Parrot 4.5.0 "Buff-faced Pygmy Parrot" | parrot.org/ | Log: irclog.perlgeek.de/parrot | #parrotsketch meeting Tuesday 19:30 UTC
Set by moderator on 8 July 2012.
dukeleto rurban: sounds awesome! 03:20
06:20 fperrad joined 06:43 wagle joined 07:24 brrt joined
dalek kudo/nom: 65e2aae | moritz++ | src/core/Str.pm:
start to &fail typed exceptions from Str.substr
08:26
kudo/nom: 1fe39c7 | moritz++ | src/core/Str.pm:
another typed failure from Str.substr
08:48
d_parrot: b49449c | (Bart Wiegmans)++ | pudding/psgi.pl:
niner++ for fixing green threads and invocation
08:51
brrt nine++ for fixing it
09:08 schm00ster joined 09:52 Khisanth joined 10:20 Psyche^ joined 10:50 JimmyZ joined 11:19 dukeleto joined 11:44 brrt joined
brrt good afternoon #parrot 11:44
nine Good afternoon, brrt 11:51
11:51 dukeleto joined
brrt nine++ good fix 12:05
it worked
however, don't worry if this doesn't get to be merged soon
because i'm going to run scripts in their own interpreter
(which is good for cleanup etc) 12:06
nine brrt: well I hope to get it merged with the rest of threads. Just have to fix one test file which probably ist just using too short timeouts... 12:08
brrt ok :-) very well 12:13
JimmyZ nine: did you see blogs.perl.org/users/rurban/2012/07...sting.html 12:21
12:35 lucian joined 13:14 drift joined
dalek Heuristic branch merge: pushed 211 commits to parrot/coke/rm_pasm by coke 13:21
rurban good morning #parrot
Maybe I have to add to my qemu blog post that a CPAN update to build TAP::Harness::Archive lasts about two days :( 13:22
Coke hopefully you can save the results of that, though? 13:23
rurban Sure. t/archive is there
dalek rrot/coke/rm_pasm: 3361489 | coke++ | tools/dev/pbc_to_exe.winxed:
Track PASM -> PIR change for constants.
13:25
nine JimmyZ: yes, I will try it when I'm at home 13:26
Coke coke/rm_pasm's pbc_to_exe.winxed rewrite needs to have all its PASM removed. 13:38
Coke will get to this tonight if JimmyZ doesn't beat him to it.
13:41 bluescreen joined 13:48 drift joined 13:50 tobym joined
moderator Parrot 4.6.0 in progress, master locked until Thursday | parrot.org/ | Log: irclog.perlgeek.de/parrot | #parrotsketch meeting Tuesday 19:30 UTC 14:00
rurban will name 4.6.0 after the Wild Parrots from Telegraph Hill 14:01
Coke thursday? 14:02
rurban Tuesday I thought
oops
moderator Parrot 4.6.0 in progress, master locked until Tuesday | parrot.org/ | Log: irclog.perlgeek.de/parrot | #parrotsketch meeting Tuesday 19:30 UTC 14:02
rurban I still need ssh access to parrot.org. Can someone add rurban to the parrot group there? 14:04
ftp-osl.osuosl.org is done
Coke seen cotto? 14:07
aloha cotto was last seen in #perl6 2 hours 22 mins ago joining the channel.
Coke msg cotto "<@rurban> I still need ssh access to parrot.org."
aloha OK. I'll deliver the message.
14:25 PacoAir joined
dalek p: 8cbe162 | moritz++ | tools/build/PARROT_REVISION:
downgrade parrot to 4.4.0

there was no reason to require 4.5 over 4.4, so we can depend on an older parrot (but it still works on the newer one)
14:32
14:46 drift joined 14:52 dmalcolm joined
rurban added a parrot remark on the PyPy blog morepypy.blogspot.com/2012/07/hello...ryone.html 15:27
dukeleto rurban++ 15:36
rurban github.com/rurban/parrot/commit/19f3bd5d82 looks good on debian linux, debian powerpc and cygwin
dukeleto rurban: hilarious quote, as well :) 15:37
rurban "...the birds are perfectly capable of taking care of themselves. The viewer probably has a little less confidence in Mark though, as he says he has no idea what he will do..." 15:38
dukeleto rurban: your blog post about using qemu for testing is very useful! thanks. 15:42
rurban The next big thing will be how to run macosx and classic on linux, but this will have to be a private matter because of apple licensing terms 15:44
first I'll cover sparc32 and sparc64 with solaris, and then probably parrot on armel 15:45
and I'll make the test images available for download also 15:46
15:55 drift joined 15:59 benabik joined 16:01 dukeleto joined
benabik ~~ 16:02
dukeleto benabik: howdy! 16:06
benabik dukeleto! How's things?
16:08 davidfetter joined 16:16 jashwanth joined
dukeleto benabik: getting back to normal after a few crazy months 16:17
benabik dukeleto: I think I can sympathize, although it's only been a couple weeks for me.
dukeleto benabik: having a baby is probably just as disruptive (or more) as getting married and buying a house within 9 days of each other. I hope you are having the appropriate amount of fun! 16:18
benabik dukeleto: Yes. "Fun". ;-)
dukeleto: I knew about the house, but had missed the marriage. Congratulations to you and Duchess Leto. ;-) 16:19
dukeleto benabik: thanks! 16:21
rurban windows doesn't look too good: smolder.parrot.org/app/projects/rep...ails/30135 31 failed 16:22
16:58 tuxit joined
nine rurban: is this on master? 17:39
rurban nine: This is the 4.6.0 release, master + 1
See github.com/rurban/parrot/commit/19f3bd5d82 17:40
Ah you mean windows, sorry. Windows is master, yes. not on any branch.
benabik Well it looks like some of those are caused by a lack of IPv6 support. 17:41
rurban yes, this is easliy fixed.
I wrote a blog post once how to setup ipv6 on windows 17:42
benabik But the opsc failure is a crash somewhere, and there are a lot of NCI "why does this fail"...
rurban Unfortunately too late for 4.6, but I guess no big problem. 17:43
blogs.perl.org/users/rurban/2011/12...ndows.html 17:44
on solaris auto::snprintf - Test snprintf...ld.so.1: test_1990: fatal: libncurses.so.5: open failed: No such file or directory is fatal. on other failing libncurses.so.5 tests not. 17:47
The solaris problem is that the system perl is inherently broken. Only a self-compiled perl works fine. 17:53
benabik Solaris utilities are... quirky 17:56
benabik was so glad when RIT put Ubuntu on all the Sun boxes 17:57
18:02 PacoAir joined 18:12 dukeleto joined
rurban solaris is really pretty quirky. The exact same linker cmdline as seen with verbose-step works fine. There must be some env side-effect... 18:13
18:20 dukeleto joined
dukeleto jashwanth: pong 18:20
jashwanth hello dukeleto 18:21
I had found a linker error on some symbols 18:22
dukeleto: especially "*"
dukeleto: so am writing code for multiplication of matrices 18:23
18:41 drift joined 18:46 lucian joined
dukeleto jashwanth: make a gist of the full error 18:57
rurban jashwanth: you mean duplicate symbol warnings? 18:59
jashwanth winxed: symbol lookup error: /home/jashwanth/parrot/lib/4.3.0-devel/dynext/linalg_group.so: undefined symbol: cblas_dgemm
this is the error
19:03 drift joined
dukeleto jashwanth: does that function actually exist? 19:07
jashwanth: also, did you dlfunc that function, if that is the correct name?
jashwanth dukeleto: no i did not do that 19:09
19:12 brrt joined 19:27 dukeleto joined 19:33 brrt1 joined
jashwanth dukeleto:is there anything I need to do to fix this error ar know more about it 19:36
19:40 brrt joined
jashwanth dukeleto: and also the function result for inverse calculation looks wrong 19:41
brrt is there a way to introduce variables that belong to a namespace? 19:47
dukeleto brrt: yes 19:49
brrt: what are you working in? C, PIR, Winxed ?
jashwanth: do you have a test for the function inverse?
jashwanth: have you pushed all your code to github?
jashwanth dukeleto:yes 19:50
brrt dukeleto: winxed
dukeleto brrt: namespace Foo { var x; } 19:52
brrt: if you want to do Foo::Bar you might need: namespace Foo { namespace Bar { var x; } }
brrt: there is also a "class" keyword
brrt: namespaces and classes have distinct meanings
brrt ... lets see if that works 19:55
dukeleto jashwanth: do the tests work for you? 19:57
jashwanth: i get : elements() not implemented in class 'Rosella;Query;Iterable;Map' 19:58
jashwanth: but maybe my rosella is messed up
brrt dukeleto: i'd expect that to work, but it doesn't, using winxed 1.10 20:00
jashwanth dukeleto:I did not write a specific test but checked it for different inputs
20:01 sivoais joined 20:02 Timbus joined
jashwanth dukeleto:gist.github.com/3124691 20:02
dukeleto jashwanth: i understand that way, but it is not the best way. I suggest writing a test takes the identity 2x2 matrix in, calculates the inverse, and then asserts that it gets the 2x2 identity matrix back
jashwanth: does that make sense?
jashwanth dukeleto:yes on it 20:04
dukeleto did github notifications go away because github v2 api went away? 20:20
sorear dukeleto: you mean for parrot issues? 20:23
the old issues apis sucked, I think dalek was trying to parse the rss output
the v3 issues API is quite adequate to write a proper issues plugin, but I haven't gotten around to it yet 20:24
dukeleto sorear: i am pushing code to some repos, but maybe I just never setup the post-receive hooks, totally possible.
sorear dukeleto: ah
20:24 drift joined
dukeleto I create so many repos. I need a way to manage and sync post-receive hooks. 20:25
I have been pondering writing something that does that. Has anybody gotten around to that already?
20:26 alester joined
nopaste "rurban" at 208.74.121.102 pasted "PERL6LIB=lib:. ./perl6 t/spec/S32-io/IO-Socket-INET.t" (77 lines) at nopaste.snit.ch/148448 20:27
20:31 perlite joined
nopaste "rurban" at 208.74.121.102 pasted "rakudo spectests on cygwin" (51 lines) at nopaste.snit.ch/148450 20:35
brrt dukeleto: whats the PIR way of doing it? 20:39
(i.e., namespace-global variables)
dukeleto brrt: get winxed to tell you ;) 20:40
brrt: winxed can spit out PIR from a given .winxed file.
brrt: i think it is maybe --target=pir ?
brrt: nope, -c .
brrt: of course? -c stands for "compile to PIR" evidently 20:41
brrt i would but winxed won't tell me, it doesnt understand namespaces
eh, variables in namespaces
(and yes, its winxed -c :-)) 20:42
nopaste "rurban" at 208.74.121.102 pasted "SEGV in Parrot_CallContext_get_pmc_keyed_int" (61 lines) at nopaste.snit.ch/148452
brrt i.e. namespace Foo { var bar; } fails with unexpected token near var 20:45
dukeleto brrt: github.com/letolabs/parrot-libgit2...ory.winxed 20:47
brrt: i guess maybe you need to have a class nested inside the namespace or a function in the class 20:48
brrt: i am still a Winxed newbie as well
brrt: but poke around in parrot-libgit2 for some Winxed examples
brrt ooh, lets see
thanks 20:49
dukeleto msg Whiteknight do you think parrot-lapack should use PLA to marshall data (i.e. creating matrices) or should it use StructView ?
aloha OK. I'll deliver the message.
brrt .. its probably a parse error
nopaste "rurban" at 208.74.121.102 pasted "t/spec/S06-routine-modifiers/lvalue-subroutines.rakudo" (51 lines) at nopaste.snit.ch/148454 20:59
brrt dukeleto: oh, unfortunately that var is inside a class 21:02
21:24 brrt left