Parrot 2.3.0 will be released at 2010-04-20 10:00 UTC | parrot.org/ | Channel log: irclog.perlgeek.de/parrot/today | Priority: documentation sprint and pre-release testing for 2.3, fix line number annotations | Review and vote GSoC applications
Set by moderator on 18 April 2010.
00:04 zostay_m_ joined 00:14 Mokurai1 joined 00:18 mmcleric joined
Andy Who can tell me what decides when Parrot wants to regenerate imcparser.*? 00:26
00:27 davidfetter joined
cotto_work Andy, run Configure.pl with --maintainer. 00:29
Andy Yeah, I know that part, and I'm not and it's still trying to regerneate those
darbelo Also, the existance of flex and bison in your system at configure time can influence that. 00:30
cotto_work It looks like it's only a dummy target to retouch that file if you configure without --maintainer. 00:33
YACC = echo 00:34
without --maintainer
Andy I'm not sure what you're saying. 00:35
cotto_work instead of running yacc, a makefile generated by Configure.pl without --maintainer will run echo 00:36
Andy oooh, maybe it is confused 'cause I pass in --yacc 00:37
cotto_work that might be it 00:38
darbelo That will override the couce for 'echo', yes.
*choice
My solution to this is of the "Don't do that" variety. 00:39
Andy thanks, cotto_work and darbelo 00:40
I wasn't passing --maintainer, but I was still passing --yacc=bison
nopaste Someone at 192.168.1.3 pasted "my ~/bin/parrotconfig" (47 lines) at nopaste.snit.ch/20309 00:41
darbelo We should probably allow for what you are doing. But for now you shouldn't oveeride yacc (or lex) when configuring without --maintainer
Andy if anyone cares
I'd really like to get the pmc2c.pl faster 00:45
it pains me to see them take so long in aggregate
dalek kudo: f162a26 | jonathan++ | src/glue/subset.pm:
Correc the order we do checks of subtypes from most general to most specific, so we at least always consider type first before the extra conditions.
kudo: 1f9b1dd | jonathan++ | src/Perl6/ (2 files):
Implement &[+] shortcut form for infix ops.
00:47 ash_ joined 00:53 abqar joined
Andy oooh, look at this, "reallocate_string_storage" 01:01
that sounds like a candidate for disappearing.
ping bacek 01:04
01:07 kid51 joined
bacek_at_work Andy, pong. 01:08
reallocate_string_storage still used internally 01:09
Andy that seems sad.
So strings aren't really immutable until we pass them out.
bacek_at_work Andy, why?
purl Andy, why are you trying to become kudra's enemy?
bacek_at_work Ah, yes. We have to construct them before passing :)
Andy Are we going to allow strings to be passed in to be reused? 01:10
For example, we can reuse STRINGs passed to Parrot_io_readline_buffer
dalek kudo: b05155e | (Martin Berends)++ | src/cheats/setup-io.pm:
[src/cheats/setup-io.pm] implement $*PID and make the source code even more cheaty. Parrot TT#1564 requests a less cheaty feature.
 01:11
Andy I think we need some definition of when we expect immutability, and when we don't. 01:15
dalek rrot: r45812 | petdance++ | branches/immutable_strings_part1 (3 files):
fixing consts and annotations
 01:22
02:04 bacek_at_work joined 02:36 jan joined
kid51 make fulltest: linux/i386: PASS at r 45811 02:49
Coke ponders pinging andy for some help with WWW::Mechanize. 03:04
03:32 janus joined
Coke gives up and goes for something completely different. 03:32
bacek_at_work Coke, software is hard. Let's go shopping! 03:40
dalek rrot: r45813 | mikehh++ | trunk/DEPRECATED.pod:
add deprecated runcores - TT #1563
 03:49
04:19 Andy joined 04:31 tcurtis joined
dalek rrot: r45814 | plobsing++ | trunk/DEPRECATED.pod:
deprecate nci signature strings
 04:38
TT #1565 created by plobsing++: [DEPRECATION] NCI signatures 04:48
TT #1565: trac.parrot.org/parrot/ticket/1565
04:58 hicx174 joined 05:00 rurban_ joined
Coke anyone else having trouble building rakudo? 05:17
dukeleto Coke: what kind of trouble? 05:26
dalek rrot: r45815 | gerd++ | trunk/tools/docs/filename_and_chapter.pl:
Add one chapter
 05:27
Coke dukeleto: "the build fails". If the build is completing for you, you're not having it. =-) 05:36
dalek rrot: r45816 | petdance++ | trunk/config/auto/warnings.pm:
add new warnings for GCC 4.5.0
 05:43
05:57 uniejo joined
dalek rrot: r45817 | petdance++ | trunk/lib/Parrot/Pmc2c/PCCMETHOD.pm:
Consting the _call_object arg in the generated code. Removed an unused Perl method.
 05:59
06:11 he_ joined 06:20 clinton joined 06:22 aukjan joined 06:28 fperrad joined 06:31 fperrad_ joined
dalek kudo: 027ff40 | moritz++ | t/spectest.data:
[t/spec] enable S03-operators/nesting.t
 07:13
07:25 iblechbot joined 07:27 riffraff joined
GeJ clock? 07:38
purl GeJ: LAX: Tue 12:38am PDT / CHI: Tue 2:38am CDT / NYC: Tue 3:38am EDT / LON: Tue 8:38am BST / BER: Tue 9:38am CEST / IND: Tue 1:08pm IST / TOK: Tue 4:38pm JST / SYD: Tue 5:38pm EST /
GeJ release is in a couple of hours?
moritz aye 07:46
GeJ well, time to rebuild and run make fulltest. 07:47
make fulltest PASS on FreeBSD/amd64 on trunk @ 45817 08:27
sorear seen pmichaud 08:39
purl pmichaud was last seen on #parrot 12 hours, 34 minutes and 54 seconds ago, saying: basically calling dlfunc to grab getpid() from the C lib? Evil. :-)
dalek rrot: r45818 | fperrad++ | trunk (2 files):
[TAP] fix trailing plan
 08:48
08:50 alexn_org joined
mikehh All tests PASS (pre/post-config, make corevm/make coretest, smoke (#33307), fulltest) at r45817 - Ubuntu 10.04 beta amd64 (gcc with --optimize) 08:54
he_ Hm, so it appears that gcc 4.1.2 or 4.1.3 mis-compiles compiler/imcc/pbc.c for powerpc unless optimization is turned on... 09:07
09:38 gerd joined
bacek gerd, ping 09:44
gerd It is not 10:00 UCT; but I think I can already start to switch the version number to 2.3.0 09:45
darbelo gerd: Fire at will! 09:48
gerd svn commit is done
dalek rrot: r45819 | gerd++ | trunk (10 files):
update to version 2.3.0
 09:53
bacek gerd, can you replace "pcc_hackatron branch merged" in NEWS with something more meaningful for non-parroters? 10:01
Like "Parrot Calling Conventions were greatly improved and makes more sense now" :) 10:02
gerd please update NEWS like you want it; it will the the updated NEWS in the mail and WWW-Page
s/the/take/ 10:03
bacek gerd, English is my forth language. I can't formulate it properly...
darbelo "Parrot Calling Conventions are, like, totaly cool now."
That's my best attempt. 10:04
bacek darbelo, +1 :)
Next one will be "totally optimal"
moritz how about "much more consistent"?
bacek "... and follow natural semantic of passing args and handling results" 10:05
gerd Okay, should I do that add to the NEWS? 10:08
moritz gerd: committing right now 10:09
dalek rrot: r45820 | gerd++ | tags/RELEASE_2_3_0:
tagged release 2.3.0
rrot: r45821 | moritz++ | trunk/NEWS:
[NEWS] explain pcc_hackathon results, bacek++, darbelo++
bacek erm... 10:10
Looks like we have to retag release
10:13 Patterner joined
gerd Yes the old version is already tagged 10:13
Can I overwrite it with a new copy? 10:15
10:15 lucian joined
moritz I think you can 10:15
gerd r45821 is now tagged as 2.3.0 10:16
I will start to recreate the tar files with that revision 10:18
10:22 JimmyZ joined
JimmyZ gerd: tag path is wrong 10:22
tags/RELEASE_2_3_0/trunk (copied) (copied from trunk) 10:23
alexn_org guys, I think building Rakudo with the latest Parrot trunk is not working ....
invoke() not implemented in class 'Undef'
current instr.: '_block34974' pc 341244 (src/gen/core.pir:18514)
called from Sub '!fire_phasers' pc 1074 (src/glue/phasers.pir:67)
called from Sub 'perl6;Perl6;Compiler;main' pc 267852 (src/gen/perl6-actions.pir:14713)
JimmyZ gerd: It's here now. trac.parrot.org/parrot/browser/tags..._3_0/trunk
gerd: It should be trac.parrot.org/parrot/browser/tags...ASE_2_3_0/
moritz alexn_org: weird, worked for me some few revisions ago... did you reconfigure rakudo before trying to build it? 10:24
alexn_org yes ... maybe I haven't installed parrot properly, let me check 10:25
moritz alexn_org: what platform are you on? 10:26
dalek rrot: r45822 | gerd++ | tags/RELEASE_2_3_0/trunk:
tagged release 2.3.0
JimmyZ r45822 is wrong tag.
alexn_org moritz: OS X 10.6.2
moritz: If I had an older Parrot revision installed, do I need to delete it before "make install" ? 10:28
10:28 riffraff joined
moritz alexn_org: it's safer if you do 10:28
alexn_org: build works fine here with newest parrot on linux amd64 10:31
rakudo build, that is
alexn_org moritz: OK, I'm retrying the build right now 10:33
gerd The update tag seems not to have to modified NEWS file 10:35
dalek rrot: r45823 | fperrad++ | trunk/runtime/parrot/library/TAP/Harness.pir:
[TAP] workaround for tar on Windows
 10:58
11:12 muixirt joined
alexn_org guys, I'm not sure if this is the place to report this ... but building Rakudo fails on me ... (with --gen-parrot) ... paste-it.net/public/bc8442c/ 11:13
11:31 particle1 joined
moritz alexn_org: irc.freenode.org #perl6 11:31
alexn_org: looks like it built most of Rakudo, and just failed while compiling Test.pm 11:32
alexn_org it fails while running perl6
moritz how? same error?
alexn_org doesn't get to Test.pm
yes
I opened a bug here ... maybe someone else caught it too ... rt.perl.org/rt3/Ticket/Display.html?id=74524 ... at first I thought it was a conflict with my Parrot install, but with --gen-parrot that shouldn't happen, right? 11:33
moritz it should not 11:34
alexn_org: thanks for the bug report, hopefully it can be fixed before the Rakudo release on Thursday 11:35
alexn_org: could you please try 'git checkout 1f9b1dda95397e1216120352b06da116909dd92d' and then rebuild rakudo? (no need to rebuild parrot) 11:40
alexn_org: (after that you can get back with 'git checkout master')
alexn_org ok 11:41
11:42 whiteknight joined
whiteknight good morning, #parrot 11:48
I would be doing work right now, but my computer's antivirus software has helpfully decided to eat up all my CPU for a while. 11:49
Good thing too, that pesky malware I had on here was really allowing me to be too productive 11:51
bacek whiteknight, say hello to KAV! :) 11:59
alexn_org moritz: that rev works fine 12:00
whiteknight KAV?
bacek Kaspersky Anti Virus 12:02
(or however it's spelled in English :)
whiteknight oh, can't use it. The company uses Symantec
bacek or Symantec uses you! 12:03
whiteknight that's probably more true
12:05 bluescreen joined
whiteknight bacek: I would rather just use Linux than use KAV or any other antivirus software 12:05
bacek whiteknight, linux is best anivirus software so far :) 12:06
12:14 tetragon joined
dalek website: gerd++ | Parrot 2.3.0 "Samoan Lory" Released! 12:32
website: www.parrot.org/news/2010/Parrot-2.3.0
kudo: d0e93e5 | (Martin Berends)++ | src/cheats/setup-io.pm:
[cheats/setup-io.pm] set $*PID to 0 if the getpid() function cannot be
 12:48
12:48 iblechbot joined
12:59 rurban_ joined
Coke (the error that alexn_org is the same one I was having.) 13:01
13:09 atrodo joined
moritz Coke: could you please re-check with latest rakudo (revision d0e93e5)? 13:20
Coke moritz: yup, on it.
moritz Coke++ 13:21
Coke why is parrot -v an error?
13:22 patspam joined 13:36 whiteknight joined
Coke moritz: much better, danke. 13:38
moritz Coke: what does ./perl6 -e 'say $*PID' print? 13:40
13:42 Cristina joined 13:46 PacoLinux joined 13:47 ruoso joined
Coke does ./perl6 really take 866M? 13:50
moritz: 0
moritz Coke: ok, then it's not really working, just failing less forcefully than before 13:51
Coke moritz: I'm on OSX. ISTR that was on the platforms that the getpid code was complaining on.
moritz yes 13:52
dalek kudo: 5a0581f | moritz++ | t/spectest.data:
one more passing test file
 13:53
nopaste "coke" at 192.168.1.3 pasted "getpid works..." (10 lines) at nopaste.snit.ch/20314 13:54
Coke so, something with the dlfunc call, I guess.
mberends yes :-( 13:55
Coke lemme see if I can make this work in pir on my box.
mberends Coke: ironic that you gave the code that works on everyone else's OS :-/
Coke which signatures does 'dlfunc' use? 13:59
14:01 sorear joined 14:02 Andy joined
dalek kudo: e135f23 | moritz++ | docs/ChangeLog:
[docs] update ChangeLog
 14:06
mberends Coke: the signatures are probably the same as in svn.parrot.org/parrot/trunk/src/nc...thunks.nci 14:10
dalek kudo: ad01dca | moritz++ | docs/announce/2010.04:
first draft for 2010.04 release announcement
 14:23
14:25 senf_statt_oel joined
dalek TT #1566 created by jquelin++: md2 should not be used anymore 14:26
TT #1566: trac.parrot.org/parrot/ticket/1566
Coke ok. a pid_t on my box is a __int32_t - so the signature looks right. 14:27
if only jq had gotten that in a few hours ago. =-)
14:34 bubaflub joined
Coke gerd: ping 14:36
14:39 mls joined
dalek kudo: cf4eb89 | Coke++ | docs/announce/2010.04:
untypo
 14:40
mls Hi guys, 2.3.0 doesn't compile on ia64 because trace_mem_block() nowadays takes a mem_pools argument, but src/gc/system.c calls it without that in line 200 14:41
Coke gerd: let me know if you need someone to generate the docs for docs.parrot.org
is that the only problem with ia64? 14:42
it's probably not too late to scrap the tag and re-tag it.
but that's up to gerd atm.
in the meantime, if you could open a ticket...
parrot trac?
purl parrot trac is, like, trac.parrot.org/parrot/
mls redoing the ia64 build... 14:43
(Btw, the release announcement points to releases/stable/2.3.0, but the ftp server has it in the supported directory) 14:44
ia64 works after patching the call. Creating a ticket... 14:52
14:54 PacoLinux joined, theory joined
Coke AIGH. 14:56
I changed the release instructions, gerd!
;)
where is the announcement? just on parrot.org? (havren't seen anything on list)
I'll update parrot.org
mls Yes, just parrot.org 14:57
Coke parrot.org fixed to point to supported.
honest mistake, it's been 'stable' for over a year.
dalek TT #1567 created by mls++: parrot does not compile on ia64 14:59
TT #1567: trac.parrot.org/parrot/ticket/1567
Coke mls: do you have a patch for #1567? 15:00
mls Well, I just added a "mem_pools" argument without verifying that this is the right thing to do. 15:01
now I just have to find out why parrot-nqp --target=pir runtime/parrot/library/ProfTest/PIRProfile.nqp segfaults on i586 ;-( 15:02
15:05 rurban joined 15:10 zostay_m joined
dalek tracwiki: v124 | fperrad++ | Languages 15:12
tracwiki: update Parrot version 2.3.0
tracwiki: trac.parrot.org/parrot/wiki/Languag...ction=diff
rurban So the old /stable/<VERSION> uri's are wrong? 15:14
15:20 davidfetter joined
mls I hate it when adding a printf makes a segfault go away... 15:24
rurban All tests did pass on cygwin 15:31
editor/pir.vim is not generated anymore 15:32
Coke rurban: ugh, I didn't think about previous urls. 15:33
let me see if I can symlink them.
rurban please
Coke try that? 15:34
(seems ok interactively) 15:35
rurban Could not chdir to stable: server said: Failed to change directory
ftp://ftp.parrot.org/pub/parrot/releases/stable/ is was
wget ftp://ftp.parrot.org/pub/parrot/releases/stable/2.0.0/parrot-2.0.0.tar.gz failed 15:36
Coke rurban: hurm. that wget just worked for me. 15:37
rurban Is it mirrored? I was on some oregon uni server 15:38
Coke ah. right. hang on.
rurban 64.50.236.52
Coke ran the trigger script.
not sure if that's immediate or if there's a delay. 15:39
15:39 aukjan1 joined
rurban config_lib.pasm => config_lib.pir ? 15:39
thanks, ftp stable ok now
Coke rurban: np. sorry about that. 15:40
rurban Just for later
Coke updated TT #770 to reflect the ftp chagne. 15:41
rurban: (pasm -> pir) yes.
ISTR it was helpful to increase the efficiency of the config.fpmc
msg kid51 I'll let you comment on TT#1568. 15:42
purl Message for kid51 stored.
rurban no tapir manpage/pod? 15:45
dalek TT #1568 created by brianwisti++: `make smoke` reports a timeout error 15:48
TT #1568: trac.parrot.org/parrot/ticket/1568
rurban Smolder is not reacting fast enough. bogus 15:49
mls i586 segfault: don't worry, it seems to be a gcc-4.5 compiler bug 16:02
rurban Coke: trac requires a new Version 2.3.0 16:08
16:11 senf_statt_oel left
rurban library/YAML/Parser turns out empty 16:13
particle1 rurban++ # 2.3.0 added 16:19
dalek TT #1569 created by rurban++: Missing tapir docs (pod/man) 16:21
TT #1569: trac.parrot.org/parrot/ticket/1569
dukeleto hola 16:28
did 2.3.0 get cut? /me is still backlogging 16:31
16:33 alexn_org joined
Andy yes 16:38
i just got the email
now let's merge down the immutable strings branch!
Util dukeleto: yes, it got cut. To easily see the conversation, search `gerd` in irclog.perlgeek.de/parrot/today .
16:45 gaz joined, gerd joined
moderator Parrot 2.3.0 Released | parrot.org/ | Channel log: irclog.perlgeek.de/parrot/today | Priority: documentation sprint and pre-release testing for 2.3, fix line number annotations | Review and vote GSoC applications 16:47
16:49 wagle joined 16:53 alexn_org joined
dukeleto i really didn't realize Tapir was added to core until I just read the release announcement 16:57
rurban I just need a short synopsis 17:00
dukeleto rurban: i can do that 17:01
17:01 joeri joined
rurban paste it here please :) that will be faster 17:01
I just want to through out the release 17:02
dukeleto rurban: gist.github.com/372750 17:04
rurban: that is taken from the README at github.com/leto/tapir
rurban Thet's the DESCRIPTION, great. SYNOPSIS also? 17:05
got it from there, thanks
dukeleto rurban: sounds good. thanks for poking me. Are you going to commit that to Parrot core, or should I? 17:06
rurban: ah, I've answered my own question :) 17:07
rurban no. I gave up my commit bit
dukeleto rurban: yeah, just remembered. I appreciate your work on the distribution front
rurban: Parrot needs to get more organized with respect to having a process for packaging for various distros 17:08
rurban "pir based prove" isn't that shorter?
library/YAML/Parser/Syck.pbc is not generated 17:09
dukeleto rurban: it is shorter, but if someone is unfamilar with prove, then not as useful
17:11 hercynium joined
dukeleto rurban: just so you know, it seems that fperrad forked my Tapir and fiddled with it, so I don't know exactly what it is up to 17:12
rurban YAML/Parser/Syck.pbc is in MANIFEST but not in the makefile. deprecated or forgotten? 17:13
17:14 aukjan joined 17:17 szabgab joined
rurban deprecated obviously. must be deleted from MANIFEST.generated 17:19
The 2 cygwin lib files are also still missing from MANIFEST.generated
ok, the cygwin patches for 2.3: code.google.com/p/cygwin-rurban/sou....src.patch 17:25
dukeleto rurban: do you want me to apply that to trunk? 17:26
rurban someone needs to check it
I'm not familiar with the details anymore
for opengl there's a old ticket already
japhb <rez> 17:27
rurban, ticket?
rurban for random_lib.pir not
pleeease
dukeleto rurban: why did you delete runtime/parrot/library/random_lib.pir ?
rurban because it is not supported anymore. try to run it :)
17:27 pjcj joined
rurban and it is not built and installed 17:27
This is just from the first basic testing. More tests later when I try pipp and rakudo 17:29
japhb rurban, What OpenGL ticket are you talking about?
rurban that was two years ago. wait... 17:30
17:31 iblechbot joined
rurban It was NOT ticket 57006. 17:33
Sorry. I found no ticket
japhb OK. Was there something I needed to deal with, ticket or no? (Aside from ash's problem yesterday, of course.) I haven't had a chance to backlog this morning yet .... 17:35
17:39 hercynium joined
dalek rrot: r45824 | fperrad++ | trunk/config/gen (2 files):
[crypto] remove MD2, see TT#1566
 17:48
dukeleto japhb: i am in the same boat. to many things whizzing by at top speed 17:52
17:52 kjeldahl joined
japhb dukeleto, it's why I gave up on even *trying* to keep up with #perl6. Those people are insane(ly talkative) 17:52
dalek gest-dynpmcs: 8a9bd24 | fperrad++ | setup.pir:
remove MD2 (insecure, and not available with OpenSSL 1.0.0)
 17:56
TT #1566 closed by fperrad++: md2 should not be used anymore 17:59
TT #1566: trac.parrot.org/parrot/ticket/1566
18:00 alexn_org joined 18:02 uniejo joined 18:05 davidfetter joined
dukeleto rurban: just committed your POD for Tapir to trunk 18:10
dalek rrot: r45825 | dukeleto++ | trunk/tools/dev/tapir.pir:
[docs] Add some POD to Tapir, rurban++
 18:21
rurban maybe the Makefile.generated bugs also? 18:22
dukeleto rurban: MANIFEST.generated should get generated by tools/dev/mk_manifest_and_skip.pl 18:27
rurban: i am running it now to see what it says 18:28
rurban: maybe that script only fiddles with MANIFEST
darbelo removes code. 18:29
rurban I forgot what it does also. I had a branch were MANIFEST.generated was completely generated some time ago.
Coke the tool just does MAN and MAN.skip 18:30
cotto_work #ps in 120
whiteknight gerd++ # nice release 18:36
darbelo will miss ps today.
cotto_work ps will miss darbelo
darbelo I should go prepaste my report. 18:37
dukeleto darbelo: yes. 18:38
darbelo INCOMING! 18:39
purl duck!
dukeleto hobbles a report together
darbelo pokes dalek. 18:42
18:44 kurahaupo joined
dalek TT #1569 closed by dukeleto++: Missing tapir docs (pod/man) 18:49
TT #1569: trac.parrot.org/parrot/ticket/1569
TT #1467 closed by darbelo++: Deprecate the digest dynpmcs.
TT #1467: trac.parrot.org/parrot/ticket/1467
TT #1469 closed by darbelo++: Deprecate the gdbmhash dynpmc.
TT #1469: trac.parrot.org/parrot/ticket/1469
rrot: r45826 | darbelo++ | trunk (23 files):
Remove the deprecated gdbmhash and digest dynpmcs, closes TT #1467 and TT #1469.
 18:54
Coke parrotsketch? 18:58
purl parrotsketch is a status meeting for parrot core committers held every Tuesday at 20:30 UTC in #parrotsketch
19:05 smash joined
smash hello everyone 19:05
dalek rrot: r45827 | allison++ | trunk (3 files):
[cage] Apply final fix for TT #389, methods are no longer stored in
 19:10
rrot: r45828 | allison++ | trunk/DEPRECATED.pod:
[cage] Removing deprecation item for TT #389 after task completed.
dukeleto smash: hola 19:11
purl what's up, dukeleto.
dalek TT #389 closed by allison++: subs with :method should not be entered as symbols in the namespace 19:21
TT #389: trac.parrot.org/parrot/ticket/389
whiteknight allison: ping 19:29
19:30 hercynium joined
allison whiteknight: pong 19:30
whiteknight allison for TT #389, if a method doesn't have :nsentry, where is it stored and how is it callable?
I've been trying to follow your patches, but I think I'm missing something
allison whiteknight: it's stored in the class if the class is created already, or stored in a special cache in the namespace if the class isn't created 19:31
whiteknight: (but that cache isn't accessible through any of the namespace interfaces)
whiteknight ah, okay. That makes more sense
allison whiteknight: the reason my code is so simple is I was only activating features that we already put in place to make this work a year or so ago 19:32
whiteknight when I was working on this ticket, I was trying a much larger change 19:33
allison whiteknight: (though, with the mess the current namespace code is in, just unearthing those foundations felt like a herculean task)
whiteknight: yeah, I looked over the branch
whiteknight: some things like changes to find_method so it could fetch methods from the namespace object? 19:34
whiteknight I guess I was taking the "not in the namespace" requirement too literally
allison makes sense 19:35
whiteknight I'm happy to see it closed by any means necessary. Kudos
dukeleto anybody interested in implementing something like Date:Simple for parrot? 20:06
does parrot have any date/time-related stuff?
Rakudo has github.com/moritz/Date 20:07
20:14 chromatic joined 20:15 ash_ joined
Coke ... or we could just use that, compiled down to pir. 20:21
I remove a PMC... I then have to... update PBC_COMPAT? And? 20:22
moritz Coke: the source code contains links to pages the describe the algorithms very well 20:23
you can basically copy&paste them, and adapt them to PMC syntax
allison Coke: the list of compiled PMCs is automatically generated from files named ".pmc" in that directory
moritz without actually understanding what it does
Coke moritz: yes, but there's no need to /reproduce/ the implementation, is what I'm saying.
allison: yes, the list is, but PBC_compat is borked
(since the numbers change when you remove a PMC.) 20:24
yes?
allison Coke: aye
20:24 eternaleye joined
allison Coke: also, will probably have to do a realclean 20:24
Coke I think the PBC_COMPAT bump also necessitates some other step. mk_native_pbc? 20:25
20:26 alexn_org_ joined
Coke does a realclean to see if that fixes his packfilie problems. 20:26
dalek TT #1570 created by Util++: Out-of-date binaries/packages on Parrot.org 20:27
TT #1570: trac.parrot.org/parrot/ticket/1570
Coke Util: which ones are out of date? 20:28
and are they so out of date they need to be removed, or do they just need to be updated?
(I actually mostly only care about the second part.)
20:29 dngor joined
Util Coke: Cygwin, MacPorts, and Debian, at least. (well, as of last week) 20:30
Coke Util: ok. there's already a ticket about macports.
the problem isn't really with the links on that page, unless they're never getting updated again.
dalek rrot: r45829 | fperrad++ | trunk (1 files):
restore t/library/md5.t which tests Digest/MD5.pir (and not digest dynpmc)
 20:32
Coke win 2
ash_ are there any comments on my GSoC application? Just wanted to check... since tomorrow is one of the GSoC deadlines 20:33
moderator Parrot 2.3.0 Released | parrot.org/ | Channel log: irclog.perlgeek.de/parrot/today | Priority: apply deprecations, merge branches | Review and vote GSoC applications 20:38
rurban Coke: native_pbc aren't tested anyway. 20:38
Coke rurban: I'm getting failures in the packfile tests, complaining about pbc version. 20:39
(in my working copy)
rurban Sorry. I stopped maintaining that
Coke cotto_work: ping. 20:45
chromatic steps away from his keyboard again; someone else should continue #ps 20:46
20:47 PacoLinux joined 20:53 kurahaupo joined
Coke wishlist-- 20:54
chromatic I don't see the value of storing wishlist tickets in Trac.
20:55 hercynium joined
GeJ Good morning everyone. 20:55
Coke I can't see the value of storing them in the wiki. we're even.
chromatic What do you get from storing them in Trac? 20:56
Coke The ability to treat them as rfcs/todos - report on them, set priorities, list languages that would benefit from them... 20:57
chromatic We don't take advantage of that, though. 20:58
Coke chromatic: I'm going to pull an andy - it's not worth fighting about.
chromatic: we don't take advantage of most of trac (yet)
dalek tracwiki: v125 | fperrad++ | Languages
tracwiki: more update 2.3.0
tracwiki: trac.parrot.org/parrot/wiki/Languag...ction=diff
Coke ok. mk_native_pbc seems to be broken. 20:59
chromatic I don't intend to fight or argue. I'm genuinely curious.
Andy Not fighting is "pull an andy"? If there's a name I am glad to have assigned to me, that would be a good one.
Coke I bumped PBC_COMPAT. t/pmc/packfile.t fails. I cannot succesfully run mk_native_pbc because... the packfile complains.
chromatic: What do you gain from moving them to the wiki? 21:00
21:00 rurban_ joined
chromatic They're out of Trac, for one. 21:01
They're easier to revise as documents, not as email threads for two.
particle if only there was a trac plugin to wikify tickets 21:02
Coke why is having them out of trac a benefit?
chromatic They don't represent tangible units of work.
Coke what pain point is that fixing?
chromatic Bugs do.
You can't revise them as documents in Trac. You can only add comments.
Coke so if you want a list of bugs, just run that report?
chromatic I want Trac to represent tangible units of work to which we have committed. 21:03
Coke chromatic: you can certainly do that by reporting on bugs/todos only.
(document editing is not a bad reason.) 21:04
chromatic Suppose I write a wiki page that says "Here's how to fix vtable overrides in Parrot". 21:05
dalek kudo: 10b0ff4 | jonathan++ | build/PARROT_REVISION:
Bump PARROT_REVISION to 2.3.0.
kudo: 698ba87 | jonathan++ | src/builtins/Multi.pir:
Add a .name method to Multi. Fixes various issues with introspection and also role composition, which uses this introspection.
Coke my problem with wishlist items - we have no report for "show me all the wishlist items in the wiki".
chromatic Why should I add a ticket that says "Do exactly what this wiki page says"?
Coke so I, as a developer, can go to one place and say "here's what I'm working on for parrot". 21:06
not, "tickets that are assigned, and these wiki pages" (that have no way that I no of to show ownership in a searchable way.)
chromatic What value do you get from a report that says "Here's what Coke is working on"? 21:07
Coke ... so I know what I have in my queue?
chromatic I guess I don't have that problem myself. 21:08
Coke where do you keep your parrot queue? 21:09
in your head?
chromatic Yeah, I only work on a couple of things at a time.
If I have a brilliant idea (or a lousy idea), I put it on the wiki. 21:10
If i have an important bugfix, I make a branch for it.
Coke if the wiki were organized so these bits of brilliance were easier to find, that would be less objectionable. 21:11
chromatic I agree with that.
Coke but, I am not the project manager, nor do I even play one on tv.
I'll keep my list my way, you keep your list your way, and stuff will continue to get done. 21:12
chromatic Maybe we work on really different things.
bacek ~~
21:14 clinton joined
Coke could be. 21:15
allison Coke: I've got a ticket marked "roadmap" that I'm dropping off the roadmap. what should its new type be? 21:16
dalek tracwiki: v11 | allison++ | AllisonTasklist
tracwiki: trac.parrot.org/parrot/wiki/Allison...ction=diff
allison Coke: it's a planned feature
Coke: is that "feature", "todo"?
Coke just assign it to me, I'll figure it out later.
dalek kudo: aa12544 | moritz++ | (3 files):
a wee bit uncooling
 21:17
kudo: 77b52fd | moritz++ | (24 files):
Merge branch 'cool'
kudo: d3b8168 | moritz++ | docs/ (2 files):
next release will be named "Moscow"
allison Coke: well, it's assigned to whiteknight (as in, he was planning to work on the ticket)
dalek kudo: bdc5dca | moritz++ | (3 files):
Merge branch 'master' of github.com:rakudo/rakudo
Coke allison: todo, I guess.
dalek TT #1568 closed by dukeleto++: `make smoke` reports a timeout error
TT #1568: trac.parrot.org/parrot/ticket/1568
allison Coke: okay, thanks
Coke .. or move it to the wiki. =-) 21:19
allison Coke: I'm with you on not migrating a bunch of stuff 21:20
Coke: still trying to figure out when wiki versus ticket make more sense
Coke -> 21:21
dalek rrot: r45830 | dukeleto++ | trunk/lib/Parrot/Harness/Smoke.pm:
[TT #1568] Increase the timeout for Smolder, so that spurious timeouts do not occur. brianwisti++
rrot: r45831 | NotFound++ | trunk/src/debug.c:
dirty workaround for TT #1557
allison on a completely unrelated note, I just got permission to spend half my school time next year working on an LLVM JIT for Parrot (for credit)
moritz cool
Util Tickets have permanence/persistence, even after they are closed.
Wishlist items on wikis should be cleaned-up after implementation; where does the detail go?
I see advantage in "ticket as unit of work".
Multiple commits can be tied together, and have simpler commit messages, by referring to the TT#.
Wherever the wishlists live, I think a TT should exist during the actual implementation.
(I favor wiki wishlists; this is just my $0.02 on ticketing when a wish is being "granted")
allison Util: it seems like a lot of the wiki tasklists so far link to tickets on the bigger items 21:23
Util: with just bullet points for the small items, or items that aren't fleshed out yet
Util: I also think "ticket as a unit of work" works well 21:24
21:30 cotto_work2 joined 21:31 kurahaupo1 joined
dalek kudo: 5530662 | moritz++ | t/spectest.data:
run S06-operator-overloading/methods.t
 21:34
cotto_work any objections to ripping out the cgoto, cgp and switch runcores tonight? 21:37
dalek rrot: r45832 | chromatic++ | trunk (7 files):
[IMCC] Revised line number counting to start at the *start* of a line, as

tokens before the lexer reaches EOF. This should fix a lot of line numbering problems, but more tests are welcome.		 21:38
chromatic If you do it on a branch, do it.
rrot: r45833 | bacek++ | branches/compact_pool_revamp/src/gc (2 files):
Handcraft std::lower_bound for finding blocks. bsearch couldn't find them
bacek chromatic, are you merging immutable_strings? 21:39
chromatic Happy to. 21:40
I hadn't started yet.
bacek ok 21:41
chromatic Are you asking because you want me to, you want me to wait, you're doing it, or you're waiting outside my door with a cookie and a baseball bat, hoping I'll check the mail? 21:42
japhb "a cookie and a baseball bat"? That seems awfully specific .... 21:43
chromatic You've never met the Girl Scouts around here.
japhb hehe
ash_ japhb: did you ever need any other help with the opengl configure issue?
japhb ash_, I'm actually just starting on that task now. 21:44
ash_ okay, well if you need me to test something let me know
japhb I decided that yesterday was a bad idea, because I didn't want to risk perturbing the OpenGL code that close to a release.
will do
japhb wishes he could afford a computer that qualifies as OMGFAST 21:45
dalek kudo: cdb2b8f | jonathan++ | t/spectest.data:
Turn on S12-methods/multi.t.
 21:46
bacek chromatic, I will not have this morning to do proper merge. But I can do it in next 12 hours.
chromatic I could probably do it sooner if it's useful. 21:47
bacek Up to you.
ash_ japhb: you can rent one from amazon ec2 (for cheap)
bacek I can till after all other branches merged and merge immutable_strings last. 21:48
chromatic That may be wise.
I don't know if I want to rebase to trunk in one step either.
japhb ash_, that works for compiles. Not so much for doing realtime tasks. 21:49
ash_ japhb: true, generally network lag prevents most remote connections from being "snappy"
japhb ash_, your problem seems to be Apple FAIL 22:09
ash_ whats up? 22:10
japhb There is one block of entries in the glext.h header where the lines should begin with typedef, but start with extern instead.
Every extension block has an outer #if to check for the extension, and then an inner #ifdef to check whether it should typedef some pointer types, or extern some function names. 22:11
Except that the block for GL_ARB_framebuffer_object uses extern in both #ifdef branches
*sigh* 22:12
Apple has made similar errors in the past -- where every other vendor uses the ARB-approved standard headers, and Apple writes their own and gets it wrong. 10.4 did this for one of the basic GL types, even!
NIH without any rationale drives me nuts. 22:13
dalek rrot: r45834 | bacek++ | branches/compact_pool_revamp/src/gc/alloc_resources.c:
Factor out find_block_for_buffer function
japhb *huff*
japhb thinks about how to work around this dain bramage 22:14
NotFound japhb: Note that you are talking about people that wants to dictate what programming languages you can use in his gadgets.
japhb nodnod 22:15
ash_ I wonder why they made this change... it wasn't present in older version of 10.6 was it? or did something else change that caused this to show up just recently? 22:16
22:16 kurahaupo joined
japhb Good question. I thought someone had confirmed with 10.6.1 or 10.6.2, but you may be the first with 10.6.3 22:20
Maybe they decided to prepare for supporting OpenGL 3.3 and/or 4.0, and just updated a bunch of headers ...? 22:21
dalek tracwiki: v56 | cotto++ | ParrotQuotes
tracwiki: you never know where or when bacek might appear
tracwiki: trac.parrot.org/parrot/wiki/ParrotQ...ction=diff
ash_ well, beta versions of 10.6.3 had support for opengl 3.0, (currently they only support up to 2.1) but it apparently didn't go live, (at least I don't see anything for 3.0 in the headers on my computer)
japhb Ah, yeah, that sounds like they're in transition on their headers then. 22:22
Someone oopsed.
22:27 lucian joined, kurahaupo1 joined
ash_ I actually haven't ever gotten the opengl examples to work on OS X, i have a VM that I can get them to work fine in linux on, but I have always had problems with them in OS X 22:28
japhb Huh.
Odd that.
Did you install both the OpenGL and GLUT Frameworks? 22:29
ash_ OS X has its own bundled opengl and glut frameworks, i don't have an option to install anything :P 22:30
japhb Ah, they're installed by default now then. Back in 10.3 days (last time I tried it myself), you had to choose to install them as extra components. 22:31
NotFound Oh, the beautiful easy of use of Apple products X-)
ash_ in theory, it should all be working out of the box, but thats just a theory 22:32
japhb OK, looks like I have a workaround for the warnings, at least. Running some local tests before committing 22:33
ash_ most of the time its true, but then you get these edge cases, and a lot of OSS software has trouble with some of the OS X idioms, like frameworks, and fat binaries...
japhb OK ash_, give 45835 a try. 22:39
ash_ kk, checking it out 22:40
yeah, it didn't complain on the configure step 22:41
japhb Good
ash_ i am going to run the opengl examples to see if they still give me errors 22:45
japhb OK, sounds good
dalek rrot: r45835 | japhb++ | trunk/config/gen/opengl.pm:
[OpenGL] Work around bug in at least glext.h in Mac OS X 10.6.3
 22:49
ash_ japhb: gist.github.com/373190 is the error I get 22:50
i ran ./parrot examples/opengl/shapes.pir
japhb OK, let's see ...
22:52 Whiteknight joined
cotto_work hio Whiteknight 22:53
japhb All right, where is OpenGL.framework on your system, ash_?
ash_ for me: /System/Library/Frameworks/OpenGL.framework/
japhb Ur?
purl japhb: q[merlyn]'s Unix Review columns, at www.stonehenge.com/merlyn/UnixReview/ or just a hop skip and a jump away from babylonia or short for URog or a lame way to say you're or Updating a public calendar automatically (Jul 07); Export behavior, not data (May 07); The Replacements (Mar 07); Show me your references (Jan 07); Formatting reports with Template Toolkit (Nov 06) or rong
japhb ash_, could you gist an ls in there? 22:54
ash_ japhb: updated the gist: gist.github.com/373190
Whiteknight hello cotto_work 22:55
ash_ last time i looked into it i think it was trying to load /System/Library/Frameworks/OpenGL.framework/Versions/A/OpenGL
which i think should have really been: /System/Library/Frameworks/OpenGL.framework/Versions/A/Libraries/libGL.dylib but i am not sure 22:56
japhb /System/Library/Frameworks/OpenGL.framework/OpenGL is a symbolic link?
ash_ no, its a file, here, i'll give you the file commands output on it
japhb thx
ash_ updated gist.github.com/373190 again, its a fat binary dylib really 22:57
(its at the bottom of the gist)
japhb I meant, "is .../OpenGL.framework/OpenGL a symlink to .../OpenGL.framework/Versions/A/OpenGL?" 22:58
ash_ yes, well technically it points to -> Versions/Current/Opengl where Current points to A 22:59
but yes
japhb Right, got it.
ash_ it eventually resolves there
japhb sure
OK, so I'm surprised. .../OpenGL was correct in older Mac OS X results, and it's still there, and yet it's no longer correct? Harrumph. 23:00
japhb wonders if we've tripped over another damn Mac linker change
ash_ it seemed like it might be a problem opening it because it doesn't have a .dylib extension, but i could be wrong 23:01
i remember trying to track it down at one point, but i never fully figured it out
japhb ash_, In your local copy of .../runtime/parrot/library/OpenGL.pir, can you edit the .sub _load_opengl_libs (around line 106), and add 23:02
pushd libnames, '/System/Library/Frameworks/OpenGL.framework/Libraries/libGL.dylib' to the first block of GL lib names, 23:03
and so on for the other blocks (GLU and GLUT)?
Then try again and see what happens
ash_ hmmm 23:05
i see a libGL.dylib and libGLU.dylib but no GLUT
japhb It would be in GLUT.framework
ash_ hmmm no .dylib in there 23:06
japhb shrugs
well, skip that one, and just do the GL and GLU. 23:07
I want to see if it gets farther than before
ash_ do i need to re-make it?
japhb Yeah.
ash_ yea, it got passed that, now it can't find GLUT 23:08
so, GL and GLU worked
japhb OK, that's interesting. 23:09
... And good, at least we are on the right track.
So now we have to find GLUT 23:10
Coke allison: (LLVM JIT 4 CREDIT) WOOT
japhb Can you just search for GLUT on your FS?
Coke, what happened now?
23:11 tetragon joined
Coke (girl scouts) Speaking of which, I think we still have some cookies leftover. =-) 23:11
ash_ japhb: the only one I can find is under /usr/X11/lib/libglut.dylib
japhb Hmmm, I'm guessing /usr/X11/lib/ is not in the library search path on Mac OS X .... 23:12
ash_ should I be using all the libraries under /usr/X11/lib? since there are also libGL and libGLU
Coke japhb: allison mentioned it earlier in backscroll 23:13
japhb My (admittedly weak) understanding is that those versions are the GLX versions, instead of the core versions
But heck, 23:14
let's try loading that /usr/X11/lib/libglut.dylib and see what happens.
Coke, ah, about two hours ago.
EXCELLENT.
allison++
allison Coke: :) 23:15
ash_ i'd like to help with the llvm stuff, its its possible. that was my GSoC anyway, and why I am doing an independent study on compilers (which i am building on top of the llvm) 23:16
japhb: there is still no libglutcb 23:18
japhb That's supposed to be built as part of Parrot's build
ash_ oh, okay, maybe there is then
yeah, i ran it, it seems to get further
japhb Progress! 23:19
purl progress is progress
allison ash_: are you J. Harrison?
ash_ yeah
japhb: GLUT: Warning in shapes.pir: The following is a new check for GLUT 3.0; update your code.
GLUT: Fatal Error in shapes.pir: redisplay needed for window 1, but no display callback.
japhb WTF? 23:20
allison ash_: yeah, the llvm frame builder looks good
ash_ it gets that far now, also i think I realized what the difference is now, the /usr/X11/lib use X11.app not the Aqua
allison ash_: some of the interface there will be useful for the more general JIT too
japhb OK, let's ratchet back. Can you do static-triangle.pir?
ash_ japhb: same error when I do it, i tried using XTerm to open it, i still get the error that there is no display callback 23:21
allison ash_: and yes, I expect to be able to fork out parts of the project to other people
Coke odd. we have a fish named "harrison j. fish" (had, rest his fins.) 23:22
cotto_work ash_: that's good. Helping allison will ensure that everything doesn't remain in her head.
ash_ lol @ a pet with my name
japhb wondering how the display callback fails to exist 23:23
.oO( libglutcb broken on OS X 10.6.3? )
 23:24
ash_ i think one of the most impressive uses of the llvm for a jit system is macruby, IMO, it can produces native executables, and dylibs for your code, and you can have it dump the llvm-li form of your program, its pretty impressive
well, macruby does a lot more than just a jit, but it has a jit too
allison cotto: my brain needs a "dump to text" feature 23:25
ash_: have you looked at unladen-swallow?
Coke ... I'm such a juvenile. 23:26
... and that was for your feature send, btw. :P
ash_ a bit, but not as much, more a ruby buff, hence why i have looked macruby more
japhb ash_, did static-triangle.pir manage to do *anything* other than error? For instance, did it pop up a black window, or even just empty window frame? 23:27
ash_ japhb: if X11 isn't open, it opens x11 then displays the error, no frames pop up
allison Coke: it's a popular fish name www.flickriver.com/photos/whoaitsmo...500748293/ 23:28
ash_ japhb: i think i found some useful info in the system error logs 23:30
japhb ooh, do tell
ash_ japhb: gist.github.com/373190 updated the gist, it has the relevant logs from windowserver.log 23:31
Coke trac.parrot.org/parrot/ticket/1427 - I've taken that as far as I can. hopefully LHF for someone else. 23:32
ash_ japhb: i don't know much about opengl, but "Mar 29 22:15:37 [160] agc_attach: Couldn't find any matches" looks like an issue
japhb ash_, yeah, that caught my eye as well
ash, what happens if you switch the GL and GLU library paths to the X11 ones? 23:36
ash_ i did already, they are all the /usr/X11/lib versions actually 23:37
japhb is listening to the fourth hail storm today
japhb looking at src/glut_nci_thunks.c ... I wonder what pcf_i_ddddpppddpppp is for? 23:40
That's one heck of a signature
Hmmm, I think plobsing may be batter up. 23:41
ash_ i wonder why it wouldn't load the libraries as .../OpenGL and .../GLUT 23:42
japhb Good question!
purl Yeah, it is. I'm stumped.
dukeleto ash_: because Steve Jobs hates freedom
ash_ well, there's always that too, but is it not loading it because of the extension maybe? 23:43
japhb Maybe a change in how Parrot or 10.6.3 is doing dynlib?
ash_ maybe?
purl Maybe not.
ash_ its always been there for me in 10.6, as far as i can remember
dukeleto i have not used parrot under 10.6.x at all 23:44
japhb I haven't had a local Parrot-on-Mac install since 10.3
ash_ dukeleto: which version do you have? (if any?)
doesn't Coke have 10.5?
dukeleto ash_: i have 10.5.x powerpc/intel 23:45
ash_ do the opengl examples work for you? 23:46
dukeleto ash_: never have, for quite a while now
ash_ got ya, so its not just 10.6
dukeleto ash_: i think i first tested them around parrot 1.x 23:47
the reason may have changed or snowballed since then, tho.
japhb Wait, I distinctly remember having fixed OpenGL for 10.5 for someone.
So ... at some point Parrot or OS X changed, it stopped working, and until now none of the Mac guys raised a hand.
Sigh.
dukeleto japhb: i remember us talking about it, but I don't know that i ever got it to work. is there something specific you want me to test? 23:48
japhb The problem with code not in the test suite.
23:48 eternaleye joined
japhb dukeleto, if you even could get examples/opengl/static-triangle.pir to work, that would be a big step forward. 23:48
dukeleto japhb: i get this: gist.github.com/373232 23:51
japhb ash_, can you gist your fixed lines for OpenGL.pir? 23:52
chromatic bacek, I'm not going to get to the merge today.
23:52 ruoso joined
japhb It sounds like it was Parrot that changed something in dynlib handling on Mac, rather than an OS X change, since it's happening in 10.5 as well 23:53
ash_ dukeleto: an svn diff of my changes gist.github.com/373237
bacek_at_work chromatic, slacker! 23:56
dalek kudo: 7f8397d | jonathan++ | src/Perl6/Grammar.pm:
Bring our morename rule more in line with STDs.
kudo: 22805bd | jonathan++ | src/glue/role.pir:
A nudge in the direction of getting anonymous roles working (they don't yet, this is just part one :-)).
kudo: ea1760c | jonathan++ | src/ (4 files):
Get anonymous classes (mostly) working. Passes all the anonymous class tests apart from those that depend on Numeric and Stringy. Also we can remove some hacky stuff I put in earlier.
kudo: 07fd77b | jonathan++ | t/spectest.data:
Turn on S12-class/anonymous.t.
japhb ash_, dukeleto: Since you guys are forcing the X11 version of GL/GLU/GLUT, perhaps this needs to be fixed in config/auto/opengl.pm: 183: darwin => '-framework OpenGL -framework GLUT',
dukeleto ash_: i will try that 23:57
ash_ i would use the library ones if i could find a glut.dylib that wasn't in X11...
dukeleto japhb: they should become -framework GL and -framework ?
japhb dukeleto, -framework X11 I'm guessing? 23:58
ash_ if its the x11 stuff you need just add /usr/X11/include to your includes path
and remove the -framework
japhb ash_, the problem is the linker libs, being computed in config/auto/opengl.pm 23:59
ash_ if i remember right anyway, its not using a framework anymore, X11 is supposed to be compatible with other X based libraries

Powered by Cro and the Raku® Programming Language.

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

Thank you!