Please test pre0: pugs.blogs.com/pre0/Perl6-Pugs-6.2.12.tar.gz | paste: sial.org/pbot/perl6 | pugs.blogs.com | pugscode.org | pugs.kwiki.org
Set by audreyt on 12 June 2006.
00:05 jsiracusa joined
dduncan er, actually, the resmoke ended an hour ago ... 00:10
and this resmoke of the exact same build (no 'make' was repeated) showed zero errors, ... 00:11
this pretty much proves the 1 error before was transient
fishb0t what's the current state of the PIR output by pugs? 00:13
If I try to use the head parrot as a backend, I get IMCC errors... 00:14
00:55 hikozaemon joined 01:06 dudleyf joined 01:13 dudleyf left 01:28 shachaf joined 01:29 ajs_home joined
nothingmuch is anyone here from the new york area (or willing to travel a bit) interested in a 2-3 day hike around July 8th-11th (after YAPC)? 02:23
02:45 putter joined 02:47 shachaf joined
pasteling "putter" at 66.30.119.55 pasted "r10685 make test. smoke also uploaded. 51 regressions in 10 files." (18 lines, 1.1K) at sial.org/pbot/17723 02:50
putter PerlJam, etal: earlier, I shouldn't have suggested a jump straight to "what to do" about it (p6 expectation management), as further brainstorming what the problem is could be a useful prelude. ah well. 02:54
steven: moose came up in a yapc prep talk on inside-out objects. but in a rather fuzzy. if you have an inside-out story you would like broadcast, you might talk with Golden before his talk. 03:02
night &
03:06 drbean_ joined
stevan putter: I have talked with golden before about this stuff 03:35
03:35 mako132_ joined
stevan we are planning to talk at YAPC about how to improve Moose/Class::MOP's inside-out capabilities 03:36
03:51 unobe joined
unobe i 03:52
've figured out that to use PGE with pugs, I need to have one of two directories: 03:53
1) either the subdirectory src/pge with run_pge.pir in it (under the build directory), or 03:54
2) /System/Library/Perl6/darwin-thread-multi-2level/CORE/pugs/pge/ with run_pge.pir in it
03:55 lambdabot joined
unobe I think this is because of src/Pugs/Prim/Match.hs (lines 20-21). BTW, I'm running os x 03:56
so if you delete the build directory after you build pugs, then PGE rules won't work in pugs :-( 03:57
so just make sure to copy run_pge.pir into CORE/pugs/pge under whatever directory pugs -V gives as the installarchlib 04:03
04:04 unobe left 04:11 sahadev2 joined 04:14 rawr joined 04:39 dduncan joined 05:14 buetow joined 06:27 mtve joined 06:41 Aankhen`` joined 07:19 bjoern_ joined 07:25 _cnhackT_ joined 07:26 _cnhackT_ left 07:27 xerox joined 07:28 xerox joined 07:44 iblechbot joined 07:55 kanru joined 08:05 dduncan left 08:08 xern joined 08:15 marmic joined 08:28 kane__ joined 08:37 lisppaste3 joined 08:42 elmex joined
svnbot6 r10686 | audreyt++ | * LICENSE/: Add MIT and LGPL-2.1 license text. 09:15
r10687 | audreyt++ | * MANIFEST.SKIP: We are not shipping TextRegexLazy for this release. 09:24
r10688 | audreyt++ | * "make ghci" now builds its own deps without requiring 09:27
r10688 | audreyt++ | "make pugs" to be executed first.
r10689 | audreyt++ | * Properly finalize parrot on embparrot because calling parrot_exit.
r10690 | audreyt++ | * minor explicit-import style fix
szbalint audreyt: (If you've got time for a discussion) I've been wondering about roles and classes: Is there a way to intercept method calls to a class (for example for logging purposes), while the mechanism for that being in a Role? 09:31
09:33 ludan joined
ludan ciao 09:33
audreyt szbalint: look at Moose.pm :) 09:36
(and irc.perl.org #moose)
moose allows for prehandlers
I don't think it's part of canonical perl6
but it looks sensible
szbalint oh :) 09:37
ta!
audreyt np :) 09:38
svnbot6 r10691 | audreyt++ | * System.FilePath - Synchronize from Distribution.Compat.FilePath. 09:48
r10692 | audreyt++ | * Ditto for Pugs.hs. 09:58
r10693 | audreyt++ | * t/oo/proxy.t: disable test due to incoherent test results. 10:04
r10694 | audreyt++ | * README: Update third-party module notice; add Rosetta, 10:07
r10694 | audreyt++ | WTemplate and Locale-KeyedText to non-perl-license
r10694 | audreyt++ | ext/ listing; incoporate wording fixes from Allison Randal
r10694 | audreyt++ | and The Perl Foundation; Pugs is now properly copyrighted
r10694 | audreyt++ | by the "Pugs Contributors" so that I can disclaim my personal
r10694 | audreyt++ | compilation right over it.
10:11 wilx joined
gaal hey all. still loaded with $work but just popping in to report a minor problem - running make triggers a relink, even when a pugs and libpugs already exist. so 'make ; make smoke' for example takes a few seconds longer than it should. 10:26
audreyt nod. it's in build_pugs.pl 10:43
prevent build_exe with some timestampy thing?
11:09 chris2 joined 11:11 neoesque joined
szbalint Is there a possibility though that prehandlers will be part of Perl6? 11:11
audreyt you mean "before" and "after"? 11:15
well, you can do that with .wrap
but I'm pretty sure I'd like it to be part of perl6-as-implemented
not sure about perl6-as-specced
stevan will meet timtoady in YAPC::NA hackathon
they'll hash it out :)
method foo is PRE { ... } 11:24
maybe
not sure
I wonder why subtype is renamed subset... 11:25
szbalint oh okay 11:31
:)
Something unrelated, there is a most likely outdated (but passing) test in t/oo/roles about instantiating roles, while S12 says roles cannot be directly instantiated. I take S12 is correct? 11:33
Juerd It may or may not be correct, but it is the leading spec, so the test is in error. 11:37
11:40 szbalint_ joined
audreyt right. 11:43
szbalint_ audreyt: btw, is there any estimate on when pugs will have runtime role mixin ability? 11:44
11:44 szbalint_ is now known as szbalint
audreyt after NPW :) 11:45
and before YAPC::NA hackathon is over
szbalint okay, nice :) 11:46
11:53 jsiracusa joined 11:56 weinig joined
obra audreyt: are you in .no? 12:12
12:12 frederico joined
audreyt obra: yup 12:18
in hotel now
writing slides
obra heh. I need to finish that tonight
audreyt me too
obra we're in room 420
audreyt oh. maybe I'll go visit
but I'm afraid that'll kill my slidemaking
obra I'm not there now
kaia may be
audreyt ah.
obra I'm getting dragged to a dinner at 19:00 12:19
audreyt when you going to head back?
metoo, but if ENOSLIDE then not going
obra will be in the hotel after that
audreyt 'k
obra I'm a little worried about it.
is this the dinner with the uio people?
audreyt I'm doubly worried
since I'm codegen'ing to Keynote
obra ?! why?
lambdabot Maybe you meant: . v
audreyt er, fun
obra :)
ok. cool
I'll want your takahashi.key ;)
audreyt but probably going to get contents written first 12:20
(what a concept)
obra :)
what's the dinner you might go to? (Trying to figure out if it's the same as mine)
audreyt same
7pm at subway station 12:21
obra if so, come and bring your laptop. it'll give me ammo to do mine too
audreyt mmmoose
my talk is all about moose now
scary
s/all/half/
obra neat! 12:22
12:25 iblechbot joined 12:32 hugues23 joined 13:03 lisppaste3 joined 13:11 kanru joined
wolverian Juerd, why doesn't feather have vim7? 13:13
svnbot6 r10695 | fglock++ | Pugs-Compiler-Rule/temp - partial rewrite of the backtracking engine 13:35
13:41 QtPlatypus joined 13:46 rzoz joined 14:08 vel joined
Juerd wolverian: Because it hasn't been dist-upgraded in a while 14:20
wolverian: I'm installing vim 7 for you now
14:20 MicSch joined
Juerd wolverian: Done 14:21
14:31 MicSch left
gaal half moose, half elk, half caribou 14:35
and half wapiti (I could go on and on)
audreyt :) 14:36
audreyt is in JIT slide making rush
release is looking good modulo chglogging
gaal sorry i couldn't be more helpful there :( 14:37
15:01 bjoern_ joined
stevan audreyt: please post your slides when you have a chance,.. I would love to see them :) 15:04
15:07 rashakil_ joined
gaal that Calvin and Hobbes strip has forever been appropriate... 15:13
("last minute panic")
gaal hasn't finished his slides either
does the ratchet stuff change relevant slides in the peek talk? I have to look at S05 again 15:15
15:22 christopher joined 15:24 bpederse joined
wolverian Juerd, thanks. is it ok if I upgrade it too if I see it hasn't been upgraded in a while? 15:27
15:36 buetow joined
Juerd wolverian: No. 15:36
wolverian Juerd, okay. 15:37
Juerd wolverian: If you upgrade individual packages, that's fine, because then you know exactly what to test.
But after a system upgrade, often stuff like mail gets broken
And it's routine for me to check that
wolverian makes sense.
thanks. :)
15:41 sahadev joined 15:50 prozessor13 joined 16:03 buubot joined 16:17 iblechbot joined 16:27 shachaf joined 16:43 chris2 joined 16:49 Robe joined 16:56 justatheory joined
sahadev hello, isn't the shift builtin fuction supposed to shift from the @*ARGS array? 16:57
$ pugs -e 'say shift || 10'
***
unexpected "|"
expecting term
at -e line 1, column 11
@*ARGS.shift works, though
lambdabot Unknown command, try @list
16:58 bjoern_ is now known as DEpwnsPL
nothingmuch ?eval 'shift' 17:04
17:04 evalbot_10662 is now known as evalbot_10699
evalbot_10699 "shift" 17:04
nothingmuch ?eval shift
evalbot_10699 Error: unexpected end of input expecting term
nothingmuch eh, used to perl -e
uhm, it looks like you'll have to specify it explicitlty
the parser is treating it as a unary operator (i think)
17:05 ludan joined
ludan ciao 17:05
sahadev so, is it supposed to be fixed eventually?
nothingmuch probably
sahadev ok. thanks.
nothingmuch ciao ludan 17:06
17:35 frederico joined 17:55 lisppaste3 joined
svnbot6 r10700 | fglock++ | Pugs-Compiler-Rule/temp - new backtracking engine 18:02
18:39 prozessor13 left, pbelau joined
pbelau hey, can someone tell me what an unstacked call is in the perl world ? 18:40
18:47 YetAnotherEric joined 18:55 macroron joined
spinclad could you say what it is elsewhere? 18:58
what's called an inlined call in C/C++ ? 18:59
*is it what is called ... 19:00
19:14 dduncan joined
pbelau spinclad, i'm not quite sure what an "unstacked" call would in C/C++ 19:24
i figured it out though 19:25
19:30 SamB joined
svnbot6 r10701 | Darren_Duncan++ | LICENSE/ : Updated the LGPL to the current version, which includes the FSF's current Franklin Street address rather than Temple Place ... anyone else using a FSF license should likewise indicate their current address in references 19:42
19:44 wilx joined
dduncan I am now starting a smoke of r10701 20:22
20:50 DEpwnsPL is now known as bjoern_
svnbot6 r10702 | audreyt++ | * add tomorrow's talk before sleep. about 30% finished... oh well. 20:57
gaal pbelau: in Perl 5, goto-&sub reuses the current stack frame. if you mean inlining a function there's no such general purpose thing, though the optimizer happens to fold some constants, e.g. "sub pi () {3.14999}".
pbelau: in Perl 6, I suppose you could use macros for some of these things. 20:58
audreyt szbalint: to answer your earlier question. the standard translation from Moose:
gaal good night and good luck, audrey
audreyt before 'foo' => sub { $self->bar };
is this perl 6:
gaal (talking in movie titles? oops)
audreyt method foo { $.bar; call }
so I think we don't actually need another trait for PRE/POST 20:59
the "call" makes explicit the fact
and you can even change the args by saying
gaal didn't there use to be one in s06?
audreyt call(...alternate args...)
gaal ISTR BEFORE and AFTER...?
ah, there's ENTER and LEAVE in s04 21:01
PRE and POST there are design-by-contractish things.
<L04/"Closure traits">
stevan audreyt: actually with Moose, you can attach a before/around/after method to a locally defined method as well 21:03
it tends to be more useful in augmenting supermethods, but when added by Roles, there comes situations where it might apply the modifier to the local method too 21:04
audreyt: I wanna see your slides :)
pleeeeeeeaaaasseeeee :) 21:05
I'll give you kharma if you do
stevan dangles the kharma karrot
stevan ponders the kharmic impact of abusing kharma in this way ... hmm 21:06
gaal stevan: r10702 21:07
stevan doh! 21:08
stevan-- # for not backlogging
see,.. thats the kharmic impact... i am doomed :P
gaal "I want to see your slides" is not nearly as wicked-sounding as "You wanna see my modules?"
stevan or even worse.. "do you wanna pet my monkey?" 21:09
that one always scared me
gaal audreyt: the <title> is weird 21:11
not that that matters in fullscreen mode 21:13
stevan can't get the xul to work 21:14
ingy i stevan 21:15
h
stevan heya ingy
where in the world are you now ingy?
gaal in the slide with "has overdraft_account" is the "o"-ish character at the end of the line intentional?
ingy Seattle 21:16
stevan ah,.. they let you back in ... good :)
ingy heh
stevan had to save the file to his desktop to get the xul to work,.. how odd...
ingy I'll have a motorcycle in Chicago
w00t
stevan nice
gaal will you be zentaining it? 21:17
stevan ingy: nothingmuch can be your biker bitch then
ingy yeah...
he can get in line :p
21:17 weinig joined
gaal hahaha, ingy, you should look at audrey's slides 21:17
ingy url?
ingy fears 21:18
stevan doesn't hate Ingy,.. even though he wrote Spiffy
gaal ingy: svk pull ; firefox docs/talks/deploy.xul
wolverian audreyt, s/reconcilling/reconciling/ on slide 20
ingy wolverian: you have commit :P
wolverian yeah, but I have a checkout in progress :) 21:19
only 9000 revisions to go
ingy yowch
I hate the way svk won't let you do two things at once
PerlJam ingy: complain to bestpractical.com and maybe they'll pay clkao to "fix" it. 21:20
:-)
ingy PerlJam: did cl move to Boston?
PerlJam no idea if there was any physical movement. 21:21
given the nature of the people involved, I wouldn't be surprised if there was none.
stevan especially considering that clkao write svk specifically so he could hack from the .tw mountains 21:24
(or at least that is one of his oft quoted reasons)
gaal good niht & 21:25
but with a g
stevan goog niht gaal :P 21:26
gaal stevang :)
dduncan wolverian, unless you need all the revisions, you could just have svk checkout the later ones ... eg, just those from The 6.2.11 release and later 21:36
much faster
having all is useful mainly if you want to have an archive 21:37
wolverian I like doing it the hard way
dduncan getting just specific revs is something like a -n switch
wolverian yeah. 21:38
dduncan or a -<some letter>
wolverian it's -n, where n is a number, in the interactive checkout
I'm doing that now, too, on another computer, so I can compare :)
dduncan interactive checkout not necessary for what I'm thinking of
szbalint audreyt: cool. 21:39
at 'before' :) 21:40
dduncan wolverian, now that I look at it, I think it is -r 21:41
during svk update
21:42 Qiang joined
PerlJam I often have trouble getting svk to just get the head revision for some reason. 21:42
dduncan no, its -s 21:43
PerlJam (or anything other than getting all revisions)
wolverian mm. well, it's getting the 10 most recent revisions now, but it's still taking ages. :)
(I don't know how svn stores things. going through 10k patches would take some time, I guess) 21:44
svnbot6 r10703 | fglock++ | Pugs-Compiler-Rule/temp - updated TODO list; code cleanup
dduncan the first revision always pulls the whole thing at that point, then the rest are quick
svn stores diffs
but the first revision is the whole thing
wolverian mmkay
ah, done.
dduncan wolverian, specifically what I've been trying to reference all this time is ... 21:45
under svk.elixus.org/?SVKUsage
lambdabot The title of that page is "SVKUsage - Kwiki"
dduncan heading "Mirroring"
If you don't want to sync every revision from the remote, you can add a -s REV option as the first revision to sync without previous changes.
that specifically is what I was using
and controls what is copied from the server 21:46
so its not -n or -r or such
and this is not interactive, not any more than the plain 'sync'
svnbot6 r10704 | wolverian++ | typo in deploy.xul
dduncan did you get all that? 21:47
dduncan back in an hour 21:48
wolverian dduncan, I was using co directly. 21:55
but thanks, that's useful.
21:57 froh-doh joined
stevan audreyt: you have probably seen this, but... www.hlvm.org/ 22:02
lambdabot The title of that page is "High Level Virtual Machine"
stevan thanks lambdabot
22:08 ludan joined 22:18 jsiracusa joined
svnbot6 r10705 | fglock++ | Pugs-Compiler-Rule/temp - positional and named captures are ok 22:22
wolverian by the way, I was accepted to my CS dept's student research programme, which is basically an extra curriculum of research-oriented lectures and meetings. we talked about pugs in the interview, it was fun, and I think it helped :) 22:28
22:37 bpederse left 22:43 bpederse joined
QtPlatypus wolverian: Congrats 22:53
wolverian QtPlatypus, thanks :)
23:41 penk left 23:47 bpederse left
dduncan smoke of 10701 completed some time ago ... uploaded 23:55
it has 65 errors, more than the pre0
but t/pugsrun/07-dash-uppercase-v.t is not one of the failures 23:56

Powered by Cro and the Raku® Programming Language.

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

Thank you!