pugs.blogs.com | pugscode.org | pugs.kwiki.org | paste: sial.org/pbot/perl6 | <stevan> Moose... it's the new Camel ":P | .pmc == PPI source filters!
Set by Alias_ on 16 March 2006.
FB|afk Pistos, I might have seen some used Model M's through a shop window on Bloor.... 00:09
want me to ask em?
00:09 FB|afk is now known as FurnaceBoy
FurnaceBoy s'oh...wc 00:13
00:15 kisu joined 00:20 kisu joined 00:24 justatheory joined
avar FurnaceBoy: what of Model M? 00:24
00:31 clkao joined
FurnaceBoy oh, just that I may have seen one in a box in a store today through a shop window, I could go in and check 00:32
00:34 miyagawa joined
avar FurnaceBoy: you can buy Model M clones 00:44
although the company that sells them is ignoring me
00:51 christo joined 00:56 christo is now known as cjm, cjm is now known as chromo 01:01 chromo left 01:02 FurnaceBoy is now known as FB|afk 01:05 rafl_ joined 01:07 rafl_ is now known as rafl
TimToady ?eval 1+1 01:08
01:09 evalbot_9774 is now known as evalbot_9777
evalbot_9777 2 01:09
buu YAY 01:13
01:19 stennie joined, FB|afk is now known as FurnaceBoy
FurnaceBoy avar, i'm not sure, cos I didn;'t take a close look 01:20
I don't want clones, I own a couple of real ones, and they are always available on ebay
actually a usb version would be nice though :) 01:21
avar mm
I'm trying to get one from a friend
FurnaceBoy: does PS/2 => USB work?
FurnaceBoy haven't tried, I guess so
01:23 stennie joined 01:45 lavi joined 02:24 bewest left 02:47 r0nny joined 03:06 Kattana joined 03:14 kakos joined 03:53 cmarcelo joined 04:06 amnesiac joined 04:28 cmarcelo left 04:32 Khisanth joined, lavixu joined 04:38 mako132_ joined 04:42 clkao joined 04:56 vel joined 05:04 Khisanth joined 05:52 GeJ joined 05:54 kanru joined 06:06 KingDiamond joined 06:13 Quell joined 06:41 iblechbot joined 06:51 nothingmuch joined 07:01 lavixu joined 07:10 kane_ joined 07:39 drrho joined 08:03 bsb joined 08:09 nothingmuch joined 08:10 autark joined 08:22 Aragone is now known as Ara4n
nothingmuch morning 08:25
08:29 marmic joined 08:47 pdcawley joined
KingDiamond okay, so the connect() method is confusing me a bit... I want to dump some raw http data... do I just pass all of it as the first (and only) param to connect()? 09:13
I realized I'm doing something wrong when ethereal shows me wrong http data 09:14
09:15 nnunley joined
KingDiamond nothingmuch: any ideas? :/ 09:16
nothingmuch KingDiamond: what are you passing to connect? 09:18
Net::IRC might be a good thing to browse
pasteling "KingDiamond" at 202.46.19.1 pasted "POST /services/upload/ HTTP/1." (23 lines, 705B) at sial.org/pbot/16523 09:21
KingDiamond nothingmuch: I want to make a request like thus: sial.org/pbot/16523?tx=on&submi...rmat+it%21
nothingmuch: initially, I did $hdl.say("POST /foo/bar HTTP/1.1\r\n", "Host: www.foo.com\r\n") and then $hdl.say(raw stuff) 09:23
ethereal shows something's wrong
nothingmuch one sec, reading 09:33
(sorry, distractions at $work)
KingDiamond: what is it *really* sending? 09:34
and what are you trying to send?
also, $hdl.say("POST /foo/bar HTTP/1.1\r\n" will send "\r\n\n"
use print
also, have you considered using LWP/HTTP::Request via use perl5: 09:36
?
rgs: thanks for the Tie::RefHash fix =) 09:40
rgs woot 09:41
nothingmuch btw, is it a separate module too?
rgs no, it's only core 09:42
nothingmuch as in, can Random::InsideOutObjectFramework rely on a version?
hmm
shame
09:42 f0rth_ joined
rgs but you can CPAN-ize it if you want to 09:42
nothingmuch should I ask the author? p5p? 09:43
rgs say that you want to do so on P5P, and don't fork.
nothingmuch okay
rgs and your Makefile.PL should override INSTALLDIRS 09:44
nothingmuch it's not worth forking
one sec, let me write this down
rgs INSTALLDIRS => perl
(it's 'site' by default)
nothingmuch okies
rgs and get the version from latest bleadperl, since I changed stuff 09:45
there haven't been much changes to this module since 1997 09:46
nothingmuch yeah
i don't know why people keep duplicating it though
rgs at least
nothingmuch e.g. Class::Std could be unb0rked for threads very easily now, using Tie::RefHash::Weak 09:47
rgs nothingmuch: well, if you ignore CPAN, you're doomed to reinvent it, poorly :)
nothingmuch maybe it's not well known enough
s/people/module authors/; # more alarming ;-)
KingDiamond nothingmuch: k, lemme try with print 09:48
09:57 KingDiamond joined
nothingmuch welcome back 09:59
KingDiamond nothingmuch: basically, I need to post data like the following: sial.org/pbot/16523 10:01
nothingmuch: I'd prefer using connect()... but yeah, if there's no option, well...
nothingmuch yeah =/
can you dump an ethereal log? 10:02
and your code?
pasteling "KingDiamond" at 202.46.19.1 pasted "156 my $hdl = connect(Flickr::" (10 lines, 564B) at sial.org/pbot/16525 10:03
nothingmuch the code looks good 10:05
KingDiamond and the program just hangs... lemme dump the relevant ethereal log 10:06
nothingmuch oh.. hangs
i thought it was sending wrong data
what does strace tell you?
perlbot: seen lypanov? 10:07
10:12 pdcawley joined
KingDiamond nothingmuch: lemme see 10:13
darn... I'm on os x 10:14
nothingmuch KingDiamond: ktrace + kdump
KingDiamond nothingmuch: yeah, doing that :) 10:15
nothingmuch both cooler and not as cool as strace
KingDiamond: it's lunchtime 10:17
i'll try out the code myself afterwords
and maybe i can figure it out 10:18
(I'm doubtful though)
also, if you like you can use perl5:IO::Socket
KingDiamond nothingmuch: cool, thanks... I'll play around until then
nothingmuch which lets you keep perl6 space feel, but get a diff connect implementation
KingDiamond k, will take a look if I fail ) 10:19
10:24 kane__ joined 10:25 b00t joined
KingDiamond nothingmuch: there? 10:36
11:02 premshree_ joined 11:13 nothingmuch_ joined
nothingmuch_ did KingDiamond leave? 11:14
11:17 autark joined
Gothmog_ Yes. Maybe not voluntary. "13:02 -!- KingDiamond (n=premshre@nat.bangalore.corp.yahoo.com) has quit: Read error: 104 (Connection reset by peer)" 11:22
11:27 avar joined
premshree_ nothingmuch: I'm here 11:27
11:27 premshree_ is now known as KingDiamond
KingDiamond nothingmuch_: I'm usng something like this: $hdl.print("POST $upload_interface HTTP/1.1\r\n$header\r\nHost: " ~ Flickr::host ~ "\r\n\r\n" ~ $res); but without any luck 11:28
do you want me to paste the relevant kdump bits? 11:29
11:29 kanru joined
nothingmuch_ KingDiamond: yes please 11:30
11:30 nothingmuch_ is now known as nothingmuch
pasteling "KingDiamond" at 202.46.19.1 pasted "8409 pugs GIO fd 5 wrote 132 b" (53 lines, 2.2K) at sial.org/pbot/16527 11:31
KingDiamond hmm... "35 Resource temporarily unavailable"
nothingmuch: see if you're able to figure 11:32
11:33 bsb_ joined
nothingmuch . o O ( why is it doing a select ) 11:33
hmm
KingDiamond: this is really odd
does the code work on e.g. feather? 11:34
brb
KingDiamond I could try 11:36
lemme see
11:38 elmex joined
KingDiamond darn, I'll have to install a cpl modules. 11:39
meeting. will try after that.
11:47 iblechbot joined 12:18 ruoso joined
KingDiamond Juerd: I wanted to install a perl module on feather 12:27
Juerd KingDiamond: And then discovered you don't have access :) 12:28
KingDiamond: Please create a debian package and then send it to me 12:29
KingDiamond: I'd like to keep the system clean of "make install"-software. 12:30
spo0nman Juerd: hello. 12:32
Juerd spo0nman: Hello. I received your email but have so far lacked tuits. Can you wait until tonight? 12:33
spo0nman Juerd: :) thnks.
KingDiamond: please create a debian package :D 12:35
Juerd: He's a mac luser.
Juerd With access to feather...
spo0nman yes ... he could build it there. 12:36
Juerd And you don't need Debian to create a Debian package, technically ;) 12:37
It's just an ar archive with some tgz archives in it ;)
spo0nman always uses alien to create debian packages.
spo0nman ducks.
Juerd But running "dh-make-perl --build --cpan The::Distribution" should suffice.
afk again 12:38
12:53 stennie joined
svnbot6 r9778 | ruoso++ | AST-Design subtree in pX/Common. This subtree can be used to study how the AST for different languages should look like. 12:55
ruoso added a first sketch on a Perl 5 AST in this subtree 12:56
BTW, everything in this subtree was just a first sketch... please change it at will.. 12:57
svnbot6 r9779 | ruoso++ | first change in the Perl5/helloworld AST 13:01
ruoso has a meeting 13:02
KingDiamond Juerd: okay, will do 13:08
13:08 clkao joined 13:15 fglock joined
audreyt fglock: I've prepared a tarball (tested) for you at perlcabal.org/~autrijus/tmp/Pugs-Co....02.tar.gz 13:17
fglock audreyt: I sent it to CPAN yesterday :) 13:18
how do you 'make dist' on PCR? 13:19
audreyt fglock: copy it to another path (not under pugs) 13:20
then "perl Makefile.PL ; make dist"
(need to have Module::Install installed)
fglock ok
I think the full p6 compiler won't be ready by apr 1st :( 13:21
svnbot6 r9780 | fglock++ | PCR - implemented '%hash' - pmurias++
audreyt it wouldn't... but we can send off a PR that talks about lrep.p6 :) 13:22
ooooh %hash implement!
wow!
that's, like, the critical step :)
fglock cool - we still have some days to finish then 13:23
there is an infinite loop in last test in t/06-subrule.t - I can't find why 13:27
audreyt k. I need to crash now -- <3hr of sleep last night and a very long day
got a lot of hacking done though... and learned quite a bit around YARV and ruby internals 13:28
fglock audreyt: have a good night!
audreyt thanks!
13:30 Limbic_Region joined
svnbot6 r9781 | audreyt++ | * Filter::Simple::Cached support in Module::Compile 13:30
r9781 | audreyt++ | - drop-in replacement for Filter::Simple
audreyt ingy: F::S::C is in - pick it up
when you have cycles
& 13:34
fglock fixed! (weird problem with the match state) 13:36
svnbot6 r9782 | fglock++ | PCR - fixed an infinite loop in '%hash' 13:39
nnunley What's the next biggest thing to work on? 13:45
fglock nnunley: I'm always trying to find out 13:47
nnunley :)
The subrule tests passe for you? 13:48
s/passe/pass/
fglock I'm debugging '%hash' - there is a problem when the value is a subrule - it gets the match internally, but it returns undef 13:49
nnunley Ugh. 13:50
fglock Ugh indeed 13:53
nnunley Sorry, I just used Devel::CallTrace to watch the call stack fly by. 13:55
obra woot. 14:01
So glad that code was useful to someone
nnunley obra: I find myself using it about once every two weeks or so. 14:03
clkao obra: i noticed that nnunley is addicted to D::CallTrace
nnunley I can't hold the call stack in my head all the time. ;)
14:03 elmex joined
obra What would make it more useful for you? 14:04
fglock nnunley: did you find the problem?
obra hm. Jifty should have CallTrace support in halos
nnunley obra: It would be nice if there was a DWIMish output for anonymous code blocks. 14:05
FILE and line number of start of block would work.
obra *nod* That could be clever 14:07
nnunley fglock: Digging briefly to see if I can figure it out. 14:11
fglock nnunley: I think I found the problem - the separation between compile/runtime is not right - it is recompiling some things at runtime, and this breaks backtracking 14:15
nnunley Yeah, Test 21 has the wrong start. 14:18
Err, no. Brain not working right. 14:20
fglock nnunley: Brain is ok - the program is not easy at all 14:22
14:24 lichtkind joined
lichtkind sorry for this begionner style question but we had some rants in our forum about perl6 and the question arose: why := vor binding and not =: ? $A := $B seems to be more shortcut for $A = $A : $B or is it that whats wanted? 14:26
nnunley fglock: I'm finding it tempting to add local *__ANON__ notations to the emitter. 14:27
fglock lichtkind: ':=' is already used in other languages for the same thing
14:28 vel joined
lichtkind thanks fglock 14:28
but isnt it more important to be consistant? 14:29
the people here said it would be more logical to the when we use =:
fglock nnunley: isn't it expensive? the match functions are called thousands of times 14:30
lichtkind: the logic is that people don't have to re-learn, I guess 14:32
nnunley True. When is the perl5 emitter called? If you have a debug flag that indicates that it should emit anotations, then you generate code with the ANON notation.
And otherwise, not. 14:33
fglock nnunley: that's ok - you can test an %ENV variable
nnunley Or $main::DEBUG. 14:34
fglock how about DEBUG_ANON ?
bbiab & 14:35
nnunley Either works. 14:36
14:38 xinming joined 14:54 kolibrie joined
lichtkind ok i have second question is the given/when always tied to smart match or can i say given $level { 14:58
when 60 { laber.. }
when $_>40 { laber.. }
default { laber.. }
}
14:59 nronksr joined
lichtkind i have synopsys open but together with the book it is a bit misleading 14:59
PerlJam lichtkind: your syntax is correct. 15:05
lichtkind thanks perljam
but ist because the match is so smart or because its executed regularely? 15:06
PerlJam lichtkind: what do you mean?
when and if are almost synonymous
lichtkind im currently writing a perl6 summery in german for our board to convince some people to go into perl6 15:07
i mean this line when $_>40 { laber.. }
Limbic_Region lichtkind - what does your given line look like?
given $foo ~~ $bar ?
lichtkind my question is how ist this line executed
$level > 40 or $level ~~ > 40 15:08
Limbic_Region thinks the given line is crucial to answering the question but is only lurking so proceeds to shut up now
PerlJam lichtkind: $level ~~ ($_ > 40) 15:09
lichtkind: I think
PerlJam doesn't remember if $_ is aliased to an explicit named topic or not. 15:12
oh, it's neither 15:13
when $_>40 { ... } is isomorphic with if $_ > 40 { ... }
lichtkind thanks folks 15:14
PerlJam boolean expressions are recognized at compile time and the smart match semantics don't apply
(according to A04) 15:15
lichtkind your great
PerlJam wonders if pugs does this ... 15:16
15:23 jmcadams joined, jmcadams left 15:32 nothingmuch joined 15:35 hexmode joined 15:39 chris2 joined 15:44 justatheory joined 15:55 bsb joined 16:39 justatheory joined 16:42 autark_ joined
svnbot6 r9783 | fglock++ | PCR - fixed Emitter::Rule::Perl5 compile/runtime separation; %hash is 16:42
r9783 | fglock++ | broken
17:04 ghenry joined 17:09 GeJ joined
svnbot6 r9784 | fglock++ | PCR - fixed %hash (last error was a wrong test) 17:21
17:26 bernhard joined, ruoso joined
ruoso nothingmuch, I'm trying to design a Perl 5 AST (it seems such thing really doesn't exist)... my first sketches are in pX/Common/AST-Design... could you take a look at it? 17:29
nothingmuch, btw... feel free to change it at will... 17:31
17:32 stevan joined 17:37 stennie joined
fglock ruoso: h! 17:50
hi!
17:50 cjeris joined
nothingmuch ruoso: i am not sure I can help with that 17:55
and either way i'm a bit low on TUITs right now
you should poke at rgs and TimToady as they really know they're stuff
their
and also you might be able to ask on perl5-porters and ger a reasonable reply
ruoso nothingmuch, ok... thanks anyway 17:56
nothingmuch other possible victims include sky_, lathos, diotalevi, xmath (we ought to invite him here - Juerd: can you persuade?)
rgs ruoso: you can dump the perl 5 optree with B::Concise, but it's a bit far from an AST...
ruoso rgs, yes... that's why I'm trying to create an AST 17:57
nothingmuch rgs: ruoso is trhing to design an AST for perl 5;-)
rgs I've read the backlog :)
ruoso: have you looked at MAD ?
ruoso rgs, MAD?
rgs larry recently patched bleadperl to decorate optrees with extra syntactic info
ruoso hhmm... 17:58
rgs you need to compile it with the proper option
ruoso rgs, but at this moment, I'm trying to draw how the AST should look like at the end
rgs extra-bleeding edge stuff !
ruoso rgs, I've made some first sketches on px/Common/AST-Design subtree 17:59
nothingmuch ruoso: i suggest you model it after the P6 AST 18:02
but slightly less parse-ish
rgs what's the goal ?
ruoso nothingmuch, the problem is that the P6 AST seems too parse-ish
nothingmuch rgs: IIRC helping the translator efforts 18:03
ruoso rgs, svn.openfoundry.org/pugs/misc/pX/Co...ign/README
rgs which translator ? there are many
nothingmuch ruoso: you really ought to use MAD to write B::AST 18:04
rgs I think so
nothingmuch also, try to reuse code from B::Deparse
or at least learn from it 18:05
the reason MAD exists is that exactly for this purpose: making perl 5 sourcecode accessible to the perl 6 compiler toolchain
ruoso nothingmuch, I'll take a look 18:06
nothingmuch, any docs to point me?
18:07 siosiosios joined
nothingmuch ruoso: not that I know of 18:07
ruoso nothingmuch, :)
nothingmuch i'm only theroretically familiar with perl5 internals ;-)
rgs there are no mad docs, you need to check out bleadperl and dig it. (and ask TimToady for problems)
nothingmuch this means that my guesses are usually right, but I need to poke around to actually make them
ruoso: TimToady also made some YAML MAD trees of most of core perl's perl code 18:08
svnbot6 r9785 | fglock++ | PCR - backtracking into subrules disabled - t/06-subrule.t #20 18:10
18:10 autark joined
ruoso is there any chance of learning more about MAD without recompiling perl? 18:13
nothingmuch not really 18:14
MAD was written against 5.9.2 in larry's own tree, and then merged to blead about a month ago as an off by default feature
Limbic_Region ruoso - TimToady really is "The Man" when it comes to MAD 18:17
there were a few slides and what not but AFAIK, nothing about details
Nicholas might know a thing or two as well now that it is in bleed perl 18:18
ruoso ok...
I'll recompile it at home... /me at work now
ruoso will continue thinking about how the AST should looks like in the meanwhile... 18:21
18:28 autark__ joined 18:37 siosiosios joined 18:49 autark_ joined
svnbot6 r9786 | ruoso++ | another Perl 5 AST sketch 18:51
18:54 nronksr left
ruoso wonders if it is a good idea to have an special "declarations" property in the AST for each "scope" node... 18:57
18:59 nnunley joined
svnbot6 r9787 | ruoso++ | "my" is a possible value for the "visibility" property on a variable declaration 18:59
19:07 autark joined
nothingmuch ruoso: remember that lexical scopes in perl are weird WRT perl 6: 19:09
my $x = $x; # the second $x is outer
$x;
my $x;
$x;
# these are two different Xs
Limbic_Region and no implicit scoping either (last I heard)
nothingmuch so while most scopes in a block end in the same place (call it a scope frame) they don't open in the same place 19:10
theorbtwo Indeed, but it comes up less then it did in perl5.
for my $x (@list) {...} -- in perl5, that $x is scoped to the ....
In perl5 it wouldn't be, but that whole area doesn't exist anymore.
nothingmuch yay! search.cpan.org/~nuffin/Context-Handle-0.01/ search.cpan.org/~nuffin/Config-Pack...l-OO-0.01/
theorbtwo Instead, it's for @list -> $x {...}. That $x is still scoped to the ..., but that's because it's a function's parameter. 19:11
ruoso nothingmuch, take a look at svn.openfoundry.org/pugs/misc/pX/Co...k_decl.out 19:12
nothingmuch, I deal with this by aliasing the variables...
so the first $x would receive an alias from the outer scope
and the second $x would receive another alias from the inner scope 19:13
nothingmuch hmm 19:14
kinda like ANF
where all scoped variables are converted into explicit parameters 19:15
theorbtwo Another Normal Form?
nothingmuch Administrative Normal Form
PerlJam theorbtwo: I was thinking "Advanced" :)
nothingmuch there was an article a while back 19:16
on translating SSA to ANF and back
ruoso really doesn't know the names for such things...
nothingmuch and how it's easier to optimize ANF
ruoso: me neither
=)
SSA is simple though:
every name must be assigned only once
and then "combined" later if applicable
ruoso nothingmuch, hhmmm... that's kind of the idea I had... 19:17
nothingmuch so read about SSA
and read about ANF
and also, please remember to KISS
ruoso :)
what does SSA means? 19:18
nothingmuch single assignment form
or "Static Single Assignment" as the TLA
en.wikipedia.org/wiki/Static_single...nment_form 19:19
that's a good explanation
en.wikipedia.org/wiki/Intermediate_...esentation
i wish that was a category
ruoso it seems ANF stands for Algebraic Normal Form
nothingmuch nuh uh 19:20
two different things
en.wikipedia.org/wiki/Special:Searc...ext=Search
... that's a pretty good search, IMHO
read a bit about all the languages for GCC 19:21
GENERIC, GIMPLE, register transfer something 19:22
19:23 FurnaceBoy joined 19:25 autark__ joined 19:39 penk joined
nothingmuch what will we do for april fools? 19:43
maybe audreyt should change her mind and go back to being a guy?
PerlJam nothingmuch: male -> female -> asexual 19:44
Limbic_Region april fool's is supposed to be fun and I think she takes her gender very seriously 19:45
19:46 autark_ joined
PerlJam Limbic_Region: true. 19:46
ruoso thinks anyone who would understand the joke is probably reading this... 19:47
nothingmuch heh, true that 19:48
okay, what about admitting haskell was a stupid mistake, and rewriting Perl 6 in Perl 5? ;-)
it's plausible
ruoso is actually trying :)
theorbtwo Right, and not funny.
nothingmuch we could argue that lrep et all seem like a much wiser development approach
theorbtwo Admit that it was a mistake, and say that we are rewriting it in brainfuck.
nothingmuch so we could just scrap the whole haskell runtime and that is
april fools needs to at least start plausibly =/ 19:49
Limbic_Region I was thinking something along the lines of something that was funny, unbelievable, but was actually true (or soon to be true)
19:49 Ara4n is now known as Aragone
theorbtwo Exactly. Go several paragraphs before mentioning brainfuck. 19:49
Limbic_Region IOW - Pugs itself would have been a great april fool's joke 2 years ago
Limbic_Region wanders off for a meeting
theorbtwo Announce that The Perl Foundation is suing Audrey. 19:50
nothingmuch hmm 19:51
that's cool
theorbtwo++ 19:52
theorbtwo ...and has sized all his computers.
nothingmuch his?
theorbtwo Point to geeksuntie.
Er, her.
Sorry.
nothingmuch =)
geeksuntie ;-)
theorbtwo wonders why he has returned from the kitchen without dinner.
nothingmuch i could register geeksuntie 19:53
PerlJam and what would geeksuntie have to show people?
correspondence similar to what chip had with his $excompany 19:54
?
theorbtwo A centered font size=400% APRIL FOOL'S!
PerlJam chip might not find that funny at all though
gaal Pugs to replace GCC core in motion to create any-to-any compiler? 20:01
20:07 autark joined
PerlJam plausible but true might involve something with parrot being bundled with firefox, IE, Opera, etc. and perl6/pugs as the new javascript 20:12
gaal plausible but true? :) 20:18
ruoso IE and perl plausible?
PerlJam yeah, I don't know what I was thinking when I wrote that phrase
gaal ruoso: uh, sure, that actually exists already
PerlJam but plausibility is the main goal
gaal ruoso: it's called PerlScript, unfortunately it never was very popular
PerlJam and possibly true soon
ruoso gaal, is it supported by microsoft? 20:19
gaal ruoso: it's an activestate thing
ruoso ah
PerlJam gaal: "unfortunately"? I then you mean "fortunately"
ruoso on the linux world for the last 8 years...
PerlJam Because that leaves the field wide open for us!
In any case, could you see the copy? "Industry and open source leaders unite to endorse parrot as the VM of choice for their respective browsers" 20:21
<some famous javascript guy> is quoted as saying, "javascript? That's old technology! The future is in perl6 and parrot"
etc.
gaal one thing about pugs is that it's not very april fool's worthy... too many unlikely things have already happened (eg. -CJS) 20:22
PerlJam Or simpler might be "Larry Wall announces today that pugs is good enough and he won't be working on the perl6 design anymore" 20:23
gaal but I think -C.NET may be a 04-01 candidate
ah! -CPython :) 20:24
PerlJam gaal: See, that should be the real april fools joke. Just like parrot was an AFJ that came to life, so must they all! :-)
20:27 autark__ joined
Limbic_Region PerlJam - that's what I was saying earlier 20:31
Something that is funny, unbelievable but true (or becomes true)
Pugs 2 years ago would have been a great april fool's joke
also - playboy.com supporting perl by hosting a cpan mirror - funny, potentially unbelievable, but true 20:36
ruoso so the joke is to make people think it's a joke 20:37
meta-april's first 20:38
Limbic_Region well, you add icing that is obviously not true like hustler, in an attempt not to be outdone, donates a million dollars to "the other larry" or some such 20:40
the kernel of truth in it is what keeps people talking about it and perpetuating half truths and generally turning it into an urban legend
*shrug* - I guess I am just tired of the "obvious" stuff 20:41
20:53 fglock_ joined
Limbic_Region is IO::Socket::INET and IO::Select currently available or doable in Pugs ? 21:00
21:07 SamB joined
fglock ruoso: ping 21:09
ruoso fglock, pong 21:11
fglock we and audreyt started Common/lrep/Notes-Pugs - it has an (incomplete/outdated?) description of the plans for lrep 21:12
s/we/me/
ruoso looking 21:14
fglock, it's close to what I was thinking in the last days... 21:17
fglock, the difference is that I was thinking in a few more steps 21:18
like
PCR today returns a match object 21:19
a match is not an AST
it's too much parse-oriented
we could have something to turn a match into a real ast
and have this to Perl 5, Perl 6 and PIR 21:20
then we can have AST translators
from Perl 5 AST to Perl 6 AST and vice-versa
from Perl 6 AST to PIR and vice-versa
and emitters, that take the AST and emit real code 21:21
so...
fglock did you see TGE in Parrot?
ruoso not in deep 21:22
but it seems a good tool to implement the translators 21:23
maybe it's just *the* tool 21:24
but I think we'll have AST -> AST translators 21:25
Limbic_Region tries to refrain from calling any a tool as it is unflattering but being called *the* tool is a compliment
fglock the place to write about the P6 AST is docs/Perl6/API/AST.pod - how about talking to audreyt and TimToady and start working on it 21:27
21:27 larsen_ joined
ruoso fglock, I started a AST-Design subtree 21:27
in pX/Common
21:28 YetAnotherEric joined
ruoso but I'm working on a Perl 5 AST first 21:28
so we can know to what Perl 6 AST should be translated to in order to host perl 6 in perl 5 21:29
and also, my feet are much more on the ground with Perl 5 than with Perl 6
21:31 KingDiamond joined
ruoso as I know better Perl 5, it's easier to write the AST... 21:32
and it can be a base to write the AST for Perl 6 later
and TimToady can help without the fear of making Perl 6 the same as Perl 5 :)
fglock actually, my idea was to make pugs-p5 mostly the same as pugs-haskell - except in a few areas where we can implement new ideas or simplify things 21:35
svnbot6 r9788 | ruoso++ | Documenting a possible Perl 5 AST
ruoso fglock, you mean not emitting Perl 5 code but implementing a interpreter? 21:37
fglock I mean reuse pugs AST 21:40
ruoso: btw, did you see the perl5() method in PCR? 21:42
ruoso yes 21:43
but I realized that the Match is not the AST
and started thinking in the AST
fglock the capture is the AST - $match->()
ruoso and then realized that Perl 5 doesn't have an AST
fglock, the capture is too much parse-oriented 21:44
it must be transformed into something more syntax-oriented
I mean, sematic-oriented 21:46
svnbot6 r9789 | ruoso++ | the root node of the AST is always a package 21:47
ruoso fglock, see AST-Design/Perl5/block_decl.out 21:48
and AST-Design/Perl5/block_decl.in
s/sematic/semantic/
fglock btw, you can see pugs AST with: pugs -CPIL -e ' my $a = 1; print $a ' 21:55
-CPerl5 returns the AST as a P5 object 21:56
22:00 zhobson joined 22:40 ruz joined
svnbot6 r9790 | fglock++ | PCR - improved rule/subrule communication; non-working tests for subrule 22:48
r9790 | fglock++ | backtracking
nothingmuch today's 2 hrs: pasta.woobling.org:3000/list/ 23:03
(tags, broken auto completion)
oh wait, i fixed auto completion 23:04
svnbot6 r9791 | ruoso++ | documenting the Perl 5 AST... several problems to solve... what about globs? 23:17
23:54 KingDiamond joined