Check your feather email | spec.pugscode.org | paste: sial.org/pbot/perl6 | pugs.blogs.com | www.treehugger.com/files/th_images/paradigm.jpg [set by audreyt on 2006-08-29 05:26:57 -0700]
Set by Akwa|user on 14 September 2006.
putter @tell fglock Sorry for getting back so late. And that it was still scary. ;) I'll do more cleanup, just for practice. Thanks for motivating it! :) Hmm... why do the new greedy_{plus,star} sketches take min and max arguments? I'm puzzled. Abstractly at least, there are simply two range algorithms, greedy and non-greedy, x**{m..n} and x**{m..n}?. And six aliases: x* being x**{0..Inf}, x+ being x**{1..Inf}, x? being x**{0..1}, plus th 00:30
lambdabot Consider it noted.
putter 01:51
putter wonders whether to restore (accidentally?) deleted regex tests... 01:52
@tell fglock ...plus the three non-greedy versions. Or am I missing something? I've also some questions about the code itself, but that can wait. I'll try to come by tomorrow. Cheers. 01:53
lambdabot Consider it noted.
pasteling "putter" at 66.30.119.55 pasted "the new range() code, hopefully perspicuous. comments encouraged." (82 lines, 2.2K) at sial.org/pbot/20016 01:59
putter committed as r13707. 02:00
good night & 02:01
ajs_home yawn... still awake for some reason 03:07
Been looking over PGE/TGE... interesting stuff... still getting my head around the meaning of each of the passes. 03:08
I wonder, was it the intention that P6-in-P6-on-Parrot would use PIL or a TGE-based PAST/POST? 03:10
[particle] ajs: we're working towards a common past :) 03:44
think of parse trees and post as tree-ified representations of high-level language source and virtual machine opcodes, respectively 03:45
past is the boundary layer seperating them.
it is a representation that is both language and vm neutral
jdv79 what is PIL in that world then? 04:36
there's too many "representations" 04:37
audreyt PAST and POST is basically the Parrot's counterpart to Pugs's PIL and PIR trees 04:49
lambdabot audreyt: You have 2 new messages. '/msg lambdabot @messages' to read them.
audreyt they are exactly equivalent in intent and purpose 04:50
except the concrete realization is very much different 04:51
TreyHarris is svnbot gone for good? 05:06
audreyt no... restarting it now 05:19
pasting my most recent changelog :)
* Unicode support for the embedded Perl5 bridge. (This takes effect iff
the embedded Perl supports SvUTF8_on, typically Perl 5.8+.)
- Source code in eval('...', :lang<perl5>) can now contain non-ASCII
characters; they are treated as if "use utf8;" is in effect.
- Str objects passing from Pugs land into Perl 5 are always turned
into SvUTF8_on unicode strings. 05:20
- Strings passed from Perl 5 to Pugs land are always converted into
Unicode Str objects for now; this may be relaxed once we have a
native Buf type in Pugs land.
- Both named and unnamed regexes now assumes Unicode semantics.
* Use New()/Safefree() instead of malloc()/free() in p5embed APIs.
* Use svpvn and CStringLen instead of svpv and CString, so strings with
embedded nulls can be safely passed either way.
* In non-embedded builds, call binmode(STDIN, ':utf8') only on
Perl 5.7 or later.
audreyt it's back... now summoning evalbot 05:24
agentzh hey 05:34
lambdabot agentzh: You have 1 new message. '/msg lambdabot @messages' to read it.
agentzh @messages
lambdabot lanny said 14h 2m 57s ago: Not sure if this should be directed to you but Synopsis seem to be different between latest smokes on m19s28.vlinux.de and feather.
agentzh lambdabot: yes, i've just looked at the synopses at these two places and found that the ones on feather are newer. 05:35
s/lambdabot/lanny/
christopher, iblech: does cron update the synopses every hour on the smoker server? 05:36
oh, i think a p6c mail is more appropriate for such an issue. :) 05:37
skew audreyt: I've put a BSD license on the typechecker code, and made a pdf of typing rules. 05:47
agentzh @tell lanny i've sent a mail to p6c. hopefully iblech and malon will fix the smoke server soon. thanks for the catch. :) 05:55
lambdabot Consider it noted.
agentzh ingy: just to confirm: ingy was iblech? :) 05:58
christopher agentzh: nope. 06:28
agentzh christopher: nope? 06:29
christopher agentzh: they are different people, afaik.
agentzh okay. 06:30
agentzh how often does the smoke server update the synopses? 06:30
christopher iblech says every three hours. 06:31
how old is the revision you cited?
Hmm, it looks like there might be a path problem again 06:33
agentzh about 1 day after.
christopher or maybe the partition is full again so smoke-production isn't running.
agentzh ah
christopher There's his bot... 06:33
agentzh iblechbot: please get iblech to look at my p6c mail and fix the smoke server... 06:34
:) 06:35
pasteling "masak" at 130.238.83.176 pasted "I got five "undefined reference" errors when running make on pugs just now... it's nothing critical, just FYI" (13 lines, 1K) at sial.org/pbot/20019 07:18
agentzh & 07:39
audreyt masak: try rm -rf dist/build 10:59
lambdabot audreyt: You have 3 new messages. '/msg lambdabot @messages' to read them.
audreyt lambdabot: @messages
lambdabot putter said 19h 26m 22s ago: r13604 "* splatty_in_caller_position.t: It's f(|@args), not f(*@args)." looks like it deleted 15 t/regex test files. was that intentional?
nothingmuch said 13h 6m 34s ago: do you have a recursive ddrescue script from your recent "event"? my HD is occasionally blocking the USB/scsi bridge and i need to powercycle and remount it, so
ddrescue would be *very* useful
skew said 4h 29m 14s ago: I've typeset some typing rules, and put a BSD license on the code. www.ugcs.caltech.edu/~brandon/index.html
audreyt @tell putter the deletion was quasi-intentional; those were disabled parts of tests from Perl6::Rules (entirely disabled for a long time now) and of dubious applicability; if you'd like to reanimate and review them that'd be cool! 11:00
lambdabot Consider it noted.
nothingmuch audreyt: i managed with rsync + loop, so need for ddrescuue stuff 11:18
lambdabot nothingmuch: You have 1 new message. '/msg lambdabot @messages' to read it.
nothingmuch oh! i thought it worked on files only
heh
ddrescue [options] infile outfile [logfile]
that's misleading =/ 11:19
audreyt indeed 11:22
and logfile gets you arrbitrary resume
nothingmuch nevermoose
*nodnod*
nothingmuch is almost done with copying by now
i suspect only an hour
i've never felt so evil
shutting off a disk to generate IO errors
instead of a bus-level block
gaal audreyt: who's maitaining Module::Install now? 12:00
audreyt Alias 12:01
we hand out metacommitter bits like crazy though.
gaal thanks
no thanks :-p
audreyt too late :) 12:02
gaal hmm, that makes you a hypercommitter?
audreyt and you a esccommitter...
gaal bracketed committer 12:03
committed bracketeer?
reverse "commit", "ask questions" => 12:04
1. any problem adding 'mit' to inferrable licenses?
audreyt no prob at all, though I prefer uppercase "MIT" as it's abbrev 12:05
gaal 2. any problem supressing warnings in all_from etc. if a metadatum was already given explicitly?
ACK
audreyt no prob at all, surprised it didn't do that already; very surprised 12:06
gaal does serverside cpan know what to do with MIT as license?
1.03 doesn't. will fix.
gaal er 12:06
1.03? where'd I get that number from?
audreyt META.yml is not yet parsed or recognized by search.cpan.org 12:07
I mean MIT field
prod gbarr, perhaps
PAUSE doesn't care either way
gaal so it appears verbatim?
audreyt yup
gaal that's fine
i mean, it's not b0rky. will ask gbarr.
audreyt sleeps... one more $job day to go... 12:53
Juerd ?eval item "foo" xx 5 13:16
evalbot_r13710 ("foo", "foo", "foo", "foo", "foo")
Juerd ?eval item("foo" xx 5)
evalbot_r13710 ("foo", "foo", "foo", "foo", "foo")
Juerd Context is broken :)
?eval $foo = "foo" xx 5; $foo
evalbot_r13710 Error: Undeclared variable: "$foo"
Juerd ?eval my $foo = "foo" xx 5; $foo
evalbot_r13710 ["foo", "foo", "foo", "foo", "foo"] 13:17
Juerd I think xx in list context should croak.
Ehhh 13:18
I think xx in ITEM context should croak.
xx is list repetition. It makes no sense in non-list context. If someone uses xx in non-list context, they could be meaning one of three things:
1. They really wanted to use x, but don't understand the difference yet. If they only stringify the array, they will find that somehow the repetition operator added whitespace. 13:19
2. They needed to use an @array, not a $string, but don't understand the difference yet. They won't find out the difference :)
3. They really did want an array. In this case, I think they needed to be explicit about it, and use []. 13:20
To summarize:
* \W operators should not create arrays in item context. If there's no good item context behaviour, croak. 13:21
(And this includes comma...) 13:22
putter audreyt: re quasi-intentional and reanimation, ok. deleting tests because they aren't working scares me. ;) 14:39
lambdabot putter: You have 1 new message. '/msg lambdabot @messages' to read it.
putter I'll leave them deleted for this releng, as I wont have time for review before then. 14:41
putter Actually, deleting tests because they _are_ working scares me too. 14:42
maybe we could tuck a directory somewhere for old disabled veteran tests? not sure it's worth it. 14:46
vel just to let you know that jonathan let me know that i can let you know that binaries for pugs 6.2.12 on win32 are available from his site (jnthn.net/perl6/) 14:48
lambdabot Title: Jonathan Worthington :: Perl 6 / Parrot
gaal knowledge is moose 14:54
putter is anything not-moose? :)
gaal putter: mu, sss.
putter: also, 'svn cat -rSOMEOLD' is like a directory for old veteran tests 14:55
svnbot6 r13711 | audreyt++ | * INSTALL: Update Jonathan's bindist URL (thanks to vel++),
r13711 | audreyt++ | and update some make targets too.
putter gaal: re -rSOMEOLD, true, though I don't have any tools which show "shaddows of the past" when looking at the present, so at least for me, out of HEAD is gone from mind. 15:01
putter @tell fglock I'll be gone until the weekend. fyi. have -Ofun. :) 15:08
lambdabot Consider it noted.
putter & 15:09
audreyt clkao: are you still going to sao paolo? 15:12
clkao mybe not. it's bloody far away 15:15
svnbot6 r13712 | iblech++ | * Sync smokeserver with the version running on the server (one-byte-change: for 15:47
r13712 | iblech++ | debugging purposes, I commented the system("rm -rf $path_to_synopses_dir")).
r13712 | iblech++ | * Add rel="nofollow" attributes to the links pointing to the smoke reports and
r13712 | iblech++ | the synopses, as these pages are highly temporal. (Googlebot gets many 404s.)
r13712 | iblech++ | * According to svn's error messages in the error.log, the reason for the broken
r13712 | iblech++ | SYN links (empty directories) is/was most likely Openfoundry's downtime, i.e.
r13712 | iblech++ | not smokeserver's fault. :)
cdpruden the Config::Tiny in ext/ still uses 'ref' -- do most people who need something along those lines just use the version from tpe.freepan.org/repos/adamk/Config-...ig/Tiny.pm ? 15:56
lambdabot tinyurl.com/phraw
cdpruden and, following that, does anyone have a reference to what the replacement for 'ref' is?
(been out of the loop for a bit) 15:57
audreyt WHAT is it. 15:58
I mean, the answer is WHAT.
rodi hehe
audreyt er, you're asking for WHAT 15:59
rodi who's on first?
cdpruden thanks!
audreyt :) 16:00
please commit./fix away
cdpruden will try :)
audreyt I've been working from bottom of smoke matrix up, and ext/ is so far neglected
cdpruden is that in a synopsis yet?
audreyt S12 16:01
WHAT/WHICH/WHO/WHERE/HOW are specced
cdpruden oh, guess that would make sense.... have to kick the google search from feather for giving me 5 and 10 :)
audreyt and because perl6 is timeless, it doesn't have a concept of WHEN (at this time)... 16:03
Juerd WHEN could be $^T at creation. But that's a terrible waste of bytes.
rodi and because nobody else thought it was funny, WHENCE and WHITHER haven't gained any steam...
Juerd More practically, it could be __FILE__ and __LINE__ of creation. But then, that's still a terrible waste of bytes when you're not debugging. 16:04
theorbtwo Could store them iff there's a debug flag set, but probably not worth it.
audreyt rodi: I think it's funny, but WHEREFORE takes the cake on length 16:05
svnbot6 r13713 | lanny++ | * S29 smartlinks
r13713 | lanny++ | * S29/Container tests
[particle] IDONTKNOW should always return 'Third Base.' 16:07
audreyt but the uppercase-english game already had Intercal's "PLEASE DO NOT GIVE UP" as prior art 16:08
not sure we want to compete on that front
fglock putter: re simply two range algorithms: that's for backwards api compatibility - it really only needs two implementations
lambdabot fglock: You have 3 new messages. '/msg lambdabot @messages' to read them.
svnbot6 r13714 | cdpruden++ | Switch Config::Tiny from using ref to using WHAT.
TreyHarris audreyt: .WHEREFORE is the POD documentation associated with a variable 16:13
oh, that's already .WHY.... drat. 16:15
audreyt they are synonyms anyway...
so maybe .WHY is the summary as it's more brief ;) 16:16
rodi nice.
TreyHarris we need C<your> variables now, and tense (do we have indicative verbs yet?), and then we can start to work on goverment and binding theory for perl 6 :) 16:18
audreyt actually I was about to suggest "is yours"
to replace "is export" when obra asked it the other day
but it's reeally, really silly.
TreyHarris giggles
yes :)
is yours for(:DEFAULT, :MYTAG) 16:19
audreyt is DEFAULT's, is MYTAG's 16:20
TreyHarris nods
better
audreyt utter silliness, though :)
audreyt files that under the Acme:: cabinet
TreyHarris chuckles
rodi files that under having a little /too/ much fun subverting the dominant paradigm... 16:21
audreyt in other words, that belongs to the -Opun track of development... 16:22
TreyHarris I have a silly Haskell question... when you get the type of something from ghci, sometimes it uses 'a, b' and sometimes it uses 'bx, cx' as type variables. what is the difference it's articulating?
audreyt it's remembering which tyvars the original declaration used 16:23
in the typeclasses/functions/etc
otherwise serves no purpose
TreyHarris ah, ok. so if i actually declared a type and used, say, 'q', it would always display 'q'? 16:24
audreyt yup.
TreyHarris k. thanks!
audreyt for a language without builtin named arguments
that is sometimes useful
audreyt s/arguments/parameters/ 16:25
TreyHarris nods 16:26
svnbot6 r13715 | lanny++ | * added svn:ignore to some ext/* dirs missing it 16:30
r13716 | audreyt++ | * range.t: fix incorrect plan 16:36
gaal and in errors, when are a1, a2, used? and when t1 etc.? 17:29
masak audreyt: that seems to have fixed it. thx 18:44
ajs_work I'm constantly leafing through the files in docs/Perl6/Spec, and still I get pretty badly confused by the names. Would it be wrong of me to change them to S01_Overview.pod, S02_Syntax.pod, etc., or is that going to throw something off (I know I'd have to update the "update" script too)? 19:19
command_ hihi 19:51
yts hello~ 19:56
count_leto hey. 20:10
TreyHarris ajs_work: update, and smartlinks, and smokeserv... you're probably best off just grepping through the pugs tree looking for /S\d\d[.](?:pod|html/ and /(:?Overview|Syntax|Subroutine|Rule|...)[.](:?pod|html)/ to see what else might use it... 20:16
SimplySeth would be cool if dev.perl.org/perl6/ were one of the URLs in the topic 20:28
lambdabot Title: Perl 6 - perl6
wolverian spec.pugscode.org is nicer :) 20:30
SimplySeth *goes to look 20:31
yts how could i join Pugs group 20:56
avar yts: what do you mean? 21:04
[particle] yts: the only things you need to join are 1) interest, and 2) an email address. 21:05
yts i have learned Haskell, i want to develop pugs
avar:i have learned Haskell, i want to develop pugs 21:06
[particle] yts: if you /msg audtreyt a note of interest and <your email address> she'll give you a commit bit 21:07
audreyt, that is
schroeer hello?! my first visit to irc. 21:57
I am having a problem to build pugs on Linux:
The things in the third-party directory 21:58
are built to dist/build/cbits/cbits/ instead of dist/build/cbits/ 21:59
and then the build system doesn't find the anymore.
them anymore. 22:00
Goodbye 22:01
zptao granny tranny yt 23:25

Powered by Cro and the Raku® Programming Language.

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

Thank you!