6.2.12 released! | paste: sial.org/pbot/perl6 | pugs.blogs.com | pugscode.org | pugs.kwiki.org
Set by audreyt on 26 June 2006.
meppl gute nacht 00:45
ZuluOne gnight 00:47
meppl ;)
ZuluOne Random thought: Would it be useful for the smoke test result list to report arch? Maybe it already does and I am being blind. 00:50
(i.e. rather than having to go look at the full report) 00:51
The follow-up question being, would AMD64 results add anything? I could arrange for this box to run them nightly or so if it would be useful. Someone /msg me if so. 00:53
I sleep. 00:55
cmarcelo anyone up? third-party/fps' Data.ByteString conflicts with module with the same name in base package from GHC 6.5 trunk. I wonder if people using GHC 6.5 aren't yet with Data.ByteString in base, could you check with "ghc-pkg describe base | grep Data.ByteString". If that's the case it's possible prevent the conflict (which prevents build) to happen dropping references to (pugs-)fps when module already exists in base package. makes sen 02:54
se?
another thing, in third-party/fps/fps.cabal, the package Name is fps, shouldn't it be pugs-fps?
nothingmuch hola 02:56
cmarcelo olá 02:57
nothingmuch: do you use ghc 6.5? 02:59
nothingmuch nope
6.4
cmarcelo are you in YAPC::xy? 02:59
nothingmuch yes 03:00
actually 6.5.1
err
6.4.1
cmarcelo how are things going there? 03:01
nothingmuch pretty nice 03:02
just going off to the beer bof
cmarcelo =)
audreyt cmarcelo: I already renamed it as pugs-fps 03:18
cmarcelo: maybe you didnt sync it?
try rm and then svn up
svnbot6 r10938 | audreyt++ | * revert ingy's fps change 03:19
miyagawa audreyt: some perlsyck tests are failing, other than ingy's completely broken one. Is that expected? 03:21
cmarcelo audreyt: do you mind checking if your "ghc-pkg describe base" contains Data.ByteString?
miyagawa 2-scalars.t and 3-objects,t
audreyt miyagawa: 2- I just fixed 03:25
3- nothingmuch says he's wokring on it
miyagawa ok
audreyt nothingmuch: can you finish stabilizing 3-?
audreyt or was it your problem? 03:25
cmarcelo: no because my packaged GHC specifically excludes it
cmarcelo: the D.BS-in-base has another API 03:26
audreyt which totally killed interop 03:26
cmarcelo: so we either rename the module
or lost GHC 6.5 interop for now 03:27
cmarcelo audreyt: hmmm.. i think they are in process of merge now, because i manually dropped refs to pugs-fps from Pugs and HsSyck, compiled and the first minutes of make test worked...
i'm going to check libraries/base logs..
audreyt oh good
so it actually rosk
works
cmarcelo: in that case, we should restore the probe in Makefile.PL 03:29
and not use the bundled
when there's in it
cmarcelo: so maybe help me reinstall r10900? 03:30
it was a revert
cmarcelo yep. and do some processing on HsSyck.cabal too..
audreyt so revert that revert
feel free to fix it
I really need to crash now :)
cmarcelo revert revert? 03:31
audreyt svk desc
and then apply reverse patch
or something. improvise :)
cmarcelo ok
audreyt cmarcelo++
fortunately it only affects 6.5-thisweek users
so we don't need to pullback 6.2.12 or do 6.2.13 for them
:)
cmarcelo btw, just fixed one of the strange problems in HSIter.. (not the illegal instruction).. 03:32
commit soon..
spinclad smoke is back to 3:30 03:35
ext/Perl6-Grammar/t/basic.t: 40.00% (3 4 6 7 9 10 12 13 15 16 18 19 21 22 24) 03:48
t/rules/from_perl6_rules/named_cap.t: 65.38% (15..23)
t/rules/from_perl6_rules/stdrules.t: 99.52% (22)
t/rules/from_perl6_rules/word.t: 57.14% (3..5; 6 ok)
overall 99.76% 03:49
cmarcelo: pong 03:50
cmarcelo spinclad: I (think) solved the segfault problem when a lot of memory was being used before calling HSIter... 03:52
spinclad oh good. glad one of us had a handle on it, since i didn't 03:53
nothingmuch audreyt: sorry, beer bof 03:54
soon
cmarcelo spinclad: what are your plans re judy. submitting your patches for trunk? or still more changes in mind? 03:56
spinclad will submit a patch this week. i want to fill in some code doc, but no new features planned, and so far it seems stable. 03:58
cmarcelo great. OT: the original judy tree seems a little unorganized, i wonder if there's unused/deprecated stuff around.. or is just all MACRO magic.. 04:02
spinclad Judy1/JudyL ? yeah, there's a lot of macro expansion there. haven't look in detail though 04:03
do you know what caused the segfaulting? 04:07
cmarcelo my mistake. =| I wasn't proper initializing Index and Length for HSIterFirst. I suppose It worked when I was "lucky" and memory was clean, but when mem was a mess (too much memory used before calling), they contained garbage.. 04:10
nothingmuch audreyt: pingf 04:11
svnbot6 r10939 | cmarcelo++ | * Judy.Map: Fixes segfault problem when using a lot of memory before 04:16
r10939 | cmarcelo++ | calling Iter functions.
spinclad that is a tricky interface, and i'd like a better, but this is as good as i found. suggestions welcome. i'd have gladly done without an Iter structure, but it needs someplace to build the current index string in order to return it, and this structure worked for that. 04:17
spinclad *to bed* g'night all. happy pugsings & 04:30
cmarcelo bye 04:31
nothingmuch audreyt: all tests pass 05:13
ingy: fuck you man
seriously
die or something
clarification: ingy is the one that inspired me to update the tests 05:16
and then make syck actually pass them
but at least it's well behaved now 05:17
obra :)
nothingmuch ingy++ overall ;-)
i still have to write my slides though
`/ tomorrow `/
obra i need to revise my slides. but now, i need to sleep, not try audrey's crazy new multiversioning hack 05:18
nothingmuch multiversioning?
if audreyt is still awake ping for YAML::Syck moosing 05:19
svnbot6 r10940 | cmarcelo++ | * Get fps_dummy back so fps doesn't conflict with GHC-trunk Data.ByteString from 05:20
r10940 | cmarcelo++ | base package. Dons (fps maintainer) said that will merge again his fps with trunk
r10940 | cmarcelo++ | in next few weeks.
cmarcelo any good soul could check if I didn't broke "perl Makefile.PL ; make" for GHC 6.4.x people? I tested here and it was fine, but two tests are better than one.. 05:21
nothingmuch cmarcelo: will try now
cmarcelo tks 05:22
huhlig-annoyed stupid question 05:25
does perl6 have 3270 emulation capavility
nothingmuch 2370? 05:26
huhlig-annoyed 3270
tn3270
I guess thats a no 05:31
PerlJam huhlig-annoyed: That's a "yes" 05:57
huhlig-annoyed: yes, it's a stupid question :)
huhlig-annoyed: though I'm sure at some point someone may write a module to do 3270 emulation just like we have Net::Telnet and Net::FTP, etc. 05:58
cmarcelo actually, it's possible (I think) to use perl5 modules in perl6 too.. 06:01
nothingmuch: any results? 06:06
cmarcelo audreyt: build system works fine here with both GHC 6.5 and GHC 6.4.2... The only thing I ignored was a "-p" flag being used when compiling FPS_Dummy (or was the native?), since this part was moved out from Makefile.PL i didn't seem so important. 08:40
i'll try to sleep a little bit... & 08:41
mj41_ hello, how to fix win32 nmake error "expanded command line ... too long"? 11:55
Limbic_Region Juerd ping 12:06
audreyt and/or anyone working on the change log - I assume this has already been caught but 12:07
Removed support for Parrot 0.4.4 or below and added support for GHC 0.4.5
s/GHC/Parrot
Limbic_Region audreyt ping 12:23
audreyt pong
svnbot6 r10941 | audreyt++ | * oops, changelog typo caught by Limbic_Region++
Limbic_Region know anything about icfp contest this year 12:24
apparently supposed to start in a few weeks but no info updated
audreyt no idea 12:25
Limbic_Region ok - thanks
Limbic_Region will relay the info
Limbic_Region speaking of... 12:28
salutations b_jonas
b_jonas hello 12:29
Limbic_Region as I mentioned in #perlmonks, audreyt has no idea but #haskell might be a good place to check 12:30
unfortunately, I am limited to just #perl6 and #parrot from work
well, theoretically I could configure CGI:IRC to let me access any channel on any network but since I am using Juerd's implementation - I will stick to the channels he feels are appropriate for the development of p6 12:31
b_jonas there's no #parrot channel on this server, you know 12:34
or it's well-hidden 12:35
Limbic_Region b_jonas - no, #parrot is on perl.irc.org 12:38
err - irc.perl.org that is
dazjorz Hi there 12:40
b_jonas oh, that one 12:45
renormalist Hi! 13:04
renormalist In pugs/docs/Perl6/Spec/Concurrency.pod there are lines "multi foo () { ...no rebinding... }" 13:05
Is such a "multi" without "sub" or "method" or "coro" legal?
Ah, just found in S06: "If the routine keyword is omitted, it defaults to sub." 13:06
should read before ask. :-)
dazjorz b_jonas, ? 13:12
kolibrie rejoices (his wifi is working in the auditorium) 13:16
wolverian yay, syn r9718 made for syntax sane again 13:17
mj41_ my makefile pm_to_blib section is too long for nmake ... 2245B :-( 13:23
b_jonas dazjorz: pong 13:28
dazjorz b_jonas: ping 13:29
reversed world :)
b_jonas you said "?" above
dazjorz yep
you said "oh, that one" above that 13:30
b_jonas yes, that was for L~R saying "no, #parrot is on " ... "irc.perl.org"
sorry
dazjorz oh right
ok
:)
[14:40] <dazjorz> Hi there
[14:45] <b_jonas> oh, that one
sounded a bit... 13:31
lol
b_jonas in case anyone's still interested in the icfp contest thing,
dazjorz like "oh, him again" 13:32
b_jonas I haven't found much,
I've found two references on "radar.spacebar.org/", the weblog of Tom Murphy 13:33
Tom Murphy has the homepage "www.cs.cmu.edu/~tom7/", is/was a member of the CMU POP group ("www.cs.cmu.edu/afs/cs/Web/Groups/pop/pop.html") the icfp teaser site ("icfpcontest.org/index.shtml") refers to as organizer 13:34
and also Tom is the registrar of icfpcontest.org according to whois 13:35
however, both references on that blog are short and cryptic
I'd better write this on a html site instead of repeating it everywhere 13:36
dazjorz b_jonas: a html site? 13:41
svnbot6 r10942 | renormalist++ | * cperl-mode.el: coro, yield, multi without routine keyword
dazjorz b_jonas: are there any other languages to make a site in?
b_jonas oh well, don't be so picky 13:44
kolibrie good morning gaal
gaal hello hello
where you at yo? 13:45
kolibrie I'm in the auditorium, very front
center section on the left side
Perl 5.10 overview
gaal I'm in Expo, nice talk about perl -d
rgs who does that talk ? 13:46
about 5.10 overview
kolibrie yapcchicago.org/the-schedule/tuesday
lambdabot Title: "YAPC::NA 2006 &raquo; Tuesday"
rgs thanks
kolibrie rgs: Steve Peters
gaal Steven Lembark is doing the debugger talk. Must be a Steven morning. 13:47
gaal looks for moose
rgs Steve_p is very qualified to give a good talk about 5.10.
kolibrie it's very nice, he's just telling us all the cool Perl 6 things they are backporting 13:48
rgs not only backports, there are genuine improvements too
gaal use feature * 13:49
I heard last night from dragonshild that lexical pragmas are finally in?
s/sh/ch/
kolibrie I believe so
gaal yay
rgs yes, you can wrote your own ones in pure perl 13:50
perldoc perlpragma
I suppose a new 5.9.x release will be for not too long
xerox audreyt: how does one do omioutliner -> keynote as you were saying sometime? 13:58
mj41_ docs/Perl6/Makefile pm_to_blib section is too long for win32 nmake 1.50 ... 2245B 13:59
gaal mj41_: you need a newer nmake :) 14:01
mj41_ gaal: can't find newer nmake :-) 14:07
Limbic_Region wonders if mingw32-make works provided that msys is not in the path 14:10
gaal mj41_: see /msg :) 14:11
Limbic_Region tries it as it would be nice to add that to the readme for the poor schmucks on M$ that don't want to download the entire SDK just to get an nmake.exe from this century
gaal wow the network is sucking here
Limbic_Region gaal is nice enough to provide a copy for individuals on a case-by-case basis but anyone not showing up here complaining is SOL
kolibrie gaal: still okay here 14:12
mj41_ seems like we need to set makemaker max_exec_len under 2048 14:15
Limbic_Region mingw32-make is a no-go 14:19
perhaps dmake.exe?
mj41_ makemaker_args ( 14:20
....
tool_autosplit => {MAXLEN => 1500},
mj41_ grrr ... still too long :-( 14:28
svnbot6 r10943 | fglock++ | v6.pm - emitter - added '.perl' 14:33
svnbot6 r10944 | fglock++ | v6.pm - 'class Point;' compiles to 'package Point; use Moose;'; 14:59
r10944 | fglock++ | - Makefile.PL recommends 'Moose' => 0;
gaal svk st shows: M third-party/fps/fps.cabal - should this file be in source control? 15:01
kolibrie audreyt says yes 15:02
gaal the diff is -Name: pugs-fps / Name: fps
kolibrie oh, nevermind, after her fix it should be deleted 15:03
gaal is putting Best YAML in the test suite for spedups
saves almost a minute per thing-that-parses-tests.yml on my box 15:04
dasein Where is there the most need for new volunteers in the project? I'd like to contribute, if possible. 15:10
gaal ingy's talk is strating 15:24
niner dasein: AFAIK test writers are always needed. Apart from that I'd guess it depends on your talents/knowledge 15:31
svnbot6 r10945 | gaal++ | r11013@sike: roo | 2006-06-27 10:15:48 -0500 15:35
r10945 | gaal++ | * Use best available of YAML::Syck and YAML.pm in test
r10945 | gaal++ | system, for speedups during make smoke.
gaal (Bundle::Pugs::SmokeKit updated on CPAN.) 15:36
dasein niner: thanks for the reply. 15:37
mj41_ ExtUtils::MakeMaker 6.30, MM_Any, sub split_command, line 147 15:57
# set aside 20% for macro expansion.
my $len_left = int($self->max_exec_len * 0.75);
err 0.75 is my "fix" ... orginal is 0.80 :-) 16:00
0.70 work for me 16:01
svnbot6 r10946 | benedikth++ | util/catalog_tests: Make windows users happy + other bugfixes/improvements 16:22
Aankhen`` suddenly feels happier.
visq_ my pleasure 16:23
svnbot6 r10947 | fglock++ | v6/Moose - 'has $x' 16:25
mj41_ another error 16:40
subdirs :: cd ext\Algorithm-TokenBucket $(MAKE) ... cd ..
better is cd ..\.. on win32
*** Successfully built! 16:43
svnbot6 r10948 | fglock++ | v6/Moose - 'method'
fglock muses 16:44
gaal what are we doing for lunch today? 16:55
fglock bbiab &
mj41_ ExtUtils::MakeMaker::MM_Win32 17:08
sub cd ... B<NOTE> This cd can only go one level down. So far this sufficient for what MakeMaker needs.
svnbot6 r10949 | fglock++ | v6/Moose - parser - 'method' is a statement 17:19
pasteling "fglock" at 200.17.89.80 pasted "v6/Moose output" (26 lines, 398B) at sial.org/pbot/18030 17:39
prefiks fglock: nice 17:49
Aankhen`` :-o 17:51
fglock++
# sweet!
fglock it still has an enormous TODO list :) 17:58
Aankhen`` nods.
It's still cool to see what works. =)
fglock bbiab & 18:02
mj41_ I sent my Win32 MakeMaker patches to [email@hidden.address] 18:08
ext\* ... cd ..\.. fix 18:09
and command line too long .... "set aside 20% for macro expansion" "fix"
ZuluOne make smoke isn't working for me 18:39
It says I don't have the required modules and to install Bundle::Pugs::SmokeKit, but it is already installed. 18:40
TimToady ZuluOne: me too 18:46
looks like it wants something called Best. Maybe that's something new... 18:49
Limbic_Region yes, Best.pm was uploaded to CPAN yesterday 18:50
ZuluOne ah
I take it we should just wait for it to hit mirrors.
or I should anyway
Limbic_Region and has been updated twice today apparently
search.cpan.org/~gaal/Best-0.03/
lambdabot Title: "Gaal Yahas / Best-0.03 - search.cpan.org"
Limbic_Region so blame gaal 18:51
he should have probably updated Bundle::Pugs::SmokeKit
fglock nothingmuch, stevan: ping - I need a few p6-to-Moose examples to teach v6.pm 19:11
nothingmuch fglock: can't right now 19:17
talk in 3 minutes
see audrey's slides
they contain p6 and moose side by side
TreyHarris audreyt: yuval's talking about your slides for which talk? 19:19
fglock link? 19:21
svnbot6 r10950 | fglock++ | v6/Moose - '$.x'
mj41_ syntax error at ./util/yaml_harness.pl line 20, near "] qw/LoadFile DumpFile/" 19:23
cmarcelo fglock: deploying perl and MS talk have some examples p6 <-> moose.. [if that's what you are asking].. they're in pugs.blogs.com 19:25
fglock cmarcelo: found it - thanks! pugs.blogs.com/talks/npw06-deploying-perl6.pdf 19:29
svnbot6 r10951 | rootmj++ | * util/yaml_harness.pl - fix syntax error 19:30
cmarcelo i'm getting checksum mismatch trying to install Best-0.03 from CPAN. anyone had sucess installing it? 19:34
Limbic_Region cmarcelo - the checksum errors are due to \r\n versus \n 19:36
IIRC anyway 19:37
TimToady I just installed using cpanp, but it doesn't say which version it put in. But my smoke's running now. 19:38
svnbot6 r10952 | fglock++ | v6/Moose - added example 'moose.pl' 19:42
svnbot6 r10953 | rootmj++ | * another use Best syntax error fix 20:12
rindolf Hi all. 20:17
buu Limbic_Region: Hello! 20:25
Limbic_Region: Are you here?
gaal greetings kolibrie 20:31
sorry about the unBest folks. the bundle is updated to include it. 20:35
kolibrie greetings gaal, sorry we missed for lunch
gaal good news is a ~2min speedup in smokes on slow machines 20:36
kolibrie oh, very nice
kolibrie has slow machines
gaal kolibrie - that's okay, there's still dinner :)
kolibrie :)
Limbic_Region buu yes but I am headed home 20:42
well - guess it wasn't that important or else he has already left himself 20:44
TTFN all
gaal eeep, audrey's talk is covering a lot from my peek talk, which follows it immediately 20:45
buu gaal: He goes very fast =[ 20:47
gaal hopefully I won't go too slow for the audience :) 20:48
kolibrie gaal: audrey gives the intro, you get to fill in the gaps 20:50
gaal i get to inflict haskell on everyone :) 20:51
kolibrie moose!
gaal caribou!
svnbot6 r10954 | Sage++ | Single elemements of arrays and hashes are now translated correctly (i.e. $hash{$key}->%hash{$key} and $array[$i]->@array[$i]). Some new utility functions in preperation for more translations. 21:02
svnbot6 r10955 | Sage++ | Updating translate.hs to match translate-sage.hs. Adds a lot of functionality. See revisions for tranlate-sage.hs for more info. 21:06
rindolf Hi gaal 21:06
gaal hi, up for a talk now :) 21:07
rindolf IRC... there goes my productivity.
svnbot6 r10956 | Sage++ | Fixed typo in last revision that messed up a comment.
rindolf gaal: yeah.
gaal: I can say the same about you.
gaal: the first bug for two WWW-Search-* modules I wrote as a contractor was filed.
gaal: it claims that the module dies on empty results. I believe I checked them, tough. 21:08
Maybe the HTML got broken, or it's some weird variation.
My HTML::PullParser code is not very robust. 21:09
I just do a ->look_down() and often don't check for undef, etc.
Don't know if I can use XPath there.
I bitched about irc.perl.org to someone on AIM some minutes ago. 21:10
svnbot6 r10957 | fglock++ | v6.pm - moose.pl compiles; 'ro' access not implemented yet;
r10957 | fglock++ | - a change in 'my' parsing broke t/07-binding.t, t/07-range.t
rindolf Is svnbot6 written in Perl6? 21:11
kolibrie I think so
rindolf kolibrie: nice. 21:13
Maybe that's why it has a 6 appended to it.
kolibrie rindolf: examples/network/svnbot.p6 21:17
ZuluOne cool
rindolf kolibrie: OK.
kolibrie: now I'd like to write a program to analyse the log files of Freenode's #perl in Smalltalk. It would be very easy to write in Perl, but I'm trying to get fluent with Smalltalk. 21:19
Maybe I should write it in Perl first, and then translate it.
I just want a word count and stuff.
svnbot6 r10958 | fglock++ | v6 - fixed 'my' parsing - t/07-binding.t passes again 21:21
kolibrie I think stevan knows some smalltalk 21:22
rindolf kolibrie: yeah. 21:30
svnbot6 r10959 | fglock++ | p6 - t/07-range.t passes again
rindolf kolibrie: I'll also probably write it in Common Lisp which I intend to learn next.
kolibrie: did you ever design a language?
kolibrie: I did.
www.shlomifish.org/Vipe/rindolf/ 21:31
lambdabot Title: "Rindolf Specification Document"
rindolf Now I'm designing another one.
lambdabot: thanks!
audreyt fglock: I finished my talk; people seem to like it :) 21:39
clkao audreyt: it's excellent 21:43
audreyt danke 21:44
rindolf Hi audreyt 21:46
audreyt: can I quote you from IRC about being a consultant/contractor? 21:47
rindolf audreyt: the talk we had the other day? 21:47
audreyt: with/w.o. proper attribution.
fglock audreyt: great - I'm implementing v6/Moose :) 21:48
try moose.pl in P-C-P6/
rindolf fglock: wasn't Moose based on the Perl 6 object system? 21:49
Shouldn't you guys mention YAPC in the topic? Just a suggestion.
fglock rindolf: yes - I'm using it as a runtime
rindolf fglock: I see.
fglock v6 is a perl5 module
rindolf fglock: are namespaces in Perl6 first-order objects? 21:50
I suppose they are.
Are there anonymous namespaces?
fglock rindolf: anonymous classes, maybe? yes 21:51
rindolf fglock: no anonymous namespaces.
fglock: that you can clone a namespace, manipulate it and then bind it in a different place.
fglock: or use it as it is.
Don't know if it makes much sense, but it was a part of Rindolf ( www.shlomifish.org/Vipe/rindolf/ ). I don't know how useful or implementable it is. 21:53
lambdabot Title: "Rindolf Specification Document"
fglock module { my $x } -- $x belongs to no namespace, I think
rindolf fglock: I see.
fglock later & 21:54
rindolf I'd better get to sleep, but first I'd like to write some code. Refactoring - here I come.
fglock audreyt: I have to find out what I'm going to talk about in YAPC::SA 21:56
can I have subdirs in t/ (in perl5)? 21:57
audreyt fglock: sure 22:18
rindolf: sure 22:19

Powered by Cro and the Raku® Programming Language.

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

Thank you!