»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.
00:00 pecastro joined
Mouq timotimo: Indeed :) 00:01
timotimo that doesn't fix the path fuckup :) 00:03
do you have an idea of the whole makefile stuff?
Mouq me? nope.. :/ 00:04
timotimo hm.
Mouq isn't even sure where the $(DESTDIR) is defined in Rakudo 00:07
00:07 pmurias left
Mouq Ahh tools/distro/rakudo.spec 00:07
I think.... 00:08
No? Apparently I know how none of this stuff works 00:09
timotimo :) 00:10
i've been staying away from makefiles any more complex than a few lines with only very few variables in them
lue I prefer cmake whenever possible :) 00:12
00:12 dayangkun joined
timotimo i prefer someone else does it 00:15
whenever possible
but since it's jnthn who's been banging his pretty head against this piece of work ...
Mouq Okay, DESTDIR shouldn't actually be defined in the Makefile 00:17
Mouq is a newb
timotimo okay, seems to be a normal thing 00:19
00:19 fridim__ joined
timotimo how about PREFIX, though? 00:21
lue Where things are installed, e.g. PREFIX=/usr/local and then BINDIR=${PREFIX}/bin ... 00:22
timotimo right; but does it get set by something outside of the Makefile? 00:23
because i can only see PREFIX_DIR = @prefix@ there
lue timotimo: perl Configure.pl --prefix=/usr/local :)
wait, PREFIX is the one that's not used, right?
timotimo i didn't give it any --prefix, is "" the default?
PREFIX is used
Mouq PREFIX = @prefix@ in Rakudo's tools/build/Makefile-common.in 00:24
timotimo PREFIX_DIR also.
this here is Makefile.in from star, though
Mouq Then $(PREFIX) is wrong. It should be $(PREFIX_DIR) 00:25
lue timotimo: without --prefix, the script searches for where stuff is already installed, and takes that as the prefix. (at least for NQP and Rakudo; Parrot defaults to /usr/local, Moar defaults to wherever isn't typical, forget the exact path)
Mouq Or change PREFIX_DIR to PREFIX
In Star's Makefile.in
timotimo that caused badness :)
lue well, it only searches for an in-source installation, last time I tried. That's why I have to do --prefix=/usr/local for NQP and Rakudo 00:26
Mouq timotimo: PREFIX -> PREFIX_DIR causes badness? 00:27
timotimo it seemed so, aye
Mouq :?
timotimo /usr/bin/perl tools/build/bin-install.pl /tmp/rakudostar/rakudo-star-2014.03/install/bin/perl6-m /tmp/rakudostar/rakudo-star-2014.03/install/tmp/rakudostar/rakudo-star-2014.03/install/bin modules/ufo/bin/ufo-m modules/panda/bin/panda-m modules/doc/bin/p6doc-m
there is no /tmp/rakudostar/rakudo-star-2014.03/install/bin/perl6-m for example 00:28
Mouq timotimo: Are you sure that's not just because it got further in the build process and ran into another bug? 00:29
timotimo hm 00:30
mayhaps.
let me try again.
oh, hold on 00:33
something's moving 00:34
thanks for making me try again :)
Mouq timotimo: Np, any progress?
lue Apparently PREFIX = $(PERL) $(CURDIR)/tools/star/prefix.pl , so I imagine it serves a very different purpose from PREFIX_DIR
Mouq It compiles! github.com/Mouq/Archive-Tar/blob/m...ve/Tar.pm6 00:35
lue: Where did you get that? 00:36
lue tools/star/Makefile
Mouq Oh bluh
Okay :|
timotimo base64 is failing, as perhaps expected
(on jvm, that is) 00:37
Mouq lue: That's used differently though
That's used as a program prefix in tools/star/Makefile
And the Makefile.in uses it like a DIR prefix 00:38
lue Mouq: yeah, I'm just stating that trying to rename one to the other won't work :)
(or, more specifically, the reason why it didn't work before)
Mouq lue: Ah, ok
lue Yeah, Makefile.in is using $(PREFIX), not sure why that is actually. (Of course, I'm not quite sure what prefix.pl is even doing, so... :P) 00:39
timotimo Can't deserialize_stub a NativeCall object. 00:42
that's known to fail on jvm?
Mouq prefix.pl is just renaming directories to have a certain prefix :P like, all the Rakudo stuffs need to reference the rakudo/ directory
(Apparently) 00:43
lue timotimo: don't know :/
timotimo good lord, jvm module installation precompilation
is there ever a wait
BenGoldberg Just to reduce confusion, I would want to rename the PREFIX that goes to prefix.pl ... maybe a name like PREFIX_PL. 00:44
timotimo ufo doesn't have multiple -* binaries
lue That's right, star/tools/Makefile is for the tarball-maker, Makefile.in is used by star when you're compiling it. So they can each use PREFIX how they wish, I would think.
timotimo so it fails to install (with the changes i made)
(which explains why it fails i bet :P )
Mouq PREFIXIFIER 00:45
$(BEINGAMAKEFILEVARIABLEMAKESMEANGRY)
lue CHPREFIX, perhaps? :) 00:46
00:50 pecastro left
timotimo fortunately i can ctrl-c through all of the module precomp without the makefile getting shot down m) 00:51
conflicts during deserialization 00:53
on moarvm
Cannot read file 'modules/panda/bin/panda-m' for installing it: No such file or directory at tools/build/bin-install.pl line 12.
and this of course
well, this ought to get jnthn a tiny bit closer to what he needs to see without too much pain 00:54
dalek ar/tristar: 7be70fd | (Timo Paulssen)++ | / (2 files):
fix some more problems
 00:55
kudo-star-daily: 8915768 | coke++ | log/ (5 files):
today (automated commit)
 01:00
kudo-star-daily: b021630 | coke++ | log/ (5 files):
today (automated commit)
rl6-roast-data: 6e94899 | coke++ | / (6 files):
today (automated commit)
rl6-roast-data: d302cbe | coke++ | / (3 files):
today (automated commit)
01:03 hoverboard joined
Mouq timotimo: How do you test stuff in the star repo? 01:04
nvm: make -f tools/star/Makefile 01:08
Hm.. maybe not 01:09
timotimo that's about it, like jnthn explained above
you make that, then you make -f ... release VERSION=2014.03 01:10
you'll get a tarball from that, extract it somewhere and perl Configure.pl etc etc
edit the tools/build/Makefile.in and Configure.pl there and test it all out, copy back what works to the original
Mouq Ironically, it doesn't work with my version of tar
After make -f tools/star/Makefile: 01:11
tar --xform 's!parrot-[^/]*!parrot!' -xvzf src/parrot-5.9.0.tar.gz
tar: Option --xform is not supported
timotimo huh?
well, you only just started implementing tar, so maybe that's just not implemented yet :P
01:12 rurban1 joined
geekosaur mac by any chance? 01:12
Mouq geekosaur: Yuppers
geekosaur --xform is a gnu tar option; macs have BSD tar 01:13
(aka libarchive tar)
timotimo want my tarball? 01:15
geekosaur suggests if you want to do that somewhat portably, you want pax and its -s option 01:16
Mouq just pushes 01:17
timotimo it's "only" 14 megabytes
Mouq Or.. not
Mouq doesn't have permissions anyway 01:18
timotimo: Could you try: gist.github.com/Mouq/13827a05e35f5dc0ceb3
It should give the same results
:P
But I think it's de facto Makefile law that thee must use DESTDIR 01:19
timotimo only 9.5 megabytes with xz :)
isn't that the same changes i made? 01:20
oh
i made them and forgot to push
apparently >_<
huh. 01:21
did you pull before you made these changes?
Mouq timotimo: Yup
timotimo but with my changes it already ran pretty far?
Mouq timotimo: Dunno
:P
Can't run tar
Or tar --xform anyway 01:22
timotimo right 01:23
i guess i'll apply your patch and shoot a tarball your way?
can you extract .xz?
Mouq I should...
be able to 01:24
timotimo: But don't worry about it :)
I don't think I'll be of much help
timotimo i'd like you to be, though :P 01:25
01:29 clkao left
timotimo dropbox isn't connecting right now, i'll have to try something else 01:30
01:30 _thou joined, clkao joined
timotimo okay it's going up 01:31
"upload error" 01:34
Mouq geekosaur: pax won't really work because it's not integrated with gzip :/ 01:36
geekosaur: So it'll work, but I'm not sure that's a big boost in portability. And I don't know if cygwin supports pax…
geekosaur gzip -dc foo.tar.gz | pax -r ... 01:37
Mouq geekosaur: Can't pipe, not portable :P
I don't think
timotimo www.dropbox.com/s/1ttk2k8pr2m18ec/....03.tar.xz 01:38
try this
Mouq (You might be right, of course, and I'm just being disagreeable)
geekosaur pretty sure windows has had real pipes since windows nt. other platforms, well, whcih ones are we targeting? 01:39
timotimo commuting time! 01:40
geekosaur also, if you are worried about portability then you cannot rely on gtar because it uses a pipe to run gzip
01:40 hoverboard left
geekosaur it's not builtin 01:40
Mouq timotimo: Let me try getting it to work now that I've realized I can't just `curl` from the dropbox website :L 01:42
01:43 rurban1 left
Mouq timotimo: Sweet :) Thanks 01:45
Oh, also o/ timotimo :P 01:46
Mouq wasn't paying attention
geekosaur: K
geekosaur: I have no idea what I'm talking about in that case
*in most cases 01:47
01:50 logie_ left, logie_ joined
Mouq So, it's not done yet 02:07
But, with the tarball timotimo++ sent me, I'm almost done installing most of a Jakudo star… 02:08
Looks like MIME::Base64 and its dependencies failed, as well as DBDish and JSON::RPC 02:10
And Digest::MD5
Which ought to, the module it basically a single Q:PIR block :P 02:11
Oh
== Installing binaries for JVM
/Users/amoquin/perl5/perlbrew/perls/perl-5.18.1/bin/perl tools/build/bin-install.pl /Users/amoquin/Dev/Perl6/tmp2/rakudo-star-2014.03/install/bin/perl6-j /Users/amoquin/Dev/Perl6/tmp2/rakudo-star-2014.03/install/bin modules/ufo/bin/ufo-j modules/panda/bin/panda-j modules/doc/bin/p6doc-j
skids Was about to say...
Mouq Cannot read file 'modules/ufo/bin/ufo-j' for installing it: No such file or directory at tools/build/bin-install.pl line 12.
make: *** [modules-install-j] Error 2
Mouq spoke far too sone 02:12
Mouq refuses to correct sone and proclaims it a new word
lue always thought it was a bad idea to write Perl 6 modules in not-Perl 6 ... 02:13
Mouq A "sone" appears to be a "subjective unit of loudness for an average listener […] that has an intensity 40 decibels above the listener's own threshold of hearing"
Sorry for speaking so loudly guys
02:15 clkao left 02:17 clkao joined 02:22 clkao left
timotimo Mouq: 'tis okay, i was on the freeway doing 80 kph 02:22
so i didn't hear you speak that loudly
Mouq timotimo: Oh, good :) 02:23
02:23 clkao joined
Mouq running make install again with [ufo-j,panda-j,p6doc-j]>>.subst(/-j/,'') 02:23
timotimo Mouq: you know in ~12 hours i'll do a p6weekly post, you feel free to make all that stuff work until then for great karma points :D 02:24
Mouq timotimo: :D... except: school D:
timotimo if your school is anything like mine, you'll have plenty time for Perl 6 in between breaks :P 02:25
Mouq timotimo: I'm thinking my school is a little bit more "high school"-y than yours 02:27
timotimo heh.
02:28 clkao left 02:29 clkao joined, [Sno] left
retupmoca MIME::Base64 HEAD should compile fine on non-parrot (granted, I only tested on moar) 02:48
Mouq: what revision of MIME::Base64 are you trying to compile?
timotimo retupmoca: i think the modules in the star repo are quite a bit behind, actually 02:50
retupmoca ah 02:51
I fixed non-parrot compilation only a couple of days ago
timotimo i remember that fondly :)
Mouq retupmoca: 76fd9decc7c1445
02:51 klapperl_ joined
timotimo the coolest way to push a whole bunch of submodules forwards, btw, is to use git submodule foreach in conjunction with gitk --all 02:51
because then you can right-click on the commit you want and "reset master branch to here" 02:52
and when you close gitk, the next one will come right up
until you're done
retupmoca Mouq: that's branch precompile-fix 02:53
Mouq: if you update it to HEAD of master branch, it should compile
timotimo good night #perl6 :) 02:54
02:54 xenoterracide left
timotimo Mouq: maybe you can figure out how to make the modules that create binaries create binaries for each of the backends; though i think froggs' eleven branches of panda and friends have something for exactly that 02:54
Mouq timotimo: I think the simplest way to do the final line (ufo-j, panda-j, p6doc-j) is to have bin-install.pl take "j" as an argument and write both a.) bin/whatever, like it used to, and b.) bin/whatever-j that's just perl6-j calling bin/whatever
02:55 klapperl left
Mouq timotimo: The modules already have logic to handle stuff based on what perl6 executes them 02:55
timotimo ah
Mouq retupmoca: I don't have push access to rakudo/star :(
timotimo Mouq: i'm not allowed to give you access :(
fortunately, you can fork on github easily :) 02:56
with the github gem, it's just "hub fork"
Mouq timotimo: I don't know though, I don't know what stuff's going on with the eleven branches
timotimo it'll even set up the remotes for you
Mouq timotimo: I'll make a PR :)
timotimo maybe FROGGS++ will help me write up a couple of paragraphs on what eleven already can do (which was a surprisingly big amount last time i asked) and what's to be expected in the near future 02:57
Mouq: right now or some time in the future?
as in: should i go to bed now or wait for the PR?
it probably makes no big difference 02:59
Mouq timotimo: Bed with you :) I need to figure some stuff out
timotimo o/
Mouq o/ timotimo
timotimo aye
best of luck and have fun in school :)
Mouq timotimo: Thanks, sleep well :)
03:55 xragnar_ joined, xragnar left, xragnar_ is now known as xragnar 04:07 araujo left 04:17 fridim__ left 04:23 kaare__ joined 04:27 Mouq left 04:29 [Sno] joined 04:49 BenGoldberg left 04:58 SamuraiJack joined 05:08 aborazmeh left 05:29 kaare__ left 05:42 Mouq joined 05:46 Mouq left 06:03 kaare__ joined 06:10 [Sno] left 06:13 wbill joined 06:19 darutoko joined 06:38 kaare__ left, kaare__ joined, _thou left 06:47 hoverboard joined 06:54 Alina-malina left 06:55 retupmoca left 06:56 retupmoca joined 07:04 dmol joined 07:22 vendethiel left 07:26 logie_ left, logie_ joined 07:30 Mouq joined 07:35 Mouq left 07:53 hoverboard left 07:56 zakharyas joined 07:59 SamuraiJack left
dalek ar/tristar: 786ece4 | Mouq++ | tools/build/Makefile.in:
fix some less problems

DESTDIR should still be there, by de facto Makefile law, AFAICT		 08:03
ar/tristar: 4a39421 | Mouq++ | tools/build/ (2 files):
Let ufo, panda, and p6doc figure out what they're running on
ar/tristar: 469f8c0 | Mouq++ | modules/Perl6-MIME-Base64:
Update MIME::Base64
ar/tristar: d0da84f | moritz++ | / (3 files):
Merge pull request #36 from Mouq/tristar

Some updates to tristar
moritz .tell Mouq I've given you a commit bit for rakudo/star 08:07
yoleaux moritz: I'll pass your message to Mouq.
08:17 sftp_ left 08:21 skids1 joined 08:23 dmol left, takesako____ joined, skids left 08:24 xinming__ joined, Juerd_ joined, xinming_ left, takesako___ left 08:28 dylanwh left, atrodo left, Juerd left, dylanwh joined, Juerd_ is now known as Juerd 08:29 atrodo joined 08:33 dmol joined 08:39 pdcawley joined 09:03 kaare__ is now known as kaare_
masak morning, #perl6 09:12
moritz good morning, masak | #perl6 09:13
09:13 Alina-malina joined 09:14 Alina-malina left, Alina-malina joined 09:15 colors joined, Alina-malina left 09:16 colors is now known as vincent21, Alina-malina joined 09:17 fhelmberger joined
nwc10 good #perl6, morning. 09:18
JimmyZ good moar, perl6 09:24
09:31 Alina-malina left 09:32 Alina-malina joined, Alina-malina left 09:39 Alina-malina joined 09:53 lizmat joined 10:01 salv0 left, salv0 joined 10:09 xinming_ joined 10:12 xinming__ left
Ulti new python release has asyncio in case anyone is interested in what it looks like docs.python.org/3.4/library/asyncio.html 10:14
yoleaux 11 Mar 2014 18:27Z <raiph> Ulti: warn ^Inf # I love the idea of elegant golf with P6. For example, refusing to omit spaces when it would work but be ugly.
10:15 isacloud_ left, isacloud_ joined 10:18 woolfy joined 10:21 [Sno] joined 10:24 pmurias joined
pmurias jnthn: what would be a portable way of having a list of tests to run for the js-test target for the js nqp backend? 10:26
dalek p-js: 0844fa0 | (Pawel Murias)++ | src/core/NQPMu.nqp:
Temporarly add nqp-js hacks to the setting.
 10:27
p-js: 6b6e179 | (Pawel Murias)++ | tools/build/Makefile-JS.in:
Build a QASTNode for the js backend.
p-js: fb5b4b7 | (Pawel Murias)++ | tools/build/Makefile-JS.in:
Build QRegex for use by the js backend, pass more tests.
lizmat decommute& 10:28
10:28 lizmat left 10:29 woolfy left 10:32 pecastro joined 10:33 kurahaupo joined
pmurias jnthn: do we want to expose nqp::ordat and nqp::ordfirst to the user? 10:37
dalek p-js: f4c1e4b | (Pawel Murias)++ | t/nqp/59-nqpop.t:
Add a test for two argument nqp::ord.
 10:40
p-js: 6bc483c | (Pawel Murias)++ | src/vm/js/nqp-runtime-core/runtime.js:
Implement nqp::ordat and nqp::ordfirst in the js backend.
10:43 sftp joined 10:55 Mouq joined
dalek p-js: c0c1aad | (Pawel Murias)++ | tools/build/Makefile-JS.in:
Use the full nqp setting in the js backend (that makes subst work).

Also use a saner suffix for the concatenated setting (.nqp instead of x)		 10:57
10:58 aborazmeh joined 11:00 Mouq left 11:02 rindolf joined 11:08 [Sno]_ joined 11:09 [Sno] left, [Sno]_ is now known as [Sno] 11:10 denis_boyun joined 11:15 tadzik left 11:17 masak left, [Coke] left, PerlJam left, synopsebot left 11:18 Util left, dalek left 11:21 yves__ joined, tadzik joined, synopsebot joined 11:22 masak joined 11:23 Util joined, masak is now known as Guest4534, [Coke] joined, PerlJam joined 11:36 colomon left 11:40 daniel-s__ left, daniel-s__ joined 11:43 telex left 11:44 dmol left, telex joined 11:53 daniel-s__ left 11:57 aborazmeh left 11:59 daniel-s__ joined 12:05 colomon joined 12:08 skids1 left
[Coke] I wonder if I'm the only one who read "asyncio" in backscroll in the voice of someone casting a spell in HP. 12:22
colomon [Coke]: I am reading it that way now, and probably forever more... 12:23
[Coke] Infected! woot!
timotimo is deciding not to engage naysayers on twitter 12:25
colomon timotimo: pretty much always someone wrong on the Internet.... 12:26
timotimo in this case it's pretty obvious they haven't taken the time to actually look.
12:29 SamuraiJack joined 12:30 guru joined, guru is now known as Guest64406 12:31 Guest64406 is now known as ajr_
[Coke] twitter leads me to: github.com/timbunce/java2perl6 12:38
whose last commit is 4 years ago. might need some love. 12:39
timotimo yes, it seems like a very different approach to java interop than the jvmbootinterop we have in rakudo-jvm
12:41 colomon left
timotimo a friend pointed out, that if you visit perl6.org, it doesn't really tell you what perl6 is about at all. 12:41
[Coke] ENOMARKETING 12:42
or, very little, anyway.
timotimo yeah
tadzik heh
if you don't know what Perl 5 is, you'll have no idea what Perl 6 is 12:43
[Coke] seen timbunce?
.seen timbunce?
yoleaux I haven't seen timbunce? around.
[Coke] .seen timbunce
yoleaux I haven't seen timbunce around.
12:43 Mouq joined
Maddingue [Coke]: timbunce is still in Paris (QA hackathon) 12:46
act.qa-hackathon.org/qa2014/wiki?node=Travel
s/Paris/Lyon/ 12:47
timotimo the hackathon is already over, though. isn't it?
Maddingue yes, ended yesterday
12:48 Mouq left
timotimo OK 12:50
13:01 SamuraiJack left 13:10 cognominal left, pecastro left, cognominal joined 13:14 kbaker_ joined 13:26 PZt left 13:27 xenoterracide joined 13:30 araujo joined
timotimo i'm glad i could enthuse Mouq for the build system job 13:34
at least a little bit :) 13:35
13:43 sctt joined
pmurias timotimo: re java2perl6, I think the goal behind that wasn't interop but to use the java database driver api for perl6 13:49
timotimo oh, hm 13:50
nwc10 "are we nearly Star yet?" :-)
timotimo :)
13:50 sctt left
timotimo well, when i put the type-based dispatch into jvmbootinterop (knock on wood), a JDBC driver will be much nicer to make, i reckon 13:51
13:53 colomon joined 13:55 prammer left
timotimo it's already possible to make, but you're going to have to supply all the full method signatures directly in the code 13:57
well, not all of them, but most by far
13:58 me5purnd joined 13:59 sbauer left 14:03 kaare_ left 14:07 rurban1 joined
pmurias timotimo: the JDBC driver api was intented to be used on different platforms (like parrot then and moarvm nowdays) 14:08
timotimo ah
well, that's not that easy :)
pmurias with a nice perlish DBI api on top
timotimo you'll have to have a java process somewhere in there
pmurias and with drivers themself being written in Perl 6 14:09
14:09 _thou joined 14:12 sbauer joined
me5purnd perl6 devs now moving too fast ... please *slow down* development to normal speed :-\ 14:12
j/k 14:13
(bad)
timotimo i wouldn't mind Perl 6 development speeding up some more 14:14
PerlJam
.oO( I wouldn't mind a Perl 6 implementation speeding up some more ;)
 14:15
timotimo if brrt gets accepted for GSoC for the moarvm jit, that seems like a very real possibility in the near term 14:16
and rurban++ is giving parrot some love in the performance department 14:18
me5purnd rakudo runs on moarvm jvm and parrot now ? ... but not .NET
timotimo that's right
me5purnd and niecza:only run on .NET
timotimo yes, sadly niecza is falling behind a bit on spec compliance 14:19
me5purnd but higher level niecza <----> rakudo are simlar
timotimo niecza has really good performance characteristics, though
me5purnd ah ok
timotimo well, niecza has been built from the start for performance, rakudo has been developed while mostly ignoring performance
there's still lots of non-optimal things in rakudo; for example, compiling the core setting gives you a few hundred thousand copies of the string "OPER" or "dotty" :) 14:20
i tried to build an inlining solution for that, but it caused compilation failures instead of improvements :(
interning*
me5purnd hmm so both are at work improving ... and it's not like rakudo is going to use nieczca work to port itself to .NET and become the universal perl6 14:21
14:21 _thou is now known as thou
timotimo yes, niecza is not suited to become a "target" for rakudo 14:22
but at this point it would be "easy" to make a .net backend for rakudo
me5purnd :)
timotimo the backend is now very well separated from the frontend, so the backend-dependent bits are nicely compartmentalized
and you have lots of code you can look at to figure out what needs to be done
huf they only nabbed half of europe and half of asia
not nearly imperialistic :)
err. 14:23
timotimo i don't understand what you mean by that :|
huf wrong channel :)
timotimo that would explain it
was about to ask in what universe rakudo and niecza nabbed half of europe and asia
me5purnd that is good moar is better!! (oops meant to type more) why do people worry about there being too many perl6's anyway ? ... in the furture the fact that it runs on different vms etc is going to be like the fact that perl5 runs on HPUX and NeXT ... 14:24
huf :)
timotimo people on the outside of the community seem to think we're fragmenting ourselves too much 14:25
14:25 pecastro joined
me5purnd oh .. but ... mabye they don't understand VM's? 14:25
timotimo and i've seen the claim that we drop projects as soon as they are "not fun any more"
as in "parrot wasn't fun any more, so they started moarvm"
me5purnd oh
well maybe vmkit.llvm.org is -Ofun :-) 14:26
timotimo rurban is hopefully helping parrot catch up with moarvm's performance and multithreading capabilities
me5purnd wow
timotimo isn't llvm much too low-level for us?
me5purnd yeah
pmurias vmkit builds a vm on top of llvm 14:27
me5purnd for sure rurban would know :)
vmkti now has a sort of way of making a home built jvm 14:28
14:28 raiph joined
timotimo like the avian vm? :) 14:28
14:28 kivutar left
[Coke] in an ideal world, we wouldn't have needed moarvm, yes. We sadly are not coding in an ideal world. 14:29
me5purnd sometimes I have re.pl and node.js open in terminals and I accidentally type perl into node ....
timotimo we couldn't possibly have known from the start that 6model would be a thing
i can understand why the parrot team wanted to realize the "one vm for all dynamic languages" thing
it's quite sad it led to parrot becoming somewhat unideal for Perl 6 14:30
14:31 Mouq joined
pmurias how many gsoc Perl 6 proposal do we have this year? 14:32
14:34 vendethiel joined
me5purnd maybe parrot will be ideal for something else and almost ideal for perl6 ... ... real life parrots can talk any language without really deeply understanding it so ... 14:34
timotimo either that, or the analogy will break down ;) 14:35
i'm carefully optimistic.
me5purnd :)
the main thing people forget about parrots .... 14:36
14:36 Mouq left
me5purnd they fly 14:36
pmurias parrot is ideal for writing programs in the pir assembly language
;)
timotimo well ...
me5purnd yeah :)
me5purnd tells parrot something in mix of javascript, perl5 and 2001 Space Odyssey Dave Dialect ... 14:39
parrot flies off and comes back with a fish
14:40 jnap joined
rurban anyone knows what's up with dalek? is gone for #parrot 14:50
the problem with vmkit is that their GC is written in java 14:51
timotimo oh hey rurban :) 14:52
rurban and I didn't dig llvm C++ too much
bacek likes it though
me5purnd rurban: vmkit is very early days so far I gues ... I thought there was a plan eventually have non java GC ..one day ... 14:58
rurban dalek quit Mar 17 06:19:53 14:59
central time
me5purnd anyway ... viva perl6 (and 5) and perlito :) 15:00
15:00 me5purnd left
timotimo actually it's called "viv" without a :P 15:01
15:03 dalek joined, ChanServ sets mode: +v dalek 15:06 sbauer left
rurban and here he is, thanks 15:06
15:14 treehug88 joined, pmichaud joined
pmichaud good morning, #perl6 15:14
nwc10 good *, Pm 15:15
moritz good am, pm
colomon o/ 15:16
nwc10 pmichaud: how did the robot stuff go? 15:17
pmichaud My FLL team is attending the World Festival (championship) in St. Louis in April.
moritz congratulations!
pmichaud My FTC team made it as far as the South Super-Regionals (top 72 teams from 12 U.S. States) but didn't make it to the World Festival. Still, we're pretty happy about it given that this was the FTC team's rookie season. 15:18
nwc10 and then? :-)
"I've reached the top and had to stop and that's what's worrying me"? [except SpaceX is working on this]
pmichaud well, after the World Festival we start planning for next season. Oh, and there's MoonBots over the summer. :-)
nwc10 do MoonBots use a real Moon? 15:19
pmichaud made of LEGO bricks, yes. :)
the FTC team is WAAAY motivated to make it to the World Festival next year. I've rarely seen such a motivated group of individuals. 15:20
15:20 dwarring left
timotimo hello pmichaud! :) 15:21
rurban: do you have something notable for me to point out on the daily p6weekly?
froggs isn't here yet, i had hoped to ask him for a detailed explanation of all the eleven status 15:23
but i think something big is coming up this week or the next one.
so that may be a better point to write that up
15:25 jnap1 joined 15:26 BenGoldberg joined, jnap left
rurban timotime: only that the new parrot release tomorrow will support now -O2 pbc compile-time optimizations. 15:28
effects so far 3-5% perf improvements, not yet benched with rakudo
And I'm settting up a big buildbot for most architectures: blogs.perl.org/users/rurban/2014/03...ldbot.html 15:29
timotimo okay, i mentioned that -O2 thing in last week's post already 15:32
rurban the buildbot is fine then: mingw64 and accepting outside bots is still tricky
timotimo i'll post that. thanks! 15:33
rurban mozilla and google have hundreds of bot slaves
BenGoldberg only hundreds? ;)
timotimo hundreds per architecture maybe ;) 15:34
rurban trac.buildbot.net/wiki/SuccessStories (700)
raiph timotimo: did you see the feedback I added to your draft of last week's post? (I think I added them after you had already published the final.)
timotimo ah, that was you! 15:35
there is no way to identify who wrote what sadly
i think i saw it, i'm not sure if i took any of your feedback into the final post, though 15:36
rurban anamariamoz.wordpress.com/2010/11/0...ld-system/
timotimo istr it was partially redundant with what others told m
15:36 kaare_ joined
raiph i recall a couple nice things lizmat did and you didn't mention her at all 15:36
timotimo ah. that's not good at all! 15:37
this week i've seen lots and lots of spec commits from lizmat, though
so she'll be mentioned for sure this time
colomon lizmat++ 15:38
raiph timotimo: are you checking the changelogs? (she had already edited the changelog so I was surprised you didn't mention her changes) 15:42
timotimo i didn't look at the changelog for that week :(
rurban timotimo: And I started working on the parrot jit, but it will need at least a month. not yet sure if I'll use the old (tested and ok) or a new dynasm jit (luajit).
timotimo i guess i'll post about that when there's something to see :) 15:43
15:43 zakalwe left
JimmyZ rurban: I saw you reverted the old jit removing 15:44
rurban in rurban/jit: I need to bring in the perl5 Op library and the cgp runcore also. Esp. cgp will be quite fast than 15:46
15:46 Guest4534 is now known as masak
japhb__ timotimo: You had some perl6-bench questions for me, I believe. One about variants and one about components? 15:46
Did you figure them out already?
rurban we saw 20-25% perf with cgp in older days
JimmyZ rurban: Did you see the gist I msged to you? 15:47
timotimo i did not figure it out yet
do you happen to know how best to make %*ENV on parrot work with qx? 15:48
japhb__ OK, what would you like to know?
timotimo list-checkouts gives the same commit id for all components because the GIT_DIR variable doesn't stick :(
rurban JimmyZ: no
JimmyZ rurban: gist.github.com/Whiteknight/1198521
timotimo what kind of thing do we expect to be passed to --variants?
15:48 LLamaRider joined
japhb__ timotimo: Really? That's a regression. 15:48
timotimo yes, really :( 15:49
japhb__ timotimo: --list-variants will give you examples
timotimo could always call "env" instead of "git" directly
15:49 sbauer joined
japhb__ sorry, ./timeall --list-variants 15:49
timotimo ah
japhb__ That is a painful regression, actually, for anything using Perl 6 as glue. :-(
timotimo yes! 15:50
rurban JimmyZ: thanks. looks good, besides some things already done
timotimo rurban: you're a parrot person; can you help me look into the qx + environment thing?
JimmyZ rurban: yeah
japhb__ rurban: What timotimo said. :-)
rurban timotimo: ticket or link? 15:52
timotimo gimme a sec 15:53
github.com/rakudo/rakudo/blob/nom/...ol.pm#L228
we don't pass the environment here; do you know how we could?
i don't think nqp::open with pipe mode will allow us to specify env vars
rurban parrot has the env pmc, which is similar 15:56
timotimo do you feel like implementing that? or at least point me at the right way to do it? 15:57
rurban I'm trying it out... 16:00
timotimo thank you very much!
all the rurban++ for that
rurban nqp::getattr(%*ENV, EnumMap, '$!storage'); should be doable with parrot 16:01
timotimo okay; and how do we make the environment effective?
rurban openpipe
timotimo oh, i didn't realize it has an argument for that
rurban not yet 16:03
nqp::openpipe($cmd, nqp::cwd(), nqp::getenvhash(), '') is the idea 16:04
timotimo OK, do we have to map a new op for that?
well, since it's inside an #if parrot, we could use pir:: instead 16:05
rurban sub QX($cmd) {
my Mu $pio := nqp::openpipe(nqp::unbox_s($cmd), nqp::cwd(), nqp::getenvhash(), '');
fail "Unable to execute '$cmd'" unless $pio;
my $result = nqp::p6box_s(nqp::readallfh($pio));
nqp::closefh($pio);
$result;
}
I'll map it
timotimo thank you kindly 16:06
huh, latest rakudo.parrot daily run had 142 failures :o 16:10
jnthn pmurias: ordat and ordfirst are just things ord desugars to.
timotimo ah, ok S05
16:13 kaare_ left, denis_boyun left 16:15 kaare_ joined 16:19 Mouq joined
tadzik my moarvm-zavolaj test suite prints a lot of "/bin/sh: 1: PREFIX: not found". Just me? 16:20
rurban do you compile on windows with msys-mingw, strawberry-mingw or msvc? 16:23
16:24 Mouq left, kurahaupo left
tadzik nope, that's linux and gcc 16:24
timotimo tadzik: no, that's a problem with \$(PREFIX) ending up in the $*VM thingie
tadzik: feel free to investigate further :)
jnthn tadzik: Does it actually run the tests too?
tadzik: I've heard about that getting spat out as a warning. 16:25
tadzik jnthn: yes, they work fine
16:29 [Sno] left, LLamaRider left 16:30 LLamaRider joined 16:31 guru joined, ajr_ left, guru is now known as Guest83214 16:32 Guest83214 is now known as ajr_ 16:33 dmol joined 16:38 zakharyas left 16:39 vincent21 left 16:42 Rotwang1 joined 17:02 treehug8_ joined
masak Python 3.4 has landed! finally enums! :) 17:02
BenGoldberg Landed? Had it previously been in outer space? 17:04
17:05 treehug88 left
masak .oO( released? was it previously kept in a cage? ) 17:05
masak .oO( shipped? was it previously waiting in the harbor? )
masak .oO( cut? was it previously an angsty teenager? ) 17:06
enums in 3.4 look much saner and less lame than last time I looked at them in Python: docs.python.org/3.4/library/enum.html 17:08
PerlJam yeah, but who uses python 3?!? ;) 17:09
timotimo who uses Perl 6?!? ;)
rurban I still can only use Python 2.7 17:10
17:11 ajr_ left
dalek ar/tristar: ab2d701 | jnthn++ | tools/build/module-install.pl:
Actually take/use the output extension.
 17:11
timotimo huh, i wrote that exact code once before 17:12
must have gotten lost :(
jnthn aww
anyway, good news is module installing seems to work ;)
timotimo ayup :)
Mouq++
jnthn timotimo++ Mouq++ 17:14
rurban parrot openpipe with env almost done (it's open "rp" with set'able anv) 17:16
env
g d 17:17
pmurias jnthn: nqp-js currently requires a nqp-p for cross compiling, what would be the best way to handle this? 17:19
jnthn: it currently depends on an installed nqp-p
jnthn pmurias: Dunno. I mean, you can say that if you want it to work out you must --backend=parrot,js 17:20
So it will do them in that order
That doesn't solve the "installed" bit
But if it's just needed for building, does that matter? 17:21
pmurias it complicates the build process a bit 17:22
for those who don't have an installed nqp-p
jnthn True
It's temporary though? The goal is to self-host eventually? 17:23
pmurias yes
it should bootstrap soonish
but I'm not sure how fast the bootstrap will be at first
jnthn figures out why S06-macros/unqouting.t blows up 17:25
timotimo p6weekly.wordpress.com/?p=94&sh...73025224a6 - rough (and late) first draft 17:26
jnthn But heading out for dinner now, so it'll have to wait :)
timotimo: ooh, I'll read before I go :)
timotimo (if you use the feedback form to send feedback my way, consider putting your name in the text somewhere, so i can ask questions)
masak jnthn: you can explain it to me over dinner :)
timotimo: s/dilligently/diligently/ 17:27
jnthn masak: yeah...I think 'tis time? 17:28
masak timotimo: maybe link to 2048? :P
jnthn: yep.
jnthn ready in 2
masak jnthn: meed you down-in-l'-lobby
meet*
masak .oO( ooh! maybe they have mead where we're going! ) 17:29
&
pmurias should start working on a 2048 nqp-js clone ;) 17:31
jnthn &
17:35 fhelmberger left
timotimo doge2048.com/ - fantastic 17:39
tadzik this one I actually enjoyed 17:40
vendethiel pmurias: that'd be cool to see ! 17:42
rindolf Hi all. Happy Purim, everyone. 17:43
17:44 [Sno] joined
timotimo purism? 17:48
17:49 denisboyun joined 18:00 pmurias left, hoverboard joined 18:05 spider-mario joined 18:08 Mouq joined
rurban Purim: I believe he meant Bar Refaeli on Saturday in the bright yellow tiger suit 18:12
18:12 Mouq left 18:18 LLamaRider left 18:21 SamuraiJack joined
rurban rakudo fails on a lot of -Werror=missing-prototypes 18:24
18:25 LLamaRider joined, hoverboard left 18:27 dwarring joined
timotimo published 18:28
18:32 rindolf left 18:33 LLamaRider left, SamuraiJack left 18:37 hoverboard joined
rurban timotimo: that would be the perl6 fix when parrot has openpipe-env merged gist.github.com/rurban/9605572 18:40
18:45 rindolf joined 18:46 simula67 joined 18:48 darutoko left
timotimo that'll happen for tomorrow's release? 18:50
18:52 Alina-malina left 18:53 Alina-malina joined, Alina-malina left, Alina-malina joined, Alina-malina left
rurban did we ever had a pir::die missing a signature? gist.github.com/rurban/9605884 18:56
timotimo pir::die("Unable to execute '$cmd'") unless $pipe; 18:58
18:58 Alina-malina joined
timotimo that would be the one? 18:58
rurban ah, yes
timotimo can easily be nqp::die
rurban just trying the new code :)
timotimo :)
rurban timotimo: we will not touch parrot, but I want to fix rakudo for tomorrows parrot update. nqp already fine 18:59
timotimo i thought parrot needs openpipe-env to be merged?
rurban we will do that for the next release. no hurry 19:02
one by one
timotimo that's the release in ~30 days?
rurban yes
timotimo aaw
i wish i had asked you for a fix yesterday or something 19:03
rurban I'm just testing that in a branch to get earlier feedback for our parrot merge
last week would have been better. we froze saturday
timotimo ah
i didn't know about it at that time
we ought to introduce spectests that cover that particular functionality
it really ought not regress again
rurban no prob. do you need QX with env already? 19:04
timotimo Perl::Bench has a subcommand that requires it, but qx is very useful for lots of glue-ish things you could use Perl 6 for
(or: could *not* use Perl 6 for at the moment)
rurban I rather want to test if parrot -O2 is now stable enough.
I see
well, we'll need about a week to merge the openpipe-env branch 19:05
timotimo that's mostly "grace time" right?
i dare not ask you to throw out your release process for my little issue
rurban yes, request for comments. a new API. maybe we should it different
timotimo can we get a warning in place that'll try to figure out if someone modified %*ENV before calling qx? 19:07
hmm. if %*ENV !eqv CHECK %*ENV perhaps?
p: %*ENV<GIT_DIR> = "/home/foo"; say "test"; if %*ENV !eqv CHECK %*ENV { say %*ENV.kv (^) CHECK set %*ENV.kv } 19:12
camelia rakudo-parrot a24091: OUTPUT«test␤»
timotimo p: %*ENV<GIT_DIR> = "/home/foo"; say %*ENV.keys; say (BEGIN %*ENV).keys 19:13
camelia rakudo-parrot a24091: OUTPUT«XDG_SESSION_ID TERM SHELL XDG_SESSION_COOKIE SSH_CLIENT SSH_TTY USER LS_COLORS SSH_AUTH_SOCK TMUX PATH MAIL PWD LANG NODE_PATH TMUX_PANE SHLVL HOME LANGUAGE LOGNAME SSH_CONNECTION LC_CTYPE LESSOPEN XDG_RUNTIME_DIR LESSCLOSE _ GIT_DIR␤XDG_SESSION_ID TER…»
timotimo p: %*ENV<GIT_DIR> = "/home/foo"; say %*ENV<GIT_DIR>:exists; say (BEGIN %*ENV.copy)<GIT_DIR>:exists
camelia rakudo-parrot a24091: OUTPUT«===SORRY!===␤No such method 'copy' for invocant of type 'Hash'␤»
timotimo p: %*ENV<GIT_DIR> = "/home/foo"; say %*ENV<GIT_DIR>:exists; say (BEGIN %*ENV.clone)<GIT_DIR>:exists
camelia rakudo-parrot a24091: OUTPUT«True␤True␤»
timotimo hm.
p: %*ENV<GIT_DIR> = "/home/foo"; say %*ENV.keys; say (BEGIN %*ENV.keys) 19:14
camelia rakudo-parrot a24091: OUTPUT«XDG_SESSION_ID TERM SHELL XDG_SESSION_COOKIE SSH_CLIENT SSH_TTY USER LS_COLORS SSH_AUTH_SOCK TMUX PATH MAIL PWD LANG NODE_PATH TMUX_PANE SHLVL HOME LANGUAGE LOGNAME SSH_CONNECTION LC_CTYPE LESSOPEN XDG_RUNTIME_DIR LESSCLOSE _ GIT_DIR␤XDG_SESSION_ID TER…»
19:14 kurahaupo joined
timotimo p: %*ENV<GIT_DIR> = "/home/foo"; say "GIT_DIR" ~~ %*ENV.keys; say "GIT_DIR" ~~ (BEGIN %*ENV.keys) 19:15
camelia rakudo-parrot a24091: OUTPUT«False␤False␤»
timotimo oh?
p: say "foo" ~~ <foo bar baz>
camelia rakudo-parrot a24091: OUTPUT«False␤»
timotimo p: %*ENV<GIT_DIR> = "/home/foo"; say "GIT_DIR" (elem) %*ENV.keys; say "GIT_DIR" (elem) (BEGIN %*ENV.keys)
camelia rakudo-parrot a24091: OUTPUT«True␤False␤»
timotimo that'll only catch removing and adding keys, though 19:16
19:22 brrt joined 19:29 guru joined, guru is now known as Guest28326 19:30 Guest28326 is now known as ajr_
retupmoca I think I found another moar nativecall bug: gist.github.com/retupmoca/9606605 19:34
japhb__ timotimo: Why did the week # in the summary title jump from 10 to 12? Which week had the off-by-1 error? 19:41
timotimo oh, oops? 19:42
19:43 lizmat joined
lizmat is home 19:43
[Coke] hio, lizmat! 19:50
lizmat [Coke] o/ 19:51
timotimo hi lizmat
lizmat timotimo o/
19:51 prammer joined
lizmat trying to slow down from a 7 hour drive 19:51
nwc10 use more 'whisky' ? 19:52
lizmat hehe... no
use more 'chartreuse'
we also visited the Chartreuse distilleries in Voiron
and ransacked the shop :-)
nwc10 poor shop 19:54
lizmat eh... they were quite happy as we also paid them for items ransacked :-) 19:55
nwc10 mmm, you need to take lessons from the Danish, etc, on how ransacking is supposed to work :-)
19:56 Mouq joined
lizmat perhaps :-) 19:56
nwc10 although they didn't have extradition treaties and Interpol back then, so it might not work as well these days. 19:58
timotimo having "no" as a "anti-use" is kind of weird 20:00
20:00 Mouq left
timotimo it's easy to understand when it's at the beginning of a file 20:00
but right in the middle
no llama 'noooooooo!' 20:01
geekosaur no sanity 'here'; 20:04
BenGoldberg no common::sense 20:05
lizmat no use
20:07 araujo left, araujo joined 20:09 PZt joined
lizmat [Coke]: re irclog.perlgeek.de/perl6/2014-03-17#i_8448294 , I've discussed that with Tim Bunce yesterday during the QA hackathon 20:10
jave2perl predates rakudo running on the JVM
in his opinion, someone with a little Java knowledge and Perl6 knowledge 20:11
*should* be able to make that work in ~ a day
20:11 ajr_ left
lizmat again, *AND* have it also fill in the necessary NativeCall code in the generated Perl 6 code 20:11
rurban Isn't today the crazy irish day? 20:12
lizmat so that you would only need to point this script at a jar, and have it write the Perl 6 interface (when running on the JVM) for you
aka: instant access from Perl6 to all Java libraries (in theory)
20:12 ajr joined, ajr is now known as Guest8724
lizmat alas, I'm not the person to do this :-( 20:13
not without severe handholding, I'm afraid
[Coke] ok. if there's still a need, I can poke at this weekend, perhaps. 20:16
work is still exploding (it's been six months. less exploding would be good) 20:17
lizmat hopes the dust will settle for [Coke] :-) 20:24
20:26 tgt joined 20:28 woolfy joined
lizmat timotimo++ for keeping up weekly changes reports 20:40
20:43 virtualsue joined 20:46 raiph left, kaare_ left 20:52 tgt left
lizmat gets some sleep& 20:56
timotimo i try my best :P 21:01
21:02 dayangkun left, kurahaupo left
[Coke] in good/bad news, this project's deadline has just been pushed out another week. 21:04
timotimo \o/ 21:05
[Coke] ... and, oh look, that means we can do more tickets. GAH 21:06
timotimo hooray
>_>
ashleydev [Coke]: apropos?: news.ycombinator.com/item?id=7415916 21:07
21:10 jnap joined, jnap1 left 21:12 brrt left 21:17 dayangkun joined 21:18 virtualsue left
[Coke] ashleydev: nah. 21:23
but I've worked there, too. :)
21:30 Mouq joined
Mouq o/ 21:32
yoleaux 08:07Z <moritz> Mouq: I've given you a commit bit for rakudo/star
Mouq \o/ Thank you moritz++ :) 21:33
21:39 kbaker_ left 21:50 treehug8_ left
simula67 lizmat: Regarding java2perl6, did you mean Java interfaces will be written to Perl 6 roles ? 21:55
21:59 jnap1 joined 22:02 jnap left 22:12 treehug88 joined 22:34 rindolf left 22:36 dmol left 22:37 skids joined 22:42 BenGoldberg left, Guest8724 left 22:43 bjz left, BenGoldberg joined, bjz joined 22:51 hoverboard left 23:06 xenoterracide left 23:09 Rotwang2 joined 23:10 denisboyun left, Rotwang1 left 23:12 Rotwang2 left 23:15 pnu left, spider-mario left, pnu joined 23:18 Shozan joined 23:19 SHODAN left 23:20 pecastro left 23:22 pecastro joined 23:25 rurban1 left
colomon Just successfully processed two 4 MB JSON files using Rakudo-moar and JSON::Tiny. Took 15.75 minutes, alas. 23:38
23:40 hoverboard joined 23:45 hoverboard left
colomon though that is on my Mac rather than my fast Linux box 23:48
23:49 skids left

Powered by Cro and the Raku® Programming Language.

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

Thank you!