6.2.12 released! | paste: sial.org/pbot/perl6 | pugs.blogs.com | pugscode.org | pugs.kwiki.org
Set by audreyt on 26 June 2006.
00:01 turrepurre joined 00:03 hikozaemon joined 00:25 coumbes joined
audreyt rehi 00:28
nothingmuch 3rehi audreyt 00:49
Juerd Hello 00:50
audreyt moose
TreyHarris audreyt: do you use Terminal.app? I can't figure out how to type any non-ascii chars in it, i just hear beeps when i'm at the pugs prompt and try 00:53
audreyt ...I can't either. 00:54
TreyHarris ok. :-)
audreyt try iterm?
TreyHarris no beeps, but no chars either :-) 00:56
audreyt mmmm. 01:03
TreyHarris (alt-\ is left double bracket, and alt-| is right) 01:04
ah well. back in a few hours... 01:07
01:15 cognominal joined 01:25 Entonian joined 01:28 amnesiac joined 01:31 rashakil joined 01:34 coumbes joined
nothingmuch Juerd: pinfg 01:42
Juerd nothingmuch: ponfg...
Juerd wonders if it was a trick question. 01:43
A rhetorical trick question at that.
nothingmuch dfo yfou knfow whfere tfo makef fa blackf caffepressf tffshirt?
disclaimer: one handed typing is bad for your temper
Juerd Itf fwas a befta. 01:44
Dunfno iff ift'sf stfill thfere 01:45
nothingmuch www.befta.com/ doesn't seem to sell tshirts 01:46
lambdabot Title: "www.befta.com"
Juerd fhefhe 01:47
Founfd ift: www.cafyepressf.com/cp/serfvifces/newbflacfk 01:49
02:07 buubot is now known as test, test is now known as buubot 02:39 QtPlatypus joined
svnbot6 r11047 | sanug++ | chromatic's PHEME diagram. Hyoer-Linked compilation overview diagram 02:48
r11048 | sanug++ | Allison's Punie diagram, Audrey's Perl 5 diagram 02:51
02:54 niner_ joined 03:17 scook0 joined 03:19 beppu joined
TimToady we're back online... 03:28
svnbot6 r11049 | clkao++ | r14906@ab: clkao | 2006-06-30 21:52:48 -0500 03:30
r11049 | clkao++ | Data::Bind::_forget_unlocal, installing temporary symbol.
r11049 | clkao++ | This allows binding &code into subs.
r11050 | clkao++ | r14907@ab: clkao | 2006-06-30 21:53:22 -0500
r11050 | clkao++ | Merge from perl5/PC-P6, with parameter-binding fixes.
03:31 lollan joined 03:36 weinig joined 03:41 scook0_ joined 04:08 supcik joined 04:09 supcik left
audreyt ingy: www.nntp.perl.org/group/perl.perl5....ers/114522 04:17
lambdabot Title: "nntp.perl.org - perl.perl5.porters (114522)"
04:27 ludan joined 04:29 weinig is now known as weinig|zzz 04:30 scook0 joined
gaal remoose 04:30
gaal returns from a fun long bike ride
04:32 cognominal joined
audreyt yo! 04:34
gaal heya audreyt! sorry for disappearing midnode, I was abducted by fixed point comb^W^Wgear bikes
audreyt that's fine :) 04:35
we still gotta finish it somehow
but I think I need to crash soonish
gaal okay so tomorrow on trans-1hr-tz hackathoning then the day after in boston 04:36
nothingmuch: ping 04:40
svnbot6 r11051 | clkao++ | Correct packages for the v6 compiled scripts. 05:04
r11052 | clkao++ | Support some calling capture emittion with Data::Bind.
05:08 jkanter joined
svnbot6 r11053 | lwall++ | Different take on atomic syntax. 05:25
05:43 como joined
audreyt fglock: 05:46
the point example changed in my slide
TimToady said that even in the method body
you can't say $.x=0
and expect it to work with $.x is readonly
have to say $!x=0
so how about just translating $!x=0 to $self->{x}=0 05:47
clkao data::bind is on CPAN
audreyt and $.x=0 always to $self->x(0) ?
clkao: !!
obra clkao++
audreyt but not svk? :)
clkao svk was released a few hours ago
audreyt sugoi
clkao also app::cli with real sign 05:56
obra sugoi 06:00
gaal audreyt: nothingmuch thinks ast nodes for control statements􏿽xB1􏿽xC5tements should be like foreign vals 06:01
audreyt gaal: Prims, you mean? 06:02
gaal so that DSLs that have their own user defined structures produce asts that retain some sort of semantic information
audreyt but th questino is, do we call them with the full calling convention
or are then "special forms" and bypass that 06:03
or am I missing some other solution?
gaal if i understood his argument, representing them as desugared calls/applications prevents manipulators that do know aboutt he DSL in question from easily working on them 06:04
audreyt but if you have DSL you already override the statement_control:<if>
at parse time
to gen another tree
gaal so he wants something like data Expression = ExpVariable | ExpValue | ...blah | ExpControl 06:05
svnbot6 r11054 | sanug++ | Added Hyper-link for Tree Grammar Compiler
r11054 | sanug++ | Will transmit PDF each time this is updated.
gaal and data Control = ExpCond | ExpWhile | ExpForeign
audreyt that is fine with mean except each constructor will take more stuff 06:06
but sure
gaal for him doing it the way you just said is a little like constant folding happening too early
yes of course
his point about foreign here being that the new control structures must knowhow to apply themselves 06:07
audreyt and ExpForeign as trapdoor is fine
I agree. his point is floating, fixed, and power
audreyt goes preparing for sleep 06:08
gaal okay, great. So I don't mind to have ContCond | ContWhile | ContGoto etc. for all the things we can think of now
plus the foreign trapdoor
audreyt k
gaal and this being lifted means that transformers that don't care about control in detail can eaisly patmatch over them 06:09
audreyt all excellent points. please simply commit 06:10
gaal anymoose good night, I'm off to bed too soon
audreyt plsu
TreyHarris anybody know, if i want to use moose, should i use 0.05 or 0.09_03? it's marked devel, but it hasn't changed in two months... 06:15
06:16 cmarcelo joined
audreyt 009 definitely 06:20
and push stevan to releas 0.10
maybe assign him a hiveminder ticket :)
TreyHarris :-D 06:21
obra :P 06:22
TreyHarris i haven't been invited to hiveminder though 06:24
audreyt you have a commit bit? 06:25
TreyHarris yes
audreyt a hiveminder invitation is on its way to you 06:26
TreyHarris ok :-)
gaal are application/call Expression or Control? I lean to the latter 06:27
basically anything that changes flow i guess 06:28
audreyt sure, but assign/bind are still exp 06:29
gaal yeah. deref too
audreyt k. 06:30
audreyt waves and sleeps &
gaal nighty night
have a safe trip tomorrow
gaal wonders if Ctx CtxSlurpy (x::Expression)@(ExpDeref (v :: Value) is correct 06:33
but has droopy eyelids. tomorrow...
svnbot6 r11055 | gaal++ | r11148@sike: roo | 2006-07-01 01:35:07 -0500 06:38
r11055 | gaal++ | * p6ast: data Control for expressions having effect on flow.
r11055 | gaal++ | These include a trapdoor for new DSL control statements
r11055 | gaal++ | (nothingmuch++)
gaal invite ininשe accepted, to create an account? I don't have a beta password
anymoose, sleep. & 06:39
07:01 iblechbot joined 07:21 Aankhen`` joined 07:33 lisppaste3 joined
xerox gaal mismatched parens (: 07:45
08:00 Aankh|Clone joined
svnbot6 r11056 | sanug++ | Added directory to hold image-mapped HTML files with jpg's for Perl 6 Compilation diagrams - allows HyperText traversal of diagrams 08:12
08:22 Aankh|Clone is now known as Aankhen`` 08:26 shachaf joined 08:28 scook0 is now known as scook0_, scook0_ is now known as scook0 08:31 iblechbot joined 08:33 hikozaem_ joined 08:39 larsen joined 09:02 cmarcelo joined 09:09 shachaf joined 09:14 elmex joined 09:20 krychu joined 09:34 kane_ joined 09:48 nothingmuch joined 09:55 chris2 joined 09:58 froh-doh joined, larsen joined
QtPlatypus ~ 10:01
11:18 lumi joined 11:22 buetow joined 11:25 dduncan left 11:26 tacocat joined
tacocat Morning? Whats the "best" list to subscribe to for Perl6 Documentation. I was sucked into synapsis review during the hack-a-thon and would like to get my work posted. 11:26
theorbtwo tacocat: perl6-language, I'd say. 11:27
tacocat Good enough... Now to find out if it's any good... 11:28
I found reviewing the synapsis a great way to learn about some of the language but perhaps a little convoluted because the text is "raw". 11:29
theorbtwo Oh! You want to write documentation aimed at users? 11:30
You might want perl6-language.
xerox audreyt: what editor are you using in osx? 11:31
tacocat Yes, that's the idea... Considering I know little/nothing about C, C++, Perl6 and am reasonable at Perl5 it seemed about the best place.
Uri, with all his charm, convinced me this was the way to perl salvation for me. 11:32
theorbtwo Er, you might want perl6-user(s?) 11:34
11:37 nothingmuch joined 11:39 iblechbot joined
tacocat I suppose I could try them both, thanks. 11:43
11:59 mauke_ joined, ruoso joined 12:14 mauke_ is now known as mauke
ZuluOne I'm not a great fan of p6u. 12:40
But then, I never like user lists.
I'm apparently about as autistic as you can get without actually being autistic, so while I can comprehend that people know different sets of knowledge than me, it still freaks me out somewhat when they show it. 12:41
12:55 elmex joined 12:59 dolmen joined 13:11 cmarcelo left 13:18 DaGo joined 13:52 nnunley joined 13:57 weinig|zzz is now known as weinig, broquaint joined 14:12 weinig joined 14:14 cognominal joined
nothingmuch kolibrie: ping 14:29
14:39 xdg joined
svnbot6 r11057 | sanug++ | Testing html markup to allow browser to open image maps 14:44
pasteling "chromatic" at pasted "No IRC; where is the hacking this morning?" (1 line, 1B) at sial.org/pbot/18137 15:16
obra hah
15:23 scook0 joined
gaal xerox: where? 15:24
xerox Duh, nevermind, it was just a mismatched paren in your last comment or something. 15:25
svnbot6 r11058 | sanug++ | did an svn propedit to add svn:mime-type with text/html; charset=UTF-8 - audreyt had this on her file in pugs/docs/01*.html
gaal ahhh, heh 15:28
spinclad xerox too. I fix: :) ) 15:30
svnbot6 r11059 | sanug++ | modified svn types on Tree diagrams with Steve Little 15:34
marcus_ kolibrie: nothingmuch wants to know what you wanted him to photograph. 15:37
15:45 tacocat joined 15:49 justatheory joined 15:58 amnesiac joined 16:00 shachaf joined 16:05 shachaf joined 16:39 lollan joined 16:43 larsen joined 17:08 putter joined
putter audreyt: ping? 17:08
17:16 fglock joined
marcus_ log bot? 17:19
putter marcus_: colabti.de/irclogger/irclogger_logs/perl6 17:21
lambdabot Title: "#perl6 irc log"
marcus_ putter: right, we found them :) 17:22
svnbot6 r11060 | fglock++ | v6.pm - lexer fix; parse & emit '...' and ':a<b>';
r11060 | fglock++ | - t\11-capture.t compiles, but don't run yet;
r11060 | fglock++ | - runtime autoboxing
fglock clkao: ":a<b>" was a problem with selecting the longest token - fixed 17:24
svnbot6 r11061 | fglock++ | v6.pm - checks if there is a package name 17:34
putter & 17:36
fglock audreyt: in deploy.xul "Constraints" - do you mean 'or fail' ? 'err' checks if defined(), right? 17:40
17:47 lisppaste3 joined
TreyHarris for reference, one cannot trust the value of $_ at the pugs command line :-) 17:49
try "my $y = 3\n say $_\n say $_" and you'll see what I mean 17:50
17:50 kane_ joined
TimToady obviously something setting $_ to the previous result 17:53
maybe it's a feechure. 17:54
17:55 nothingmuch joined
fglock ?eval "abc".ref 17:57
17:57 evalbot_10997 is now known as evalbot_11061
evalbot_11061 ::Str 17:57
fglock ?eval "aaaa".ref eq Str 17:59
evalbot_11061 Bool::True
18:00 shachaf joined 18:02 dolmen joined
kolibrie nothingmuch: pong 18:08
18:10 jkanter joined
nothingmuch kolibrie: what picture did you want me to take? 18:31
kolibrie nothingmuch: coming from my room on the third floor, in the stairwell, I thought the view of the tunnel for the elevated train was kind of cool 18:38
nothingmuch ah 18:39
kolibrie but I forgot to take a picture on my way out
nothingmuch so this is the top of the student center
out of your window?
kolibrie no, just in the stairwell, on level three
nothingmuch so where is the tunnel? 18:40
your window == stairwell window you are referring to
kolibrie yes, the one on top of the student union
nothingmuch ah
kolibrie of course, if you can't get it, it's okay. 18:41
nothingmuch: how was your birthday party? 18:42
nothingmuch kolibrie: pretty fun 18:44
marcus got me very drunk
and quite sick (which is bad, but memorably funny)
yesterday i was zoned out
and I still feel kinda crummy today
lesson learned - do not consult vikings for drinking advice, and do not listen to them if they offer it voluntarily
18:46 nothingmuch joined
kolibrie I'm always wary of advice given voluntarily 18:47
18:47 fglock joined
TreyHarris nothingmuch: happy birthday! 18:48
nothingmuch TreyHarris: thank you =)
TreyHarris nothingmuch: what year was it? 18:49
nothingmuch 21
TreyHarris ah, a big one, at least here in the US
spinclad felicitous annuadversions, nuffin! 18:53
nothingmuch TreyHarris: yeah, i was abused on account of that 18:55
they took me out
and wouldn't let me pay
and made me drink far more than I should have =)
TreyHarris well, at least you're not in college here as you turn 21... there's a rather dangerous tradition of going to a bar immediately after midnight the morning one turns 21 and forcing the drinking of 21 drinks before last call (usually 2-3 hours later) 18:56
spinclad yow 18:57
that could kill
TreyHarris yes.
nothingmuch yeah 18:59
we didn't go that far
TreyHarris there are actually states that have changed the laws so that the drinking age is 6am after you turn 21 for that reason. it shows that human beings are more comfortable with absolute values than relative ones... or something.
nothingmuch heh 19:01
the whole US drinking law stuff is pretty stupid
i mean, few people actually refrain from drinking before
so they're not really effective
and those that wait for it are at a risk
because 21 shots is much more dangerous when you're not even used to drinking 19:02
kolibrie has never had a drink 19:04
TreyHarris kolibrie: how old are you?
kolibrie 32
theorbtwo I don't think people who actually don't drink before they are 21 are likely to do that 21 shots thing.
TreyHarris theorbtwo: and i don't think people who have drunk regularly and moderately since they were very young would, either :-) 19:05
fglock hi!
spinclad hey fglock 19:06
theorbtwo True enough.
kolibrie returns to scraping old paint on his house & 19:08
19:13 weinig is now known as weinig|bbl 19:14 merlyn joined
fglock merlyn: hi! 19:15
merlyn hey
got pugs -e 'say "Hello"' working again 19:16
I'd broken it with a reset of my darwinports trashing
19:18 marmic_ joined
TreyHarris hmm. can i get pugs to read stdin as it comes in? =<> doesn't appear to read anything until I close stdin. 19:23
gaal makes obvious smelly joke at "happy barfday" pun 19:24
nothingmuch erhoo 19:30
merlyn TreyHarris - are you doing this in a scalar context? 19:32
TreyHarris merlyn: "for =<> -> $x { say $x }" 19:33
merlyn that's a foreach loop 19:34
list context
reads all at once
I think
while =<> { say $_ }
oops... that doesn't put the var anywhere 19:35
gaal there's a chance that that's lazy, so the file isn't read all into memory 19:36
TreyHarris huh. well, "while my $x = =<>" works just fine. but audrey changed my example of that in the perl 6 wikipedia entry to "for =<>"
so i thought that was equivalent :-)
gaal on the other hand slurp was recently made eager, because otherwise $data = slurp $file; close $file may perform the close before wverything was read 19:37
merlyn Well, it's the difference betrween while (<>) and foreach (<>) in perl5
in p5, the latter is in list context
has to read the entire pile first
19:37 MenTaLguY joined
MenTaLguY hello 19:37
merlyn the former is in scalar context
TreyHarris right, but isn't it lazy list context?
merlyn maybe not implemented yet though
design != implementation
MenTaLguY I was wondering if there was a description of the semantics of maybe {} available?
gaal true, but there are lazy lists in perl 6 so even if it is a list it's not necesarily computed all in advance. 19:38
19:38 pmurias joined
pmurias hi all 19:38
MenTaLguY (c.f. pugs/examples/concurrency/stm-retry-with.pl)
pmurias nothinmuch: did you think of integrating Perl::Generate with quasiquoting?
gaal MenTaLguY: try docs/Perl6/Spec/Concurrency.pod 19:39
hmm, no, it isn't in there yet.
pmurias s/nothinmuch/nothingmuch/
TreyHarris yes, looks like lazy lists aren't there yet. i just sent pugs into an infinite loop by doing for 1... 19:40
gaal MenTaLguY: then again I have no idea what sync{} and maybe{} are supposed to mean
MenTaLguY gall: excellent, thanks
er, hrm
darn 19:41
TreyHarris whoops, and i really should've quit pugs *before* reporting that here... i'm lucky i got my computer back it was thrashing so badly :-)
MenTaLguY sync basically wraps the enclosed code in a transaction
behaves like calling an 'is atomic' Code
merlyn is nuclear 19:42
MenTaLguY maybe {} blocks are tried until one doesn't call resync
but what I'm curious about is how that interacts with scoping and so forth 19:43
and how it works with composition
if someone could direct me where the best place to look in the pugs source would be, that would be nice 19:44
gaal MenTaLguY: that would probably be Pugs.Prim:289 etc. 19:45
MenTaLguY ah, what directory? 19:46
gaal hey, interesting technique there, generating a symbol just for that block
see also line 575 (op1 "Code::retry_with")
I don't see sync{} at all implemented 19:47
nor maybe {}
MenTaLguY nor I 19:48
TimToady it's just new syntax for old.
sync was atomically
resync was retry
MenTaLguY hmm, I don't think maybe {} corresponds exactly to anything in the old syntax, though... 19:49
I certainly recognized sync and resync
TimToady maybe { ... } maybe {...} was {...}.retry_with{...}
that is, Haskell's orElse
MenTaLguY right
TimToady so it should compose fine
MenTaLguY well, here's the thing
how are the two maybe blocks alternated? 19:50
TimToady round robin, I'm given to understand
MenTaLguY aha
so if a maybe block retries it retries the whole transaction?
TimToady correct
MenTaLguY (and sets a flag not to try that maybe the next time around) 19:51
TimToady that's why I was trying to find a correspondence between sync and resync
MenTaLguY hrm
TreyHarris TimToady: can you settle the earlier question? =<> will be lazily evaluated eventually, right? so while (<>) becomes for =<>, right?
MenTaLguY I'm still not clear on maybe {}'s semantics
it's certainly different than orElse if it retries the whole transaction 19:52
Haskell's orElse establishes a nested transaction for the lhs
eh well
what time is audrey usually on (and awake)?
TimToady hard to say now--she's presumably on an airplane to Boston.
MenTaLguY ah, k 19:53
TimToady TreyHarris: that's the intent.
MenTaLguY bbl 19:54
19:54 MenTaLguY left
TimToady though it's possible that =<> is going to change to =* or some such, now that we have the Whatever type. 19:54
TreyHarris TimToady: ok, thanks. for some reason i thought lazy lists were already in pugs, so i was surprised when it didn't work
TimToady =<> was really a kludgey special case.
yeah, that's a major suckage at the moment
19:58 elmex joined 20:08 marmic joined 20:40 weinig|bbl is now known as weinig 20:46 MenTaLguY joined
MenTaLguY back 20:46
I guess to condense my earlier question, if I understand correctly: 20:51
{ maybe {...} maybe {...} } ; { maybe {...} maybe {...} } 20:52
is equivalent to:
{...}.orElse {...} ; {...}.orElse {...}
I'm curious how that would work
TimToady I would imagine that two independent roundrobins in the same transaction would need to try (eventually) the cross product of all the possibilities. 20:59
not sure how that would be implemented, or how it composes though.
but maybe doesn't autowrap itself in a transaction in the current syntax. That's explicit with sync. If we violate composability of maybe inside that, then it can be an error. 21:04
but I'd rather have the atomic scope explicit.
MenTaLguY fair enough 21:09
if sync { maybe { ... } ; maybe { ... } } ; sync { ...e tc ... } is required that makes sense 21:10
TimToady my earlier remark was a misread of your example.
that's more like 21:11
sync { maybe { ... } ; maybe { ... }; OTHERSTUFF; maybe { ... } ; maybe { ... } }
MenTaLguY hmm 21:12
TimToady that is, what if the maybes aren't contiguous within a particular sync
MenTaLguY right.
TimToady are they separate roundrobins that do a search for a solution
like /(a|b)(c|d)/
only they just keep cycling instead of failing at the end 21:13
actually, I think the contiguous maybes aren't allowed to have ; between. 21:15
it would be like saying if 1 { 123 }; else { 456 }
MenTaLguY aha 21:16
then that answers my question I guess
contiguous maybes are grammatically a single statement
TimToady yes
MenTaLguY okay, no problem then 21:17
TimToady good. this is something I'm not a complete expert on...
I'm just the user interface guy... :) 21:18
gaal is there a linux program that can display .graffle files? 21:21
MenTaLguY I've never been able to find one, no. 21:25
nothingmuch gaal: i can reexport as $foo for you
gaal nothingmuch: thanks, just want to look at `find docs -name \*ffle`... 21:28
21:32 mj41_ joined 21:42 ramzi joined 21:44 lisppaste3 joined
svnbot6 r11062 | fglock++ | v6.pm - unmatched '}' now finishes the expression parser, allowing the migration of Yapp statements to plain Rule 21:50
r11062 | fglock++ | - t\10-moose1.t is now broken - awaits for new statements parser
21:50 Debolaz joined
fglock recursive calls between Perl6 and Rule compilers should now be possible (p5 backend) 21:54
also, terms can have Perl6 inside 21:55
21:57 jdv79 left 22:07 dduncan joined 22:52 larsen joined
svnbot6 r11063 | clkao++ | pull from perl5/PC-P6. 22:57
r11064 | clkao++ | * Re-enable the : prefix for named signatures. 23:21
r11064 | clkao++ | * Hard code some builtins that we don't use data::bind callconv.
23:38 MenTaLguY left 23:39 justatheory joined
svnbot6 r11065 | clkao++ | :$infix<~~> and update TODO. 23:54