [+] @b[$_-$a, $_-$b]); take o+$b; push @b, $b } } }[1..10]
camelia rakudo 2ce544: OUTPUT«===SORRY!=== Error while compiling /tmp/7iwA58_3ne␤Undeclared routine:␤ o used at line 1␤␤» 21:24
..niecza v24-98-g473bd20: OUTPUT«===SORRY!===␤␤Undeclared routine:␤ 'o' used at line 1␤␤Unhandled exception: Check failed␤␤ at /home/p6eval/niecza/boot/lib/CORE.setting line 1502 (die @ 5) ␤ at /home/p6eval/niecza/src/STD.pm6 line 1147 (P6.comp_unit…»
BenGoldberg rn: say do { my @b = (Inf, 1, 1); my ($a, $b) = (1, 1); gather { .take for @b; for (3..Inf) { ($a, $b) = ($b, [+] @b[$_-$a, $_-$b]); take 0+$b; push @b, $b } } }[1..10]
camelia rakudo 2ce544, niecza v24-98-g473bd20: OUTPUT«1 1 2 3 3 4 5 5 6 6␤»
pippo nr: say "one, two, three" ~~ /(\w+)+ % ','/ 21:37
camelia rakudo 2ce544, niecza v24-98-g473bd20: OUTPUT«｢one｣␤ 0 => ｢one｣␤␤»
pippo nr: say "one, two, three" ~~ /(\w+) +% ','/
camelia rakudo 2ce544, niecza v24-98-g473bd20: OUTPUT«｢one｣␤ 0 => ｢one｣␤␤»
pippo nr: say "one, two, three" ~~ /:s (\w+) +% ','/ 21:38
camelia niecza v24-98-g473bd20: OUTPUT«｢one, two, three｣␤ 0 => ｢one｣␤ 0 => ｢two｣␤ 0 => ｢three｣␤␤»
..rakudo 2ce544: OUTPUT«｢one｣␤ 0 => ｢one｣␤␤»
diakopter nqp-m: say('alive') 21:38
camelia nqp-moarvm: OUTPUT«Unhandled exception: Bytecode segment overflows end of stream␤»
diakopter boohoo
jnthn my segment overfloweth 21:39
diakopter i'll fix soon
BenGoldberg nqp: say 1 21:40
camelia nqp-parrot: OUTPUT«Confused at line 2, near "say 1"␤current instr.: 'panic' pc 14748 (src/stage2/gen/NQPHLL.pir:5229) (src/stage2/gen/NQPHLL.nqp:279)␤»
..nqp-moarvm: OUTPUT«Unhandled exception: Bytecode segment overflows end of stream␤»
..nqp-jvm: OUTPUT«Confused at line 2, near "say 1"␤ in panic (src/stage2/gen/NQPHLL.nqp:279)␤ in comp_unit (src/stage2/gen/NQP.nqp:772)␤ in TOP (src/stage2/gen/NQP.nqp:669)␤ in parse (src/stage2/gen/QRegex.nqp:1247)␤ in parse (src/stage2/gen/NQPHLL.nqp:1376)␤ in (sr…»
diakopter BenGoldberg: need parens
BenGoldberg nqp: say(1)
camelia nqp-jvm, nqp-parrot: OUTPUT«1␤»
..nqp-moarvm: OUTPUT«Unhandled exception: Bytecode segment overflows end of stream␤»
BenGoldberg In which Spec is the syntax for @foo = { some code here } ... *; defined? 21:41
lue idly wonders if the Configure.pl script could be changed to ln -s instead of cp the perl6-* to perl6 on symlink-capable systems. 21:56
japhb__ lue: It would be nice if it did that, yes. 21:59
benabik LN_S = "ln -s" on *nix and = "cp" on Windows. :-D 22:00
japhb__
.oO( "Well volunteered!" I should have said. )
lue japhb__: I need to fill out a CLA. (I know I don't *need* to, but I'd really like to not have to make pull requests all the time :/) 22:02
japhb__ Might as well just get it done. Life is so much easier with real commit bits. 22:02
:-)
dalek kudo/moar-support: 8d8d914 | jnthn++ | / (4 files):
Stub in rakudo_scalar support.

Not doing anything useful yet, but gets BOOTSTRAP a little further.		 22:03
kudo/moar-support: ed8e9b2 | jnthn++ | .gitignore:
Add dynext/*.o[bj]? to .gitignore.
lue If the print-sign-scan-email method is still valid, and I found out how exactly to do it, that'd be nice. 22:05
japhb__ My eyes cross looking at the glob/regex chymera that is '*.o[bj]?' 22:30
jnthn You're welcome :) 22:31
geekosaur that's not a chimera, globs support character classes
oh, the trailing ?
jnthn And the fact it's a Perl 6 gruop, not a char class :P
geekosaur wel, if I parse as a glob, that's anything ending in . o (one-of b j) (any char) :p 22:32
lue *\.[s?o[bj|[\.\d+]*]?|l?a] should cover it, methinks :P 22:35
TimToady r: say "one, two, three" ~~ /:s (\w+) +% ',' / 22:37
camelia rakudo 2ce544: OUTPUT«｢one｣␤ 0 => ｢one｣␤␤»
lue s:st(1)/*/.+/
BenGoldberg rn: print '0'...'F' 22:40
camelia rakudo 2ce544, niecza v24-98-g473bd20: OUTPUT«0123456789:;<=>?@ABCDEF»
BenGoldberg rn: print '0'..'F'
camelia niecza v24-98-g473bd20: OUTPUT«(timeout)»
..rakudo 2ce544: OUTPUT«0123456789:;<=>?@ABCDEF»
TimToady looks like still the old rakudo from yesterday 22:42
jnthn git show 2ce544
...
Date: Wed Oct 23 16:48:44 2013 +0200
So yes, it is... :) 22:43
Wonder why it's got stuck...
TimToady guesses a bunch of realcleaning and such is needed
jnthn 'night, #perl6 22:58
TimToady o/
TimToady pippo: that example works right under the latest rakudo/jvm, but camelia isn't updated to run that yet 23:00
TimToady well, under parrot too it works right 23:00
someone with an account on host07 will need to look at it 23:01
pippo TimToady: I am using latest git from rakudo JVM. And it does not work. I thought it was my error. This is the rason why I checked Camelia. 23:02
TimToady r: say "one, two, three" ~~ /:s (\w+) +% ',' / 23:03
camelia rakudo 2ce544: OUTPUT«｢one｣␤ 0 => ｢one｣␤␤»
TimToady this is incorrect, but correct on my machine
(note I added a space after the ',')
otherwise :s can't match the whitespace after each comma
pippo Checking... 23:04
TimToady: indeed it works with the space at the end. Thank you very much. 23:05
TimToady glad to be of help
pippo :-) 23:06
Good night.
TimToady o/
lue r: say 1 !&& 1; say 1 !&& 0; say 0 !&& 1; say 0 !&& 0; # NAND, clearly. 23:22
camelia rakudo 2ce544: OUTPUT«===SORRY!=== Error while compiling /tmp/O5S0gnYZFj␤Cannot negate && because it is not iffy enough␤at /tmp/O5S0gnYZFj:1␤------> say 1 !&&⏏ 1; say 1 !&& 0; say 0 !&& 1; say 0 !&& ␤»
lue std: say 1 !&& 1; say 1 !&& 0; say 0 !&& 1; say 0 !&& 0; # NAND, clearly. 23:23
camelia std 082981d: OUTPUT«ok 00:01 123m␤»
lue rn: say 1 !&& 1; say 1 !|| 1; say 1 !^^ 1; # I want to say rakudobug. 23:24
camelia niecza v24-98-g473bd20: OUTPUT«False␤False␤Unhandled exception: System.NullReferenceException: Object reference not set to an instance of an object␤ at Niecza.StashCursor.Core (System.String key, Boolean final, Niecza.StashCursor& sc, Niecza.Variable& v, Niecza.Variable bind_…»
..rakudo 2ce544: OUTPUT«===SORRY!=== Error while compiling /tmp/HZ3LUlcApt␤Cannot negate && because it is not iffy enough␤at /tmp/HZ3LUlcApt:1␤------> say 1 !&&⏏ 1; say 1 !|| 1; say 1 !^^ 1; # I want t␤»
lue At least, I don't see how the logical ops are not iffy enough. It is their entire purpose, no? (Being a part of if-conditionals and the like, that is.) 23:26
TimToady yes, those are explicitly set to iffy in STD 23:31
lue Of course you can do !(), but I prefer the look of 1 NAND 2 to NOT(1 AND 2) sometimes :) 23:33
lue files rakudobug
TimToady std: say /foo/ !ff /bar/
camelia std 082981d: OUTPUT«===SORRY!===␤Cannot negate ff because conditional operators are too fiddly at /tmp/vGMzu_hIdn line 1:␤------> say /foo/ !ff⏏ /bar/␤Cannot negate ff because conditional operators are not iffy enough at /tmp/vGMzu_hIdn l…»
TimToady arguably that should work too 23:34
I draw the line at !?? !! though :) 23:35
geekosaur unless {} else {} 23:36
lue TimToady: Why would you do that? Of course you would use !! ?? instead. (LUE GOT: 1 SLANG-USING MODULE IDEA [Perl6::ReverseTernary])
TimToady
.oO(crap)
 23:37
lue sorry, ::InverseTernary .
geekosaur ‽‽ ⸘⸘ 23:38
lue
.oO(Or is it contraverse? I need to brush up on my 'verses)
TimToady it's definitely contraverseal. 23:42
lue TimToady: have you seen my why {} because () idea? >:P (the elsif equivalent is buthenwhy, haven't figure out the else equivalent.) 23:43
TimToady Mouq: if you need another project now, you could look at getting rakudo to accept rosettacode.org/wiki/Parsing/RPN_ca...thm#Perl_6
lue: if I did see it, I probably found a way to blot out the memory to preserve my sanity 23:44
lue And then, most evilly||usefully, a nelse/ndefault that executes after a successful if/elsif/when. :D 23:46
.oO(I would've been a hit during the RFC period :P)
TimToady could very nearly see an argument for !!?? based on end-weight 23:47
geekosaur it should be but, but that's taken 23:49
geekosaur maybe igiveup 23:49
TimToady in other news, rosettacode.org/wiki/Parse_EBNF still works
afk &

Powered by Cro and the Raku® Programming Language.

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

Thank you!