Perl 6 - the future is here, just unevenly distributed

IRC log for #perl6, 2006-06-12

Perl 6 | Reference Documentation | Rakudo

| Channels | #perl6 index | Today | | Search | Google Search | Plain-Text | summary

All times shown according to UTC.

Time Nick Message
00:03 cognominal joined perl6
00:10 qwr audreyt: thanks.
00:16 audreyt np :)
00:20 hikozaemon joined perl6
00:21 b00t joined perl6
01:13 dduncan joined perl6
01:14 dduncan so, is there an ETA for a prelaunch release candidate that needs testing?
01:14 dduncan tarball that is
01:18 audreyt yes. 5hr from now I think
01:19 audreyt or less
01:19 dduncan okay
01:21 dduncan depending on timing issues, such as whether I want to stay up late, I may only download and build/test this starting at about 8 hours after that (mid-morning my time) ... with results c 3-5 hours later
01:21 dduncan but whatever happens, happens
01:25 audreyt sure
01:27 ludan joined perl6
01:52 svnbot6 r10628 | audreyt++ | * Relax the parsing rule for RHS-as-list somewhat:
01:52 svnbot6 r10628 | audreyt++ |     $x.foo = 1,; # RHS is a List
01:52 svnbot6 r10628 | audreyt++ |     $x.foo = 1;  # RHS is a Int
01:52 svnbot6 r10628 | audreyt++ |   again, principle of least surprise.
01:52 svnbot6 r10629 | audreyt++ | * "do {block}" now captures lexicals created at the block body level correctly.
01:59 mako132_ joined perl6
02:45 QtPlatypus joined perl6
02:50 viirya joined perl6
03:34 lambdabot joined perl6
03:45 audreyt another perf refactoring hits the trunk :D
03:45 audreyt bbiab &
03:45 svnbot6 r10630 | audreyt++ | * Pugs.AST: We now install a multi nullary sub "&Foo" whenever
03:45 svnbot6 r10630 | audreyt++ |   the package/class/module/role/grammar "::Foo" is introduced.
03:45 svnbot6 r10630 | audreyt++ |     Foo.new;
03:45 svnbot6 r10630 | audreyt++ |   is now always commpiled to:
03:45 svnbot6 r10630 | audreyt++ |     Foo().new;
03:45 svnbot6 r10630 | audreyt++ |   this removes the primary stateful parsing ambiguity, resulting
03:45 svnbot6 r10630 | audreyt++ |   in a nice 20% speedup in "make smoke" (15min -> 12min).
03:58 lambdabot joined perl6
04:20 dduncan audreyt, assuming those last few "repository snapshot / darwin" smokes are yours, your smoker is broken
04:20 dduncan it says r10590 for all the last smokes
04:20 dduncan at least a day out of date
04:21 dduncan or the version build is broken
04:21 particle_ joined perl6
04:22 dduncan just checking now, a "pugs -v" on my system shows "r10629", which means the problem is either specific to your install, or it is specific to the smoke process
04:23 dduncan I'm basing my judgement on what http://smoke.pugscode.org/ says
04:26 dduncan its probably your install, since someone's Linux smoke shows a r10627
05:22 audreyt nod
05:23 audreyt I think it's a svk problem
05:23 audreyt i.e. pugs not picking up the svk checkout number
05:27 cognominal joined perl6
05:31 svnbot6 r10631 | audreyt++ | * Pugs.Parser: Implement correct OUTER:: compilation:
05:31 svnbot6 r10631 | audreyt++ |     my $x; { $x and my $x = 1 }
05:31 svnbot6 r10631 | audreyt++ |   is compiled to
05:31 svnbot6 r10631 | audreyt++ |     my $x; { $OUTER::x and my $x = 1 }
05:31 svnbot6 r10631 | audreyt++ |   but there is no OUTER for this:
05:31 svnbot6 r10631 | audreyt++ |     my $x; { my $x = 1 if $x }
05:31 svnbot6 r10631 | audreyt++ |   it's according to spec, but a significant divergence from Perl 5.
05:31 svnbot6 r10632 | audreyt++ | * scopt.t: correct syntax for global package.
05:34 svnbot6 r10633 | cmarcelo++ | * Simple tests on JudyHSIter FFI are working now.
05:34 svnbot6 r10633 | cmarcelo++ | * First try at Cabal package. Works here.
05:36 cmarcelo joined perl6
05:36 bjoern_ joined perl6
05:41 audreyt cmarcelo: hi
05:43 cmarcelo audreyt: hello. I malloc'ed some time to work on hsjudy =)
05:44 audreyt good good
05:45 cmarcelo iter bindings seems to be ok. so the low-level stuff is pretty much done.
05:47 cmarcelo audreyt: I'm going for that Pugs.Prim change before looking for new APIs, if that's ok..
05:47 audreyt sure it's okay
05:47 audreyt though please nopaste it as a patch first :)
05:48 audreyt btw, have you seen the Collection API?
05:48 audreyt I think IOCollection or MCollection is a nice fit with it
05:48 cmarcelo not yet..
05:49 audreyt it's a bit more closer to judy than Edison I think
05:49 audreyt http://hackage.haskell.org/trac/ghc/wiki/CollectionClassFramework
05:52 audreyt ByteString is probably natural as JudyHs keys
05:52 svnbot6 r10634 | audreyt++ | * Pugs.Parser.Export: Importing via "use" is now treated
05:52 svnbot6 r10634 | audreyt++ |   as an in-scope "state" declaration for hoisting/shadowing purposes.
05:54 cmarcelo (i'm taking a look on collection stuff..)  didn't get your ByteString comment =P
05:54 cmarcelo ByteString from fps? It has stuff to export CString?
05:55 cmarcelo s/stuff/functions/
05:55 cmarcelo or must do a BS -> String -> CString?
05:56 audreyt        unsafeUseAsCStringLen,  -- :: ByteString -> (CStringLen -> IO a) -> IO a
05:56 audreyt which is "safe" for all JudyHS uses
06:01 cmarcelo class Collection c i o | c -> i o      -- means knowing c type we can derive/infer i and o types?
06:01 audreyt yes.
06:01 audreyt one idea is simply MCollection
06:02 audreyt class CollectionM c i o m | c -> m i o
06:02 audreyt where m == IO is inferred
06:02 audreyt from the instance
06:06 cmarcelo got it..
06:10 cmarcelo JudyHS seems to be more likely to instance Data.Collections.Map..
06:11 cmarcelo but Collection is good enough for BitSet
06:11 svnbot6 r10635 | audreyt++ | * unTODO passing tests.
06:12 audreyt right, but those functions are all pure in the class head
06:12 audreyt    member :: k -> c -> Bool
06:12 audreyt whilst we would need
06:12 audreyt    memberM :: k -> c -> m Bool
06:12 audreyt most likely
06:13 cmarcelo yes yes =)
06:14 cmarcelo is the Freezable-stuff from Meacham still insteresting to keep? I mean, is it actually useful?
06:14 audreyt personally I think it's useful
06:15 audreyt freeze :: (Ix i, MArray a e m, IArray b e) => a i e -> m (b i e)
06:15 audreyt it's analogous to the MArray's freeze
06:15 audreyt so arguably we can turn a CollectionM to a Collection
06:16 audreyt freezeM :: (CollectionM c i o m, Collection c' i o) => c -> m c'
06:16 audreyt makes sense?
06:18 dduncan audreyt, regarding my earlier comment ... what's happening with your versions has happened to me before ... the fix is possibly to work from a clean checkout ... or alternately just following a make realclean ... for the record, myself and at least another person are getting newest versions into pugs from svk at build time, so the problem isn't specific to the code it seems ... but your box
06:19 cmarcelo yes, it makes. i got the types, but i don't get when this kind of thing is used. "just" to jump out from IO monad later (like tests on Freezed BitSets)?
06:19 dduncan or even simpler, explicitly run the Makefile.pl then do a make, perhaps
06:19 audreyt yes. the idea is that we may want to use a BitSet as regular Set
06:20 audreyt in nonmonadic code
06:20 audreyt but we want to construct it very efficiently in IO monad
06:21 cmarcelo ok
06:45 spinclad hi audreyt, cmarcelo
06:46 spinclad audreyt: S02 r9538: "Conversely, put parenthesis around the lvalue expression to force inner context to list:" does not seem to be the case any more.  also, S03 r9575: s/ are allowed.$/&)/ : keep the ending parenthesis.
06:46 cmarcelo spinclad: hey. my iter tests in HsJudy bidings are working now. thanks =)
06:47 spinclad getting my code to meet its own spec helped
06:48 spinclad glad you haven't seen further bugs
06:52 audreyt spinclad: nopaste a S02 patch?
06:52 cmarcelo spinclad: my tests now are very simple -- just to check if the bindings are ok.. next weeks there'll be APIs on top of these low-level bindings, making tests easier to do and bigger..
06:53 audreyt spinclad++ cmarcelo++ # soon Pugs may become Fast Enough For Production (tm)
06:55 dduncan following a new pull+build, my pugs -v reads r10635
06:55 dduncan that's it for today ... in about 9 hours I will check for a preflight tarball and test it
06:55 audreyt okie
06:55 audreyt thanks, dduncan
06:56 dduncan did you solve the version thing at your end? ... it is at your end only
06:56 dduncan for me, certain refreshing of files did the trick
06:57 audreyt yup, I think I grok the reason
06:57 audreyt thanks for the hint
06:57 dduncan you're welcome
06:57 svnbot6 r10636 | audreyt++ | * Perl6::Container::Array - revive much of the functionality and unTODO
06:57 svnbot6 r10636 | audreyt++ |   the now-passing-because-member-variables-really-works tests
06:58 cmarcelo svk users: how do you people organize svk? in which svk-path you put openfoundry mirror, for example?
06:58 audreyt /pugs/trunk and /pugs/local for m
06:58 audreyt me
06:59 audreyt was //mirror/pugs ad //pugs
06:59 audreyt both works quite well
06:59 cmarcelo audreyt: why change?
06:59 audreyt /pugs/ is so that I can copy the depot to other people without inflicting other paths to them
06:59 audreyt i.e. ~/.svk/pugs is the repo by itself
07:00 dduncan for me, I put it in //pugs_try2/mirror
07:00 dduncan when I used separate local repos, that was in //pugs_try2/local
07:00 dduncan I also use svk for something else, which is named //utsl_try1/mirror
07:00 dduncan so first level is the name of the main server, second level is mirror or local
07:01 dduncan the tryN thing is if certain things go wrong and I want to do a clean re-mirror of the main server
07:02 cmarcelo tks!
07:02 dduncan I'm up to #2 or #3 now, but started at #1, fyi
07:03 iblechbot joined perl6
07:04 cmarcelo dduncan: what did "go wrong" to need a clean re-mirror?
07:05 dduncan I was setup with a dual mirror+local system, so that I could commit locally even if the main server was down, and push those commits later ...
07:05 dduncan as some people here do
07:05 dduncan though I wasn't using part of the reason for that feature, which was keeping a local branch ... I stuck to the mainline
07:05 dduncan the problems ...
07:07 dduncan I often got spurious "merging" going on, where svk says that it was merging specific file differences from my local to the mirror, as if there were third party changes to those files since my checkout and checkin, but in truth I was the only one working on them ...
07:07 dduncan sometimes svk pushed the same change multiple times to the main server, where they show up in pugs' logs as multiple commits, but I only actually committed that once
07:08 dduncan sometimes those spurious things were like undos of previous commits, and then later redos, and I couldn't find any easy way to control this ... it seemed to be random
07:08 dduncan there were some other issues, but that was what came to mind
07:09 dduncan actually, here's another problem:
07:09 cmarcelo dduncan: scary problems...
07:10 cmarcelo dduncan: which version of svk were you using then?
07:10 dduncan when I wanted to "svk mv" a directory or file or something, the main server is supposed to receive the move, which is very light operation in svn, and preserves file history from the old to the new ... but what actually happened is, while this happened correctly on 'local', the copy from local to mirror was done as a merged, and pushed to the main server as a full add plus delete, so those files lost their history, and took up more spac
07:11 dduncan I was using the latest svk and svn at any given time, or exactly 1 version behind
07:11 dduncan all fresh installs
07:11 dduncan now, I found that all the problems went away when I used just a mirror only, and checked out / committed to, that mirror
07:12 dduncan fyi, using just the mirror, the svk is working more like the underlying svn, but that it keeps a local mirror copy of the repo on my machine
07:13 cmarcelo hmm (svk) checking out the //pugs/mirror don't give you offline commits then? i.e., when you commit on a checkout from //mirror, it commits into the mirror server itself?
07:13 audreyt yup
07:13 dduncan you can recheckout previously mirrored server content, but any commit tries to push it live
07:14 audreyt see my diagram
07:14 dduncan generally this isn't a problem for me, as the server is up most of the time I need it, and I don't really travel right now
07:14 svnbot6 r10637 | audreyt++ | * don't generate OUTER:: when we are in toplevel.
07:14 audreyt http://commons.wikimedia.org/wiki/Image:Svk-overview.png
07:15 dduncan so the mirror is still helpful when you want to clean-refresh your checkout dir ... it doesn't have to pull it from the server at all
07:16 audreyt also it makes operations such as "log" much, much, much faster.
07:16 svnbot6 r10638 | audreyt++ | * err, typo
07:16 dduncan yes
07:16 mjk joined perl6
07:17 cmarcelo audreyt: ask svk devel(s) to update the diagram link in http://svk.elixus.org/
07:18 audreyt it's a wikie...
07:18 cmarcelo didnt find "Edit" =P
07:18 audreyt click "Preferences"
07:21 marmic joined perl6
07:22 pasteling "spinclad" at 209.94.133.172 pasted "suggested diff to S02.pod" (13 lines, 666B) at http://sial.org/pbot/17685
07:23 cmarcelo audreyt: no "edit" for HomePage, it's common for wikis to lock (or restrict) main page to prevent trivial "defacement"...
07:23 audreyt oh, k.
07:24 audreyt clkao: fix "autrijus, diagram: architecture diagram" to point to http://commons.wikimedia.org/wiki/Image:Svk-overview.png and fix my id?
07:25 svnbot6 r10639 | audreyt++ | * HTML-Entities: Change this:
07:25 svnbot6 r10639 | audreyt++ |     %hash{chr $x} //= 1;
07:25 svnbot6 r10639 | audreyt++ |   to this:
07:25 svnbot6 r10639 | audreyt++ |     %hash{~chr($x)} //= 1;
07:25 svnbot6 r10639 | audreyt++ |   since otherwise the 1-slice is defined and therefore not initialized.
07:25 svnbot6 r10639 | audreyt++ |   Not sure if it's sane, but oh well.
07:31 svnbot6 r10640 | audreyt++ | * repair Perl6::Value::List too
07:38 spinclad (audreyt: diff pasted)
07:44 audreyt committed
07:44 audreyt spinclad++
07:45 elmex joined perl6
07:47 cognominal joined perl6
07:52 dduncan since I was up longer due to the last chat, I confirm that r10640 also builds and shows up in pugs -v fine
07:52 dduncan &
08:07 cmarcelo time to sleep. see you &
08:07 cmarcelo left perl6
08:10 drrho joined perl6
08:18 bjoern_ joined perl6
08:19 bjoern_ joined perl6
08:24 kanru joined perl6
08:25 lisppaste3 joined perl6
08:39 xern joined perl6
08:43 xerox joined perl6
08:44 xerox joined perl6
08:49 buetow joined perl6
09:05 revision17_ joined perl6
09:45 theorbtwo joined perl6
09:46 svnbot6 r10641 | audreyt++ | * getprint.t: @INC needs to be written as @*INC now.
09:49 svnbot6 r10642 | audreyt++ | * my.t: This test was bogus:
09:49 svnbot6 r10642 | audreyt++ |     try { $undeclared }
09:49 svnbot6 r10642 | audreyt++ |   as it is a compile-time, not runtime, failure.  Changed to use 'eval'.
09:49 svnbot6 r10643 | audreyt++ | * Change inlinePerform* back to unsafePerform* when
09:49 svnbot6 r10643 | audreyt++ |   NOINLINE is in effect.
09:49 svnbot6 r10644 | audreyt++ | * Prelude: s/$seek/$self/.
09:51 Aankhen`` joined perl6
09:52 svnbot6 r10645 | audreyt++ | * Pugs is now capable of doing parse-time analysis on
09:52 svnbot6 r10645 | audreyt++ |   unbound lexical variables:
09:52 svnbot6 r10645 | audreyt++ |     $ ./pugs -c -e '$x'
09:52 svnbot6 r10645 | audreyt++ |     *** unexpected global symbol "$x" requires explicit package name
09:52 svnbot6 r10645 | audreyt++ |         expecting "::"
09:52 svnbot6 r10645 | audreyt++ |         at -e line 1, column 3
09:52 szbalint_ joined perl6
09:53 svnbot6 r10645 | audreyt++ | * Also, named subroutines are now expressions instead of statements:
09:53 svnbot6 r10645 | audreyt++ |     # This is valid Perl 6
09:53 svnbot6 r10645 | audreyt++ |     my @subs = (sub f { 1 }, sub g { 2 });
09:53 svnbot6 r10645 | audreyt++ |   consequently, all unqualified symbol lookup (&f and &g above) are
09:53 svnbot6 r10645 | audreyt++ |   now lexical instead of dynamic.
09:53 svnbot6 r10645 | audreyt++ | * Private subroutines are now exportable:
09:53 svnbot6 r10645 | audreyt++ |     module Foo;
09:53 svnbot6 r10645 | audreyt++ |     my sub bar is export { ... }
09:53 svnbot6 r10645 | audreyt++ |   this allows import from Foo but disallows &Foo::bar.
09:54 audreyt another step closer to a Real Language Implementation...
09:56 xerox audreyt++ !
09:57 xerox audreyt, do you know how to get happy compiled? I couldn't compile darcs without it. Or maybe some important dependences.
09:57 Juerd export!
09:57 Juerd audreyt++ :)
09:59 audreyt xerox: it Just Worked for me
09:59 audreyt alex maybe?
10:00 audreyt I did needed to create a UFS partition to get alex darcs checkout
10:00 audreyt due to case sensitivity
10:00 theorbtwo audreyt++
10:00 xerox Happy-1.15 says Unrecognised platform: i686-apple-darwin8.6.1
10:00 audreyt xerox: use darcs trunk
10:00 audreyt also, maybe autoreconf
10:01 xerox audreyt: may you paste somewhere the output of `ghc-pkg list' ?
10:01 audreyt ALUT-2.0, Cabal-1.1.4, GLUT-2.0, HGL-3.1, HUnit-1.1, OpenGL-2.1, QuickCheck-1.0, X11-1.1, base-1.0, fgl-5.2, (ghc-6.5.20060608), haskell-src-1.0, haskell98-1.0, mtl-1.0, network-1.0, parsec-2.0, readline-1.0, rts-1.0, stm-1.0, template-haskell-1.0, time-0.3.1, unix-1.0
10:02 xerox Good.
10:02 audreyt my happy is 1.16
10:02 xerox Ah.
10:02 xerox Hmm.
10:02 audreyt installed with cabal
10:02 audreyt http://darcs.haskell.org/happy
10:02 szbalint_ is now known as szbalint
10:03 xerox No darcs without happy, no happy without darcs...
10:03 xerox Eek.
10:03 audreyt wget -m ;)
10:05 svnbot6 r10646 | audreyt++ | * Pugs.Parser: make special allowance for "@INC" because
10:05 svnbot6 r10646 | audreyt++ |   Test::Harness uses it.  Sigh.
10:05 svnbot6 r10647 | audreyt++ | * more s/@INC/@*INC/
10:05 xerox Hmpf.
10:06 xerox It went into mirroring darcs.haskell.org/* before darcs.haskell.org/happy/*/*
10:06 audreyt wget -m -np
10:06 audreyt sorry
10:07 xerox Don't be sorry, thank you very much.
10:07 * xerox crosses fingers
10:07 xerox I hate wget manpage, did I say that? :)
10:07 audreyt mm np :)
10:07 iblechbot joined perl6
10:08 xerox (Seems like it will work)
10:40 cdpruden joined perl6
10:46 GeJ audreyt: commiting from the airport?
10:51 audreyt no, it's 24hr from now
10:54 ruoso joined perl6
11:19 chris2 joined perl6
11:23 elmex joined perl6
11:35 ludan joined perl6
11:36 ludan ciao
11:44 buetow joined perl6
11:45 tifo joined perl6
11:51 jsiracusa joined perl6
11:54 b_jonas joined perl6
12:05 b00t joined perl6
12:17 weinig is now known as weinig|work
12:33 svnbot6 r10648 | audreyt++ | * Unify named and unamed subroutines.
12:33 svnbot6 r10648 | audreyt++ |   (note - for releng, I'll rollback this thread of changes
12:33 svnbot6 r10648 | audreyt++ |    and hold until 6.2.12 is released, to investigate more
12:33 svnbot6 r10648 | audreyt++ |    into the serious perf hit caused by it.)
12:38 svnbot6 r10649 | audreyt++ | * unpush r10645 and r10648 as promised.
13:12 buetow joined perl6
13:26 svnbot6 r10650 | audreyt++ | * cherry-pick the unsafePerformSTM patch, as well as other misc. cleanups
13:29 svnbot6 r10651 | audreyt++ | * allow redecl of $! and $/
13:29 svnbot6 r10652 | audreyt++ | * er, typo
13:39 rafl joined perl6
13:41 vel joined perl6
13:41 svnbot6 r10653 | audreyt++ | * when we enter a block at parser level, mark envOuter properly.
13:41 svnbot6 r10654 | audreyt++ | * ...and put it back too
13:47 kolibrie joined perl6
14:00 svnbot6 r10655 | audreyt++ | * retionalize the OUTER:: generation logic somehow; if it's
14:00 svnbot6 r10655 | audreyt++ |   not visible in outer scope either, don't bother.
14:00 svnbot6 r10656 | audreyt++ | * Do not generate $_ as implicit method invocant.
14:00 svnbot6 r10656 | audreyt++ | * Amazingly, anonymous hash literals are still closures, and
14:00 svnbot6 r10656 | audreyt++ |   so OUTER:: should work:
14:00 svnbot6 r10656 | audreyt++ |     my $x; my $hash = { 1 => $OUTER::x };
14:09 svnbot6 r10657 | audreyt++ | * corrected various tests testing ::Class etc for truth,
14:09 svnbot6 r10657 | audreyt++ |   which has been defined as False in the Synopses ever
14:09 svnbot6 r10657 | audreyt++ |   since the Arbel Hackathon.
14:12 svnbot6 r10658 | audreyt++ | * ditto.
14:12 svnbot6 r10659 | audreyt++ | * ditto.
14:15 svnbot6 r10660 | audreyt++ | * pairs.t: Now that list assignment is listop, this no longer works:
14:15 svnbot6 r10660 | audreyt++ |     ok @pairs = $pair.pairs, "pairs on a pair";
14:15 svnbot6 r10660 | audreyt++ |   we have to write this:
14:15 svnbot6 r10660 | audreyt++ |     ok((@pairs = $pair.pairs), "pairs on a pair");
14:15 svnbot6 r10660 | audreyt++ |   sad but true.
14:16 masak left perl6
14:17 b_jonas does that mean that @foo = 3, 5 will be equiv to @foo = (3, 5) ?
14:17 fishb0t dumb question maybe, but are pack and vec going into pugs?
14:17 svnbot6 r10661 | audreyt++ | * substr.t: prototype objects are no longer true.
14:19 audreyt probably not as such, but as compact structures in S09
14:20 audreyt someone proposed doing that a while ago... forgot who
14:21 b_jonas yes, I proposed it on perlmonks and TimToady said they considered it but decided against so it probably won't work
14:21 b_jonas I mean TimToady said they _had_ considered it before that
14:21 b_jonas so someone else has proposed it
14:22 b_jonas http://www.perlmonks.com/?node_id=488991
14:22 audreyt ?eval @foo = 3,5; @foo
14:22 evalbot_10625 is now known as evalbot_10661
14:22 evalbot_10661 Error: Undeclared variable: "@foo"
14:23 audreyt ?eval my @foo; @foo = 3,5; @foo
14:23 evalbot_10661 [3, 5]
14:23 audreyt look, itworks
14:23 b_jonas cool
14:23 b_jonas ?eval my($foo, $bar) = 3, 5; [$bar, $foo]
14:23 evalbot_10661 [5, 3]
14:24 b_jonas ?eval my $foo, $bar = 3, 5; [$bar, $foo]
14:24 evalbot_10661 Error: Undeclared variable: "$bar"
14:24 buetow ?eval my ($foo, $bar); $foo, $bar = 3, 5; [$bar, $foo]
14:24 evalbot_10661 [3, undef]
14:25 b_jonas ?eval my $foo = 3, 5; $foo
14:25 evalbot_10661 \3
14:25 svnbot6 r10662 | audreyt++ | * -CPIR: catch up the VType change by treating it as VUndef (for now)
14:25 b_jonas ?eval my $foo = (3, 5); $foo
14:25 evalbot_10661 [3, 5]
14:26 b_jonas I wonder how'd it get \3
14:26 buetow Whats \3 ?
14:26 b_jonas the eval said that. I guess it's a ref-to-scalar
14:26 buetow Reference to 3?
14:26 buetow Ah, ok
14:27 fglock joined perl6
14:28 audreyt means it's actually bound to 3 at lvalue
14:29 audreyt instead of a 3 at rvalue
14:29 b_jonas ?eval ref(\3)
14:29 evalbot_10661 ::Int
14:30 fglock audreyt, TimToady: I need to know in which city will you be right before YAPC::SA, so that we can arrange the plane tickets with $sponsor
14:30 audreyt Taipei.
14:30 b_jonas I don't quite understand that.
14:30 fglock ok
14:31 svnbot6 r10663 | audreyt++ | * another round of triaging...
14:32 b_jonas ?eval my ($foo = 3), 5; $foo
14:32 evalbot_10661 is now known as evalbot_10662
14:32 evalbot_10662 \undef
14:32 b_jonas ?eval (my $foo = 3), 5; $foo
14:32 evalbot_10662 \3
14:32 b_jonas ?eval my $foo = 3; $foo
14:32 evalbot_10662 \3
14:33 b_jonas I don't understand this.
14:33 b_jonas ?eval my $foo = 3; ~$foo
14:33 evalbot_10662 "3"
14:37 svnbot6 r10664 | audreyt++ | * more triaging and definedness.
14:37 audreyt which one do you not understand?
14:40 b_jonas the '?eval my $foo = 3; $foo' one. why does it print \3
14:40 b_jonas not 3
14:40 svnbot6 r10665 | fglock++ | rm empty dir Pugs-Grammar-Perl6-Lib
14:40 svnbot6 r10666 | audreyt++ | * intentional regression: sacrifice the macro expansion of
14:40 svnbot6 r10666 | audreyt++ |     state $x = 42;
14:40 svnbot6 r10666 | audreyt++ |   into
14:40 svnbot6 r10666 | audreyt++ |     state $x; FIRST { $x = 42 };
14:40 svnbot6 r10666 | audreyt++ |   because currently the declarator cannot look beyond the signature
14:40 svnbot6 r10666 | audreyt++ |   it takes.  Suggestions welcome...
14:47 audreyt preflight in ~20 minutes.
14:47 svnbot6 r10667 | audreyt++ | * TODOize assign.t.
14:48 audreyt b_jonas: because $foo is bound to a rw \3, so to speak
14:48 audreyt actually the \ is not the correct notation either, should be something like Scalar(3)
14:48 audreyt or even Scalar(Int(3))
14:49 Qiang joined perl6
14:56 svnbot6 r10668 | audreyt++ | * because "toplevel" in Parser now has an implicit block,
14:56 svnbot6 r10668 | audreyt++ |   extend the inTopLevel test for OUTER:: generation by one level.
14:56 svnbot6 r10669 | audreyt++ | * $a[$y xx $z] is not a scalar lhs, unlike $a[$y x $z].
14:58 Teratogex joined perl6
14:59 svnbot6 r10670 | audreyt++ | * final triaging and TODOizing. ready for preflight!
14:59 Teratogex is Perl 6 going to be certified to (at least theoretically) be approved for use in critical systems like nuclear reactor control software and avionics and such?
15:01 elmex joined perl6
15:04 bpederse joined perl6
15:04 ushi joined perl6
15:05 b_jonas certified?
15:05 b_jonas what do you mean certified?
15:06 stevan Teratogex: I *highly* doubt that
15:07 stevan there are very strict real-time requirements for any avionics work, and I highly doubt Perl 6 will be able to meet them
15:07 stevan maybe a subset of Perl 6 could be used in some way
15:07 stevan but you would have to write your own compiler for it
15:07 tifo \bye
15:08 particle_ or maybe an implementation of perl6 in ada, if someone were to write it
15:08 cognominal so far most language implementation come with a disclaimer that says too bad if you use it in the real world whatever the effect intended or not
15:08 rafl joined perl6
15:08 stevan particle_: :)
15:09 stevan even with Ada, they sometimes have to use a strict subset of the language for certain highly critical tasks
15:09 b_jonas cofnominal: it's not a question of certification but of warranty
15:09 stevan ada is a BIG language though
15:09 cognominal we should start by forbidding any natural language that allow to formulate bad thoughts
15:10 cognominal and certify the natural language used to define Perl6
15:10 particle_ without bad thoughts there would be no perl 5! ;)
15:10 stevan b_jonas: no, the FAA has a set of test that a system must pass
15:10 b_jonas does it?
15:10 stevan b_jonas: the real-time tests the FAA has are insane
15:10 * stevan had a link at one point to information on it
15:10 stevan you could probably google it
15:10 stevan they have a vested interest in keeping planes in the sky ;)
15:12 kolibrie and when a pilot has about two-seconds to make a decision, all the instruments better be working right
15:13 stevan b_jonas: see this link - http://www.lynuxworks.com/solutions/milaero/do-178b.php3
15:14 stevan seems to me that anyone who is selling a product which is to be used in avionics needs to have this certification
15:14 stevan or somesuch
15:15 Juerd We're not selling :)
15:15 stevan Juerd: true
15:15 stevan then they company writing the software needs to be certified
15:16 stevan if they use Perl6/Parrot, they must make Perl6/Parrot pass the certification
15:23 svnbot6 r10671 | audreyt++ | * final round of unTODO and TODO before preflight #0
15:25 Juerd Because of certification, open source is always a problem for these branches of business
15:25 Juerd Some try to get oss certified, some even manage to get through it in one piece
15:25 Juerd I personally think certification, almost all of it, should be done by governments, and for free.
15:26 Juerd But that's not a realistic view for the time Perl 6 is released
15:26 nothingmuch joined perl6
15:26 jsiracusa joined perl6
15:29 svnbot6 r10672 | audreyt++ | * bump version to 6.2.12; getting ready for preflight.
15:34 * gaal meows
15:34 gaal have fun with the release everybody, I'm loaded with $work...
15:36 audreyt poor moose
15:36 audreyt I've been consciously ignoring $work...
15:36 audreyt ...and plans to catch up on plane
15:40 audreyt uploading
15:40 svnbot6 r10673 | audreyt++ | * don't include third-party Judy in MANIFEST for this release.
15:49 audreyt http://pugs.blogs.com/pre0/Perl6-Pugs-6.2.12.tar.gz
15:51 audreyt SIZE=3550098 SHA1=402a344b0187e69d2e597992e6d910c316675dbe
15:52 audreyt changes topic to: Plese Please test pre0: http://pugs.blogs.com/pre0/Perl6-Pugs-6.2.12.tar.gz | paste: http://sial.org/pbot/perl6 | pugs.blogs.com | pugscode.org | pugs.kwiki.org
15:52 audreyt changes topic to: Please test pre0: http://pugs.blogs.com/pre0/Perl6-Pugs-6.2.12.tar.gz | paste: http://sial.org/pbot/perl6 | pugs.blogs.com | pugscode.org | pugs.kwiki.org
15:52 FurnaceBoy joined perl6
15:52 theorbtwo joined perl6
15:54 qwr left perl6
15:57 audreyt unfortunately, I can't stay up anymore. will check backlog as soon as I wake up
15:58 audreyt some help at changelogging would be most appreciated :)
15:58 * audreyt waves &
15:58 FurnaceBoy n8
16:00 rindolf joined perl6
16:00 rindolf Hi all.
16:11 macroron joined perl6
16:17 iblechbot joined perl6
16:21 iew0tja joined perl6
16:50 chris2 joined perl6
16:59 tifo joined perl6
17:10 buetow joined perl6
17:33 bernhard joined perl6
17:37 hexmode joined perl6
17:38 mj41_ joined perl6
17:46 clkao audreyt: sure, remind me again later
17:58 rindolf Hi clkao
18:41 FurnaceBoy_ joined perl6
19:02 iblechbot joined perl6
19:09 SamB joined perl6
19:12 dduncan joined perl6
19:18 larsen joined perl6
19:33 dduncan I have now downloaded the preflight and am running the makefile ...
19:36 dduncan now running 'make'
19:40 vel joined perl6
19:42 frederico joined perl6
19:46 baest joined perl6
20:15 dduncan make done, now starting smoke
20:18 dduncan note that this smoke has concurrent=4, on my 2-cpu machine, so both cpus are maxed, and speed hits from blocking should be lessened
20:18 dduncan still lots of / enough ram free beyond that
20:19 dduncan so we'll see if that helps with my past smoke speed problems
20:24 baest joined perl6
20:50 svnbot6 r10674 | fglock++ | Pugs::Compiler::Rule - added 2 SKIP'ped tests for parsing errors in t/04-rule.t
20:53 dduncan this is probably known, but I observe that the changelog in the preflight is not ready to go on cpan ... but pre0 is probably just for testing the code anyway
20:57 pasteling joined perl6
21:05 nothingmuch joined perl6
21:08 mj41_ joined perl6
21:15 svnbot6 r10675 | fglock++ | ChangeLog - removed XXX note in pX section
21:28 fglock left perl6
21:34 leo audreyt: make test results of  10673
21:34 leo (linux/x86)
21:34 leo Failed Test  Stat Wstat Total Fail  Failed  List of Failed
21:34 leo -------------------------------------------------------------------------------
21:34 leo t/oo/proxy.t               18    6  33.33%  4 6 9 13 15 18
21:34 leo I have these since quite a time - nothing else now
21:46 stennie joined perl6
21:51 iblechbot joined perl6
21:56 kanru joined perl6
22:49 dduncan smoke complete
22:49 dduncan looks like a 2.5 hour runtime
22:53 dduncan my smoke reports all okay except 1 failure at the end of t/pugsrun/07-dash-uppercase-v.t
23:01 dduncan I don't know what the problem is, though, since when I run t/pugsrun/07-dash-uppercase-v.t directly, it passes all 59 tests
23:01 dduncan maybe the harness didn't receive the 59th ok ? ... it says 'stub' in the smoke report
23:01 dduncan is this test failing for anyone else?
23:09 froh-doh joined perl6
23:21 dduncan left perl6

| Channels | #perl6 index | Today | | Search | Google Search | Plain-Text | summary

Perl 6 | Reference Documentation | Rakudo