Please test pre1: pugs.blogs.com/pre1/Perl6-Pugs-6.2.12.tar.gz | paste: sial.org/pbot/perl6 | pugs.blogs.com | pugscode.org | pugs.kwiki.org
Set by audreyt on 24 June 2006.
00:34 cognominal joined
spinclad .oO{ the quiet before the storm } 00:39
00:39 Ymmv_ joined 00:57 dduncan joined
kolibrie has arrived at the campus for YAPC 01:10
01:22 pdcawley joined 01:30 pdcawley joined 01:36 tup joined 01:51 pdcawley_ joined 02:01 scook0 joined 02:02 pdcawley joined 02:13 pdcawley joined 02:18 macroron joined 02:19 jiing_mIRC joined
kolibrie sleep & 02:20
02:20 pdcawley joined 02:22 tcliou joined 02:24 tcliou joined 02:25 tup joined 02:28 christopher joined 02:31 miyagawa joined 02:32 rashakil_ joined 02:35 miyagawa joined 02:37 Ymmv_ joined 02:38 pdcawley joined 02:55 pdcawley joined 02:58 rashakil__ joined 03:02 cognominal joined 03:15 pdcawley joined 03:21 weinig|away is now known as weinig 03:22 weinig is now known as weinig|sleep 03:27 pdcawley_ joined 03:28 nothingmuch joined
nothingmuch hola from SSV 03:28
obra hey nothingmuch 03:29
where are you?
nothingmuch room 2.2
with stevan and robkinyon 03:30
where are you guys?
(north building)
miyagawa my room is North 2.7
nothingmuch commits YAML-Syck local branch 03:31
audreyt yay
I think we can roll a release of Pugs tonight
nothingmuch obra: i fixed yout test 03:32
03:32 stevan joined
nothingmuch it was a weirdness of ok accepting code refs and invoking them 03:32
i also added leak tests
and partially plugged a leak in Storable
when the eval fails
obra nothingmuch: 4.2 center 03:33
pop up and say hi
nothingmuch will do shortly
nothingmuch has been moving for 24 hrs now
stevan hey obra
nothingmuch after ~3 hrs of sleep the night before
so shower is first priority
obra hey stevan
03:43 svnbot6 joined
audreyt pre2: 03:45
pugs.blogs.com/pre2/Perl6-Pugs-6.2.12.tar.gz
Preflight! pugs.blogs.com/pre2/Perl6-Pugs-6.2.12.tar.gz | paste: sial.org/pbot/perl6 | pugs.blogs.com | pugscode.org | pugs.kwiki.org 03:46
stevan hey audreyt 03:46
audreyt stevan: hi 03:47
svnbot6 r10910 | fglock++ | v6.pm - fixed 'sub' ast
audreyt going to visit here a bit? :)
fglock: do we get a joint release? :)
pugs 6.2.12 and 0.01 of v6
stevan audreyt: are you in chicago? 03:49
audreyt I'm in room 4.2 center 03:51
want to come over and chat a bit?
stevan ah,.. with obra
yes, I would love too
I will come by in a couple of mintutes
audreyt woot
stevan has to shake the travel cobwebs from his brain 03:52
03:52 cognominal joined 03:55 fglock joined
fglock audreyt: I'm cleaning up the tests - there are still 3 warnings to go 03:58
audreyt ok
fglock also s/# TODO/# skip TODO/ which is easier to see
but it looks ok, tested in windows & linux 03:59
do you think it can be released as-is? would you review v6.pm POD? 04:00
04:02 pdcawley joined
svnbot6 r10911 | fglock++ | v6.pm - s/TODO/skip TODO/ 04:02
fglock uploaded Pugs-Compiler-Rule-0.06 to cpan - adds Precedence.pm, needed by v6.pm 04:11
audreyt yay 04:14
fglock it's ready, I think - make test emits some warnings, but it pass 04:15
svnbot6 r10912 | fglock++ | Pugs::Compiler::Rule - version 0.06; updated v6.pm Makefile.PL
fglock I should probably sleep before release - this gives some time to fix any problems 04:17
audreyt yup 04:18
I should do that too
fglock :) 04:19
04:19 pdcawley joined
fglock later & 04:20
04:20 fglock left 04:26 pdcawley joined
dduncan I just noticed pre2 ... will download and test now ... after a restart, though that power conditioner is coming soon 04:32
04:34 justatheory joined 04:35 dduncan left
audreyt ok... 04:38
04:40 Qiang joined 04:43 Qiang joined, justatheory joined 04:44 dduncan joined 04:45 pdcawley joined
dduncan Makefile.pl in progress ... 04:49
'make' in progress 04:54
05:03 buetow joined
svnbot6 r10913 | audreyt++ | * add GHC to path when building pugs 05:06
05:07 mtve joined
svnbot6 r10914 | audreyt++ | * repair JS compilation since the macro is no longer needed 05:18
05:22 pdcawley joined
audreyt dduncan: I'll re-upload pre2 with the JS compilation fixes; however that will not affect "make smoke" result 05:23
so please keep the smoke cycle going 05:24
I'm going to post on my blog for final round of -pre2 testing
and then release it tomorrow during conference
dduncan okay
audreyt (reuploaded)
(I'm running "nmake test" in Win32 via Parallels now)
dduncan I see
I ordered that program at the pre-order price, so I'll have it when I get an Intel Mac 05:25
audreyt cool
dduncan is your copy the pre-final or final version? 05:26
audreyt well 05:27
svnbot6 r10915 | audreyt++ | * proxy.t: disable the test due to inconsistent reports due to missing feature
dduncan almost done make (at prelude) ...
audreyt the one currently at the pre2 url 05:28
is final unless someone find showstopping bugs in it
2691084 bytes
dduncan er, my question was about Parallels
audreyt sha d996445c7c35b3a52e668a3c42dd7c1880ce2a31 05:29
oh
dduncan they had a number of free previews
audreyt sorry. I think it's trial
free previews
dduncan I ordered the first GA at pre-release $40 price
now begun 'make smoke' 05:30
fyi, I ordered a RAM upgrade for my 4 year old machine as well as the power conditioner, so with the then 1.75GB of RAM, I should freely be able to set smoke-concurrent to a number much higher than 4, if it would help significantly (the machine is dual-cpu) 05:31
but maybe it won't ... going from 2 to 4 had little effect 05:32
so the smoke should be done in about 2.5 hours
fyi, after you've got the final 6.2.12 on CPAN, I will feel free to commit some major improvements to my ext/ modules prior to the hackathon, assuming I have time away from $job 05:33
audreyt excellent
dduncan among other things, it involves changing Relation into a Role, and altering its interface to allow a wider bredth of implementations
audreyt the final release is expected to happen same as v6
v6.pm
and then I can move to Role support 05:34
as well as other OO things
but the speedups was felt as more important
(and indeed it may be more helpful in the long run for test-driven development is possible again :))
dduncan I have a language question along those lines ...
audreyt yes? 05:35
dduncan if one declares a Role 'Foo', and a Role or Class 'Bar' which does() 'Foo', and then a Class 'Baz' that does() 'Bar' ...
I would like in that situation to declare a set of more generic multi-methods in Foo, then let Bar and/or Baz override those with more specific versions ... 05:36
so that, for example, ...
say that 'Foo' declares a method quux() that takes a 'Foo' argument, then Bar declares a quux() that takes a 'Bar' argument, and so on ... 05:37
could it be said that each successive role is more specific than previous ones ...
eg, that Bar is more specific than Foo ... 05:38
so that if quux() is called with something that does Bar, then Bar's method will be called, otherwise Foo's will, assuming it does Foo?
audreyt yes.
dduncan my rationale is this ...
audreyt you don't need methods
just regular multisubs
will do
dduncan sure, and I may do that ... 05:39
Ideally, all 'Relation' subs can be used like a non-method operator
and hence will also have forms that vary in number of arguments, some all the way from zero to infinity (in reason) 05:40
audreyt *wave* *sleep* &
dduncan okay, I"ll stop
audreyt maybe ->p6l :) 05:41
good night!
dduncan you too
but you answered my main question, assuming that you know what I was asking
05:44 fglock joined 05:48 tup joined 05:50 pdcawley joined
fglock audreyt, ingy: v6.pm will need an updated Module::Compile release 05:52
just remembered: v6.pm will need an updated Module::Compile release
heh 05:56
I think v6.pm may already be able to run pil2 - will try later 05:58
06:08 tup joined 06:09 pdcawley joined 06:19 pdcawley joined 06:21 fglock left 06:22 tup joined, kanru joined 06:35 visq joined 06:45 masak joined 07:06 iblechbot joined, Aankhen`` joined 07:11 kane__ joined 07:15 scw joined 07:22 pdcawley_ joined 07:25 christopher joined 07:26 SM_ax joined 07:33 dasein joined 07:42 pdcawley joined 07:49 Aankh|Clone joined 08:02 mjk joined
dduncan smoke of pre2 complete and uploaded; results are, under normal runcore, release/darwin, all tests pass, no failures, no unexpected successes 08:20
runtime 143 minutes on a dual-1ghz-g4 with concurrent=4
I look forward to tomorrow's official release of this 08:21
good night
08:21 dduncan left 08:29 pdcawley joined 08:30 lisppaste3 joined 08:31 Skyddsman joined 08:51 pdcawley joined 08:56 elmex joined 08:59 didie_p6 joined 09:05 pdcawley_ joined 09:21 cognominal joined 09:24 como joined, pdcawley joined 09:27 kanru joined 09:34 pdcawley joined 09:38 pjmm joined 09:42 pdcawley joined 09:56 drrho joined 10:01 pjmm left 10:03 paultcochrane joined 10:16 pdcawley_ joined
svnbot6 r10916 | Aankhen++ | * ext/libwww-perl/lib/HTTP-Status: aliasing &RC_MOVED_TEMPORARILY to &RC_FOUND should probably be a run-time operation. 10:23
Aankh|Clone wonders why / is listed as modified. 10:29
10:29 Aankh|Clone is now known as Aankhen`` 10:32 cognominal joined 10:42 SM_ax joined
svnbot6 r10917 | audreyt++ | * correct two more test failures on Win32 10:53
11:04 chris2 joined 11:10 pdcawley joined 11:18 b_jonas joined 11:24 weinig|sleep is now known as weinig 11:33 weinig is now known as weinig|work 11:37 iblechbot joined 11:38 pdcawley joined
kolibrie good morning all 11:41
webmind good morning 11:42
kolibrie wonders when breakfast is
webmind before lunch ?
kolibrie ah, probably
11:46 orafu joined 11:47 cognominal joined 11:50 szbalint joined 11:52 renormalist joined 12:12 marmic joined 12:28 bpphillips joined 12:35 christopher joined
svnbot6 r10918 | audreyt++ | * Make precompiled Test.pm part of the build process to speed up testing. 12:39
12:42 Limbic_Region joined
svnbot6 r10919 | renormalist++ | * cperl-mode.el: better indenting for class blocks and proto|multi sub(methods) 12:45
12:45 kane-xs joined 12:46 pdcawley joined 12:48 lumi joined 12:56 pdcawley joined 12:57 hcarty joined 12:58 rzoz joined 13:03 shlomif joined
shlomif Hi all. 13:03
Any YAPCers here? 13:04
TimToady: hi.
13:04 shlomif is now known as rindolf
kolibrie rindolf: yes 13:04
rindolf Hi kolibrie
kolibrie: kolibrie is a kind of bird, right? 13:05
kolibrie correct
hummingbird, to be exact
rindolf kolibrie: OK. 13:06
kolibrie: are you a girl?
kolibrie not that I know of
rindolf kolibrie: OK.
kolibrie: so how's YAPC so far?
audreyt: here?
kolibrie nothing has happened yet, as far as I know 13:07
except I ate breakfast
rindolf I don't recall exactly how audreyt makes a living?
Something like Erdos I think.
kolibrie: OK.
What time is it in .il.us?
kolibrie 8:07
rindolf kolibrie: I see. Pretty early. 13:08
Limbic_Region rindolf - she is hacker for hire AFAIK
rindolf TimToady is up.
Limbic_Region: I see.
Hi Limbic_Region
Limbic_Region: are you at YAPC?
Limbic_Region no rindolf
rindolf Limbic_Region: OK.
Maybe someone should change the subject here. Someone with chanops. 13:09
kolibrie well, except preflight is what is going on
audreyt yo 13:14
kolibrie good morning audreyt
audreyt rindolf: I work as a consultant 13:15
kolibrie audreyt: is your room super freezing cold?
audreyt by-hour telecommuting etc
kolibrie: there is a dial on the wall. it's defaulted to freezing
rotate it the other way gives you nonfreezing room
kolibrie I haven't found mine yet 13:16
kolibrie looks
13:16 pdcawley joined, SamB joined
kolibrie mine must be very hidden, cause I'm not seeing it 13:17
audreyt use the DORM explorer API
kolibrie I don't think I have that installed 13:18
audreyt mine is on the left side of the shared restroom
rgs and file a bug report to get saner defaults
kolibrie in the restroom? /me looks there
audreyt no, just outside
door.prev_sibling
Limbic_Region kolibrie - if you are paying for this room then talk to management - service industry should provide service 13:19
kolibrie yep, mine is missing 13:20
Limbic_Region: good idea
13:20 rashakil_ joined
kolibrie audreyt: I might like your help with wireless, too. I see four routers, but dhclient doesn't ever get a response 13:21
13:23 macroron joined 13:24 fglock joined
Limbic_Region anyone know why make realclean is intermittently necessary to build pugs after check-out despite not having done anything other than checkouts and builds? 13:25
13:26 pdcawley joined
audreyt kolibrie: why are you online then 13:27
kolibrie wall jack
audreyt kolibrie: I'm on "iit" which requires a login first; check your room handout for how
going to the conf now & 13:28
kolibrie I have the handout, which I followed for the wall jack
fglock audreyt: should I release v6.pm?
kolibrie audreyt: I'll head there too
fglock not sure about the state of Module::Compile though
audreyt fglock: the M::C support for v6 has been restored 13:29
I'll find ingy today and tell him to release M::C
fglock I mean, is it in cpan yet?
ok
thanks!
audreyt so mark the dep at 0.16
and you'd be good to release
0.00_01 or 0.01 is both fine 13:30
I'll bbiab
fglock 0.001 - as it will need a lot of releases :)
13:31 vel joined, nothingmuch joined, colomon joined
colomon pre2 make test: 13:32
Failed Test Stat Wstat Total Fail Failed List of Failed
-------------------------------------------------------------------------------
t/perl5/class.t 1 256 ?? ?? % ??
1520 subtests skipped.
Failed 1/521 test scripts, 99.81% okay. 0/9469 subtests failed, 100.00% okay.
pre make install:ghc-6.4.2: unknown package: pugs-fps-0.7 13:33
Build failed: 256 at util/build_pugs.pl line 99.
hcarty colomon: Did you build pugs with embedded perl5? I wrote that test, and it's my first attempt at writing a pugs test, so it's quite possible that I mixed something up 13:41
13:42 rindolf joined
rindolf Hi all. 13:42
Sorry, got disconnect. Again.
audreyt: so what do you consult about?
Limbic_Region she's on her way to the conference 13:46
audreyt colomon: cool, I've fixed that since 13:47
hcarty: you put 'use' outside eval
which triggered an error
hcarty ack, gotcha. Thanks for the tip. 13:48
audreyt colomon: so "make install" still fails? hm
which platform?
hcarty I can submit a fix to the test if it hasn't been yet?
audreyt hcarty: it's fixed, just svnup
colomon Suse linux (10?), x86-64 13:49
audreyt ok. will look into it
did you use PREFIX or --prefix or anything like that?
did you run "make install" as root? 13:52
colomon: also is this make install on "pre2"?
colomon I did not use prefix or anything like that. I ran make install as root. 13:57
13:57 pdcawley joined 13:58 frederico joined
audreyt rindolf: mostly jifty, some haskell, some pugs 13:59
14:00 hexmode joined
rindolf audreyt: I see. 14:01
audreyt: how long did Jifty exist?
audreyt: what was your last day job?
After I quit working for szabgab, I played with the idea of becoming a consultant. I placed a "Hire me" page on my site, but no-one would hire me. 14:03
I guess you need the right connections.
audreyt rindolf: the last day job was building a jifty-based project for random financial institutions
rindolf Well, at least no one would finalise this.
audreyt the first year is always hard :)
14:03 chris2 joined
rindolf audreyt: yeah, I guess. 14:03
audreyt: now I'm working for a PHP 5 shop.
It might be somewhat premature for me to become a consultant. 14:04
audreyt it's mostly not connections; it's mostly bringing something --even a prototype-- that address someone's pressing need quickly
*nod* it'd be better if you have at least one (or more) gigs lined up before you quit your office job.
rindolf audreyt: did you give the Portland.pm Jifty presentatin eventually.
audreyt rindolf: it's on pugs.blogs.com. 14:05
rindolf audreyt: I had a few gigs. My recent ones were WWW-Search-AOL and WWW-Search-MSN.
audreyt: OK.
audreyt colomon: try this one for me
colomon: run Makefile.PL again as root and then make install as root
(I'll fix, but that will identify the problem)
rindolf Which I got some money for each for my PayPal account.
colomon okay, give me a sec
Limbic_Region consulting in the US can be a very profitable business but it can also be quite dangerous
Limbic_Region has preferred to stay a contractor and/or full-time employee 14:06
rindolf Which reminds me that I'd like to blog about it.
Limbic_Region: by "consultant" I also mean "contractor".
At least in Israel the terms are diffused.
Limbic_Region oh - well they are definately different from my experience (in the US)
colomon audreyt: that appears to have worked! 14:07
Limbic_Region a contractor works for a company who gets gigs for them
audreyt colomon: okay, thanks, that identified the problem.
thanks!
Limbic_Region a consultant is solo freelance
rindolf Limbic_Region: hmmm.... I wasn't aware of the difference.
Limbic_Region a contractor may get benefits like health insurance and time off
svnbot6 r10920 | audreyt++ | * fix "make install" sequence so it doesn't register Pugs as a Haskell lib by default
Limbic_Region a consultant is completely on their own
audreyt there are gray areas... consultants morph into contractors and vice versa from time to time 14:08
Limbic_Region in the US - being a government contractor has the best of both worlds (high pay and job security)
14:08 ruz joined
Limbic_Region audreyt's statement also fits with my experience 14:08
14:08 pdcawley joined
b_jonas I see perl6 will have hashes that can have aggregate hash keys 14:10
that's nice
Limbic_Region IT consultants can make mad mad money in the US but finding gigs can be hard and independently paid for healthcare is uber expensive 14:11
audreyt kolibrie: mystery is explained
16:10 <@waltman> bjo1: there seems to be only one thermostat for each pair of rooms
b_jonas I hope it will also have an easy way to create a hash whose hash keys are compared by reference, not by contents
audreyt kolibrie: so contact your roompairmate
fglock audreyt: I see a lot of .tcc files under t/ - not a big problem, but I don't know where they come from
b_jonas well, perl5 will now have fieldhashes which is almost the same as this latter thing but also the keys are weak 14:12
audreyt fglock: me tooo, too, that's weird
b_jonas so I guess perl6 will have to have it too
svnbot6 r10921 | fglock++ | v6.pm - tests cleanly
fglock otherwise, I guess it's ready for release
audreyt b_jonas: fieldhash? 14:13
b_jonas audreyt: fieldhashes => www.perlmonks.com/?node_id=557420 14:14
lambdabot Title: "Perl core support for inside-out classes"
b_jonas links to this which gives the technical details => www.tu-berlin.de/zrz/mitarbeiter/an...dHash.html
lambdabot Title: "Hash::Util::FieldHash - Associate references with data"
b_jonas is this bot new? 14:15
Limbic_Region must be new (to the channel at least) as I have pasted plenty of perlmonks links here without ever having the title coming up like that 14:16
fglock releasing... unless someone got failing tests 14:26
...done! 14:30
14:41 hcarty left
svnbot6 r10922 | audreyt++ | * Normalize the CRLF before calculating checksum; noticed by fglock++ 14:49
14:49 pdcawley joined
renormalist Hi all! I'm thinking about ';' after blocks like in gather, map, grep, etc., Is there a rule, when a finishing ';' is needed? Sometimes I thought only in some structures like "while", "for", "if", etc I could omit the ';'. Can I always omit it after such blocks? 14:50
14:51 buetow joined
renormalist And which synopsis contains such "semicolon" stuff? 14:52
14:56 justatheory joined
fglock btw, is this valid: for map {~$_} @x { $_.say } 14:57
# for @x.map:{~$_} { $_.say } 14:58
lunch & 15:00
15:04 kanru joined
audreyt fglock: no, comma is mandatory after {} now 15:10
for map {},@x { .say }
fglock: btw I fixed the .tcc bug 15:11
look at perl5/M-C
(in larry's talk, network connectivity is spotty, may be off and on for a moment) 15:12
15:12 pdcawley joined
audreyt ooh, Hash::Util::FieldHash indeed looks like a Good Layout 15:15
I like it
svnbot6 r10923 | renormalist++ | cperl-mode.el: indenting+highlighting for async/atomically blocks found
r10923 | renormalist++ | in pugs/examples.
audreyt any chance of getting it backported? I wonder
15:23 stevan joined 15:29 lisppaste3 joined 15:31 rzoz joined
audreyt colomon: I fixed the problem ou reported 15:35
please retest
(r10924) 15:36
smoke here on macintel and win32 are both fine 15:37
svnbot6 r10924 | audreyt++ | * Fix the "make install as roto can't pick up fps" bug
r10924 | audreyt++ | reported by colomon++ by always re-registering the third-party/
r10924 | audreyt++ | dependencies as part of "make" instead of "Makefile.PL".
audreyt so if install works, going to release today
s/roto/root/ 15:40
15:46 Weems joined
rindolf audreyt: do you read Joel on Software? 15:46
audreyt a bit, not regularly 15:47
rindolf audreyt: I see.
audreyt: do you use an RSS reader?
audreyt Endo 15:48
15:49 pdcawley joined
fglock audreyt: let's make a single announce for Pugs & v6.pm? 15:57
audreyt fglock: yes, definitely 15:58
ingy is fixing another M::C bug
will release soon
fglock: please draft the announcement
and nopaste it somewhere and I'll pickitup
(still in the TPF talk)
fglock I think v6.pm POD can be used - it's a rewrite of previous announcement attempts 16:03
audreyt 'k 16:06
16:06 pdcawley joined 16:08 r0nny joined
svnbot6 r10925 | fglock++ | v6.pm - added README, ChangeLog 16:08
r0nny hi
anyone knows how the stats are for the real perl6 ? 16:10
bpphillips stats?
r0nny how much of it is allready completed 16:11
pmichaud depends on what "real perl6 means"
r0nny the not-golfing implementation of it made to run directly on parrot 16:13
pmichaud at the moment, we have simple subs in place, but not arrays or hashes
I'm hoping to do that this week
r0nny how long till there are all the oop things ? 16:15
Limbic_Region r0nny - even "not-golfing implementation of it made to run directly on parrot" can have multiple answers depending on what you mean since there are several initiatives to get p6 targetting parrot
16:16 rzoz_ joined
r0nny hmm 16:16
wich one tends to be the fastest one, and wich one is the most developed one ?
Limbic_Region I would probably say the pugs one is the most developed but that's just guessing 16:17
and yes, I mean pugs targetting of p6 on parrot
not pugs targetting of p6 on p5 or javascript or haskell or any of the other backends 16:18
r0nny hmm
Limbic_Region as far as speed - the important thing is not how long it takes to convert p6 to something parrot can run but how fast parrot runs it once it does
interestingly enough - one route would be p6 to p5 to ponie 16:19
r0nny hmm - i meant runtime speed, not compilation speed, sorry i wasn's specific enough
audreyt currently our parrot backend doesn't run quite as fast as we'd like 16:20
trying to fix that too
(also trying to improve the haskell-native runtime and the v6/Moose runtime)
Limbic_Region r0nny - well, that's 3 moving targets 16:21
r0nny hmm
Limbic_Region 1 - the ever changing p6 specification, 2 - the side targeting parrot - 3, the ever changing parrot
audreyt you get those "oop" things via Moose/P5 (which is the most complete at the moment), less so on JS/Haskell runtimes, and the Parrot runtime we havn't been exposing the object model to p6 land yet 16:22
Limbic_Region any one of which can impact the others
16:22 hcarty joined
r0nny hmm 16:22
fglock be careful with "which is the most complete at the moment" :)
audreyt if you are already coding in p5, I'd suggest playing with Moose.pm first (or P::C::R or one of the Perl6 runtime stuff)
but Moose/v6 isn't yet here
so you'll have to write in pseudo-p6 syntax in p5 :) 16:23
fglock: I'm very careful :) Moose/P5 is one thing and Moose/v6.pm is another
fglock oh you said P5
ok :)
Limbic_Region audreyt - I think r0nny is only interested on running on Parrot 16:26
16:26 gugod_ joined
Limbic_Region that is - with as much of p6 as possible 16:26
16:27 gugod_ is now known as gugod
r0nny hmm 16:28
16:29 rashakil__ joined
r0nny btw - anyone here got a p6 lib helps for depency tracking/solving 16:29
Limbic_Region r0nny - I can only guess at your reasons for asking such a question - but they make me think you are interested in writing production code in p6 today
that seems like a scary thing to do if that indeed was your reason for asking
r0nny Limbic_Region: yeah
Limbic_Region as I said - 3 moving targets, with none of them promising stability any time soon 16:30
fglock audreyt: does it make sense to try to make v6.pm to execute PIL^N ? 16:31
is it meant to be valid p6
r0nny well - its either perl6, or python - and i'd rather use perl6
16:31 pdcawley joined
Limbic_Region r0nny - if I were your employer and you were considering using p6 at this time - I would likely tell you to find another job 16:36
16:36 lollan joined, lollan_ joined
Limbic_Region but don't let that discourage you - pioneers and pilgrims are very valuable indeed 16:37
Limbic_Region wanders off for lunch
svnbot6 r10926 | fglock++ | Pugs-Compiler-Perl6 - s/compiler/experimental compiler/ 16:38
r0nny Limbic_Region: its a fun project - i like neat stuff for my own work, i'd never consider perl6 at this point, if it was important work for a employe 16:40
16:40 Ymmv joined
fglock v6.pm POD - search.cpan.org/~fglock/v6-pugs-0.001/lib/v6.pm 16:40
lambdabot Title: "v6 - an experimental Perl 6 implementation (under way) - search.cpan.org"
svnbot6 r10927 | fglock++ | v6.pm - added 'abstract' to Makkefile.PL 16:43
Limbic_Region r0nny - I understood "you are interested in writing production code" to not mean "fun project" but something for professional work - sorry we got our wires crossed 16:55
svnbot6 r10928 | fglock++ | v6.pm - added links to dev.perl.org, parrotcode, pugscode
Limbic_Region enjoys his food
16:55 jabbot joined 17:00 chris2 joined 17:03 shachaf joined 17:05 hexmode joined
audreyt hm, to backport the fieldhash we need to override Perl_sv_kill_backrefs 17:10
not impossible, but probably too costly
oh well.
# www.gossamer-threads.com/lists/perl...ers/205608
lambdabot Title: "[PATCH] Hash::Util::FieldHash | Perl | porters"
rgs yes ?
audreyt rgs: I just pointed fieldhash to stevan as a possible Moose layout 17:11
thinking about the best trategy to maintain some compat with <5.9 perl
rgs that won't go in 5.8.x
r0nny Limbic_Region: i want to dig together a build system for use on one of my projects needs a build-modularity semilar to the linux/bsd kernel
audreyt rgs: yup, so either some emulation
or some XS loadable magic
the latter is actually quite hard
so probably emulation 17:12
lunch, bbiab
17:12 esammer joined
rgs yep 17:12
17:27 penk joined 17:30 pdcawley joined 17:31 frederico joined, jabbot joined 17:32 visq_ joined 17:35 hcarty left 17:44 Eimi joined 17:57 larsen joined 18:04 rzoz_ is now known as rzoz 18:09 DaGo joined
visq hi; I was looking at multim. dispatch and the current behaviour seems a bit ... strange to me 18:10
maybe I didn't get something, but what happens if no multi declaration matches ?
called "distance Infinity" in A12
no tests for that in t/subroutines/multi_sub.t ? 18:11
pasteling "visq" at 62.178.138.213 pasted "Multimethod Dispatch - No Match" (9 lines, 329B) at sial.org/pbot/17993
audreyt hi 18:12
when there's no match, an error should be thrown
18:13 pdcawley joined
visq Should that be in the test file ? 18:13
audreyt however, it's not specced if strings with numbers are in fact coercible to numbers 18:14
for the purpose of dispatch
so pugs tries a last-ditch coercion
but whether (or how) to do that is unspecced
visq String to Rat ?
s/String/Str
audreyt it went to Int I think
18:15 prefiks joined
wolverian type inference with implicit coercion oughta be fun.. 18:15
audreyt indeed, indeed. 18:16
the JS way is to distinguish
sub f ($x does Int) { ... }
from
sub f ($x as Int) { ... }
the latter triggers a coercion call if needed
wolverian which is Int $x?
audreyt the first 18:17
wolverian works for me
audreyt but again, this is unspeccedland
visq ok, but for multis ? you want to choose the best coercion ?
sounds funny
audreyt visq: there is no best coercion...
if it can be coerced into multiple things
then it's an error too
unless it can be resolved at marking a variant as default (unimplemented) 18:18
visq ok. in the example I posted the Str was coerced to Rat instead of Int (i think)
audreyt it says 1 no? 18:19
visq yes (calling ($f+1).int in the Rat multi)
audreyt oh right. sorr
y
I misparsed
visq not a very sensible example, though 18:20
audreyt the reason is our distance algorithm is currently weird
if it can't be upcasted
(Str isn't Rat or Int)
then it's converted to the common base
visq which is Num I guess
audreyt no, Str isn't Num either
so it's the scalar value basetype 18:21
and then it does downcast
encoutering Rat first
and go with it
visq :)
audreyt again, this is completely unspecced, so this implementation is purely conjectural
visq maybe an error would be best until specced ? 18:22
or a test :unspecced ?
audreyt or a p6l post :) 18:24
I think ew should implement the same semantics as this
www.mozilla.org/js/language/js20/co....html#type
lambdabot Title: "JavaScript 2.0 Concepts"
audreyt also this
www.mozilla.org/js/language/js20/li...types.html
lambdabot Title: "JavaScript 2.0 Types"
audreyt where implicit coversion ("of") and explicit conversion ("as") are distinguished 18:25
and then throw in the Any type
to stand for the "typechecker, stay away from this"
type
which is the default for all values with no annotations
...which is distinctly VB 18:26
only that the annotation attitude is reversed
18:26 vel joined
audreyt (they require explicit "as Object" annotation to get dynamic typechecker-stayaway variables) 18:26
18:27 gaal joined
gaal moose! 18:28
kolibrie moose!
18:28 cognominal joined
gaal the wonders of high-latency and high packet loss networks! 18:28
audreyt moose!
audreyt has quite a bit of S* commits for TimToady to review
but release 6.2.12 first, today, for real 18:29
fglock: so should i just take v6.pm synopsis and run with it?
(for the announcement)
gaal audrey, will there be some hackage this afternoon somewhere? 18:30
audreyt I don't know, do you want one?
where are you?
gaal sure, why not
releasathon
I'm in the alumni classroom 18:31
with nothingmuch, sevan, kolibrie
audreyt say hi to rkinyon
gaal he's pondering the lighting, I believe
audreyt I think I've grokked ::Deep, so I'm harrassing clkao instead
gaal getting ready to talk
audreyt maybe meet at the sofa place you said hi to me after lembark's talk? 18:32
Limbic_Region anyone know what file(s) on an AIX server to look for to tell named not to try and go to the internet for SOA outside of internal domain?
gaal sure, there's a some 'lectricity there 18:33
Limbic_Region would have assumed named.conf or data.conf or reverse.data.conf
gaal l~r, no, but I'm sure you can find it in SMIT?
gaal hasn't been on AIX for, ugh, eight years
audreyt aaaaargh *flashback of painful SMIT fights*
18:33 bernhard joined
Limbic_Region gaal - unfortunately I don't have access to the box - an admin is claiming they can't find it and I was being asked to confirm that not finding the file should be expected 18:34
I don't think it should be expected but I am not familiar with AIX (only HPUX, Solaris)
gaal l~r, I think you really need to do things via the management system on AIX otherwise your changes are ephemeral 18:35
Limbic_Region bah - I thought *nix was supposed to not try and be so "helpful" 18:36
thanks - anyway - please return to your normal scheduled programming 18:37
gaal AIX just exposes a unixy interface. it's helpful sometimes not to try to use it. :) 18:38
18:42 bpphillips left 18:43 bpphillips joined 18:46 pdcawley_ joined
gaal nothingmuch: ping 18:49
nothingmuch pongf 18:50
gaal heya, see /msg 18:52
audreyt I don't see it
have you regnicked?
gaal audreyt: it was to nothingmuch :)
audreyt heh :)
gaal
.oO( merged antlered audreyt nothingmuch)
 18:53
18:55 pdcawley_ joined
fglock audreyt: re announcement - v6 synopsis is ok, I think 18:57
bbiab &
audreyt ok... anyone up for a quick "make test" on perlcabal.org/~audreyt/tmp/Perl6-Pu....12.tar.gz ? 18:59
if it tests clean I'll release it :) 19:00
obra Grabbing 19:01
not usre I have a new enough haskell
audreyt danke
6.4.1 will do
in fact it's the only thiing I havn't tested yet ;)
(so far all smokers are on 6.4.2 it seems)
gaal I would but a smoke takes too long on this box 19:02
audreyt that's fine
19:08 pdcawley_ joined 19:16 dduncan joined
audreyt ok, I'm in the lobby now 19:18
the power strip seems a bit occupied
in any case I'll wait for the lambdamoose to come and maybe find a new location
kolibrie come to our corner room 19:19
audreyt where's that?
kolibrie south west?
gaal audrey, stay put
Limbic_Region imagines mice in a maze with walkie-talkies 19:20
gaal a maze of walkie talkies, all alike 19:21
19:23 SamB joined
Aankhen`` audreyt: What revision is the release? 19:28
audreyt Aankhen``: get the url
Aankhen`` Bleh, okay. 19:29
audreyt perlcabal.org/~audreyt/tmp/Perl6-Pu....12.tar.gz
Aankhen`` I guess pre-compiling Test.pm would make sense... 19:31
audreyt ...that's now done by default as part of "make pugs" 19:32
Aankhen`` Ah.
How come?
I mean, is it necessary to do it by default? o_O 19:33
19:34 jiing joined, f0rth joined 19:35 _SamB_ joined 19:36 _SamB_ is now known as SamB, baest joined
dduncan so my pre2 smoked with seemingly perfect results yesterday 19:39
and how's things today?
19:39 _visq joined
audreyt should be fine 19:50
svnbot6 r10929 | audreyt++ | * gaal++ notices that there's :todo<spec> and :todo<hardfail>
r10929 | audreyt++ | markers; fix them to say :todo<unspecced> and :todo<bug>
audreyt going to release after obra's smoke finished
19:53 f0rth joined
spinclad I'll try it too 19:54
btw Sunday's smoke (pre2 +-) was 5 hr where Saturday's (pre1) was 3:30 here 19:55
don't know what was different but maybe Test didn't get precompiled on Sunday 19:56
audreyt yup
spinclad (Saturday: r10874, Sunday: r10891) 19:57
audreyt should be back to normal smoketime now 19:58
19:58 xerox joined
audreyt so, anyone running make test? 20:01
20:03 Zemyl joined
Aankhen`` is. 20:05
I'm on t/builtins/io/io.t so far.
io_edge_error_cases.t now.
And moving on...
fglock audreyt: about the v6.pm release note - I've been asked "it doesn't pass 100% tests, right? how much is actually implemented?" - it passes about 90% of 01-sanity.t/ tests; it currently implements a thin semantics layer, which means a non-oo, untyped subset of Perl6 20:07
spinclad audreyt: not for a while, make will take a couple hours here 20:10
audreyt okie
"it implements perl 4."
(or something like that)
20:11 marmic joined
spinclad "Current version is 0" 20:15
(/usr/bin/perl -Iinc util/version_h.pl src/Pugs/pugs_version.h)
audreyt right 20:16
which is fine
(it's used to determine the display of pugs -v only)
for releases we don't display svn revs
20:17 nnunley joined, marmic joined
gaal should this work? 20:18
<a b c> >>.say
?eval <a b c> >>.say 20:19
20:19 evalbot_10889 is now known as evalbot_10929
evalbot_10929 Error: unexpected ">" expecting comment, operator, postfix conditional, postfix loop, postfix iteration, ";" or end of input 20:19
audreyt yup, but not currently parsed
gaal thanks
20:19 nnunley joined
svnbot6 r10930 | audreyt++ | * This be Pugs, version 6.2.12. 20:19
gaal whoa! 20:20
gaal cheers
audreyt :))) 20:21
Aankhen`` breaks out the pom-poms!
clkao hurrah!!!
is it beer time yet
spinclad champagne
Aankhen`` Gimme a P! Gimme a U! Gimme a G! Gimme an S! WHAT'S THAT SPELL? I dunno either. :-(
audreyt 6.2.12 released! | paste: sial.org/pbot/perl6 | pugs.blogs.com | pugscode.org | pugs.kwiki.org 20:21
audreyt PUGS! 20:21
Aankhen`` Yay!
dduncan the pugs is dead, long live the pugs 20:23
20:24 marmic joined
audreyt starts writing announcement 20:24
spinclad a toast! 'to absent friends' ... no, that's not it... 'in conclave assembled, in meatspace and farspace, may the pugs let us golf till the perl6 is born'
Aankhen`` applauds!
20:24 Zemyl joined
Aankhen`` raises a glass. 20:24
fglock a toast! 20:25
20:25 _bernhard joined
audreyt spinclad: may I quote you on that in the announcement? 20:25
spinclad sure! 20:26
dduncan so is anyone going to be interjecting into a YAPC::NA talk that the pugs has landed, or something?
spinclad counts on it
dduncan I remember at OSCON last year that someone interjected into Ingy's (?) talk to quote audreyt's comment on this list 20:27
they beat me to it, alas
spinclad audreyt: s/farspace/far/ and it scans a bit better maybe 20:28
audreyt sure
clkao is anyone working on making 'use Test' work in -BPerl5 ? 20:30
audreyt certainly fglock
that's I think his next target
clkao fglock: what's the plan? 20:31
20:32 bpphillips left
audreyt ingy++ 20:32
fglock: new M::C is on CPAN now
fglock: you going to do another v6.pm?
fglock audreyt: it already requires the next version 20:33
requires('Module::Compile' => '0.16'); 20:34
clkao: I'm working through 01-sanity - it's imported to v6.pm t/ 20:35
audreyt cool
20:35 marmic joined
clkao oh wow 20:35
mostly passing, i am seeing 08 failing because of use()
audreyt particle: speed up, v6-pugs is about to beat parrot/perl6 :) 20:36
(on getting Test.pm to pass)
oh wait, he's not here 20:37
fglock clkao: the few remaining tests require more runtime support - multis, more complex signatures, and a few fixes in the grammar 20:38
btw, if you re-run the tests you can see the actual speed, as the compiled code gets cached 20:39
svnbot6 r10931 | ingy++ | r713@ingy: ingy | 2006-04-14 23:47:37 -0500 20:41
r10931 | ingy++ | Fixed so that pod2html works
r10932 | ingy++ | r714@ingy: ingy | 2006-04-15 06:42:01 -0500
r10932 | ingy++ | S26 first sweep of changes
20:41 rashakil_ joined
fglock of course, if someone simplified Test.pm, that would help a lot :) 20:43
ingy thanks! 20:45
clkao fglock: ya 20:46
fglock (my plus-plus shows as spaces for some reason)
dduncan so, to clarify, is r10930 the actual 6.2.12 release for CPAN now, with later commits not going in, like ingy's last 2? 20:49
clkao fglock: so for use Test in 08, test.pm has to be compiled to somewhere to avoid v5 collision right?
dduncan if that is so, then I'll commit a template for the subsequent revision to the changelog 20:50
spinclad dduncan: i believe you have it's essence true 20:52
(ack! s/it's/its/) 20:53
dduncan okay, so I will commit post-6.2.12 changes shortly
I'm going to assume the next release will be 6.28.0, and that it won't happen in June 20:55
if not, the figure can be adjusted 20:56
spinclad safe bet /me thinks
perl6 is getting ready to eat a Moose 20:57
svnbot6 r10933 | Darren_Duncan++ | updated ChangeLog to add a template for the next release, probably 6.28.0
spinclad (present Møøsen excepted. please excuse the language) 20:59
svnbot6 r10934 | Sage++ | =~ -> ~~ now correctly applied, as well as s:P5:g///. $hash{someword} now translating to %hash<someword>. Added support for some escaped characters in uni strings, but still very much beta (ugly code). TestInit.pm.yml now correctly parses as presented, chomp modifier support still needed. 21:09
21:14 theorbtwo joined
fglock clkao: Test.pm should be look for in PERL6LIB env variable 21:16
looked
Aankhen`` Er... I hate to say this, but I've got 9 subtests failing. 21:17
pasteling "Aankhen``" at 203.101.1.39 pasted "make test output" (8 lines, 548B) at sial.org/pbot/18005 21:18
fglock clkao: and 'use perl5:Test' is a plain perl5 'use' 21:22
clkao *nod*
of course ;)
which is pretty much what it is doing now. we just need to fix the exporting thing
21:23 dasein joined
Aankhen`` Well, I'm off to sleep. 21:29
Congrats on the release. :-)
G'night.
fglock clkao: just setting @INC and then calling a plain 'use' should work, since v6.pm takes care of the compilation 21:30
$work.die & 21:32
svnbot6 r10935 | ingy++ | r1931@ingy: ingy | 2006-06-26 15:29:00 -0500 21:34
r10935 | ingy++ | perl5/
21:34 cognominal joined 21:37 fglock left 21:41 zgrim joined
zgrim hello folks, is ghc 6.4.1 good/recent enough for building pugs ? 21:44
spinclad yes
though 6.4.2 or 6.5 would be better if you can get them
zgrim i'm trying to build it like this: GHC=/usr/bin/ghc-6.4.1 PATH=/home/zgrim/src/perl6/parrot:$PATH PUGS_EMBED="perl5 parrot" perl Makefile.PL 21:45
but i'm getting the strange: make[1]: Entering directory `/home/zgrim/src/perl6/pugs/ext/File-Find'
Perl v6.0.0 required--this is only v5.8.8, stopped at lib/File/Find.pm line 3.
this only happens on my desktop machine, the laptop, which is also debian sid, built pugs just fine 21:46
spinclad i haven't seen that failure myself (on debian sarge/etch mix) 21:47
sounds like some perl6 code with 'use v6' in it got fed to perl5
zgrim can you see the reason why this would happen. i'm not familiar enough with pugs' build system 21:48
spinclad not offhand. i see that ext/File-Find/lib/File/Find.pm does use v6... 21:51
you have run '... perl Makefile.PL' and then 'make', yes? 21:52
zgrim yes, ofc
spinclad let me check my build logs 21:53
zgrim i did see that File::Find requires v6 :), i was asking what could be wrong there in the build system though...
or maybe in my setup
spinclad odd.. the only thing my make does in ext/File-Find is 21:55
cp lib/File/Find.pm /data/perl6/pugs/build/trunk/pugs/blib6/lib/File/Find.pm
and later, in testing, the test harness runs ext/Find-Find/t/01-use.5 21:56
s/5/t/
21:57 rindolf joined
rindolf Hi all. 21:57
pasteling "zgrim" at 195.7.14.25 pasted "build regarding File::Find v6 || V5" (12 lines, 725B) at sial.org/pbot/18006 22:02
22:02 hcarty joined 22:06 unobe joined 22:11 unobe left
audreyt zgrim: you have "lib" in your PERL5LIB somehow? 22:15
or -Ilib in your flags?
zgrim ]$ env | grep -i perl5lib 22:16
PERL5LIB=:/home/zgrim/lib/perl5:./lib:/home/zgrim/lib/perl5:./lib
aha, i see now
thanks 22:17
nothingmuch 1/win13 22:20
does anybody know how to run a BOF? 22:23
does anybody want to volunteer 22:24
(to run it)
22:29 Limbic_Region joined 22:31 pjcj joined
audreyt miyagawa: fixed 22:36
nothingmuch: I'm seeing 22:37
--- ! >-
{ 42;
}
not !perl/code 22:38
from the trunk
of perlsyck
nvm, fixed. 22:40
xerox__ Teh coding machine.
22:41 rashakil__ joined 22:42 gaal joined, Ymmv joined
Limbic_Region audreyt - on a machine where I can do make test - is it still needed? 22:43
22:46 mako132_ joined 22:47 ingyfoo joined
gaal l~r: see topic... 22:48
22:48 ingy joined
Limbic_Region gaal - thanks 22:51
would have replied sooner except my kbd inexplicably stopped functioning
or at lease, it was originally inexplicable
how exactly a kbd becomes unplugged mid keystroke is beyond me
nothingmuch audreyt: i have no idea what that was
gaal whee, Best.pm uploaded to PAUSE 22:52
kolibrie gaal++
gaal }:) 22:53
nothingmuch Best.pm?
audreyt the best module
well, best perl module, anyway
gaal nothingmuch: use Best [ qw/YAML::Syck YAML ];
but with another / somewhere 22:54
nothingmuch awesome
gaal hopefully Module::Compilable 22:55
nothingmuch how so?
gaal presumably Best.pm isn't going to be core, so if I don't know if my run site has YAML::Syck, I don't know if it has Best, either :) 22:56
this module has two usages: XS-vs-non, and special-vs-core (e.g. CGI::Simple / CGI) 22:57
well, except for the general fallback (e.g. the yaml example)
clkao sub ::Y { require YAML::Syck; YAML::Syck::Dump(@_) } 23:02
"You'll never see me"; # End of Best
gaal oops. 23:04
0.02 coming up :)
well, the you'll never see me is correct, because we propagate the successfully loaded module's return value. 23:05
::Y and ::YY are just favorite debugging tools :)
clkao ya, i just found it funny, not saying it's wrong ;)
audreyt nothingmuch: you caused stack corruption 23:06
nothingmuch audreyt: where?
trunk is a bit broken right now
audreyt line 694
clkao audreyt: i am going to have dinner with sussman and kfogel, want to come?
audreyt syck_emit_scalar(e, OBJOF("tag:perl:code:"), SCALAR_UTF8, 0, 0, 0, SvPV_nolen(text), len-1);
nothingmuch what does that do?
audreyt at this point "tag" is set to '{{?' on my computer
23:15 unobe joined
unobe i'm having a problem building the latest Perl6-Pugs-6.2.12.tar.gz 23:17
any pointers? 23:18
pasteling "unobe" at 198.37.24.106 pasted "build failing with latest Perl6-Pugs-6.2.12.tar.gz" (78 lines, 5K) at sial.org/pbot/18008 23:19
svnbot6 r10936 | clkao++ | TODO for now. 23:21
unobe i searched the logs and only found two instances of others being given similar errors, but running ./Setup --configure user didn't work 23:22
23:23 hcarty left
unobe i mistyped actually. i tried ./Setup configure --user, but to no avail. 23:40
nm...i forgot to change permission for ~/.ghc (I have to run root to make install) 23:45
svnbot6 r10937 | fglock++ | v6.pm - the emitter supports ` %ENV<PERL6LIB>[].push( '.' ); `
23:46 unobe 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!