Check your feather email | spec.pugscode.org | paste: sial.org/pbot/perl6 | pugs.blogs.com
Set by TreyHarris on 9 October 2006.
lambdabot TimToady said 3h 36m 31s ago: that is, we should have the warning, but it probably shouldn't apply to blank lines. 00:00
audreyt as in segfaulting it?
TimToady no, as in confusing the harness somehow
(I think)
00:00 lollan joined
TimToady however, if you remove the =cut it dies violently 00:01
audreyt after sleeping on it some I think lone =ut should indeed be fatal
=cut, even
TimToady dies as in segv 00:03
svnbot6 r14283 | audreyt++ | * Pugs.Parser.Literal: Proper warning (with expected indent,
r14283 | audreyt++ | the original line, and the line number) about inconsistent
r14283 | audreyt++ | q:to indentation line, and make an exception for blank lines.
audreyt 'k will check 00:05
workshere though
svnbot6 r14284 | audreyt++ | * sort.t: repair lone =cut
nothingmuch @tell Limbic_Region rain in israel is seasonal, i meant that it was the first rain of the year 00:07
lambdabot Consider it noted.
00:09 lumi_ joined
nothingmuch night folks 00:09
audreyt nite nothingmuch 00:16
audreyt preps for $work
00:21 nekokak joined
TimToady ?eval say ~"mĆøĆøse".perl.split(/<null>/) 00:24
00:24 evalbot_r14276 is now known as evalbot_r14284
evalbot_r14284 OUTPUT[" m ƃ Āø ƃ Āø s e " ] Bool::True 00:24
00:34 mjk joined 00:35 avarab is now known as avar
audreyt TimToady: see if the segfault still occur with r14385 00:51
$job &
svnbot6 r14285 | audreyt++ | * ext/ and t/: Repair various lone =cut. 00:52
r14286 | audreyt++ | * Pugs.Parser.Doc: Lone =cut is now fatal; also, tolerate
r14286 | audreyt++ | trailing spaces in POD directives.
01:08 vel joined
TimToady audreyt: got segfault first time I reran, but since then it doesn't fail. so appears to be intermittent, or dependent on something that prove6 cleared out (I've been going back and forth between running directly and using prove6). Failure is after test 17 with no other warning. 01:12
so does not get to: not ok 18 - parse trait on block closure
anyway, have to wander off myself, so should have new smoke in .5 hr. & 01:14
audreyt anyone on win32? sigbjorn would like tester for haskell.org/ghc/dist/6.6/ghc-6-6.msi 01:40
and I'd like to know if it builds pugs fine, too 01:41
TimToady: does tr[][] still work? 01:42
or is it always tr/// tr!!! now? 01:43
I don't think pseudoassignment works for tr[]=
trans.t currently tests for $japh ~~ tr(a-zA-Z)(n-za-mN-ZA-M); 01:45
01:46 christopher joined 01:50 Eimi joined 02:15 cognominal joined 02:17 kanru joined
audreyt TimToady: also, "s[...] = .moose" and "s[...] .= moose" behaving differently is a big weird 02:33
(the first talks abuout $_, the second talks about $())
s/big/bit/
but I'll keep it implemented that way for now 02:38
02:47 mako132_ joined 02:50 weinig|food is now known as weinig 02:54 weinig is now known as weinig|zZz
obra audreyt: you around? 02:59
03:09 lanny joined 03:10 kanru joined
lanny *: checked for instances of =begin/=end/=cut by grepping=cut and searching previous 7 lines for all t/*. Either someone handled this already and I missed the svnbot notice or sort.t was only example. (I based sort.t off another files example though so a bit confusing on that score.) 03:10
03:25 mdiep joined 03:32 SubStack joined
TimToady Don't see the problem. $foo = .moose and $foo .= moose means two different things as well. 03:51
we could use a tr[] = q[] syntax if we restrict the right side (for now) to q 03:52
but tr(a-zA-Z)(n-za-mN-ZA-M) is wrong two ways not counting any right-side changes. 03:53
tr[a..z A..Z][n..z a..m N..Z A..M] is more like. 03:54
but I'd be just as happy doing away with tr/// sugar entirely and making people write .trans(q[a..z A..Z], q[n..z a..m N..Z A..M]) or some such 03:57
but maybe we better keep the shell of it for old times' sake, even if we s:g/-/../ 03:58
04:00 mdiep joined 04:11 mdiep joined 04:16 justatheory joined
TimToady audreyt: I'm still getting the intermittent set-infinite segfault too with 6.6 04:18
svnbot6 r14287 | lwall++ | tr/// allows whitespace and uses .. for ranges. 04:20
04:34 mdiep joined 05:10 mdiep_ joined 05:15 BooK_ joined
svnbot6 r14288 | lwall++ | More tr/// whacking. Still need to make lexer not do qq for us though... 05:27
r14289 | lwall++ | forgot to change testplan 05:30
05:32 kanru joined 05:33 mdiep joined 05:44 iblechbot joined 05:48 buetow joined 05:59 agentzh joined
agentzh @messages 06:02
lambdabot audreyt said 5d 18h 20m 38s ago: excellent talk :)
lanny said 1d 10h 39m 54s ago: Feel free to shoot the messanger at any time cause I'm feeling like a broken record myself :) but when I look at latest smoke SYN and Perl 6 docs on feather they still
don't match up. (I use S29 as a test.)
agentzh audreyt: thanks :)
lanny: yes, indeed you're right. S29 on the smoke server still looks odd... 06:03
svnbot6 r14290 | audreyt++ | * split.t: Fix plan and unTODO. 06:06
agentzh iblech: would you please update your Pod::Simple::HTML from cpan and check whether S29 is *actually* up-to-date? the "Any" subsection right below the "Function Packages" section is missing in the smoke S29. weird.
agentzh doesn't have access to the smoke server so he can't effectively diagnose the problem. 06:08
06:10 marmic joined 06:14 lollan joined
svnbot6 r14291 | malon++ | Spec/update - add more synopses to %Spec hash 06:14
r14292 | lwall++ | Further tr/// fixups
TimToady @tell audreyt I fixed things up to allow tr/// with whitespace and .. but you might well find a better way to rewrite Literal.hs to make tr/// not run its args through qq 06:15
lambdabot Consider it noted.
06:15 lollan joined
audreyt sure 06:15
lambdabot audreyt: You have 1 new message. '/msg lambdabot @messages' to read it.
audreyt I implement s[] etc
TimToady you were probably fiddling with the same rule...
the rest of the tr processing is down in Prelude, but you can hoist some of that into the parser as well if you like. 06:16
audreyt @messages
lambdabot TimToady said 1m 36s ago: I fixed things up to allow tr/// with whitespace and .. but you might well find a better way to rewrite Literal.hs to make tr/// not run its args through qq
TimToady namely the removal of whitespace and qb// 06:17
obra audreyt: hi
TimToady hi obra, long time no see, just hear...
obra grins at TimToady. Indeed. 06:18
TimToady audreyt: any clues on the intermittent set-infinite coredump?
obra Pity I didn't talk myself into Brazil.
audreyt TimToady: yeah. another Jufy/unsafeEvalExp thing I think
but at $job atm
obra $job! /me swoons. be sure to gather lots of requirements for templating ;) 06:19
audreyt yeah :)
obra audreyt: I think we've found a hotsprings resort for the pdx hackathon. see your mail 06:21
audreyt obra: see your mail 06:22
obra woo
obra waits for filters
06:23 Daveman joined
svnbot6 r14293 | audreyt++ | * Implement "s[...]=..." and deprecate "s[...][...]". 06:26
obra decries his greylisting
svnbot6 r14294 | lwall++ | comb now works better 07:08
TimToady smoking r14294 07:09
07:14 takesako joined 07:19 zakharyas joined 07:28 agentzh left 07:29 elmex joined 07:30 theorbtwo joined 07:32 takesako joined
audreyt ?eval $_ = 'a b c d e f g'; s[\w] .= ord; $_ 07:55
07:55 evalbot_r14284 is now known as evalbot_r14294
evalbot_r14294 \"97 b c d e f g" 07:55
audreyt ?eval $_ = 'a b c d e f g'; s:g[\w] .= ord; $_ 07:56
evalbot_r14294 \"97 97 97 97 97 97 97"
audreyt hm, that looks wrong.
TimToady you have to delay the . as well... 07:58
svnbot6 r14295 | lwall++ | [inplace.t] It helps if you use P6 character class instead of P5...
audreyt fixed 08:02
svnbot6 r14296 | lwall++ | fix a couple of old s{}{} instances.
r14296 | lwall++ | comb potentially faster using lvalue substr instead of copying string.
r14297 | audreyt++ | * Pugs.Parser.Literal: Build .= expression correctly for s[].= 08:03
TimToady btw, the multi method dispatch of comb didn't successfully distinguish a Regex from a Str arg. That's why the current Prelude impl uses a switch. 08:04
audreyt hm. isolate it into a test of its own? 08:06
I'll look tonight
meanwhile, $job calls... &
TimToady well, bed calls &
audreyt sleep well! 08:11
TimToady I added a test to multi.t 08:12
It fails appropriately. :) 08:13
svnbot6 r14298 | lwall++ | [t/oo/methods/multi.t] Added test for misdispatch of Regex to Str
TimToady Now I shall fail appropriately...
zzz & 08:14
TimToady is sleepsmoking again... 08:15
08:32 nwc10 joined 08:33 nwc10 left 08:46 jferrero joined 08:49 andara joined
nothingmuch audreyt: is the new YAML::Syck bug a syck problem? 09:20
rt.cpan.org/Public/Bug/Display.html?id=22043]
lambdabot rt.cpan.org/Public/Bug/Display.html?id=22043]
nothingmuch eek, how did that bracket sneak in there
audreyt nothingmuch: close that ticket please? 09:24
nothingmuch i'm not sure how
nothingmuch doesn't know syck
audreyt the yaml spec says, pretty clearly, ["foo", "bar"] is fine but ["foo","bar"] not
nothingmuch where can i find docs?
audreyt so it's wontfix
nothingmuch ah
gaal also, is expect indented? 09:25
nothingmuch in the email i got i think it was
gaal ysh is putting it inside test
nothingmuch i think that the RT web interface is not indenting 09:26
gaal what's the rationale for the spec requiring whitespace tere?
nothingmuch aye, when viewing the source this is evident
audreyt gaal: I guess same reason why 09:27
foo's
mo"o"se
are all literals 09:28
as is
nothingmuch omigawds
i didn't know that
audreyt blah,blah,blah
demerphq hey audreyt, did you ever actually do anything about building pcre with perl? you mentioned something about it a while back...
gaal so why isn't ["foo","bar"] a singleton list with '"foo","bar"' its element? 09:29
audreyt demerphq: it's svn.openfoundry.org/pugs/perl5/re-override-PCRE/
gaal: well it would be if it's
[_"foo","bar"]
but an opening double quote also carries special meaning
lambdabot Title: Revision 14298: /perl5/re-override-PCRE, tinyurl.com/ewktx 09:30
audreyt I think the YAML team, in their json interop mode discussion, considered relaxing it
not sure about how it went though
09:30 ruoso joined
nothingmuch so this is really a bug in YAML ;-) 09:31
demerphq pokes around 09:32
audreyt well, parsing "foo","bar" as one single string
is surprising as well
so bailing out seems sensible to me
pasteling "nothingmuch" at 84.108.53.74 pasted "YAML bug" (22 lines, 465B) at sial.org/pbot/20376
nothingmuch is implicit concatenation legal?
that is "mooose"bar
== "moosebar" ? 09:33
demerphq how much of that is autogenerated code, and how much of it is you?
audreyt well YAML.pm is broken
nothingmuch both of them have a syntax error for "moose"bar
audreyt demerphq: pcre.i is me. pcre_wrap.c is genned from it 09:34
by swig
demerphq ok 09:35
gaal this is an "interesting" feature...
demerphq im a little unfamiliar with this stuff
umm, did it work? 09:37
nothingmuch i'm notsure how foo,bar is ever useful 09:40
for example: a sentance: the good, the bad, and the ugly
sentance in a list: [ the good, the bad, and the ugly ]
the second is ["the good", "the bad", "and the ugly"]
audreyt demerphq: it works for me 09:41
nothingmuch it's really confusing that at times part of a literal is a meaningful token
oh well
audreyt demerphq: you need to check out the pugs repo in its entirety
nothingmuch will write tests for perlsyck & YAML.pm
audreyt demerphq: then cd into perl5/re-override/, run makefile.pl and make install
demerphq: then cd into perl5/re-override-PCRE etc 09:42
the reason is that it copies the pcre sources from ../../src/pcre/
demerphq ok.
audreyt afterwards you can "make dist" from there
demerphq im just trying to figure some stuff out in what i can see. I see $^H{regcompp} = sub { ... 09:43
but i dont recall ever seeing $^H{regcompp} before. 09:44
audreyt there's much more info in re-override/
it's all putter's work essentially
all I did was some tidynig up
demerphq ahah.
now the dots are connecting.
right he hand hacked a perl 5.6 to experiment with right? 09:45
audreyt not sure of the details (and I need to decommute RSN) :) he'll be up in a few hours I think...
demerphq ok, thanks a lot. this is very interesting. 09:47
audreyt glad to hear that :)
demerphq parallel evolution or something.
perl 5.9.5 has something like what he did there in it.
audreyt please salvage any ideas and/or bits :)
demerphq im gunna try.
i dont have your compiler foo tho. 09:48
er, compiler-fu
audreyt I don't even know C...
gaal ha. 09:49
audreyt afk, bbl... & 09:50
10:00 baest joined 10:02 lumi joined 10:15 norageek2 joined 10:20 ivas joined 10:32 jferrero joined 11:09 chris2 joined 11:19 weinig|zZz is now known as weinig 11:25 kanru joined 11:31 ludan joined 11:33 b_jonas joined 11:43 mauke joined 11:47 weinig joined 11:58 xinming joined 12:11 markstos joined 12:14 Limbic_Region joined
Limbic_Region salutations all 12:15
lambdabot Limbic_Region: You have 1 new message. '/msg lambdabot @messages' to read it.
Limbic_Region @messages
lambdabot nothingmuch said 12h 7m 24s ago: rain in israel is seasonal, i meant that it was the first rain of the year
Limbic_Region @tell nothingmuch WRT rain I was being quite facetious
lambdabot Consider it noted.
nothingmuch rly? ;-)
lambdabot nothingmuch: You have 1 new message. '/msg lambdabot @messages' to read it.
nothingmuch @moosages 12:16
lambdabot Limbic_Region said 27s ago: WRT rain I was being quite facetious
b_jonas oh no, I've started a stupid perl6 flame on #perl 12:17
lambdabot b_jonas: You have 1 new message. '/msg lambdabot @messages' to read it.
12:42 b_jonas joined
stevan @tell audreyt nothingmuch's $work task for the next few weeks is now officially MO & Perl 6/5 meta stuff, please (ab)use him for your every meta-whim :) 12:48
lambdabot Consider it noted.
nothingmuch =) 12:49
12:49 elmex joined
andara hi all. Compiling pugs r14289 with GHC 6.6 fails on my linux box: 13:02
ghc-6.6: unknown package: base-1.0 (dependency of HsJudy-0.1)
Limbic_Region andara - hrm, anytime I see something wrong with Judy I rm third-party, svn up, and remake and it usually fixes it 13:12
drastic measure and likely unnecessary
assuming you know exactly what file in third-party that needs to be removed
Patterner doesn't base belong to ghc? 13:13
andara Limbic_Region: thanks, I can just rm -Rf third-party, no problem
Limbic_Region don't forget the svn up that goes with it 13:14
andara Limbic_Region: of course :-)
Limbic_Region: yup, that did it. Thanks a lot! 13:17
Limbic_Region andara - no worries 13:20
Patterner I just compiled r14298 with ghc 6.6 13:24
andara Patterner: me too, running smoke now ... 13:26
Limbic_Region has been using 6.6 for over a month now
of course I was building it from source on Win32 prior to the official release
but that's just because i'm a masochist 13:27
andara is a mere mortal ...
Limbic_Region andara - I'm probably more of a mortal
but you know the saying "the squeeky wheel gets the greese" 13:28
whine long enough to people with the capabilities to fix something and they will fix it just to get you to shut up
andara I'll keep that in mind *grin*
Patterner ponders ignoring andara :P 13:29
andara Limbic_Region: did you by any chance compile 6.6 on OS X as well?
Patterner my mac is too slow for that... 13:30
Limbic_Region nope but it was also done by another #perl6 person not too long ago
@tell audreyt ext/Set-Infinite/t/set-infinite-recurrence.t is still causing the windows level error "pugs has encountered a problem...."
lambdabot Consider it noted.
Limbic_Region ok, must wander of for $work - happy hacking all 13:31
what's the svn parameter to revert to a previous rev? 13:44
"revert" seems just to undo local changes 13:45
13:45 weinig is now known as weinig|bbl 13:47 elmex joined 13:52 hexmode joined 14:01 [particle] joined 14:10 vel joined 14:11 cjeris joined
gaal Limbic_Region: svn up -rREV 14:14
or svn cat -rREV file > file, if you want to merge off an old r 14:15
svnbot6 r14299 | andara++ | [runpugs] 14:33
r14299 | andara++ | -added info page, small change to front page
r14300 | lwall++ | [t/xx-uncategorized/unicode.t] added test for "m?\195?\184?\195?\184se".perl.chars == 7
TimToady heh, looks like svnbot is also unicode confused. that should read "mĆøĆøse".perl.chars == 7 14:44
b_jonas TimToady: so is this channel utf8? 14:45
TimToady @tell audreyt I've added a test for the "mĆøĆøse".perl problem 14:46
lambdabot Consider it noted.
14:46 iblechbot joined
b_jonas I wonder how I can set a per-channel encoding in xchat 14:46
TimToady do you see slashed o's?
b_jonas no, I see two-character sequences
because my irc encoding is set to 8859_2
I see something like an A with an u accent then nothing with a hook below 14:47
TimToady yes, that's utf8 encoded in 8859
b_jonas yes, exactly
TimToady maybe you can set up a window with utf-8 for unicodey channels. 14:48
b_jonas I dunno, I think the encoding is per-connection 14:49
TimToady most of the Linux terminal emulators will happily do utf-8 these days if LANG is set appropriately
b_jonas yes,
TimToady I'm not sure irssi even knows what it's passing.
b_jonas but this is xchat, so it's not in a terminal
wolverian xchat supports per-channel encodings, I think. (except the gnome xchat interface, which doesn't expost that functionality) 14:50
b_jonas I set the encoding in the connection settings for each server
wolverian s,expost,expose,
b_jonas wolverian: does it? how do I set it?
wolverian let me check.
b_jonas this is Xchat 2.4.1 btw
wolverian (installing xchat now.. I only had xchat-gnome :) 14:51
14:54 Eimi joined
wolverian b_jonas, er, sorry, can't find it. I suppose I was thinking of some other client 14:55
(use irssi ;)
b_jonas well, I sometimes use irssi anyway but for other reasons
but I don't really care because nonascii is rare on these channels 14:56
14:56 theorbtwo joined, andara left, chris2 joined
kolibrie we get occasional chinese, hebrew, and mĆøĆøse in this channel 14:57
}:)
wolverian and perl6.. 14:58
b_jonas yes, but I wouldn't understand chinese and hebrew anyawy
now perl6, yes, that's a problem
wolverian for @foo ā†’ $bar { ... }
(although I don't think that's specced :)
b_jonas double angles and yen signs are already in use 14:59
wolverian I think set operators, too 15:00
(though I don't think I can type those easily)
b_jonas are they? I didn't see those in S03
wolverian oh, I was thinking I've seen them in some tests. maybe it was examples.
b_jonas is there an ascii for them, like /\ and \/ and (+) ?
wolverian yes. something like that.
b_jonas the folks on the other channel think it's not possible in xchat either. 15:06
15:11 b_jonas joined 15:14 b_jonas is now known as b_jonas_, b_jonas_ is now known as b_jonas__ 15:15 b_jonas__ is now known as b_jonas
b_jonas ok, I did something 15:17
could you type moose in utf8 again now please
and, is this an e with an acute in utf8: Ć©
wolverian it's an e with a right-slanting accent 15:18
(I'm not sure what acute means)
b_jonas that's it
wolverian I'd type moose but I can't find the letter on my keyboard, apparently :) mĆøĆøse!
svnbot6 r14301 | audreyt++ | * t/builtins/io/dir.t: Use $FindBin::Bin instead of "."
wolverian found it!
svnbot6 r14301 | audreyt++ | to test dir.t, because on concurrent smoke, this is a
r14301 | audreyt++ | race condition against temp files created by other processes.
r14302 | audreyt++ | * Pugs.Prim: ".perl" now returns Unicode strings properly.
b_jonas nice, thanks 15:19
it works
now I have two different connections to freenode
one with utf8 and one with 8859_2
wolverian I suppse that works too :)
s,suppse,suppose,
b_jonas I just have to take care to join the right channels from the right connection
wolverian can you automate it in xchat? 15:20
I think it wants you to connect only once per network if you want to automate it.. same with irssi, I think
kolibrie irssi can connect multiple times to the same network 15:22
kolibrie has done it accidentally a few times
but perhaps not automatted 15:23
TimToady ?eval "mĆøĆøse".perl.say
b_jonas wolverian: yes
15:23 evalbot_r14294 is now known as evalbot_r14301
b_jonas I can automate it 15:23
evalbot_r14301 OUTPUT["mƃĀøƃĀøse" ] Bool::True
wolverian well, that's broken. 15:24
svnbot6 r14303 | audreyt++ | * Pugs.Eval: Allow "/.../" to match against $_ immediately
r14303 | audreyt++ | not only under (Bool $) context, but also (Num $), (Str $)
r14303 | audreyt++ | and () (aka void) contexts.
b_jonas maybe I should write an asciifying ircbot and connect it to #perl6
TimToady ?eval 42 15:26
evalbot_r14301 42
Limbic_Region is audreyt fixing things in her sleep again or is she around? 15:28
audreyt ping
TimToady ?eval "mĆøĆøse".perl.say 15:30
evalbot_r14301 OUTPUT["mƃĀøƃĀøse" ] Bool::True
TimToady ?eval die
15:30 babusri joined
evalbot_r14301 Error: Died 15:31
TimToady ?eval "mĆøĆøse".perl.say 15:32
evalbot_r14301 OUTPUT["mƃĀøƃĀøse" ] Bool::True
TimToady is there any way to kick evalbot to the next version from within evalbot?
15:32 babusri joined
audreyt it's still compiling 15:33
theorbtwo It should do it all by itself, though it won't reload until the next time somebody says something to it after it finishes recompiling.
TimToady how often does it recompile?
hmm, /tmp/ghc26651_0/ghc26651_0.hc:8:24: error: Parrot_hsc.h: No such file or directory 15:35
going for walk &
15:41 hexmode joined
svnbot6 r14304 | audreyt++ | * Pugs.Eval.Var: Infer (Syn "//" ...) as Regex, not Any, 15:45
r14304 | audreyt++ | thus closing the MMD bug TimToady++ pointed out.
audreyt ?eval "\x[7777,8888,9999]" 15:46
15:46 evalbot_r14301 is now known as evalbot_r14303
evalbot_r14303 "ē·č¢ˆé¦™" 15:46
audreyt ?eval "\x[7777,8888,9999]".perl
evalbot_r14303 "\"ē·č¢ˆé¦™\""
audreyt ?eval "\x[7777,8888,9999]".perl.say
evalbot_r14303 OUTPUT["ē·č¢ˆé¦™" ] Bool::True
b_jonas I wanna try that too 15:47
?eval "ƁrvĆ­ztűrő tĆ¼kƶrfĆŗrĆ³gĆ©p".perl.say
evalbot_r14303 OUTPUT["ƁrvĆ­ztűrő tĆ¼kƶrfĆŗrĆ³gĆ©p" ] Bool::True
theorbtwo What language is that, b_jonas? 15:48
15:48 kanru joined
nothingmuch .no? 15:49
no
b_jonas theorbtwo: .hu
nothingmuch .hu mabe?
ah
b_jonas yes
it's a common pangram-like thing which has all the accented letters of hungarian so it's used for testing 15:50
en.wikipedia.org/wiki/Pangram#Only_...ic_letters 15:51
audreyt "\x[041d]\x[043e] \x[0441]p\x[043e]\x[043a] \x[043d]\x[0430]\x[0441]\x[0442]\x[0430]\x[043b], \x[0412] \x[0434]\x[0430]p \x[043f]\x[043e]\x[043b]\x[0443]\x[0447]\x[0438]\x[043b]".say
nothingmuch cool
audreyt ?eval "\x[041d]\x[043e] \x[0441]p\x[043e]\x[043a] \x[043d]\x[0430]\x[0441]\x[0442]\x[0430]\x[043b], \x[0412] \x[0434]\x[0430]p \x[043f]\x[043e]\x[043b]\x[0443]\x[0447]\x[0438]\x[043b]".say
15:51 evalbot_r14303 is now known as evalbot_r14304
evalbot_r14304 OUTPUT[ŠŠ¾ сpŠ¾Šŗ Š½Š°ŃŃ‚Š°Š», Š’ Š“Š°p ŠæŠ¾Š»ŃƒŃ‡ŠøŠ» ] Bool::True 15:51
theorbtwo sees lots of white boxes.
nothingmuch sees cyrillic 15:52
theorbtwo is on the console.
(I really ought to figure out why X hangs this machine.)
b_jonas I see cyrillic too except it's ugly because it shows up in a different font but the 'p's show up in the normal font as they're not typed as ciryllic
16:04 araujo joined 16:08 buetow joined 16:09 weinig|bbl is now known as weinig 16:10 Psyche^ joined, Psyche^ is now known as Patterner
nothingmuch Duane Allman++ 16:21
[particle] (slide guitar)++ 16:22
nothingmuch when tasteful =)
16:24 babusri joined 16:31 justatheory joined
[particle] nothingmuch++ for inspiring me to create an allman brothers pandora station 16:31
nothingmuch =)
how many of t hem died on motorcycles? 2 or 3? 16:38
[particle] 2 16:42
like a year apart, in the same town 16:43
nothingmuch aye
Duane and the bassist, i forgot his name
anymoose, dinnertime
16:52 mdiep_ joined
svnbot6 r14305 | lwall++ | Use newline surrogate instead of space on evalbot output. 16:55
TimToady ?eval say "hi" 17:01
17:01 weinig is now known as weinig|bbl
araujo hello TimToady 17:01
TimToady hi, araujo. Thanks for sticking around after I invited you. :) 17:02
17:02 evalbot_r14305 joined
araujo TimToady, hah, it's ok , i am slowly knowing this perl6 language :-) 17:02
TimToady I'm sure you'd rather have been invited by audreyt. 17:04
araujo TimToady, actually, i am starting to like it ;-) 17:05
TimToady Hopefully Perl 6 will be prettier than me and uglier than Audrey.
araujo hahaha
araujo liked TimToady's HWN recent quote
TimToady For a "quote of the week" I thought it was rather ancient... 17:06
araujo hah
TimToady I don't even remember when I said it.
svnbot6 r14306 | audreyt++ | * Set.pm: Use $!members for stringification, instead of the slightly misleading @.members. 17:09
TimToady ?eval 42 17:12
[particle] evalbot has left the building
17:12 evalbot_r14306 joined
TimToady I find it difficult to believe that my one-character change killed evalbot... 17:15
[particle] 14304 signed off 17:16
and has bounced every rev since 17:17
TimToady does evalbot leave a log somewhere on feather? 17:18
ajs I'd like someone to confirm that I'm not making something up. Will P6 rules be able to do this: rule A { <before <B> <!before <C>>> <D> } 17:19
S05 isn't too clear on the extent to which before will be able to encapsulate abitrary rule content.
TimToady sure
ajs Ok, thanks
TimToady <foo blah> is short for <foo(/blah/)> 17:20
17:20 evalbot_r14306 joined
ajs Someone was asking me about the relation between PEGs and P6 rules, and I wanted to be sure that <before> and <!before> could build arbitrary predicates. 17:20
[particle] i don't know about any evalbot log, but i don't know much about evalbot
TimToady me either--I just cargo cultishly changed it to join output lines with ā¤ instead of space. I find it difficult to fathom that other unicode is handled in the same result string, but that isn't. 17:22
audreyt no, it's failing because of Set.pm doesn't handle "method prefix:<~> is export" gracefully in the current Pugs
will fix it properly tomorrow but for now I've reverted to multi *prefix:<~>
TimToady whew!
audreyt also, replicated the Set::Infinite segfault. 17:23
Limbic_Region yeah
yay even
TimToady segfault--++
audreyt strangely here it only shows up on pugs-prof
Limbic_Region was going to mention that all_parse.t is not working again
audreyt and never ./pugs
yes, I saw that
replicated tha ttoo
TimToady oh, well, I can get pugs-prof to segfault on all sorts of things.
at least 6.4.2's -prof
audreyt I'm building GHC 6.6 from source on feather with nice -n 20 17:24
17:24 evalbot_r14306 joined
audreyt ?eval say "hey"; say "hi"; $_ = 'a b c'; s:g[\w] .= ord; $_ 17:25
Limbic_Region going to run a smoke in a few minutes 17:26
TimToady maybe the ghc compile is messing up evalbot's dynamic linkage? 17:28
hmm, you wouldn't think it'd get even that far though.
audreyt also, ghc doesn't do any dynamic linkage 17:29
by default, anyway
17:29 araujo joined
Limbic_Region hey TimToady, did you see where demerphq had gotten named captures working in p5 re? 17:29
TimToady no, I didn't. 17:30
is this a p5p thing?
Limbic_Region not sure - I know he is taking a lot of input from you and others here in #perl6
thought he might have shared his results here too
17:31 evalbot_r14306 joined
Limbic_Region doesn't follow p5p but he is addicted to PerlMonks which is where demerphq usually hangs out 17:31
audreyt ?eval say "hey"; say "hi"; $_ = 'a b c'; s:g[\w] .= ord; $_
evalbot_r14306 OUTPUT[heyā¤hiā¤] \"97 98 99"ā¤
Limbic_Region ?eval say "hello TimToady"
TimToady evalbot++
audreyt whew.
evalbot_r14306 OUTPUT[hello TimToadyā¤] Bool::Trueā¤
audreyt I can rest in peace now :)
Limbic_Region sleep well audreyt
TimToady I can go primp for $job in peace. :)
wolverian what's the [?] character it outputs now? 17:32
TimToady a newline surrogate. supposed to say NL only on the diagonal in one char
audreyt on the irssi transliteration here it says "NL" 17:33
19:31 < evalbot_r14306> OUTPUT[hello TimToadyNL] Bool::TrueNL
svnbot6 r14307 | audreyt++ | * Set.pm: Remove the "method is export" for prefix:<~>,
r14307 | audreyt++ | as it causes evalbot failure. Maybe this release isn't
r14307 | audreyt++ | ready for exporting methods after all... Will investigate
r14307 | audreyt++ | in more depth tomorrow.
17:33 bernhard joined
TimToady It's U+2424. 17:34
audreyt which font of yours can display this?
[particle] eew, here in says a with carat, box, little box with dots at outside corners
TimToady my terminal is using sazanami, I think.
Limbic_Region the CGI:IRC on feather is displaying it correctly too 17:35
that just has the character set at Unicode (UTF-8)
TimToady the basic iso-10646 font on Linux also seems to have it. 17:36
audreyt nite :) & 17:37
wolverian very weird. generally pango falls back to whatever font it is that has the character. I find it hard to believe that _none_ of my fonts have it.
TimToady firefox seems to handle it fine
ah well, I think it's worth the temporary visual breakage anyway, though we should chop the final NL 17:38
17:39 mako132_ joined
TimToady which, interestingly, would have happened automatically with P5's split semantics that delete trailing null fields. 17:39
well, shower &
17:46 larsen joined 17:57 jferrero joined, weinig|bbl is now known as weinig 18:10 norageek joined 18:12 renormalist joined
svnbot6 r14308 | lwall++ | chomp the last newline from evalbot results. 18:16
18:34 aufrank joined 18:35 elmex joined 18:36 weinig is now known as weinig|away 18:50 ron__ joined 19:28 andara joined 19:29 rindolf joined
rindolf Hi all. 19:29
TimToady hi 19:34
agh, p5 just used ?+ *+ ++ for possessive quantifiers, where p6 uses ?: *: and +: 19:36
19:36 justatheory joined
PerlJam TimToady: submit a patch :) 19:36
TimToady maybe to prevent confusion p6 shouldn't use ?+ *+ and ++ for greed overrides.
PerlJam the perl5 version still hasn't made it out into the wild yet, it's still just as mutable as the perl6 19:37
TimToady unfortunately use of + for "possesive" is not suffiently general
seen demerphq? 19:38
jabbot TimToady: demerphq was seen 9 hours 50 minutes 14 seconds ago
b_jonas PerlJam: not exactly, it has to be backwards compatible, so it can't use +:
TimToady or enable via pragma 19:39
b_jonas TimToady: what do you mean it's not sufficently general?
TimToady can't put it on other things like [a|b|c]:
(a|b|c)+ means something else
b_jonas TimToady: rather via a regex flag like imsx than a pragma
yes, you can only write that as (?>a|b|c)
integral dmq was just talking on rhizo #p5p
TimToady I understand the compatibility issues, but we just can end up using ++ for two misleadingly similar things. 19:41
*can't
demerphq hu 19:42
huh
SamB you can't?
SamB would never have guessed that it was impossible to do so
TimToady if x++ means one thing in p5 and something incompatible in p6 it'll be rather confusing
demerphq er, i would have thought we cant use : in perl5. 19:43
and the ++ thing is established elswhere from what i understood.
b_jonas isn't p6 rules syntax sufficently distinct from p5 regexps that it won't cause confusion?
demerphq also, what b_jonas said.
TimToady that's correct, not without pragma or /x support. that's why I'm also considering changing it in p6 instead
demerphq :-)
TimToady do you know where else ++ is possessive?
b_jonas (yeah, I know you don't like me considering p6 a new language)
demerphq its written in MRE that the '+' is the possessive form. 19:44
Maybe TCL, id need to check.
TimToady b_jonas: that doesn't bother me--I can break things if I want, but this is the sort of niggly breakage that isn't obvious enough to stand out.
so is likely to be confusing.
demerphq anyway, i used plus cause thats what friedl said was right, so shoot him if anybody needs shooting. =)
TimToady so maybe greed override in p6 should be ?! *! +! or something. 19:45
well, he should know, but I think possessive quantifiers aren't really the general solution to backtracking control...
b_jonas well, it's true that ? * + ?? *? +? are the same in p5 and p6 res, so it might be confusing
TimToady as long as our quantifiers start with something reserved we can use other chars after them as "tokens", so ! might work. 19:46
b_jonas I'll look up if vim regexps have a syntax for atomicity (I know it has the equivalent for ?? *? +? {}?)
demerphq wait, no, they are from java. sorry
java.util.regex
PerlJam demerphq: clearly it's wrong then ;-) 19:47
demerphq no doubt. why did i ever both.
bother.
:-)
TimToady I suppose we could make ++ a synonym for +:
demerphq Friedl you TRICKED ME!
Anyway, im sorry this has caused a problem. Should i change the implementation timtoady? 19:48
PerlJam demerphq: you didn't cause a problem, you hilighted an issue :)
TimToady no, if other languages are using ++ for possessive, we should probably just avoid it in p6, or make it a synonym for p6's +: backtracking control. 19:50
b_jonas the vim equivalent of ? * + ?? *? +? are \= * \+ \{-,1} \{-} \{-1,}
so there's no special syntax for ?? *? +? but only for {}?
TimToady the thing I mentioned before you got here was that ++ doesnt generalize to (a|b|c)+ 19:51
demerphq friedl makes a big deal out of them.
TimToady whereas +: does generalize to (a|b|c):
they're a partial solution to the issue of backtracking, I think
demerphq agreed 19:52
b_jonas and \(foo\)\@> might be the vim equivalent of an (?>foo)
I'm not sure I read it correctly
PerlJam demerphq: friedl makes a big deal about them in mre3?
demerphq yep
TimToady I like the visual correspondence of +: and +! too.
demerphq no, 2nd
Limbic_Region I thought the major changes in 3rd edition was PHP stuff anyway? 19:53
PerlJam hmm. I only have mre1 around here
demerphq anyway, i have to go in a minute
but before i go, is there anything i should know or do? 19:54
btw, i did (?(DEFINE)....) as well just recently.
b_jonas yes, \@> is atomic match in vim
demerphq syntactic sugar for (?(?{0})...)
TimToady i think you're okay now, was just a little panicked at the mismatch
b_jonas so a++ is \(a+\)\@> in vim 19:55
PerlJam demerphq++ btw for the recent p5 patches :)
demerphq ok, well my apologies for the panic. thanks perljam. :-)
TimToady I'll do something different in P6
demerphq for define? no doubt, that was just a workaround.
TimToady +! or +- or +<> or something 19:56
demerphq oh. i get it. duh.
TimToady no, for greed override
PerlJam likes +! fwiw
demerphq right.
TimToady since tokens turn off greed
tokens and rules default to possessive
demerphq ok, well, i must trundle or the boss will get mad. shes waiting in the pub
TimToady thanks
or ?* ** and ?* unless friedl has claimed those as well... :/ 19:58
b_jonas I can't find "++" anywhere in S05 -- what does it do? 19:59
TimToady: but ** is used for another conflicting thing in perl6 rules, so will that improve anything?
TimToady it's at line 1234 20:00
only for **{...}
which could be *{...} or just use space to distinguish ** {...} 20:01
b_jonas I'm looking at a (possibly old) html version at "dev.perl.org/perl6/doc/design/syn/S05.html" so I can't go to a certain line
TimToady but if we need space we might as well go with *{...} and require space for * {...} 20:02
I picked **{...} because it was otherwise illegal.
b_jonas but it's also nice because it implies power 20:03
which is how a mathematician would write repetition
TimToady but **{0..*} is a lot of stars
Limbic_Region speaking of dev.perl.org being outdated - is there a reason why S26 isn't there either?
is there a sync issue?
TimToady so only just added to pugs repo, I think
pmichaud_ "My god...it's full of stars" 20:04
b_jonas isn't there a monadic suffix .. now, so you can write **{0..}
TimToady no, we don't allow infixes to just leave off their argument.
throws the whole term/operator expectation thing into disarray
b_jonas As an unrelated thing, will the ^ prefix operator give numeric context to its argument so I can write ^@array instead of ^+@array ? 20:05
yes, that sounds right
once the rules for parenthesis-less function calls fixed as is, they don't cause the parsing ambiguity they would otherwise 20:06
so we shouldn't ruin that
TimToady it will just send a ^ message to the object and the object will do whatever it likes with it. For most non-class objects it should probably ^+
b_jonas I see
TimToady and we're relying very heavily on term/term juxtaposition illegality to keep things sane 20:07
svnbot6 r14309 | fglock++ | v6/Emitter/Perl6/Perl5.pm(2) - the emitter does subroutine parameter binding by generating & compiling perl6 code; this way it gets typechecking and coercions for free.
20:08 SubStack joined
TimToady if ^@foo is 0..^@foo then arguably ^%hash is %hash.keys 20:08
b_jonas I also like ruby's rules for parenless method calls
which is that if the call is a statement then it's ok to leave out the parens, otherwise it's not
PerlJam But I thought ^ was the "meta" sigil :)
b_jonas (of course ruby excepts statements at lot more places than perl5)
TimToady I looked at that a long time, and decided I didn't like it, compared to the : thing we have 20:09
b_jonas sure
TimToady ^ is the meta sigil on metathings like prototypes
but I think only prototypes respond that way to ^
otherwise you should use .HOW 20:10
b_jonas the disadvantage of the ruby thing is that as (foo;bar) is legal, it's difficult to catch unclosed parenthesis errors
especially because a newline also implies a semicolon
in ruby
TimToady yeah, ; are also useful in self-clocking
b_jonas ?self-clocking
TimToady natural sync points, though arguably ; is not natural 20:11
b_jonas as for semicolon, I like the way Mathematica uses it
TimToady but say like, in Japanese, where most words start with kanji and end with hiragana
PerlJam EOL works that way too (though invisibly)
er, *can* work that way
(thus ruby's rule about implicit ; ) 20:12
b_jonas that is, 'foo;' evaluates foo and then returns void
TimToady that seems like to small a character to carry that kind of weight. 20:13
*too
especially for people who are used to ignoring ;
PerlJam so ... was there an agreement on the possessive quantifier modifier character?
(try saying that 3 times fast)
TimToady agreement? what is this thing? 20:14
20:14 andara left
PerlJam er, concensus? 20:14
lack of discord?
b_jonas PerlJam: wonder why I call it atomic?
PerlJam fiat?
TimToady ah, that I understand... :)
b_jonas PerlJam: as demerphq is gone, I don't think there could be a final word on it yet 20:15
PerlJam And here I thought Guido was the BDFL
TimToady I gave him permission to stay with ++ in p5
I'm mostly wondering if there's prior art with ** 20:16
and ?* and +* and *{...}* of course
or whether I should just grab some other character like ! 20:17
PerlJam *{...}* is oddly symmetrical (while the others aren't)
TimToady * would work nicely, all other things being equal, which they never are
b_jonas what's the problem with the colon then?
TimToady colon is implicit in tokens and rules, so you need an anticolon 20:18
b_jonas oh,
TimToady and I don't want to make *: mean opposite things
b_jonas you want a 'do backtrack' statement
TimToady yeah
the extra + was that, but it looks like a bad idea now. 20:19
so I wonder if anyone (besides me) as made an extra * mean anything...
[particle] :(0)
if only that would work... or i gues ::(0) 20:20
TimToady is that a smiley or a suggestion?
[particle] like :i(0)
b_jonas is there an anti-space for when the :s modifier is in action?
[particle] of course, it won't work
!:
TimToady :!s
[particle] :!:
b_jonas sure, but, for only one place 20:21
or does it act only locally?
I mean between two non-space things
TimToady I guess ! already is a meta after !, except it means "not" there rather than "dammit!"
[particle] it's lexically scoped
TimToady after : I mean
[particle] is ; used? 20:22
b_jonas is "dev.perl.org/perl6/doc/design/syn/S05.html" a recent version of the synopsis? if no, are there up-to-date versions easily reachable via http anyehere? 20:23
TimToady no, but I think there's a lot of languages that would be unhappy if ; became a metachar...
20:23 penk joined
[particle] b_jonas: that should be the latest 20:23
b_jonas isn't the point to make all nonword chars metachars in p6 rules?
:)
[particle] well, in some languages ; is whitespace 20:24
TimToady and ; is probably too easy to confuse with : visually
b_jonas (poor fonts)--
[particle] yes indeed
how about =
TimToady um, S05 has a last-modified date which will tell you how up-to-date it is...
b_jonas = is already used, isn't it, as named captures 20:25
TimToady no, :=
since : is a meta
b_jonas ah indeed
it says oct 2006
TimToady maybe :! should mean "don't not backtrack"
[particle] perhaps it's time for real smileys 20:26
:P
:-* 20:27
b_jonas :(
TimToady :O
:D
[particle] :X
TimToady "I would be very happ if you
Din't not backtrack here
need to learn to type... 20:28
b_jonas = sounds nice for an anti-colon though
TimToady sec..
[particle] :-K
"that's what i think of backtracking right now..."
b_jonas oh, so + is the anti-? quantifier-modifier, I see 20:29
pcre lacks that
TimToady I think *:! is superior to ** in the same way that *: is superior: is superior to ** 20:32
you can say [a|b|c]:! and turn bcktracking back on 20:33
[particle] i'd really like to be able to meta-enable backtracking in tokens 20:34
TimToady sorry about the lousy typing, but I'm mostly typing blind because the link between $job and $home is very slow for some reason...
ingy seen stevan
jabbot ingy: stevan was seen 7 hours 45 minutes 27 seconds ago
b_jonas oh, = wouldn't work because := is used already
[particle] so, [a|b|c]:!? would still work?
and is that the same as [a|b|c]?:! ? 20:35
TimToady er, huh?
[particle] optional a or b or c 20:36
TimToady I suppose, but that is making a good argument for * instead of :! I think... 20:37
dunno, need to think about it all some more. 20:38
will do so over lunch.
so, lunch. &
20:49 dolmen joined 20:52 buetow joined, Aankhen`` joined 21:00 weinig|away is now known as weinig 21:03 aufrank left
ajs <?>...> is already taken? ;-) 21:20
21:23 MacVince joined
rindolf ingy: ping 21:27
21:28 leo joined 21:39 Limbic_Region joined
nothingmuch @seen cmarcello 21:46
@seen cmarcelo
rindolf Hi nothingmuch 21:49
nothingmuch hi
21:54 lambdabot joined 21:55 avarab joined 22:01 spedri joined 22:09 cjeris left 22:25 Aankhen`` joined 23:02 weinig is now known as weinig|bbl
svnbot6 r14310 | fglock++ | v6/emitter(2) - positional arguments work 23:32
r14310 | fglock++ | - passes t/01-sanity/05-sub.t 23:33
23:36 mako132_ joined 23:40 demerphq joined, jferrero 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!