Summer of Perl | 6.2.12 released! | paste: sial.org/pbot/perl6 | pugs.blogs.com | pugscode.org | pugs.kwiki.org
Set by ingy on 4 July 2006.
TimToady It's also possible that not many folks read p6c. 00:00
dduncan one detail I put forth that would help for more discussion is where and what nature the test harness would go
TimToady compared to p6l
dduncan true
I could post it to p6l, but audrey suggested p6c
TimToady p6c is probably sufficient public notice before the bulldozers move in to tear up the wetlands. :) 00:01
dduncan on a related note, at some point the test suite should be moved to a different repository-project from pugs, if reasonable, just as the perl 6 specs are already separate 00:02
TimToady some of 'em 00:03
I like having S29 still under pugs for now
dduncan the idea is that people's experience with SVN layouts somewhat matches the CPAN distros
eg, someone sees a distro, they can find a SVN that more or less matches it 00:04
TimToady and anything else that's primarily still implementation driven
dduncan of course, the SVN separation can happen later ... it is the CPAN separation thats more important short term
merlyn TimToady - dunno if you saw the scrollback about sub vs multi sub
TimToady perhaps we need to distinguish pugs from pugs 00:05
yeah, still thinking about it. 00:06
00:06 surfdue joined
surfdue what is this? 00:06
dduncan what are you?
merlyn a channel for people building perl6
surfdue interesting
merlyn yeah, what are you? :)
surfdue so a new version of perl
or a fork?
merlyn if you have to ask, this may not be the place for you. :) 00:07
surfdue im the freenode magical flying pixie horse
I go around and check our random channels
TimToady we don't think this is a very random channel, except when it is.
00:07 surfdue is now known as MagicalFlyingPix
MagicalFlyingPix .. 00:07
TimToady, it is when I have never been here 00:08
*POOF* you are now in /list
happy channel list-b-day!
00:08 MagicalFlyingPix is now known as surfdue
surfdue :P 00:08
dduncan so does that mean #perl6 will show up on freenode.net/primary_groups.shtml now? 00:10
lambdabot Title: "Primary Groups"
surfdue maybe
*POOF*
surfdue disappears
dduncan well it ought to ... considering this is a major place for stuff to happen
surfdue you iseth correct young master. 00:11
00:11 prefiks left, bennymack joined 00:13 bennymack is now known as bennymack[linuxn 00:14 bennymack[linuxn is now known as bennymack
surfdue bennymack you mispelled linux 00:14
dduncan I will defer further t/ changes for the moment, as the fate of Test.pm partly determines how it will organize
and because I should really get to $job stuff
surfdue anyways, if there is anyone in here that codes perl (thats a givey), and has time on there hands (another givey), and would like to help code a new upload filesharing site. Please feel free to ask questions 00:15
merlyn surfdue - highly inappropriate in this channel 00:16
just for future reference
surfdue ok 00:17
merlyn it's like walking into a car factory design meeting and saying "anyone have a used car here?"
surfdue merlyn, ive talked to you before :)
merlyn, i do that all the time over at ford ?
merlyn whether they do or not, it's the wrong place to ask
gaal dduncan: ping 00:24
obra 00:25
bennymack hi everybody. I'm here because I was pulled into Perl6's Reality Distortion Field at YAPC
i'm interested in trying it out. If I get the 6.2.12 will it have the features that Larry and Damian talked about?
TimToady welcome to the madhouse
bennymack *new features
TimToady don't remember what we talked about. :) 00:26
any particular ones you care to name?
bennymack one sec
the argument objects for example 00:27
TimToady ?eval \(1,2,3) 00:28
evalbot_11274 [1, 2, 3]
bennymack I wish I had that presentation: Perl6 The Final Ascent
TimToady they're still probably partially implemented
most things having to do with objects are getting a refactor after 6.2.12.
gaal bennymack: Captures will probably be available in the next release 00:29
we're working on refactoring the internals a bit, which is a prerequisite.
TimToady on the other hand, 6.2.12 is just the Haskell implementation, and a lot of the exciting stuff is going on with v6.pm nowadays.
which bypasses the Haskell and compiles down to Perl 5. 00:30
bennymack TimToady: ahh I was wondering about that too. Thanks!
dduncan as I've also said, my current prediction is that v6.pm will be the first Perl 6 implementation ready for production use 00:31
gaal dduncan: re: tests, a few thoughts.
merlyn I'm not sure how, unless you move a lot more functionality into perl5
TimToady mmm, depending on how you define "ready for production".
dduncan and/or used in production
TimToady my guess is that it will be pretty darn slow.
gaal first, please let's not call it certification :)
merlyn the mismatch between perl6 semantics and perl5 semantics will kill ya
TimToady at least to begin with.
dduncan merlyn, have you looked at audrey's tower? 00:32
merlyn yup
but there are fundamental hamming principles at work
gaal second, obra's company is putting some effort into abstracting out the testing tools to a reusable perl 5 smoke system
it's still in development, but I've seen it and it's looking good.
dduncan sure it will be slow at first, but it should be functional and useful in the real world
merlyn there's no amount of code I can type as a human at perl5 to make it do everything perl6 will do
SamB hamming principles...
hmm.
merlyn which means you have to emulate the perl6 world for some operations
dduncan why is that merlyn? 00:33
SamB trys to find a really really bad pun...
merlyn which means slow, and possibly bandwidth mismatches to handle the translations
dduncan Perl 6 is being implemented over other things that probably didn't do what it does
merlyn Well - the perl6 world is emulated on haskell
this is why it's slow. :)
and parrot is being build so that the perl6 world is correctly inside of it 00:34
that's why it will be faster
gaal they've been developing it with their needs in mind, which aren't precisely a p6 compiler's needs, but the overlap between them (as well as $random_perl_project) are pretty big, so there's great potential for happiness
dduncan yes
merlyn but targeting perl6 to java-vm probably would bite
because you'd have to build an emulated perl6 world in java
dduncan parrot will be faster, but v6 will be ready earlier to do real work with, is my prediction
merlyn for the parts of perl6 that are not exceeding the semantics of Perl5, sure
TimToady for problems where correctness is more important than speed, sure. 00:35
dduncan initially thats fine for me
merlyn and vice versa, yes
gaal dduncan: most of t/ doesn't contain much pugsisms
dduncan my initial use of Perl 6 will be mostly as syntactic sugar
anyway
merlyn syntactic brown whole sugar. :)
dduncan incidentally, I wonder if a Pair type is being backported to bleadperl5, like a few other things are backported ... that would be useful 00:36
TimToady I think we'll see a certain amount of P5-ish looking P6 code in the interests of mapping more directly to the P5 engine.
dduncan that's how I will code initially
merlyn right
gaal re: splitting the release, I'm not sure it's time, just yet, since no project passes any reasonable percent of the test suite except pugs yet
TimToady but MMD is gonna be an interesting optimization problem. 00:37
gaal but I'm sure now that Test.pm + 01-sanity/ passes on v6, this well get a lot better fast
TimToady I'm not suggesting splitting the releast that way, just that the word
dduncan gaal, I don't see that as a problem
TimToady "pugs" means two different things right now.
SamB does it mean a kind of dogs?
obra having a make target to package the test suite could be very cool.
TimToady It's referring both to the Haskell project and the larger umbrella project that is not necessarily in Haskell. 00:38
gaal obra: echo "testbundle:\n\ttar czvf p6-test.tgz t" >> Makefile :)
TimToady Maybe the Haskell Pugs should be rename Hugs or something.
SamB the name is a pretty clear reference to hugs, fwiw...
hugs is already taken, TimToady ;-) 00:39
obra grins at gaal
bennymack is installing v6 now
00:39 surfdue left
TimToady Phugs. 00:39
pughs.
SamB oooh
obra At least with phugs, it would have theme music
SamB second one is better
TimToady phughs
gaal gotta go & 00:40
dduncan a lot of my like of Perl 6 is just its improved well-thought syntax, the big features are nice too
and the little things help a lot too, like the Pair type
TimToady .thanks for @Larry;
well, I'm being called to go grill some poor defenseless kielbasa... 00:41
dduncan cook them until they don't cry no more, right?
TimToady I just don't want to know what they're made out of. 00:42
Later &
bennymack it seems like it would be difficult to get a lot of minds on the haskell wagon
dduncan ping gaal 00:48
00:48 Jedai joined
Jedai Hello 00:48
dduncan please reply on p6c what you said about my tests suggestoin
if obra's company is doing something that my suggestion may duplicate, please say so there for the benefit of others 00:49
bennymack so i'm installing v6 and it seems like the requirements are the same as the YAPC talk schedule
dduncan hehe
audreyt rofl
dduncan coptor
obra dduncan: we're building a generalized smoke testing infrastructure
that works well for projects in svn repositories 00:50
dduncan so, that will accomplish what I mention on p6c, where the tests are distributed separately from pugs and v6 on cpan?
obra it's not a packaging or distribution system.
url to your pc6 post?
p6c
dduncan its the last one
audreyt www.nntp.perl.org/group/perl.perl6.compiler/1434
lambdabot Title: "nntp.perl.org - perl.perl6.compiler (1434)"
dduncan yes, thats it 00:51
audreyt dduncan: I think separate svn is largely orthogonal
it means one more commit-bit arrangement
dduncan yes, I agree ... see the post
Jedai ?eval "Hello world !".reverse.reverse.say; 00:52
dduncan I quote: As to whether any of the above are still maintained within Pugs SVN or moved to separate SVN, that is a possibly orthogonal matter.
evalbot_11274 OUTPUT[! dlrow olleH ] Bool::True
obra dduncan: I think it's the bit between () that we're doing
audreyt so I'd prefer even radical reorganization of the tree to actual split svn
dduncan second-last paragraph
obra I think that moving the tests out of pugs svn would be a bad plan
the tests are successful because they're so close to the code with the most hackers.
dduncan well, as I said, its the appearance of being completely separate on CPAN that I consider important 00:53
obra why is the appearance important?
audreyt dduncan: I agree very much.
obra The availability is important.
dduncan I think it should be possible for one to download just v6.pm and the Perl 6 tests, without downloading anything pugs-specific
audreyt separate cpan distros that multiple other cpan distros can depend on = very important
yes.
dduncan but the focus of my post is on what CPAN users see, not developers
audreyt yes. 00:54
dduncan everything can stay in the same SVN
audreyt dduncan: so write t/Makefile.PL :)
obra audreyt: gaal did a page ago ;)
audreyt we also need a reflection API that moves utils/run-smoke.p into the t::
obra: that isn't a cpan dist
obra audreyt: fair enough
audreyt and basically say "run these *installed* test suite using this impl"
i.e. the v6::tests will have the tests somewhere in sharedir
using Module::Install::Share 00:55
installed into sitelib
dduncan: where are you based in the meatspace, and will you be at OSCON or some other conf? 00:56
Jedai ?eval "abc".reverse.reverse.say; my Str $s = "abc"; $s.reverse; $s.say; $s.reverse; $s.say;
evalbot_11274 OUTPUT[cba abc abc ] Bool::True
obra dduncan: svn.bestpractical.com/cgi-bin/index...mps/trunk/ is what we've been up to
lambdabot Title: "browse: /bps/Test-Chimps/trunk/ (Rev: HEAD, via SVN::Web)"
dduncan audreyt, I would love to be at various conferences, but I can't afford the $currency at the moment
audreyt Jedai: .= reverse
wolverian ?eval "abc".reverse.reverse 00:57
evalbot_11274 ("cba",)
Jedai Yes but I was most intrigued by the .reverse.reverse ?
wolverian that looks strangeish
:)
Jedai (my other test were just to check if it was in place)
dduncan I figure late this year or next year should be a different story though 00:58
audreyt Jedai: what's your email addr?
Jedai chaddai.fouche@ens-lyon.fr
svnbot6 r11324 | audreyt++ | * fix "foo".reverse.reverse to return "foo" by declaring
r11324 | audreyt++ | that reverse's scalar form always return a string, not a list.
r11324 | audreyt++ | Spotted by Jedai++.
audreyt Jedai: can I send you a commit bit and for you to commit that test into t/builtins/strings/reverse.t?
oh good. also add yourselve to AUTHORS. commit bit sent
Jedai Thanks 00:59
I'll do it :)
audreyt yay :)
cognominal Jedai, are you in contact with the Lyon mongers?
Jedai Not yet but I plan to (next year as I'm not in Lyon right now) 01:00
dduncan audreyt, the answer to the other part of your question ... I am within meatspace at Victoria, BC
cognominal Jedai: Lyon will orgainize either the european yapc or the French Perl workshop in 2007 01:01
Jedai Wow :D At least one YAPC I won't miss ! 01:02
cognominal probably the second because Vienna has an headstart for YAPC 2007
audreyt dduncan: oh, you're in .ca?
dduncan yes
audreyt cool. maybe I'll visit .bc.ca next year 01:03
Jedai cognominal, you're in the Lyon mongers ?
dduncan pacific ocean, 48th parallel
merlyn the land I cannot enter.
audreyt yes, the land forbidden to merlyn
dduncan merlyn, why not?
merlyn you don't let felons enter canada
dduncan last I recall, Canada was easier to enter than the US
cognominal no, I am in Paris suburb (Arcueil) so I am part of the parisian mongueurs
merlyn well - I'm in the US, so that's not a problem to go to the US. :)
cognominal rgs (a perl5 pumpking) works in Paris. He is divorced and visit his kids at Lyon almpst every WE 01:04
merlyn but the previous 4 YAPCs have been alternatingly (a) in canada or (b) during a geekcruise
which is why this was my first YAPC since 19101
dduncan as I recall, no YAPCs take place on the west coast of NA, because OSCON is in the general area
hence there are no conferences that are both cheap for me to attend and travel to 01:05
merlyn No - it's just that nobody pitches it
dduncan really?
merlyn yeah, it's not "vs oscon"
it's a different market
oscon is no longer a perl conference
yapc is
dduncan so Larry's not giving his onion there any more?
merlyn He will
TimToady ?eval my Str $s = "abc"; $s.=reverse; $s.say; $s.=reverse; $s.say;
evalbot_11274 Error: unexpected "=" expecting ".", subroutine name, array subscript, hash subscript or code subscript 01:06
merlyn O'Reilly makes a token nod toward perl
but that's not OSCON's major player any more
audreyt ?eval "abc".reverse.reverse
evalbot_11274 ("cba",)
merlyn why the extra comma?
obra dduncan: iirc, there's a free conference organized around oscon
and 80% of the value to an active perl hacker is the hallway track, which is free ;) 01:07
merlyn yeah, there's another os "con" near oscon
audreyt merlyn: it's a bug. and fixed
merlyn ahh
audreyt merlyn: evalbot_11275 will have it right
(it's currently rebuilding on feather)
?eval my Str $s = "abc"; $s .=reverse; $s.say; $s .=reverse; $s.say;
merlyn heh - you broke it. :)
audreyt hm, someone renamed evalbot helper.
that must be dduncan++ :)
audreyt goes rerun the helper
dduncan it is very possible 01:08
TimToady prolly me.
dduncan there may be the odd file name ref / file name or shebang or whatever that got out of sync
merlyn Hmm. I still need to hack at a linuxmag column
merlyn wanders back to #perl 01:09
dduncan brb
TimToady I distinctly remember renaming an evalbot.p6 to evalbot.pl
audreyt it's TimToady++ then :)
01:09 fglock joined
TimToady I even fessed up to it at the time on #perl6, but nobody was listening... 01:10
audreyt I'm still amazed that the entire renaming happened practically overnight 01:11
01:11 Jedai joined
audreyt would easily take me weeks to do it :) 01:11
TimToady audreyt: you are not a good liar.
audreyt no, seriously. I'll be distracted massively 01:12
and wander off to write some sort of automator
and then hack svk
etc etc.
obra "as it renames, it programmatically transforms the code"
audreyt yes.
since that was what I was about to do :)
TimToady yes, but the problem of automating it would have fascinated you sufficiently that you'd evenutally have something that would do it in 5 minutes
audreyt yeah, but the "eventually" can take weeks 01:13
dduncan audreyt, the actual file renames I did manually, with the help of Spotlight to find the files
the references inside the files were the easy part ... 01:14
bennymack ok, here's my first p6 noob question. sorry for the 2 line paste
ben@slaqfu:~$ perl -Mv6 -e 'for 1,2,3 -> $x { say $x }'
Perl v6.0.0 required--this is only v5.8.4, stopped.
dduncan my text editor, BBEdit, has built-in abilities to search and replace, with regexps, on whole folders of files
TimToady yeah, well, the strengths of this community are in recognition of the multiplicative results of including both ADD and Autistic roles.
dduncan I never wrote any actual script to do those 500 or 300 shebang line changes at once 01:15
but each actual file rename I did one at a time
and TimToady helped with those
audreyt bennymack: we are still trying to find a way to make -Mv6 -e work
it's... nontrivial
as at -M time, it still doesn't see -e yet 01:16
TimToady well, we tend toward the autistic side of the typeset...
dduncan but audreyt, I can understand the idea of being distracted from such a task when you see something else to work on
bennymack ok thanks
TimToady tchrist always claimed I have "task switching deficit disorder"
audreyt bennymack: here is workaround:
perl `perldoc -l v6` -e 'for 1,2,3 -> $x { say $x }'
dduncan similarly, I have a lot of difficulty in multitasking 01:17
audreyt but it's so wrong. better ideas welcome :)
TimToady tracking 3 or 4 conversations at once on IRC is a growth action for me...
audreyt installing bin/v6-alpha is one way to do it
but I wonder if there's some other way.
bennymack hoohoo ouch 01:18
time perl v6.pm -e 'for 1,2,3 -> $x { say $x }'
real 0m11.426s
audreyt well, it gets faster if you save it to a file :)
echo 'use v6-alpha;' > test.pl
echo 'for 1,2,3 -> $x { say $x }' >> test.pl
time perl test.pl
to see the difference 01:19
it's currently slow because the parser is still using backtracking regexes
which, from first-hand experience, I know it's a bad idea
fglock is converting it to use non-backtracking :ratchet :)
bennymack: and help is always welcome -- if you like hacking regexes, that is. want a commit bit? :) 01:20
bennymack audreyt: I must admit I'm not good at regexes 01:21
fglock I'm trying to document the file structure, to help people help
audreyt fglock++ 01:22
01:22 ludan joined
audreyt goes downstairs to try to locate dinner 01:22
markstos fglock: What should be done if we notice something is broken, when there is currently a lot of tests that are known not to pass? I don't suppose adding another failing test helps in that case. For example, in the recent CPAN releases, hashes aren't parsed: my %h = ( a => 'b' ); 01:25
fglock markstos: adding simple tests to v6/t helps the development - it helps to sort out what is more important to implement next 01:28
audreyt merlyn: also, for perl4ish things like this, consider simply writing more 01-sanity/ tests 01:29
er sorry
markstos: also, for perl4ish things like this, consider simply writing more 01-sanity/ tests
basically what fglock said
fglock another option is to add t/02-sanity to pugs/t - what are the most basic things that should be supported 01:30
after we get Test.pm running 01:31
markstos So in other words, if I squeak with a test, I might get some some grease?
"The squeaky wheel gets the grease". 01:32
fglock and then you can help make the test pass too
audreyt hm
markstos fglock: For that, I might wait for your docs about how to help people help.
audreyt tries a trick
fglock: I think one way to do this is to list the tests to import in MANIFEST
and for Makefile.PL to parse that 01:33
and copy from ../../t/ when needed
if you think it's sane I'll go ahead and do it
01:33 justatheory joined
fglock audreyt: sure 01:33
markstos audreyt: makes sense to me.
fglock audreyt: how about to use p5 Test.pm for the basic installation? 01:34
markstos unless there is a way to use the existing SKIP/TODO tricks to declare what might pass for v6 or not.
audreyt fglock: sure
markstos: dduncan was thinking that a "capability" flag is the way to do it
dduncan hello 01:35
markstos audreyt: To somehow declare what the implementation is capable of? 01:36
audreyt markstos: yeah.
it's a bit fuzzy though.
markstos If using SKIP/TODO environment variables allows you to skip/todo tests within a test file, it be better for the finer granularity. 01:37
s/it be/it might be/
obra killall -SIGFOOD audreyt
markstos I must admit I haven't looked at how SKIP and TODO can be manipulated in Perl6 much.
dduncan within the current t/ ... 01:38
I was thinking that sanity-1 should not use any modules at all, meaning the last one is moved out
then sanity-2 tests that Test.pm works and/or the language features that Test.pm requires to run 01:39
fglock markstos: see the v6 tests for simple SKIP and TODO, implemented with try and eval
dduncan then after that are tests for features that Test.pm doesn't require to run
markstos fglock: OK. 01:40
dduncan otoh, it appears that 01-sanity is already testing what Test.pm requires
audreyt I checked non-working Makefile.PL
please carry on
but I'd prefer 02-basic over 02-sanity if possible 01:41
svnbot6 r11325 | audreyt++ | * v6 Makefile.PL + MANFEST hack to copy t/ -- work in progress
dduncan but of course
I will add it now
fglock what's a good filename for the hacking docs - is HACKING ok?
audreyt sure
dduncan actually, is 01-sanity even a good name ... should that be renamed? 01:42
audreyt Test.pm has its own tests
under ext/
dduncan yes
but those should be run after 01-sanity and before everything else
audreyt in any case, I think cherrypicking tests from t/ for v6 is maybe more interesting...
dduncan: the thing is that if it's testing "what Test.pm requires" 01:43
then by default you can't use test.pm to test it
so I still wonder what's 02 for :)
dduncan yes
audreyt but in any case... feel free to try and see
dduncan but 01-sanity doesn't use Test.pm
audreyt I need to dinner now before obra and/or nothingmuch kills me
audreyt waves - bbiab
svnbot6 r11326 | fglock++ | v6 - base HACKING file 01:47
Jedai ?eval 'ōæ½xCB'.ord 01:50
TimToady er, looks like we distracted Audrey at the wrong time... 01:56
the larrybot returns 203 for that.
svnbot6 r11329 | Darren_Duncan++ | added t/02-test-pm/ which consolidates tests for Test.pm itself ... t/01-sanity/08-test.t became 0-compile.t, all of ext/Test/t/* was otherwise copied in ... I didn't yet remove the copies under ext/Test/t, pending whether or not it is a good idea 02:02
r11330 | fglock++ | v6 - HACKING - most files explained 02:04
markstos thanks, flock. 02:07
02:12 fglock joined
spinclad dduncan: considering Test (the harness, iiuc) as a separate CPAN module from the test suite, it needs its own test suite within its distro tree. i would think ext/Test/t needs to stay where it is. 02:14
02:16 evalbot_11330 joined
dduncan so I will leave it there 02:21
audreyt dduncan: re t_pugs 02:22
dduncan but the same tests are also useful to run among the normal Perl 6 tests
audreyt why no t/run/ ?
and then change ./pugs to $*EXECUTABLE_NAME in the tests
dduncan the idea here was that t_pugs contained contained tests only applicable to Pugs
svnbot6 r11331 | fglock++ | v6 - HACKING update
dduncan or do those arguments work with v6 et al too? 02:23
or should work
audreyt other than -C and -B
which is indeed questionable
I think they should eventually all work
a perl that doesn't support -e or -M or -I
is... probably not good for interchangeable deployment. 02:24
dduncan well, that was a rough cut, things can be moved more
the idea is that Pugs probaby has a small handful of tests specific to it, which go in t_pugs/
those that are generic go in t/
I don't know enough about the details yet to decide which go where
audreyt I think I'll first move them back to t/invocation 02:25
and then work out which ones are indeed not universal
dduncan sounds good
so then, t_pugs/ will just be an empty folder for now
audreyt and also it hurts autocompletion
dduncan but an existing folder
audreyt I'd much prefer t/implementation-specific/pugs
or something along that line
dduncan if that works out 02:26
the thought I had was that t/ as a whole could become the test-only distro
and presumably a test-only distro shouldn't know anything about specifics of individual implementations
audreyt that I agree. 02:27
dduncan except that individuals can influence the structure of the generic framework
svnbot6 r11332 | audreyt++ | * move t_pugs/invocation back to t/run/.
dduncan to better accommodate them
audreyt I'd argue your notion of t_pugs should always be empty
if there's something ./pugs can do that cannot be expressed as a capability
then we need to add some declaration to say that capability 02:28
dduncan so then just remove t_pugs/
audreyt already done
but I agree the name "pugs" should not appear anywhere in the t/ hierarchy names
dduncan:++ 02:29
dduncan++
dduncan since you removed the whole dir, refs to it in docs need updating ...
audreyt also t/bugs 02:31
svnbot6 r11333 | fglock++ | v6 - HACKING - "Implementing new syntax"
audreyt is misnamed
obra suggests t/uncategorized
which is good. I'll do the mv.
dduncan I know that t/bugs needs to be broken up, but on short notice it was better than t/pugsbugs 02:32
audreyt Juerd: may we remove t/Dialects/ ? I think it's now better suited to a module that implemnts those dialects, should one exist 02:33
fglock please let me know if there is something more I should write in HACKING - it's just the basics right now
audreyt dduncan: yeah, lookat the passing ones especially in t/xx-uncategorized/ and sort them -- that'd be wonderful too 02:35
?eval "reverse".reversed.reversed 02:36
02:36 evalbot_11330 is now known as evalbot_11333
evalbot_11333 Error: No such method in class Any: "&reversed" 02:36
audreyt ?eval "reverse".reverse.reverse
svnbot6 r11334 | audreyt++ | * rename t/bugs to t/xx-uncategorized.
evalbot_11333 "reverse"
02:37 play9999 joined
Jedai :) 02:41
02:46 miyagawa joined
svnbot6 r11335 | fglock++ | v6 - HACKING - all files are described; copyright/license 02:51
r11336 | Darren_Duncan++ | brought lib/Pugs/Doc/Hack.pod and some other docs up to date with changes to t/ et al ... other parts of Hack.pod still need updating
dduncan it occurs to me that lib/Pugs/Doc/Hack.pod has stuff that wasn't true since 2005 ... I will address the obvious parts now 02:52
nothingmuch has anybody seen my card reader? 03:15
spinclad ?eval "reverse".reverse.reverse.reverse 03:16
03:16 evalbot_11333 is now known as evalbot_11337
evalbot_11337 "esrever" 03:16
obra I haven't :/ what style is it?
nothingmuch obra: usb, cf (no others) 03:17
it's gray 03:19
i think it's by fuji
hmm
obra nod. I'll keep m,y eyes open
nothingmuch okay
thaks
*n
found it 03:25
well, remembered where I lost it =)
obra: do you have windows in any form on your macbook? 03:27
obra yes 03:28
XP in parallels
nothingmuch hmm wait
i think autostitch may have a mac version by now
oh, nope 03:29
03:29 xdg joined
nothingmuch anyway, i'm going to import the NEF raw images, and exports jpegs for the panoramas... can you run it for me? 03:29
obra sure. 03:30
url?
nothingmuch www.cs.ubc.ca/~mbrown/autostitch/autostitch.html
lambdabot Title: "AutoStitch"
nothingmuch (awesome software, btw, too bad it isn't opensource)
obra tell me when i cna get images 03:32
looks like www.kekus.com/download/index.html calico is the mac version 03:33
lambdabot Title: "Kekus Digital - Panorama Tools for OS X :: Download"
nothingmuch obra: will do. is it free? 03:35
nevermind, checking right now
god aperture can kill a computer
<3 UV filters 03:37
the pictures look far less hazy then reality did 03:38
markstos I'm starting to pick up where audreyt left off on the Pugs-Compiler-Perl6/Makefile.PL refactor. In particular, I'm removing the test files that are now being copied over from the main test tree as part of "perl Makefile.PL" 03:48
03:52 diotalevi joined
nothingmuch ola diotalevi 03:52
diotalevi Hi.
I hoped I'd just run into you.
Lucky me.
nothingmuch if i told you you had a beautiful body, would you hold it against me?
diotalevi Heh. You ask so nicely. 03:54
So what kind of format is that Module-Compile/t/fold_blocks.t thing written in? 03:55
I'd write a failing test for it but he didn't write it in any perly language.
nothingmuch is guessing Test::Base 03:56
checking
yes
basically
test base is data driven testing
=== Fold a heredoc whilst preserving ~s is the test case
diotalevi Yay lisp?
nothingmuch --- perl is the type of test
it also contains the input 03:57
oh wait
filters({ perl => 'process' });
no_diff;
run_is perl => 'folded';
that's the part that determines how things are tested 03:58
--- perl and --- folded are just the fields
diotalevi I wish this stuff were written in something sane.
nothingmuch it filters the perl entry of every test record with the process method, and then tells Test::Base to run_is (That is make sure that is( process( $entry->{perl} ), $entry->{folded} ); 03:59
foreach my $entry split_it_according_to_ingy_rules(<DATA>) )
it makes sense in a very ensugared kind of way
i think the block folding is there to make sure that source filters that don't care about heredocs, comments, etc get these magic hex strings
which are then reinflated 04:00
diotalevi K.
So have you noticed that empty tt blocks are a syntax error?
nothingmuch nope, I haven't 04:01
i only used tt for very specific code
diotalevi Didn't work for me. It threw errors.
nothingmuch and haven't really tested rigorously
nothingmuch takes a look
diotalevi C<use tt;\nno tt;>
audreyt ...Test::Base, like most ingyware, is actually a very good idea, nicely executed, but the initial learning curve can be a bit strange :-)
markstos I deleted some files, but svn didn't notice. How do I get svn to notice some files are gone? (I'm used to darcs, which notices automatically... ) 04:02
audreyt svk ci --import 04:03
nothingmuch audreyt: could you help with the MC silliness?
audreyt oh wait. you are using svn ;)
well... "svn up" to revert and "svn rm" to delete
markstos I don't mind switching to svk.
audreyt ok. in svk then, just rm, then "svk ci --import"
nothingmuch yay svk
markstos Ok. I think for the moment I'll use "history" to figure out what I deleted. 04:04
SamB you know what darcs doesn't notice? 04:05
when you do find -name '*.hs' | xargs rm, thats what...
nothingmuch HOLY CRAP I LOVE MY MUM"S CAMERA
i took a picture the a-10s that ruined our (clkao and me) lunch while hiking
you can see the pilot's breathing tube
audreyt SamB: which is why I vastly prefer ~/.svk/ for metadata.
nothingmuch SamB: you can still co that darcs repo to reconstruct it, methinks 04:06
diotalevi zowie. That thing you had with you in Chicago?
nothingmuch diotalevi: yep 04:07
audreyt nothingmuch: not if you _are_ the darcs repo...
nothingmuch audreyt: the patch files shouldn't be affected, just the pristine tree
IIRC 04:08
i may be wrong though
audreyt oh hm. point.
nothingmuch but ~/.svk vs. $PWD/_darcs both have tradeoffs
SamB I'm sure you can reconstruct it from patches
nothingmuch i'd rather have a choice of using either according to my preference
but that is bloatful software
SamB as long as you didn't have any important changes that you hadn't recorded yet
I meant to do find -name '*.hi' | xargs rm 04:09
nothingmuch SamB: so what did you mean by "doesn't notice" ?
SamB it doesn't notice anything is wrong right away
nothingmuch "wrong"
?
why does it care about .hi files? 04:10
SamB well, I typoed it to ".hs"
so it did the wrong thing...
nothingmuch uh
i still don't get it
please don't say "wrong"
say what you expected and what actually happenned =) 04:11
04:11 fglock joined
SamB well, I was expecting to have typed "*.hi" correctly 04:11
but I actually typed it as "*.hs" 04:12
so it deleted all my .hs files...
(including the ones in the _darcs/ directory)
nothingmuch okay 04:13
so what did darcs do/not do?
notice that they were gone and let you revert that?
SamB thats what it didn't do, yeah
nothingmuch hmm 04:15
did you submit a bug?
i can reproduce and it seems pretty nasty
i can repair it by using darcs get --no-pristine 04:16
but that really sucks
markstos What's happening is that darcs is comparing your working copy against the pristine tree and noticing there is no difference, because the reference copy of the file is gone, too. 04:17
svnbot6 r11338 | markstos++ | v6 - Remove test files in favor of having them be copied over by "perl Makefile.PL"
r11338 | markstos++ | by analyzing the manifest.
r11338 | markstos++ | Also, add new t/data_types/hash.t. It should fail, but unfortunately,
r11338 | markstos++ | the ".tc" test file is poorly formed, so all tests are skipped instead.
markstos It should be fixable by "darcs repair", which rebuilds the pristine tree from the patch set.
04:18 shachaf joined
nothingmuch oh right, i forgot that even exists 04:18
markstos It's avoidable in the future by using "darcs query manifest" or "ack" to list files in the current dir, and exclude _darcs.
nothingmuch ack++ 04:19
lovely tool
all it needs is -A -B and -C
markstos Yes, I just installed it to avoid seeing ".svn" directories....
nothingmuch well, for svn there is really no reason to use svn =) 04:20
svk++ for 1 ...
but i still prefer darcs' lightweight funkiness
markstos My reason to use svn is that svk seems to be more complex. It ask me more questions and has more behaviors. 04:21
nothingmuch that's my reason to use darcs ;-)
markstos It really bothers me that sometimes "svk commit" sends files upstream, and sometimes it commits locally.
nothingmuch you get used to it 04:22
audreyt markstos: my initial proposal for "svk switch" has
"svk offline"
"svk online"
as to train people into thinking in POP3 client terms
nothingmuch heh
SamB it isn't a bug, though...
audreyt i.e. in offline mode when you hit "send"
it's actually in outbox
markstos I'm sure, but I prefer the simplicity of darcs. It has record, and push, and the difference is clear.
audreyt and I agree :) 04:23
audreyt thinks darcs's win is mostly due to not having to be bug-for-bug compatible with svn, and therefor cvs, and therefore rcs
markstos I initially tried to use darcs to commit to Perl6, but it seems like it's not used at all in actually.
nothingmuch yup =)
audreyt otoh, there's something to be said about compatibility...
nothingmuch 's #2 reason for darcs is that anything is a repo 04:24
http, ssh, etc
markstos audreyt: I think it was a great social decision to host Perl6 in svn.
SamB svn is supposed to be bug-for-bug compatible with CVS? I don't think so...
audreyt SamB: the command set
svnbot6 r11339 | fglock++ | v6 - HACKING - links; added to MANIFEST
audreyt not the semantics
the surface feel 04:25
SamB oh, sure, wart-for-wart, mostly
audreyt in that regard svk is not unlike perl6 and darcs is not unlike haskell :)
nothingmuch aye 04:26
SamB only with even worse algorithms...
audreyt i.e. what svk did to svn is comparable to what v6.pm is doing to perl5...
nothingmuch svk is a revision control chainsaw
darcs is like a katana
real ninjas can also kill with a chainsaw quite cleanly
SamB a katana with some bad big-O complexity
nothingmuch but occasionally they fry the catalyst repo
and then a truer ninja must rollback everything
*cough*
audreyt when you kill with a chainsaw, I hear it's called "frag" 04:27
nothingmuch has never been burnt by darcs' conflict behavior
SamB I thought that was rockets
I ran into it when Lemmih did massive changes to Conjure...
audreyt maybe. the last time I played FPS it was Doom 2
SamB well, I dunno, this was SIM Car 04:28
okay, okay, Streets of SimCity
audreyt and truth to be told I didn't enjoy it, but I did prefer chainsaw, and giggled a lot. :)
nothingmuch hey 04:29
girls don't play shoot-em-ups
nothingmuch is an unreal-tournament-assault-sniper-rifle guy
audreyt I was 14 or so, and it was the only game available in our .com office for a while...
svnbot6 r11340 | markstos++ | v6 - Add section on "writing tests" to HACKING, to explain about the new "perl Makefile.PL" hackery.
nothingmuch hmm 04:30
when I was 14 i didn't even know how to program
i was just getting started with html
audreyt ponders playing katamari for a bit. 04:31
diotalevi So with this M::C stuff, doesn't Ingy eventually have to know whether the next thing being looked at is an expression or an operator and whether something is in a string or not?
I just got << in some strings to get M::C to fail. 04:32
In the test suite that is.
audreyt diotalevi: yes. PPI or MAD is prolly the only sane way out of this
a patch to use PPI to implement M::C folding would be very nice. 04:33
(or MAD.)
diotalevi yeah. I was just about to say that but didn't when you already did.
TimToady eep. 04:36
Use of uninitialized value in subroutine entry at /usr/lib/perl5/site_perl/5.8.6/Test/TAP/Model/File.pm line 80.
04:36 justatheory joined
TimToady "... error: The submitted smoke has an invalid format! 04:36
nothingmuch aaaaaaaaaaaaaaaaaaah what did I do? 04:37
my line 80 doesn't look like it could generate that error 04:38
svnbot6 r11341 | Darren_Duncan++ | updated lib/Pugs/Doc/Hack.pod to bring the 'Source tree map' section up to date
r11342 | fglock++ | v6 - HACKING - more info
nothingmuch my line 81 might possibly
TimToady my $scalar = List::Util::max(@values);
nothingmuch that's my line 89
audreyt running smoke here to reproduce
TimToady repeats that message about 15 times at the end of the smoke. 04:39
smoke.html seems well formed enough for firefox. 04:40
but only did about 3000 tests
so I don't think the smoke tester has tracked all the directory changes. 04:41
I might be a little out of date on that though.
now there's a new one. svn up produces: 04:42
svn: Working copy 'perl5/Sub-Multi/inc' is missing or not locked
nothingmuch TimToady: the routine that generates the test case objects (one per block) seems to gnenerating that warning
err error
which seems failry bogus 04:43
because the sub in question is the filter (i think)
which is always hard coded
so something is calling _c (the convenience construct method guy) in a naughty way (i think)
likely cause is malformed TAP output
TimToady got a fresh 'svn up' so I'll try it again. 04:48
diotalevi I'm new to svn. If I have a checkout is it normal that subdirectories will get reported as not being under version control even though it was originally provided in the co? 04:49
Tried svn diff src/dynpmc/Makefile from within a parrot checkout and got "svn: 'src/dynpmc/Makefile' is not under version control" which is wrong, AFAIK. 04:50
SamB probably it is a generated file 04:51
audreyt it is a generated file.
nothingmuch jupiter and saturn
dduncan what about them? 04:52
nothingmuch just singing
www.lyricsfreak.com/p/pink+floyd/as...08806.html 04:53
lambdabot Title: "Pink Floyd | Astronomy Domine Lyrics"
diotalevi bummer.
SamB this is how you do it: o/` yodo, yo-yo-yo-yo yoda o/` 04:54
nothingmuch is paying $40 for a *$@%!% panorama stitching app 04:55
if *ANYBODY* wants their panoramas stitched
=)
luckily making good panos is just a function of CPU time
diotalevi got the *right* makefile this time. thanks for the pointer audreyt. 04:57
audreyt np :) 04:58
markstos wonders of hugin is any good at Panorama stitching. hugin.sourceforge.net/
lambdabot Title: "hugin - Panorama Tools GUI"
nothingmuch markstos: i tried that a while back, IIRC
www.cs.ubc.ca/~mbrown/autostitch/autostitch.html
lambdabot Title: "AutoStitch"
nothingmuch this is what i want 04:59
unfortunately it's a govnermnet funded university project
so it's only got commercial licensing
</cynical>
obra: i'll actually take you up on the macbook 05:00
svnbot6 r11343 | fglock++ | v6 - TODO update
nothingmuch the reg key will take ~24 hrs to arrive
markstos AutoStick does look neat.
nothingmuch it's got very good output... i've used it several times before 05:01
but using virtual pc silliness
markstos fglock: thanks for adding hash and array to the TODO list. :) 05:02
audreyt smoke completed and uploade 05:05
fglock markstos: not sure what's the best way to implement the runtime yet - maybe translate all @xxx to $_ARRAY_xxx may get closest to the p6 semantics - but hash/pair is a bit more difficult 05:06
audreyt I don't seen any problems here
fglock: hm, what's the problem with reusing native hash and array types? 05:07
just contexts? 05:08
fglock audreyt: not sure yet - will need some tests to see how it works 05:09
05:10 cmatei joined
audreyt ok. something like Perl6::Contexts is always possible 05:11
fglock %a = (1,2,3) works already
a block containing a list is always a hash or a pair, right? 05:13
audreyt uh 05:15
no
If the closure 05:16
appears to delimit nothing but a comma-separated list starting with
a pair (counting a single pair as a list of one element), the closure
will be immediately executed as a hash composer.
also, {} is a hash 05:17
fglock is '{ }' also a hash (with spaces inside) 05:21
audreyt yes 05:22
heh. Simon Cozens, in the POD of B::Generate: 05:23
This module can obviously be used for all sorts of fun purposes. The best one will be in conjuction with source filters; have your source filter parse an input file in a foreign language, create an op tree for it and get Perl to execute it. Then email me and tell me how you did it. And why.
this sounds eeriely familiar...
obra *laugh* 05:29
nothingmuch mentions something about talking to rgs about this in OSDC 05:30
all we need is a volunictim to actually execute this
audreyt mm volunictim. 05:35
victeer
nothingmuch i don't like victeer 05:38
it sounds like it involves winning
05:38 amnesiac joined
audreyt ...and sneering 05:39
svnbot6 r11344 | Darren_Duncan++ | removed some stuff from MANIFEST.SKIP that actually is in version control, sorted the rest 05:45
r11345 | fglock++ | v6 - hash composer: '%a = { }; %b = { 1, 2 }; %c = { $a }' emits hash,hash,code; 05:48
r11345 | fglock++ | - not implemented yet: '%d = { $pair, }; %e = { 1=>2 }'
nothingmuch ughghghgghhgg 05:49
Darren_Duncan--; # perl5/Data-Bind should *not* be rolled into the next release of pugs even if it is under version control 05:50
etc etc
dduncan so MANIFEST.SKIP isn't just for generated files etc? 05:51
and .svn and stuff
audreyt it's a list of things we don't bundle with Perl6-Pugs.
dduncan sometimes its hard to tell whats cruft when things are routinely out of date 05:52
I can revert that file then
a related question ... 05:53
fglock sleep &
dduncan if you say "svk status" , does it look in MANIFEST.SKIP so to not display various generated files that aren't in version control with '?'s 05:54
if so, I'm getting somewhat inconsistent results; eg, some Makefile show up, others don't
05:55 scook0 joined
nothingmuch nothingmuch.woobling.org/DSC_3487.jpg 05:56
dduncan quick question ... the order of items in the file doesn't matter, does it? 06:01
while I revert the items, I'm inclined to leave them sorted so they're easier to search in
nothingmuch dduncan: probably not 06:02
ExtUtils::Manifest
since they're actually regexes they don't have an "order" 06:03
if a file matches any of these it'll be skipped
obra -> bed
dduncan yes
I saw a redundant one, I think: ^throw_away/ and ^throw_away 06:04
will keep shorter
nothingmuch nuh uh
keep the one that is more explicit
=)
it doesn't like the throw_away directory
dduncan fine then
nothingmuch not any file whose path starts with throw_away 06:05
but meh, it doesn't really matter
dduncan eg, we might have a throw_away_not_really
06:05 Khisanth joined
dduncan but as I said, both were there before, so the less specific subsumes the other anyway 06:05
svnbot6 r11347 | Darren_Duncan++ | restored MANIFEST.SKIP entries removed in r11344, but made them sorted in groups by [full match, match start, match end, match middle] 06:12
nothingmuch gallery.woobling.org/album57 06:51
lambdabot Title: "Woobling :: Untitled"
07:06 Aankhen`` joined 07:11 dduncan left
nothingmuch clkao_: ping re link 07:12
good night
clkao_ hi 08:12
08:31 f0rth joined 08:44 xinming joined 08:48 Khisanth joined
clkao_ pushes the bos-lhr branch 08:51
08:51 pingu joined
svnbot6 r11348 | clkao++ | r15253@ab: clkao | 2006-07-09 01:11:42 +0100 08:55
r11348 | clkao++ | disable the "is" sub emit cleverness.
r11349 | clkao++ | r15254@ab: clkao | 2006-07-09 02:08:34 +0100
r11349 | clkao++ | add_multi interface for incremental declaration.
08:57 nnunley joined
svnbot6 r11350 | clkao++ | r15255@ab: clkao | 2006-07-09 02:13:24 +0100 08:57
r11350 | clkao++ | die when there is extra incoming positional parameters.
r11351 | clkao++ | r15256@ab: clkao | 2006-07-09 02:28:36 +0100 09:00
r11351 | clkao++ | Pass simple-multisubs.
r11352 | clkao++ | r15257@ab: clkao | 2006-07-09 03:06:35 +0100
r11352 | clkao++ | Support "my sub foo {}", which is local *foo = sub {}.
r11353 | clkao++ | r15258@ab: clkao | 2006-07-09 04:26:42 +0100 09:03
r11353 | clkao++ | Optionaly Pugs::Compiler::Regex->compile cache with Cache::FileCache.
r11353 | clkao++ | Before:
r11353 | clkao++ | % time perl -MPugs::Compiler::Perl6 -e1
r11353 | clkao++ | 13.221u 0.067s 0:13.66 97.2% 10+23038k 0+0io 0pf+0w
r11353 | clkao++ | After: (note the second run)
r11353 | clkao++ | % time perl -MPugs::Compiler::Perl6 -e1
r11353 | clkao++ | 13.835u 0.231s 0:14.54 96.6% 10+24234k 0+0io 0pf+0w
r11353 | clkao++ | % time perl -MPugs::Compiler::Perl6 -e1
r11353 | clkao++ | 3.153u 0.134s 0:03.36 97.6% 10+13086k 0+0io 0pf+0w
r11354 | clkao++ | r15259@ab: clkao | 2006-07-09 05:14:50 +0100
r11354 | clkao++ | ->arity.
r11355 | clkao++ | r15260@ab: clkao | 2006-07-09 05:25:45 +0100 09:06
r11355 | clkao++ | Pass builtin/arity.t
r11356 | clkao++ | r15261@ab: clkao | 2006-07-09 06:30:16 +0100
r11356 | clkao++ | doc fixes.
09:08 chris2 joined
svnbot6 r11357 | clkao++ | r15262@ab: clkao | 2006-07-09 06:57:31 +0100 09:08
r11357 | clkao++ | Cache for yapp too, this makes Pugs::Compiler::Perl6 load time down to 1sec.
r11358 | clkao++ | r15263@ab: clkao | 2006-07-09 07:02:48 +0100
r11358 | clkao++ | src_to_blib - touch pmc files at the end, so they don't get recompiled.
clkao_ seatguru++ # finding me power on economy 09:10
monkey_ um, the paste page isn't working :( 09:11
so anyway
svnbot6 r11359 | clkao++ | Fix PCR bootstraping problem. use p5 version of iterator-engine.
monkey_ does anyone have any idea why the following doesn't seem to terminate: 09:12
multi sub foldr (&f, $x) {$x}; multi sub foldr (&f, *$h, *@t) {&f($h,foldr(\&f,@t))}; foldr(sub {@_[0] + @_[1]},(1,2,3));
the last line just never returns 09:13
i'm a complete p6 noob thoug, so maybe i've got bad syntax or i'm missing something
clkao_ is there try(...) ? i thought there's only try {block}
t/perl5/basic.t, that is 09:15
09:15 elmex joined
clkao_ ?eval try(1) 09:18
09:18 evalbot_11337 is now known as evalbot_11359
evalbot_11359 Error: Cannot cast from VInt 1 to Pugs.AST.Internals.VCode (VCode) 09:18
clkao_ ?eval try({1})
evalbot_11359 1
clkao_ ?eval try {1}
evalbot_11359 1
clkao_ mmm
meet superfast v6 compile ;) 09:25
09:39 larsen joined 09:42 coderhasan joined 10:03 Jedai joined
Jedai Hello 10:03
svnbot6 r11360 | clkao++ | * Move eval from prefix op to a runtime function. 10:06
r11360 | clkao++ | * Support eval(..., :lang<perl5>).
10:12 mj41_ joined
mj41_ hello, 10:13
Could not find module `Data.ByteString': use -v to see a list of the files searched for (imported from src/Pugs/AST/Internals/Instances.hs)
win32, ghc 6.4.2 10:14
svnbot6 r11361 | clkao++ | * Make try emit a do block, so "if try { ... }" would work. 10:15
r11361 | clkao++ | * Pass perl5/basic.t
10:18 Jedai is now known as Jedai666 10:22 Jedai666 is now known as Jedai, coderhasan left 10:39 elmex joined 10:43 Provito joined 11:05 penk joined 11:09 Jedai joined 11:33 mauke_ joined 11:40 DaGo joined, iblechbot joined 11:47 lumi joined 11:48 mauke_ is now known as mauke 12:27 Corion joined, Corion left 12:28 prefiks joined 12:43 elmex joined 12:56 snowstalker joined
meppl guten morgen 12:58
Jedai Hello
13:16 nothingmuch_ joined
nothingmuch_ morning 13:34
13:35 nothingmuch_ is now known as nothingmuch
gaal hey nothingmuch 13:58
14:03 theorbtwo joined
nothingmuch hi gaal 14:07
whazzaaaaaaaaaap
nothingmuch waits for all the slackers to get up so that we can have breakfast
gaal we went to see Rocky Horror last night, in the theater where "it all started" 14:15
nothingmuch oi 14:28
one of my exes was really into that in a scary way
nothingmuch has bad connotations since
gaal two amusing things, to me: first, the back talk has been being updated all along: there were mentions of cell phones and Bush 14:47
second, most of the cast were under thirty, meaning they were born after this particulat venue had stopped putting up the show (this was the first time in years) 14:48
(omg present perfect progressive passive again!) 14:50
nothingmuch heh 15:00
gaal++; # nice silliness
15:14 ludan joined 15:15 rashakil joined 15:26 FurnaceBoy joined 15:31 nothingmuch joined
clkao_ nothingmuch! 15:37
15:37 Aankh|Clone joined
ajs_home What do I have to do in order to acquire this mythical commit bit that Larry said was incoming? Does an account need to be created somewhere that I give as a command-line option to svn? 15:39
nothingmuch clkao_: ! 15:40
did you see your pics?
15:40 cognominal joined
clkao_ nothingmuch: yes! very nice! except the one i am sleeping 15:41
nothingmuch clkao_: you're up already 15:42
upload yours
i need to link to my winning entry in the hideous hiker competition 15:43
nnunley grins.
15:55 Aankh|Clone is now known as Aankhen``
nothingmuch clkao_: !!!!!!!!!!!!!!!! 15:57
gaal ajs_home: can we use the email you used in the post to p6-l? 15:58
clkao_ nothingmuch: zzz 15:59
nothingmuch clkao_: UPLOAD
clkao_ lazy 16:01
gaal ajs_home: actually I see that an invitation has already been sent to you; did you not receive it?
nothingmuch clkao_: oh come on 16:02
i wanna see that picture
ajs_home Sorry was away a bit
I'm a back, and yep, [email@hidden.address]
gaal resending
ajs_home k, thanks 16:03
Mail may take a bit. I greylist.
I'll check it later, after I see A Scanner Darkly
gaal resent: with this metabit you can invite others, too.
heh, I just saw that last night - 1 16:04
enjoy
ajs_home gaal: Thanks... oh, and what did you think? Good adaptation?
gaal I wouldn't want to spoil anything for you so let's talk after you see it :)
ajs_home k, wilco 16:05
monkey_ gah, don't suppose anyone has any thoughts on the foldr problem i posted earlier? 16:07
hmm 16:08
?eval 1
16:08 evalbot_11359 is now known as evalbot_11361
evalbot_11361 1 16:08
nothingmuch gaal: is meta-bit-meta-bitting now std?
i wasn't sure if i should invite as admin or not
monkey_ ?eval multi sub foldr (&f, $x) {$x}; multi sub foldr (&f, *$h, *@t) {&f($h,foldr(\&f,@t))}; foldr(sub {@_[0] + @_[1]},(1,2,3));
evalbot_11361 (no output) 16:09
monkey_ hmm
gaal nothingmuch: ajs already had a metacommitter invite set for him
nothingmuch gaal: i meant like, new invitees 16:10
in general
gaal audrey said a couple of times that she's happy with more metacommitters, and ajs has been doing S29 work
nothingmuch is ajs special for getting meta comitter access or not?
ah
16:16 integral_ joined 16:27 integral_ is now known as integral 16:29 larsen_ joined 16:37 justatheory joined 16:39 putter joined
putter hi all 16:39
re mmd dispatch
clkao_ greetings 16:40
putter hi clkao :)
my next line is "oh, never mind, I get it". :)
clkao_ putter: sub::multi is for you, fill in the vapourware detection loop ;)
putter ok :) 16:41
are there partial orders on classes/roles? 16:42
the two parts which have remained stable for a while are the phase 0 "ignore variants who's signatures are not compatible with the call capture", and type narrowness comparison is inclusion-based. 16:44
TimToady To the first approximation, the basic criterion is that people you've met personally can have metacommit access. 16:45
(that's what Audrey told me) 16:46
basically we just want to make sure metacommiters are real people. :)
clkao_ me . o O (anna) 16:47
TimToady (so that we can come and break their kneecaps if they do something stupid. :) 16:50
svnbot6 r11362 | markstos++ | simple code formatting/alignment fix.
r11363 | nnunley++ | Quick extraction of all the calls to *::Internals::* functions.
putter are the variants of a multi sub, subs, and of a multi method, methods? 16:53
or simply a Code,Signature tuple. 16:54
clkao_ nothingmuch: you are not on #svk, we are loving moose there
szbalint How is a meta commit bit different from a regular commit bit?
putter not a pressing question, I've just been wondering
meta lets you add commiters 16:55
szbalint ah
I think I've only met TimToady from here. 16:58
TimToady so you could give me a metacommiter bit. :) 17:02
especially since you already have one.
17:05 nothingmuch joined
szbalint :) 17:09
17:11 lisppaste3 joined
putter anyone willing to hear a mmd story? 17:12
mmd = multimethod dispatch algorithm. i'd explain it all as the story goes. 17:14
svnbot6 r11364 | markstos++ | v6 - update docs for Sub::Multi and Data::Bind. peer review requested. 17:17
r11365 | markstos++ | v6 - start 'CHANGES' file for 17:20
r11366 | markstos++ | My last commit was an oops. This adds another entry Sub::Multi CHANGES.
17:22 diotalevi joined
gaal airport & 17:23
putter gaal: have a good flight. 17:25
markstos: ping?
ah well. I need a user to soundboard the user story on mmd with. bbl & 17:30
Jedai Could someone send me a commit bit (normal ;) ), audrey said she would send me one but I think I missed it 17:46
17:53 shachaf joined
szbalint Jedai: It is already sent. Did you not receive it? 18:07
Jedai I did not :( You sent it on chaddai.fouche@ens-lyon.fr ? 18:09
18:10 hexmode joined
svnbot6 r11367 | nnunley++ | Adding a missing dependency. 18:17
18:21 penk joined 18:22 larsen___ joined
nothingmuch mĆøĆøsage 18:40
18:40 xdg joined
audreyt Jedai: check your spam box? 18:41
Jedai: or just register on rt.openfoundry.org/ yourself
and tell us your id
18:47 Bit-Man joined 18:48 larsen joined
svnbot6 r11368 | audreyt++ | * First draft of Feb 2006 Perl 6 Summary from Ann Barcomb (kudra). 18:53
r11369 | audreyt++ | * add kudra to AUTHORS.
r11370 | audreyt++ | * move to summary. 18:56
r11371 | audreyt++ | * 02-28.
19:11 markstos joined
svnbot6 r11372 | jedai++ | * Adding a test for reverse in t/builtins/string\n* add jedai to Authors 19:18
Jedai It seems I was added automagically after creating a "jedai" account since you already added me as a commiter
Is it normal ? o_O
19:28 bpederse_ joined
audreyt that is normal 19:33
very nice :)
Jedai++
19:38 bpederse_ left 20:32 larsen_ joined 20:35 fglock joined 20:36 stephanepayrard_ joined
fglock wow - the irc daily logs are getting huge again :) 20:42
clkao_: ping 20:51
ajs_home When I submit the registration page on openfoundry it just comes back blank... is that typical? 20:52
clkao_ fglock: pong 20:53
fglock clkao_: I saw the last commits - nice!
clkao_ fglock: ;) it's now superfast
fglock: i was going to make smoke work. but it seems there's some fun with PERL5LIB 20:54
20:56 kanru joined
audreyt ajs_home: just refresh POST 20:56
hit ctrl-r 20:57
20:57 khisanth_ joined
ajs_home Worked nicely, thanks audreyt 20:57
clkao_ audreyt: how goes 20:58
20:59 ruz joined, khisanth_ joined
audreyt working on docs/summaries/2006/02-28.pod with nothingmuch 20:59
20:59 khisanth_ is now known as Khisanth 21:00 trym joined 21:02 tcliou joined, knewt joined, obra joined, unobe joined, spinclad joined 21:05 xdg joined, buubot joined 21:08 elmex joined, scw joined 21:10 larsen joined, nothingmuch joined
clkao_ moose 21:11
svnbot6 r11373 | ajs++ | Updates to S29 / Functions based on p6l disucssions and some general cleanup.
r11373 | ajs++ | Fleshed out or otherwise modified:
r11373 | ajs++ | * constants
r11373 | ajs++ | * sprintf
r11373 | ajs++ | * grep/map
r11373 | ajs++ | * pop/push/shift/unshift/splice
r11373 | ajs++ | * wait/fork
r11373 | ajs++ | * sleep
r11373 | ajs++ | * sign
r11373 | ajs++ | * defined/undefine/undef
r11373 | ajs++ | * zip
r11373 | ajs++ | * substr
ajs_home Ahem...
svnbot6 r11373 | ajs++ | * join
r11373 | ajs++ | * bless
r11373 | ajs++ | * ord/chr
r11373 | ajs++ | Also, some gereral cleanup around optional arguments syntax. and removing
r11373 | ajs++ | code-as-documentation where I've provided actual documentation.
r11373 | ajs++ | Also moved waitpid to obsolete section, since named arguments on wait 21:12
r11373 | ajs++ | subsume its functionality. It's not clear if we need or want an alias...
ajs_home er ... sorry about the long comment :-/
svnbot6 r11374 | nothingmuch++ | uri shortening script 21:14
21:16 SamB joined
svnbot6 r11375 | audreyt++ | * 02-28: joint proof-reading and link fixes with nothingmuch+ 21:16
r11376 | audreyt++ | * properly utf8'ing 21:19
21:21 fglock joined
nothingmuch kudra++ # real work behind summaries 21:21
svnbot6 r11377 | nothingmuch++ | add -u option to shorten script - to unshorten uris
audreyt kudra++ # indeed
fglock how about creating an AUTHORS file for v6.pm? 21:23
clkao_: please let me know if there is anything I can do to help you - I'm now working on array/hash 21:24
clkao_ fglock: there's some t files failing compiling, even more, it's infinite loop when parsing 21:31
nothingmuch clkao_: photos
fglock clkao_: the infinite loop may actually be the parser backtracking over some error - I'll check 21:32
clkao_ nothingmuch: working on i. i think i forogt my card reader, where i now have the usb calbe 21:35
svnbot6 r11378 | nothingmuch++ | stub for 03-31.pod
r11379 | nothingmuch++ | stubs and props for other months 21:38
r11380 | nothingmuch++ | small fixes to the summary and summary stubs 21:45
fglock clkao_: in t/01-pod.t - the error is because the eval was moved out of the lexical context - it happens in a subroutine now
eval needs to be inlined 21:46
clkao mm
svnbot6 r11381 | nothingmuch++ | explain wtf is parrot-porters 21:48
21:48 lisppaste3 joined
fglock clkao_: it's incredibly faster 21:50
svnbot6 r11382 | nothingmuch++ | revedrt nothingmuch--'s change, and s/perl6-internals/parrot-porters/ where appropriate due to list rename
clkao fglock: ;))
nothingmuch video.google.com/videoplay?docid=-3...49704& 21:52
i'm surprised they didn't make fun of the segway incident 21:55
szbalint oh they finally renamed the list? interesting++ 22:03
obra szbalint: chip decreed that parrot-porters was a part of resetting the culture. 22:04
audreyt amen to that 22:05
svnbot6 r11383 | fglock++ | v6 - added infix:{'=>'} 22:11
clkao fglock: we also need to emit things respecting $infix:<XXX> 22:19
fglock clkao: example? 22:20
clkao if you override infix:<+> somewhere
fglock I was leaving this for later - it makes optimizing much more difficult, because all ops need to be tested before executing 22:21
localizing the grammar is possible, but it's not implemented yet 22:22
but it is in the TODO list 22:23
however, objects implementing ' ' can be translated to overloads 22:26
'plus'
22:28 Khisanth joined
clkao i think we need a new set of tests 02-prelude that supports the syntax and semantics used by prelude that is to be shared 22:29
fglock where does Cache::FileCache stores files in windows? 22:33
clkao i don't know. <tmp directory>/FileCache 22:34
fglock thanks! 22:36
22:41 dduncan joined 23:03 lumi joined
TreyHarris anyone good with figuring out why Build scripts are installing where they are? I install everything in my home directory, but so far, even with prefix, install_base, and all the makemakerflags set to my home dir, it's still trying to install DateTime::Format::MySQL into /usr/local/man and /Library/Perl/5.8.6/darwin-thread-multi-2level/auto/DateTime... 23:03
bennymack perlbot cpan 23:07
perlbot Comprehensive Perl Archive Network (where to download perl modules) is at www.cpan.org/, and can be searched with search.cpan.org/
bennymack hmm
svnbot6 r11384 | fglock++ | Pugs/Grammar/Precedence.pm - (1,) parses as ('1',null) instead of '1' 23:17
r11385 | fglock++ | v6 - {1=>2} is Code; {1=>2,} is Pair/Hash 23:29
clkao fglock: can you make t/builtins/arrays/elems.t parse fail gracefully? 23:30
it's currently looping
nothingmuch clkao++ # some photos 23:32
fglock clkao: looking 23:34
dduncan following suggestions posted here yesterday, wherein I said I couldn't afford to attend any conferences short term (the cheap ones, YAPC, require plane fare, those close by, OSCON, are expensive), it was suggested that I could try the "hallway track" of OSCON ... having looked more into that event, I've come to realize that there is a lot going on there that is free, and those parts are generally more valuable than any sessions or tutor 23:48
despite the short notice, I am now strongly considering going and doing this ...
so besides those listed as speakers, and hence would be in the area, I'm interested in knowing who will be there
and also, where you will be staying at the time 23:49
clkao nothingmuch: my camera date is fscked
nothingmuch =( 23:52
23:55 pasteling joined

Powered by Cro and the RakuĀ® Programming Language.

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

Thank you!