»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.
timotimo indeed :| 00:02
dalek p/awesome_bindattr: bd80cda | (Timo Paulssen)++ | src/vm/parrot/QAST/Operations.nqp:
this actually works!
00:06
timotimo (oops)
damnit, what am i doing wrong now? >_> 00:07
lue
.oO(And then they tell you about a cool, effectively-essential matching paren highlighter! How else could you ever figure out that mess o' parens?)
00:08
00:11 eternaleye left
timotimo where do i accidentally generate a repr_bind_attr_int_p_p_sc_s_ic?! 00:11
ah, of course 00:12
00:12 Psyche^ joined
timotimo i'm quite surprised it didn't blow up earlier 00:12
00:13 xenoterracide left
timotimo like, before i commited the code publically and made a total ass of myself ;) 00:15
lue then your next commit should start with "... in breaking nqp." :) 00:16
00:16 Psyche^_ left
dalek p/awesome_bindattr: 1e8a295 | (Timo Paulssen)++ | src/vm/parrot/QAST/Operations.nqp:
confusion about argument order cleared up.
00:16
timotimo aaw, i saw that just a second too late! :(
that's why i shouldn't get a rakudo commit bit, i break stuff :) 00:17
i hope my changes are stable and result in a faster nqp+rakudo
that would be hella nice
the changes from the very first commit already cause a 3% speedup when compiling the setting :)
okay, now to build a rakudo and time it :)
i have 12 timings without the optimisations
lue on the upside, you don't seem to be committing to nqp/master . 00:18
timotimo that's right 00:19
tomorrow (logically; physically it's "today" to me) is rakudo-release-day
good news 00:27
the first timing is 3:21.8s wallclock vs 3:27.0s wallclock without the optimization 00:28
00:28 fridim__ joined
timotimo (but this is rm + gen-cat.pl + perl6 compile-ze-setting + parrot compile-ze-pbc) 00:28
(so the measurement is a bit off, especially the last step takes quite a bit of time, too)
i should really measure those steps individually and subtract them from the overall run time 00:29
else these measurements are useless
00:30 berekuk left 00:34 stevan_ left
diakopter timotimo: you're being too hard on yourself :P 00:34
00:34 eternaleye joined, stevan_ joined
timotimo ah, very good. i can subtract 30 seconds from each measurement just for the parrot pbc step. 00:34
the gen-cat takes just 0.02s, so that makes no difference 00:35
00:42 berekuk joined
timotimo so, is a win of 6% when compiling the setting worth merging my branch? 00:43
probably closer to 3% actually 00:56
timotimo runs spectests 00:57
01:03 zwut00 left
timotimo spectests are clean 01:06
01:09 FROGGS joined, berekuk left
timotimo ah, github pull requests are b0rked still 01:11
.tell masak would you consider mergint awesome_bindattr before cutting the release today? 01:14
yoleaux timotimo: I'll pass your message to masak.
timotimo bedtimey times
01:14 FROGGS left 01:19 eternaleye left 01:21 jnap joined 01:37 jnap left, FROGGS joined 01:46 FROGGS left 01:48 FROGGS joined 01:49 BenGoldberg left 01:52 crab2313 joined 01:59 kbaker joined 02:08 FROGGS left 02:09 FROGGS joined 02:25 FROGGS left 02:47 FROGGS joined 02:54 kbaker left, crab2313 left 03:00 FROGGS left 03:05 preflex left, preflex_ joined, ChanServ sets mode: +v preflex_ 03:06 preflex_ is now known as preflex 03:12 cjwelborn joined 03:20 EvanTeitelman left, FROGGS joined 03:38 FROGGS left 03:40 FROGGS joined 03:45 FROGGS left 03:50 FROGGS joined 04:08 eternaleye joined 04:09 eternaleye_ joined, eternaleye left, eternaleye_ is now known as eternaleye 04:13 FROGGS left 04:15 SamuraiJack__ joined, daniel-s joined 04:24 xinming joined 04:25 FROGGS joined 04:27 fridim__ left 04:33 baest left 04:36 FROGGS left 04:39 cjwelborn left 05:08 xinming left 05:09 xinming joined 05:24 rindolf joined 05:33 xenoterracide joined 05:35 agathangelos_ joined 05:40 raiphvb joined 05:41 domidumont joined 05:42 risou_awy is now known as risou 05:44 raiphvb left 05:46 domidumont left 05:47 domidumont joined 05:51 denis_boyun joined 06:02 agathangelos_ left 06:09 trnh joined, trnh left, trnh joined 06:10 dmol joined 06:13 xenoterracide left 06:15 denis_boyun left 06:21 rindolf left 06:24 flussence left 06:25 flussence joined 06:26 FROGGS joined, iSlug joined 06:28 baest joined 06:30 dakkar joined
hoelzro good morning #perl6! 06:36
TimToady o/
JimmyZ \o 06:37
06:37 trnh left
TimToady actually, here in california it's about time for werewolves, with the full moon and all 06:38
06:45 [Sno] left, [Sno]_ joined 06:49 trnh joined 06:51 kaleem joined 06:55 domidumont left 07:05 sqirrel joined
moritz timotimo: is it an nqp-only branch? or some things required in rakudo too? 07:07
07:08 domidumont joined
arnsholt hoelzro: I think changing return in NQP to being a token should be pretty straightforward. If you want a bit of help with it, I'm pretty sure I can talk you through it 07:14
07:16 zakharyas joined
lizmat good *, #perl6! 07:25
moritz o/ lizmat, arnsholt, hoelzro, * 07:27
lizmat moritz o/
07:35 daxim joined
hoelzro so my allocations test showed that allocations make up about 2 seconds of runtime when building the core setting 07:58
arnsholt: I can give it another shot tomorrow 07:59
07:59 risou is now known as risou_awy, fhelmberger joined
arnsholt hoelzro: Cool, gimme a prod if you have any questions 08:00
hoelzro hands arnsholt a cattle prod
lizmat so what was the question? 08:01
hoelzro I'm sure I have them; I just don't know what they are yet =)
arnsholt s/if/when/, in that case =) 08:03
hoelzro I find it interesting that many languages (at least among the ones I've studied) don't really distinguish between 'if' and 'when' 08:04
either that, or I haven't studied said languages enough =)
08:05 mberends joined
lizmat hoelzro: I'm pretty sure said languages *do*, but their *users* don't 08:09
hoelzro ah
08:21 iSlug left 08:33 betterwo1ld joined 08:42 berekuk joined
arnsholt lizmat: The distinctions a language makes is mostly dictated by its users though 08:45
Prescriptivists may rant and rave, but correct language is still socially constructed 08:46
08:51 berekuk left 08:53 DarthGandalf left 08:57 daniel-s left
dalek rlito: 657299a | (Flavio S. Glock)++ | / (3 files):
Perlito5 - grammar - format() tweak
09:03
nwc10 [Coke]: On the Friday, does the caterpiller eat "Apfelsinen" and the next day is called "Sonnabend"? :-) 09:04
09:06 salv0 left
hoelzro how is EXPR implemented, again? 09:07
moritz with magic :-) 09:08
hoelzro iirc it's specially generated via the operator precedence table or something
moritz it's a regular method
but it uses the operator precedence table
FROGGS hoelzro: there is a EXPR method in nqp
09:08 berekuk joined
hoelzro ah! 09:08
FROGGS nqp/src/HLL/Actions.nqp:77: method EXPR($/, $key?) {
ahh, this one: nqp/src/HLL/Grammar.nqp:403: method EXPR(str $preclim = '', int :$noinfix = 0) { 09:09
09:13 berekuk left 09:14 berekuk joined, pernatiy joined 09:17 risou_awy is now known as risou 09:21 DarthGandalf joined 09:37 salv0 joined 09:52 Maddingue left 09:57 eiro joined 09:58 Maddingue joined 10:02 ssutch left 10:06 rindolf joined 10:16 dmol left 10:19 krokite left 10:22 berekuk left 10:24 berekuk joined 10:55 sqirrel left 10:58 flussence left, flussence joined 10:59 FROGGS left
arnsholt hoelzro: The EXPR rule in the grammar is inherited from QRegex (and it does OPP using the infix, prefix etc rules). That rule then calls EXPR in the Actions class normally, just like any other rule 11:01
And now that it's written in NQP, it's a lot simpler to understand than it was in PIR =) 11:03
11:03 sqirrel joined 11:19 xinming left 11:21 xinming joined
lizmat r: my %h; my $a; $a ?? %h<a>:k !! %h<a>:v # what's wrong with this? 11:21
camelia rakudo 6f8668: OUTPUT«===SORRY!=== Error while compiling /tmp/MbN0_LQyVO␤You can't adverb that␤at /tmp/MbN0_LQyVO:1␤------> my %h; my $a; $a ?? %h<a>:k !! %h<a>:v⏏ # what's wrong with this?␤ expecting any of:␤ pair value␤»
lizmat std: my %h; my $a; $a ?? %h<a>:k !! %h<a>:v # what's wrong with this? 11:22
camelia std 7c17586: OUTPUT«===SORRY!===␤Precedence of :k is too loose to use between ?? and !!; please use parens around inner expression at /tmp/ZcCwZyyk5Q line 1:␤------> my %h; my $a; $a ?? %h<a>:k⏏ !! %h<a>:v # what's wrong with this?␤ expectin…»
lizmat LTA error message is what is
11:24 [Sno]_ is now known as [Sno]
timotimo moritz: nqp-only 11:27
11:28 berekuk left, iSlug joined 11:29 berekuk joined
timotimo hoelzro: now the question is: even if the allocations only take up 2 seconds in total, does reducing the number of existing cursors to half/a third/two thirds/whatever by using a singleton failure cursor relieve some GC pressure or something like that? 11:29
hoelzro hmm 11:30
I should do some tests on the size of objects that are being allocated
timotimo actually, i could try the singleton failure match myself, i'm kind of intrigued now 11:31
hoelzro if you have the time, please do
I'd like to see how it does
my schedule is full until tomorrow night =/ 11:32
timotimo OK 11:33
hm, the cursor is a role that gets mixed into a grammar class, i wonder if i can get away with just mixing the role into an empty class to pun it?
can i has your allocation counting code for this? 11:35
compiling the first attempt 11:37
hoelzro timotimo: it's at home =/ 11:39
I might be able to reproduce it quickly, though 11:40
timotimo not that important
i'll just measure time to see if it makes a big difference
hoelzro ok 11:41
11:42 berekuk left
timotimo seems like i compiled it, that's pretty good 11:43
11:44 berekuk joined
hoelzro \o/ 11:44
compilation!
timotimo FINISH IT!
corecatcher :>
timotimo now for an actual timing 11:48
now it's at 117s for parse 11:52
i don't know what exactly it was yesterday on nqp-master
[Coke] nwc10: (apfelsinen, sonnabend) ,ja
timotimo but i do know how much time it cost all in all
11:52 salv0 left
timotimo shaves off about as much as the awesome_bindattr optimisation; like 6 seconds out of 170 11:54
[Coke] r: say 5/170 11:55
camelia rakudo 6f8668: OUTPUT«0.029412␤»
[Coke] r: say 6/170
camelia rakudo 6f8668: OUTPUT«0.035294␤»
lizmat timotimo: another 3%
cool
timotimo doesn't mean it both adds up to 6%, though
also, 6% is still kinda pitiful :(
where are my 50% improvements? :(
lizmat I'll take those 6% 11:56
moritz currently tries to optimize some regex-heavy perl 5 code. What a PITA
11:56 risou is now known as risou_awy 11:57 xinming left
timotimo moritz: imagine it had been perl6 code initially, then you'd have gotten a huge win by porting it to perl5! ;) 11:57
11:58 jnap joined
dalek p/singleton_fail: 3314cf5 | (Timo Paulssen)++ | src/QRegex/Cursor.nqp:
use a singleton FailCursor for all parsing.
12:00
timotimo here's your commit, sirs and madames 12:01
12:01 jnap1 joined, xinming joined 12:02 jnap left
hoelzro timotimo: what about memory usage/ 12:03
?
arnsholt Does it change the time it takes to compile CORE.setting appreciably?
timotimo no big change i can see
hoelzro grr
I was hoping to see a huge win =/
oh well
6/170 seconds isn't a bad start =)
timotimo let me calculate some averages and shit 12:04
the average maxresidentk of unoptimized nqp is 1_732_984 12:05
i need to make a few measurements now.
12:13 pmurias joined, FROGGS joined
dalek kudo-js: 68ac723 | (Pawel Murias)++ | src/QAST/Compiler/JavaScript.nqp:
Implement if ... -> $value {...}.
12:13
kudo-js: b9a56df | (Pawel Murias)++ | runtime/hash.js:
Fix bug.
kudo-js: 76c3337 | (Pawel Murias)++ | runtime.js:
Add .to_s conversions to ops.

Implement nqp::markcodestub. Stub nqp::freshcoderef. Make nqp::where work on all objects with an _STable.
kudo-js: 56b789b | (Pawel Murias)++ | runtime (2 files):
Implement nqp::objprimspec.
12:14 berekuk left 12:19 salv0 joined 12:20 skids left 12:31 benabik left
pmurias can I use git diff to compare untracked files? (I want to use the --color-words option which GNU diff does not have) 12:39
yoleaux 18 Sep 2013 23:46Z <[Coke]> pmurias: - re smartlinks - let's start with checking to see if every variant listed in the docs has a test. (and that every variant in the code has doc)
12:43 grondilu joined
moritz pmurias: you can just git add them, compare, and then unstage them again 12:44
or use wdiff (not color, but word-based diff)
$ wdiff a b
[-A-]{+B+} B
grondilu had to run 'cd /usr/include; ln -s ./i386-linux-gnu/unicode .' as root in order to compile rakudo with ICU. Doesn't know why exactly.
12:45 SamuraiJack joined
hoelzro ugh 12:45
--target=jast output is not intended for human consumption, is it?
moritz nope 12:47
nwc10 [Coke]: the choice of those two words is quirky. They are colloquial or obscure 12:48
12:48 SamuraiJack__ left
pmurias moritz: git diff seems to work fine without adding them :) 12:49
moritz pmurias: huh. It always comes out empty for me 12:51
12:53 xinming left
pmurias with the --no-index option? 12:55
12:55 xinming joined
moritz ah, that works 12:55
13:00 SamuraiJack left, SamuraiJack_ joined 13:04 dmol joined 13:06 nyuszika7h left, nyuszika7h joined 13:09 berekuk joined, grondilu left 13:10 ajr joined 13:11 ajr is now known as Guest93640, Guest93640 is now known as ajr_ 13:12 trnh left
arnsholt pmurias: vimdiff does colors (but maybe not the kind you want) 13:16
hoelzro is there a way a trait could affect a Grammar's underlying regex implementation? 13:18
could one, for example, do something like this: grammar MyGrammar is RE2 { regex something { ... } }
moritz hoelzro: you'd have to give it a different meta object, which will probably be done with a different keyword than 'grammar' 13:20
hoelzro hmm 13:21
that would probably be ok
lizmat away until much later tonight / tomorrow& 13:33
pmurias git diff --no-index --word-diff does what I want 13:36
timotimo is --no-index like git diff ... HEAD? 13:37
FROGGS moritz: you can even override the grammar keyword lexically... it is somewhere in the RaNIW material
timotimo indeed, the EXPORTHOW thingie, no?
FROGGS hmm, dunno :o)
dalek kudo-js: 86893fc | (Pawel Murias)++ | runtime/null.js:
stringifcation for nulls
13:40
kudo-js: c3ffffb | (Pawel Murias)++ | runtime.js:
fix nqp::existskey.
13:43 imarcusthis- joined 13:47 imarcusthis- left 13:58 thou joined, skids joined
pmurias can the current file position on a file handle move by itself? node lacks seek/tell in it's standard libary so I could either emulate that myself or use an external library 14:04
hoelzro it can 14:05
if you're sharing a file descriptor with another process
how are they missing seek and tell? 14:06
=(
could you add bindings for them?
timotimo wow, that's a bad miss.
but node isn't meant to do any process sharing or threading or any similar thing anyway
hoelzro timotimo: what if you inherit a file from your parent process, though? 14:07
timotimo node will probably not let you use it 14:08
hoelzro I mean, it's stupid to have the parent affect the file offset if it's handed a file descriptor off to its child
but it's possible
timotimo somehow, i think of node as a toy with training wheels everywhere
maybe i should try to use it some time?
i guess i will when rakudo runs on it
14:09 benabik joined
pmurias hoelzro: re add bindings, I can import seek from a fs-ext module (and implement tell using seek) 14:16
they don't have a function to read a line also :/
timotimo does seek return what tell would?
hoelzro with SEEK_CUR and offset = 0, yes 14:17
14:18 patspam joined, Juerd joined, rs0 joined, pmichaud joined, jfried joined 14:19 pernatiy left 14:20 `patch` joined 14:23 trnh joined, trnh left, trnh joined, denis_boyun joined 14:31 ajr_ left 14:32 sqirrel left, ajr joined, ajr is now known as Guest59671 14:35 Guest59671 is now known as ajr_ 14:36 pernatiy joined 14:45 Woodi left, tba left 14:46 tba joined
pmurias so how is nqp::readlinefh supposed to work? read a line (and a bit more) using explicitly explicitly specified offsets and the seek to the exact end of line? 14:51
14:51 Woodi joined, dalek left 14:52 dalek joined, ChanServ sets mode: +v dalek
diakopter pmurias: see it in moarvm 14:54
JimmyZ did it I think 14:55
FROGGS and me
there is also a good test in moarvm
diakopter i mean FROGGS
FROGGS t/nqp/19
15:00 risou_awy is now known as risou 15:04 xinming left, xinming joined 15:09 sqirrel joined
pmurias what's the default encoding? 15:10
15:10 dmol left 15:14 xenoterracide joined
pmurias doe nqp::readlinefh work on utf16 files (it doesn't seem to looking at the code) 15:15
JimmyZ pmurias: there is an setencoding 15:22
pmurias: but looks like utf16 encoding doesn't work on moarvm
15:27 FROGGS left 15:31 zakharyas left 15:35 colomon left
pmurias :/ node.js doesn't support the latin1 encoding 15:38
not that latin1 doesn't deserve to die ;)
15:42 kaleem left 15:50 colomon joined
dalek kudo-js: 03810bc | (Pawel Murias)++ | runtime.js:
Make nqp::primspec handle nqp::null().
15:55
kudo-js: 24a4cb2 | (Pawel Murias)++ | / (2 files):
Pass test 19.

Implement nqp::tellfh,nqp::readlinefh,nqp::chdir,nqp::rmdir, nqp::mkdir,nqp::getstdin,nqp::getstdout,nqp::getstderr and nqp::unlink.
kudo-js: a95feca | (Pawel Murias)++ | run_tests:
Add test 19 to run_test.
15:56
PerlJam Anyone want to vet some changes I'm thinking of making to the release guide? See gist.github.com/perlpilot/6625489 16:01
colomon oooo, release day! does this mean we're finally getting a bump in the NQP version?
PerlJam I moved the list of releases that have already happened to the bottom of the document and removed some parrot-specific bits and re-emphasized some other parroty bits. 16:02
moritz will be gone until late Sunday or so 16:04
have a nice Rakudo release day, and a nice weekend!
16:05 coax75ohm joined
PerlJam moritz: you have fun in whatever you're doing too! :) 16:05
16:05 xenoterracide left 16:06 FROGGS joined, ajr_ left 16:07 ajr joined, kaare_ joined, ajr is now known as Guest85591 16:08 Guest85591 is now known as ajr_ 16:09 dmol joined
[Coke] nwc10: your truth is a downer. 16:19
[Coke] thinks pmurias is doing an awful lot of work during the pencils nearly down period. shhhh. ;)
16:22 xenoterracide joined
daxim remember 2008? use.perl.org/use.perl.org/_Ovid/jou...36776.html 16:24
16:24 domidumont left
pmurias jnthn: we have both a NQPRegex class and a NQPRegex knowhow? 16:30
16:31 trnh left 16:32 panchiniak joined, iSlug left 16:33 daxim left
pmurias [Coke]: it's better then after the pencils down period ;) 16:33
16:36 denis_boyun__ joined, denis_boyun left 16:38 panchiniak left
dalek kudo-js: 1431cb5 | (Pawel Murias)++ | / (3 files):
Misc fixes attempting to get grammars to work in standalone nqp.

Implement nqp::push_s. Fix the !dba regex subtype. Set the custom-regex-lib nqp-compiler.nqp.
16:38
16:40 ajr_ left 16:42 coax75ohm left 16:44 lambdacat joined, domidumont joined
masak oh hai, #perl6... 16:45
yoleaux 01:14Z <timotimo> masak: would you consider mergint awesome_bindattr before cutting the release today?
masak ...from a train! :D
16:46 fhelmberger left
masak timotimo: yes, I'll consider it. 16:46
geekosaur entrainced by perl6?
masak starts by backlogging 16:47
timotimo: [backlog] "Luminosity" 16:51
pmurias jnthn: I'm confused by the fact that we have a NQPRegex in the setting and a different one in our regex library 16:54
17:01 pernatiy left
benabik Isn't jnthn on vacation? 17:08
[Coke] Yes. 17:09
17:13 sqirrel left 17:15 iSlug joined 17:18 lambdacat left
TimToady pmurias: and he said to use .tell becuase he didn't expect to backlog 17:18
*au
pmurias .tell? 17:25
ahh, the bot command
17:26 krokite joined, krokite left, krokite joined
pmurias I was implement the nqp::tellfh(...) today so I got confused ;) 17:26
benabik Is jnthn a file handle? 17:27
diakopter yes, a buffered one 17:28
nwc10 [Coke]: I don't know. Be amused that the choice of wording is slightly quirky, and happy that you (like me) have learned a couple of obscure words
(just be aware that they aren't the usual ones)
and be thankful that they aren't Austrian dialect :-)
17:29 pmurias left, ssutch joined
jnthn Well, that was eventful... 17:33
TimToady eh? 17:34
nwc10 but did it eventually end up with you in the right place, with your luggage?
can they use the plane again?
masak jnthn! \o/
TimToady jnthn! /o\
jnthn nwc10: I'm pretty sure they will use that plane again, but it'll need the multiple things wrong with it fixed... :S
nwc10 oh, one of those?
did the airport fire engines come out to meet you? 17:35
jnthn Yeah. 15 minutes in... "So, we have this landing geer issue..."
nwc10 oh :-/
jnthn *gear
17:35 dakkar left
jnthn Made it down safely, but the plane lost its ability to be steered, so we nicely clogged the runway for a while until we could be towed away... 17:35
nwc10 last year, the "problem" was "fixed", and then after quite a while we arrived on the runway. But the meet and greet from all the fire engines was highy uneventful.
jnthn Yeah, the landing in the end was thankfully uneventful. 17:36
nwc10 aha. that sounds a bit too much "interesting times"
TimToady maybe we shouldn't allow jnthn to fly on the same plane as jnthn...
jnthn Yes, the pilot spoke very reassuringly but I'm not sure how much, if anything, he left out. :) 17:37
Anyway, ended up rather late. And on a different plane. :)
nwc10 were the cabin crew, um, bricking themselves? 17:38
17:38 sqirrel joined
jnthn The one I had chance to observe seemed quite composed in the air, but decided I was a good person to talk a lot to after we landed. :) 17:39
masak read that as "seemed quite composed of air"
nwc10 aha - OK. Yes, on that flight the crew didn't want to tell us what had happened, but they did seem shaken. Kept it all together while in the air. 17:40
so, I'm buying into your assumption that some stuff was left out.
jnthn Well, the fact the thing couldn't be steered off the runway after we landed suggests there was a slightly more interesting failure than "the wheel door won't close properly" 17:41
nwc10 you landed back where you took off from?
(with that plane) 17:42
TimToady well, 15 minutes in sounds like they couldn't crank it up, so maybe just a general hydraulics failure in the area
jnthn Yeah
After a lot of circling because we were too heavy to land.
diakopter "we don't dump fuel; we just burn it" 17:43
benabik Well, dumping fuel ends with flammables falling on people's heads. 17:48
TimToady and if you try to flare it off, and it doesn't ignite immediately, you have a lovely fuel-air bomb
jnthn I'm pretty sure burning it off was by a good way the best option here 17:49
TimToady thinks they reserve dumping for when you have to land Right Now
17:49 thou left 17:50 thou joined
jnthn I'd imagine so. It doesn't sound like a great option. 17:50
TimToady but acutally, knowing how they design hydraulics, it's unlikely to be hydraulics cause this kind of issue, though it could be mechanicals closely associated with the hydraulics
generally with simple hydraulic failure you just lose 25% of your slew rate or so, assuming four systems 17:51
and if you lost all your hydraulics, you'd be in the news
jnthn Aye... Anyways, all worked out well in the end. 17:52
Think I'll go find some dinner while there's still places open :)
TimToady well, er, continue to have the appropriate amount of fun!
masak timotimo: luminous.elcenia.com/story.shtml 17:53
jnthn :P
bbl
17:56 colomon left 17:59 colomon joined
masak presently running the spectests on my Linux install on the laptop. getting lots of "Dubious" with test files returning 1 (wstat 256, 0x100). 18:00
hm. reading the output more closely, seems 'make spectest' failed to fast-forward t/spec properly. false alarm -- so far. 18:03
18:08 thou left 18:17 iSlug left 18:18 dmol left
masak heh, this second run of 'make spectest' looks much better. :) 18:19
18:20 dmol joined 18:22 cjwelborn joined 18:26 pecastro_ joined 18:28 pecastro left 18:30 denis_boyun__ left
Tene masak: Yes, luminosity was a fun read. 18:31
18:38 denis_boyun joined
masak Tene: moritz suggested it. I haven't started reading it myself yet. 18:39
Tene: though today (through the hpmor subreddit), I found parahumans.wordpress.com/table-of-contents/ which is supposedly very much worth reading.
18:41 iSlug joined, sftp_ joined
Tene It was fun, but somehow I never got around to finishing the latest third or so. 18:42
18:43 sftp left
masak huh. 18:49
18:54 Rotwang joined 18:55 risou is now known as risou_awy 18:57 domidumont left 19:06 sftp joined 19:08 zwut00 joined, zwut00 left 19:10 sftp_ left 19:12 denis_boyun left 19:21 skids left 19:26 SamuraiJack_ left 19:31 ksh left 19:45 pernatiy joined 19:47 lichtkind joined 19:51 iSlug left 19:52 rindolf left, denis_boyun joined 19:53 denis_boyun left 19:54 spider-mario joined 20:01 skids joined
lue hello world o/ 20:14
masak hellue \o
timotimo's origin/awesome_bindattr is a fast-forward merge from nqp's origin/master. it passes all tests. should I merge it into master?
or, to put it this way: is there any compelling reason not to?
I haven't tried spectesting Rakudo on top of the merged version. maybe I should. 20:15
nwc10 that would seem wise 20:16
masak :)
nwc10 how long does it take to build and run the tests of what goes into Star?
20:24 kaare_ left
masak it will take another ~30 minutes until I get home and can start the release engineering. 20:27
it would be very useful if someone could build Rakudo on top of nqp's awesome_bindattr branch, to see if Rakudo passes all the spectests.
if that doesn't happen before I get home, I may simply decide on the safer option of releasing without merging. 20:28
(because it's late and I'm tired)
nwc10 I'm not in a position to do this 20:29
(It is late and we all had a bad night last night, and we may get another one)
diakopter "we all"? 20:31
masak guesses dad + mom + child
nwc10 correct, although I wouldn't spell mom that way :-)
actually, "bit of a bad night" is more accurate, but it's enough 20:32
masak .oO( you Brits, you mummify words )
as for me, I speak and write International English, which is a euphemism for scavenging BE and AE for the best parts and using a pidgin of those 20:33
(as taught by Swedish television) :) 20:34
detrain & 20:39
20:41 spider-mario left
timotimo masak, in the backlog i wrote i spectested it 20:41
diakopter jdk7 skipped from update 25 to update 40? 20:46
meh, okay.. 20:48
timotimo one increment per security hole plugged 20:56
20:57 djanatyn left
skids
.oO(Because "Mom" better damn well be capitalized! :-)
20:57
21:00 zwut00 joined 21:03 skids left, benabik left 21:07 djanatyn joined 21:09 Rotwang left, alester left, Tene joined, Tene left, Tene joined 21:10 sqirrel left
hoelzro is there a (working) way to get a file's modified time? 21:10
21:11 dmol left 21:13 BenGoldberg joined
FROGGS nqp: nqp::statfh() 21:14
camelia nqp: OUTPUT«Error while compiling block : Error while compiling op statfh (source text: "nqp::statfh()"): No registered operation handler for 'statfh'␤current instr.: '' pc 52004 (src/stage2/QAST.pir:18966) (src/stage2/QAST.nqp:3186)␤»
FROGGS nqp: nqp::stat)
camelia nqp: OUTPUT«Confused at line 2, near "nqp::stat)"␤current instr.: 'panic' pc 14693 (src/stage2/gen/NQPHLL.pir:5223) (src/stage2/gen/NQPHLL.nqp:279)␤»
FROGGS nqp: nqp::stat()
camelia nqp: OUTPUT«Error while compiling block : Error while compiling op stat (source text: "nqp::stat()"): Operation 'stat' requires 2 operands, but got 0␤current instr.: '' pc 52004 (src/stage2/QAST.pir:18966) (src/stage2/QAST.nqp:3186)␤»
FROGGS r: say ".".IO.^methods
camelia rakudo 6f8668: OUTPUT«open close eof get getc lines read seek tell write opened t print slurp spurt copy chmod IO path flush encoding d e f s l r w x z modified accessed changed say Str gist perl <anon> <anon>␤»
hoelzro ah, thanks FROGGS 21:15
FROGGS r: say ".".IO.modified
camelia rakudo 6f8668: OUTPUT«1379320960␤»
masak timotimo: oh, you spectested it? well, then I don't see a reason not to merge. 21:20
yoleaux 21:10Z <PerlJam> masak: Have a look at gist.github.com/perlpilot/6625489 and tell me your thoughts. I've de-emphasized Parrot a little bit and moved some stuff around. I'm waffling on actually changing the release guide because I can't quite decide if it's not better to just wait until moarvm/jvm are the standard backends.
masak PerlJam: is this something I should have a look at before I start release engineering? because I was thinking of doing that... now. 21:22
PerlJam masak: nah, it can wait
masak: or you could follow it and see if you get lost :) 21:23
dalek p: 1af1098 | (Timo Paulssen)++ | src/vm/parrot/ (3 files):
WIP on bind/set/getattr with hints
p: bd80cda | (Timo Paulssen)++ | src/vm/parrot/QAST/Operations.nqp:
this actually works!
p: 1e8a295 | (Timo Paulssen)++ | src/vm/parrot/QAST/Operations.nqp:
confusion about argument order cleared up.
masak heh :)
timotimo: awesome_bindattr merged. feel free to remove the branch.
Rakudo release process starts... now. 21:24
do not adjust your set.
PerlJam masak: good luck! 21:25
masak this month's release name: Shanghai.
why? because I say so, that's why.
PerlJam be sure to put that in the release announcement :)
masak (they have a .pm group, and I visited Shanghai in July.)
that's way less crazy than some of the past code names. :) 21:27
"Bicycle"!?
:P 21:28
timotimo masak: you could have merged the failure branch thingie, too
i think
Tene Just don't name it "release name"
timotimo oh, actually i have not spectested that
dalek rl6-roast-data: 60a3879 | coke++ | / (5 files):
today (automated commit)
21:30
masak timotimo: then let's wait.
timotimo sure, no problem 21:31
masak Tene: is "release name" a .pm group?
timotimo maybe hoelzro's future work on it will create better performance gains
PerlJam recently created CorpusChristi.pm (just saying ;-)
Tene masak: Not currently.
I'd advise against changing that, however. 21:32
masak oy vey, the ChangeLog hasn't been updated much lately, has it? 21:33
timotimo oh my :(
have we even changed anything?
i think we've all just been sitting on our butts
all the pull requests i see relate to jvm, chdir, shell and panda for jakudo 21:36
dalek kudo/nom: a8ea2a5 | masak++ | docs/announce/2013.09.md:
[docs/announce/2013.09.md] draft announcement
21:38
masak eyeballs welcomed.
timotimo: I can't imagine that with all the lizmat activity, we havn't changed anything!
21:39 donaldh joined
masak but let's find out. ChangeLog is next. 21:39
masak pops open a stout
PerlJam masak: commits 5caf9288d38b2225df879f58acf003f9ecbcc60b and 283f763a5e670e66951ec837a761088f26e9cad9 should make it into the ChangeLog IMHO :)
(just reading the git log right now)
masak ok. 21:40
masak .oO( people are so full of suggestions, but do they commit to the ChangeLog? noooooo ) :P
dalek ast: 28715ba | (David Warring)++ | S05-mass/charsets.t:
more charset tests, including unicode classes
21:41
PerlJam masak: if you look at the release_guide.pod, you'll see that step one is to remind people to update the ChangeLog a few days before the release ;) 21:42
21:42 xinming left
masak well, shoot. 21:42
timotimo hehe
masak will a retroactive reminder do?
or is it enough if I berate you for failing to remind me to remind you to update the changelog? :P 21:43
21:44 xinming joined
FROGGS masak: there is no parrot release, no? 21:46
PerlJam heh
FROGGS because there actually is a patch which would be nice having 21:47
PerlJam FROGGS: you should read my reformulation of the release_guide too: gist.github.com/perlpilot/6625489
:-)
21:47 krokite left
PerlJam FROGGS: which patch? 21:47
FROGGS my patch :P 21:48
timotimo JVM sockets could go in, too
FROGGS PerlJam: a bugfix about utf8 character sequences being split by chunksize
PerlJam aye, I've already found it in the parrot logs :) 21:49
FROGGS++
FROGGS hehe, I guess it wasn't too hard to find *g*
PerlJam (I've stopped paying attention to parrot, so I hadn't noticed that you'd even submitted a patch)
masak FROGGS: I will simply follow the current release_guide, and I suspect that will end up without your patch this time :/ 21:50
FROGGS: I'm sorry there wasn't a Parrot release.
it would've been nice to get your patch in.
FROGGS masak: well, as it sounds only you and me had the issue anyway 21:51
masak troo
21:59 trnh joined, trnh left, trnh joined
dalek kudo/nom: d93f1da | masak++ | docs/ChangeLog:
[docs/ChangeLog] update since last release

Highly subjective; please add stuff I missed, probably unintentionally.
22:04
p: 931fd2b | duff++ | docs/release_guide.pod:
minor wording change
masak PerlJam++ 22:05
masak makes the corresponding change in the release announcement
PerlJam afk & 22:06
dalek kudo/nom: aa1417b | masak++ | docs/announce/2013.09.md:
reword slightly
22:07
kudo/nom: 733c824 | masak++ | docs/announce/2013.09.md:
[docs/announce/2013.09.md] some of the changes
22:09
22:09 trnh left
diakopter masak: learnperl6.com makes me cry 22:10
*headfloor*
masak that has absolutely nothing to do with release engineering.
focus! :)
diakopter first I cry from laughing so hard.. then I cry from the sadness of it 22:11
dalek kudo/nom: dbaa372 | masak++ | docs/announce/2013.09.md:
[docs/announce/2013.09.md] add contributors
22:16
masak diakopter: while crying, please review the release announcement.
masak .oO( "...presently" ) 22:17
22:19 BenGoldberg left 22:23 lichtkind left
dalek kudo/nom: 965ee80 | masak++ | docs/release_guide.pod:
[release_guide.pod] move current release up
22:23
kudo/nom: 84b953e | masak++ | docs/release_guide.pod:
[release_guide.pod] add some future release dates
22:24 FROGGS left
masak descending into NQP release sub-process. 22:24
22:25 pmurias joined
masak indeed, by the instructions of the NQP release guide, I should *not* bump the Parrot revision. 22:26
dalek p: 9441e1c | masak++ | VERSION:
bump VERSION to 2013.09
22:27
pmurias .tell jnthn why is NQPRegex defined twice? (as a knowhow in the setting, and as a class in the regex lib) 22:28
yoleaux pmurias: I'll pass your message to jnthn.
22:39 skids joined
masak NQP tag pushed. 22:40
dalek ast: f7cb1e8 | (David Warring)++ | S05-mass/charsets.t:
:Punct => :Punctuation
22:41
masak nqp-2013.09.tar.gz uploaded. 22:42
popping the call stack; back to your regularly scheduled Rakudo release engineering. 22:43
diakopter returns from biweekly twitter turn-the-other-cheek-to-p6-bashing
dalek kudo/nom: 337cf72 | masak++ | tools/build/NQP_REVISION:
[release] bump NQP revision
22:44
kudo/nom: f86a1a8 | masak++ | VERSION:
[release] bump VERSION
diakopter except lately it's been more like turn-the-other-cheek-to-raiph-reactionism 22:45
er, reactionism-to-raiph 22:46
masak doy++ # twitter.com/doyster/status/380009297998282752
REPLeffect diakopter: out of curiousity, what about learnperl6.com makes you cry (with the caveat that I have only glanced at the main page) 22:47
masak REPLeffect: it's either the CGI or the PIR, I'll bet. 22:48
REPLeffect what about the CGI?
masak CGI ceased to be a good idea at least a decade ago. PIR was never a good idea.
22:49 markov left
diakopter well. 22:49
lue
.oO(But... CGI is the future!)
yoleaux lue: Sorry: that command is a web-service, but it didn't respond in plain text.
REPLeffect :D 22:50
lue w-what?
.oO(Webservice)
diakopter
.oO(But...
yoleaux diakopter: Sorry: that command is a web-service, but it didn't respond in plain text.
diakopter
.oO(But
.oO(But..
yoleaux diakopter: Sorry: that command is a web-service, but it didn't respond in plain text.
diakopter
.oO(But.
yoleaux diakopter: Sorry: that command is a web-service, but it didn't respond in plain text.
diakopter
.oO(GET.
yoleaux diakopter: Sorry: that command is a web-service, but it didn't respond in plain text.
geekosaur o.O 22:51
REPLeffect oh, geez. Has that page not been updated since 2009?
22:52 markov joined
lue .TestingYoleaux. 22:52
yoleaux lue: Sorry: that command is a web-service, but it didn't respond in plain text.
lue .TestingYoleaux 22:53
diakopter REPLeffect: well, it describes itself as "document as I learn Perl 6, for others who want to learn Perl 6"... but then it teaches some PIR
REPLeffect anyone know who's domain it is? 22:54
pmurias heh, learnperl6.com uses the horrible parrot Env PMC I had the pleasure to remove from nqp
masak pmurias++ 22:55
diakopter REPLeffect: whois says Lyle Hopkins
REPLeffect was just getting there :D
masak it's after midnight here. do I mark the release date on Wikipedia as 2013-09-19 or 2013-09-20. 22:56
I know LylePerl. he's here occasionally.
he was at YAPC::EU.
diakopter: he was the one who lightning talked about helping a designer.
REPLeffect being that's a hosting site, I doubt Lyle spends much time with Perl 6 (or that he's ever spent any) 22:57
lue masak: I recommend consulting UTC.
masak well, then I have another hour. should be fine.
REPLeffect: correct. 22:58
diakopter masak: orly 22:59
23:01 donaldh left
masak yarly 23:05
rakudo tarball uploaded. 23:08
p6c announcement email sent. 23:11
23:11 Mouq joined
masak Wikipedia article updated. 23:17
done. standing down release mode.
conveniently, the beer is already finished -- so I can skip the celebrating, too :) 23:18
...and go directly to bed.
diakopter gnaight
masak 'night, diakopter, #perl6
23:23 risou_awy is now known as risou 23:30 Pleiades` left 23:31 Pleiades` joined 23:38 BenGoldberg joined 23:46 jnap1 left 23:47 jnap joined 23:52 EvanTeitelman joined 23:54 Mouq left