6.2.2 is airborne! | pugscode.org <Overview Journal Logs> | pugs.kwiki.org
Set by autrijus on 1 May 2005.
robkinyon any lamda'ers around? 00:57
Juerd I'd pretend to be one, but I think you'd notice the difference very quickly. 00:59
robkinyon is dragonchild, fyi
i am trying to fix a cygwin bug
Juerd How about epsilon?
robkinyon i think i know how to do the haskell changes, but i'm leery
Juerd Heh - I hadn't noticed before that kinyon == dragonchild. 01:00
Funny. :)
robkinyon arghs
be back soon (wife called)
Juerd Why not just change and commit, by the way?
If it's wrong, it can be reverted.
That's the most important of version control wonders :) 01:01
3am here
Time to get some sleep
Good localtime!
afk
stevan robkinyon: how big is the change? 01:02
meppl gute nacht 01:11
stevan I hate it when I do something cool, and no one is awake to share it with 03:17
for anyone who is awake,... I just got inside out objects to work :)
horray for OO without OO :)
see ext/Tree-Simple
although since I cannot bless anything, it kind of fragile 03:19
(my "opaque" object is an Int)
robkinyon what did you do? 03:21
stevan LOL
hey Rob
robkinyon heyla
stevan inside out object :)
robkinyon still haven't fixed my doesFileExist to work with cygpath yet
the change will be like 4 lines
to Prim.hs
stevan line number?
robkinyon lemme make sure my copy is latest 03:22
stevan ok
robkinyon just search for doesFileExist
there's three instances
look for the one in the op for require
line 204 03:23
stevan ok
robkinyon I think I can add a line before it that looks something like:
jabbot pugs - 2842 - adding Tree::Simple inside-out object; H
robkinyon if ( $OS eq 'cygwin' ) then pathanem <- cygpath pathname 03:24
where cygpath is some function that calls out to the cygpath program
s/pathanem/pathname/
but, i have no idea how to make that work. 03:25
and, since this bug is preventing the test suite from running, i can't verify my change
stevan well line 309 ( the system() built in ) should give you the code to call out 03:26
boolIO system "cygpath"
maybe?
(BTW - I am totally cargo culting here)
robkinyon basically, i need to convert the following to haskell:
chomp( $pathname = `cygpath $pathname` ) if $^O eq 'cygwin'; 03:27
everywhere in src/* that doesFileExist is used.
(I'm pretty sure other functions beside doesFileExist will need that treatment, but that's as far as I've gotten.) 03:28
stevan Pugs/Compat.hs is I think where stuff like that goes
robkinyon ok. 03:29
robkinyon reads Compat.hs 03:30
stevan getConfig "osname"
that seems to get the value put into $?OS
line 206 of Eval.hs 03:31
robkinyon Ok ... I should be able to do something like os <- fromVal =<< readvar "$?OS" ... right?
stevan uhm
sure
only one way to find out :) 03:32
robkinyon that how incs is done on line 196 of Prim.hs
stevan - I. Don't. Have. A. Test. Suite.
stevan robkinyon: if it compiles, then you know you are closer then nothing :)
robkinyon *laughs*
stevan if you really want, you can nopaste and I can try 03:33
but it wont tell you if cygwin worked :)
robkinyon i'll try it
TIAS and all that
here goes nuthin 03:38
robkinyon sets the elevator music timer for 20 min.
stevan :) 03:39
good luck robkinyon, time for me to go to sleep 03:52
adios &
jabbot pugs - 2843 - improved the inside out instance; it now 03:53
gaal good morning. 03:54
robkinyon gargh 03:55
gaal hi, rob. i encountered the weirdest of bugs the other day. 03:56
robkinyon what's up, gaal? 03:57
gaal a little out of things recently :/
do you happen to have more than one cpu on your box? 03:58
robkinyon nope. 03:59
i barely have a complete hamster inside the thing
robkinyon is giving up
gaal oh, you've been doing cygpat in haskell? i can imagine that's tiring :/ what have you tried? 04:00
btw you might make use of Pugs::Internals::runCommand which is in Prim.
robkinyon how I do convert the following Perl?!? 04:01
if ( $^O eq 'cygwin' ) { $pathName = `cygpath $pathName` }
that's literally the whole change on line 206 of Prim.hs
jabbot pugs - 2844 - changing $instance to $self because its 04:03
gaal first separate it to newpath = $*OS ~~ 'cygwin' ?? "moose" :: oldpath 04:04
so you know you got the basic if syntax right
do you know trace? took me a while before i heard about it? 04:05
robkinyon nono - not in P6
gaal yes i understand
robkinyon I'm fixing a bug within Pugs
robkinyon is confused
gaal trace in haskell is for debug printing. helps some. :)
look for example usage in Eval.hs
robkinyon ok ... 04:06
i think i'll come back to it tomorrow 04:07
thanks for the tip
gaal can you nopaste what you do have? i'll see if i can get it in.
perlbot nopaste 04:08
perlbot Paste your code here and #<channel> will be able to view it: sial.org/pbot/<channel>
robkinyon what i have doesn't compile
gaal that's okay
robkinyon no - i've spent an hour trying to get that P5 snippet I showed you to compile in Haskell
I haven't gotten past "if then x = y"
gaal welcome to the club :) 04:09
stevan isnt haskell fun :)
gaal you don't have x = y in haskell. variables are immutable.
robkinyon immutable
gaal yeah, which makes the name variables, uh, interesting! 04:10
robkinyon i can't just replace the value stored in the variable?
gaal not with usual vars, no
robkinyon gaah!
i really need to just go to sleep now, before I do something to the stupid Prim.hs i'll regret. 04:11
gaal "that does put a damper on our relationship", as another Rob would put it.
robkinyon oh, well. 04:12
there's always tomorrow
gaal sec, before you leave
you wanted to cygpathify "pathName", right?
robkinyon yah? 04:13
yes
gaal right after the let? or only if ok is false?
robkinyon actually, i need to cygpathify anything that gets send to doesFileExist
i don't care when it happens
gaal okay.
(we should probably assimmilate cypath into pugs, really.) 04:14
(but that's step #3)
robkinyon i have pugscc "working"
gaal cool!
robkinyon (insofar as pugscc actually calls the right stuff)
it still crashes, but I think that's cause of other cygwin issues
i just want to get "use Test;" to work before going anywhere else
gaal sounds like a good idea :) 04:15
robkinyon does this mean you'll patch doesFileExist?
gaal i thinki i'll change
let pathName = 04:16
to pathName <- cypath $ ....
where cypath is a noop on non-cywin platforms.
robkinyon could you change doesFileExist instead?
or, rather, wrap it?
and have something in Compat.hs to DWIM
actually, ALL the filetest actions need to be wrapped in Compat.hs 04:17
for example, "-e /tmp/foo" will fail unless cygpath-ified
gaal shelling out to cygpath all the time is insane though :(
robkinyon and it's the only solution GHC provides
gaal but as i said, step #3.
robkinyon cygwin users expect a 20-50% slowdown in everything they do, anyways 04:18
apache, mysql, perl ... it's all slower on cygwin
if pugs were different, you'd be giving people conniption fits! :-)
gaal :) 04:19
robkinyon cool - thanks gaal
i appreciate it
gaal sure thang. i'm a novice haskeller though - i only think i know what i'm doing :) 04:20
robkinyon heh
g'nite 04:21
robkinyon waves
gaal bye.
05:02 saorge_ is now known as saorge
gaal i just realized this cygwin compat thing can't happen at compile time :( this is not good - it means a runtime hit on all platforms, or at least we can limit it to all windows platforms, with some ugliness. 05:09
also, the current $*OS is wrong. 05:11
it is what should be in $?OS.
and now i have to go to $work :/ 05:12
xinming anyone here can tell me where to get online doc for perl6 please? 05:27
revdiablo search.cpan.org/~ingy/Perl6-Bible-0.16/ 05:28
xinming revdiablo: thanks 05:32
I've download the archive. could you please tell me how to use it? :-S I extracted it, But seems nothing useful. :-S 05:41
./updater doesn't work
Khisanth probably install it and then use perldoc 05:45
bsb Hi, can anyone tell me if the "-- #include <sys/types.h>" in src/RRegex/PCRE.hs is a pragma or just a comment? 05:57
integral bsb: I'm pretty sure that's a comment 07:31
bsb Thanks integral, haddock thinks it's a directive 07:36
jabbot pugs - 2845 - haddock compatibility tweaks 07:44
pugs - 2846 - change cut-n-pasted test comment
nothingmuch i love good uptime 07:56
unminimizing windows you opened a week ago is a nice way to continue where you left off
ingy hola 08:10
Corion Hmmm. I don't understand the backlog about cygpath. From how I understand cygwin/cygpath, any (cygwin-)native application can use /tmp and /etc and /whatever directly, and the cygwin1.dll translates. If Pugs is compiled under cygwin, there should be no need to munge any paths further, unless the cygwin setup is wrong. 08:36
gaal|work Corion, a cygwin pugs doesn't depend on cgdwin.dll 08:40
it uses the msys ghc and all 08:41
just before work today i realized we got $*OS all wrong for cygwin.
the main advantage of building with cygwin and not msys is that you have a reasonble perl(5) around. 08:42
Corion gaal|work: Still, the path things should Just work
What's the matter with $*OS on cygwin?
gaal|work but you get the same pugs you'd get building under msys.
$*OS should be the one you're running on, not the one you built on. 08:43
Corion gaal|work:Oh. Yes.
gaal|work actually $*OS is wrong everywhere.
Corion Well, partly. If you're built for cygwin, you'll always be cygwin, even if you run on Win32
gaal|work it just happens to be accidentally right alsmot everywhere too. 08:44
Corion (because on Win32, /tmp has no meaning, while on cygwin, it has)
gaal|work Corion, no: a cygwin pugs is not really a cygwin pugs.
it's the same as an msys pugs. 08:45
Corion gaal|work: Ah, it was merely built on cygwin, but is "native", as ghc outputs non-cygwin code?
gaal|work yes.
Corion Ugh.
So we should ditch cygwin. I don't want to deal with the cygprefix stuff.
gaal|work see www.haskell.org/ghc/docs/5.04/html/...build.html and www.reed.edu/~carlislp/ghc6-doc/use...11221.html 08:46
Corion (what good is building on cygwin, if you don't get a cygwin executable?)
gaal|work its significantly less sucky perl5.
("less sucky" than msys')
Corion Hmmm. ActiveState has a good enough Perl for building Pugs :)
If you're end up with a native Win32 executable anyway, why not go all-native? 08:47
gaal|work i don't know that they do, Corion. Does the smoke toolchain have no XS dependencies at all?
Corion Not that I know of. But then, these are all transparent to me, as I have a working C compiler setup here... But thinking about it, the fanciest thing the build uses is Module::Install, which isn't XS 08:48
gaal|work petal requires mkdoc::xml or something like that 08:49
i don't remember if it's xs.
htmlmatrix wants petal
make smoke wants htmlmatrix.
Corion Hmm, that's true - Petal also is fancy
gaal|work if someone ports petal to p6, great.
Corion ... and mkdoc might be ugly. It's been a long time since I installed it ;)
gaal|work but that isn't easy. 08:50
Corion True, we should get a templating system ported ;)
Well, it would be easier if we had rules (as Petal needs a compiler anyway :) )
gaal|work don't we almost have them already?
i've been out of things lately. 08:51
Corion I don't know, I've been out of things as well (or rather, into different things)
gaal|work the latest mention is use.perl.org/~autrijus/journal/24571 (6 May) 08:53
looks like the ball is in pm's hands
but it's not like i'm running off to port petal myself right when that's done. 08:54
so: you're right, cygwin isn't really a target platform at all. it is, however, useful to be able to build pugs on.
Corion True... But I want the path munging in one central place, and I'm not even sure it should happen in Haskell... Fudging @*INC at BEGIN time (or Just before it) should be enough ... 08:55
gaal|work well, dragonchild was missing -e </tmp/x>; and the pugscc script is broken.. 08:56
anyway, i'm off to lunch.
see you later :) & 08:57
jabbot pugs - 2847 - * EOLs at EOFs in ext/. 09:44
nothingmuch has a vegetable overdose 10:16
jabbot pugs - 2848 - mklivecd.pl now copies our Perl 6 module 10:24
pugs - 2849 - Changed LWP::Simple to use HTTP/1.0 inst 10:34
pugs - 2850 - * Added svnbot.p6, a bot which informs # 11:44
iblech ?uptime 11:47
:) 11:49
scw pulling 504 revisions 11:51
gugod heh 11:53
jabbot pugs - 2851 - This is a dummy commit to test if svnbot 11:54
lightstep is svnbot really fast? 12:00
iblech it checks for new revisions every 240s (but you can change it using @*ARGS). *But*: Because readline() is blocking, svnbot6 has to wait for server input before it can check for new revisions 12:03
lightstep so why not immediately? that's cause it to block until there's a new revision 12:04
iblech Therefore, I sent it to #enrc.wikipedia, too, as that's a channel with high traffic
It has to control two things -- the IRC socket and svn. In the IRC handling code of Net::IRC there's a readline(), which'll block until the server says something. svn_check() is called after each readline() 12:05
In Perl 5, I'd solve this problem by either using POE or by using %*SIG<ALARM> (but %*SIG doesn't work yet in Pugs)
lightstep stm can be used to solve this. but i don't know is autrijus exposed the stm primitives to pugs 12:07
iblech he did not yet, IIRC 12:08
lightstep sequence . repeat $ readIRC >>= handleCommand `orElse` waitForSvn >>= showRev 12:09
meppl guten morgen 12:28
Juerd moin 12:49
iblech ?uptime 13:13
jabbot pugs - 2852 - More Haddocks for Bind 13:14
scook0 svnbot6++ :) 13:17
13:52 chady__ is now known as chady
luca32 hi everyone there! 14:07
how long it will take to perl6? 14:08
sorje 20 minutes ;-)
luca32 come on!
really
what's the status of parrot
and of pugs?
misc_ pugs works 14:09
luca32 is it complete?
misc_ luca32: see svnbot6 , a perl6 irc bot
luca32 cool!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
sorje have a look at svn.openfoundry.org/pugs/examples/ 14:10
luca32 how many people will join tha austrian acaton
and parrot?? 14:11
i guess perl6 is becoming the best language
it will superceed c#
and java
autrijus is he online now? 14:12
how can i contribute?
i have very strong skills in c,c++
plus some skill in perl5 14:13
misc_ luca32: you need some skill in haskell
luca32 and strong skills in .net and c#
misc_ luca32: or write doc, and test
luca32 well
actually i don't have time to learn haskell
everyone here knows haskell
?
gaal|work misc_, hi, sorry about the disappearing act from yesterday. did you manage to get the smoke running?
misc_ gaal|work: no 14:14
gaal|work: but it hink the problem comes from my side
gaal|work got tuits to try now?
luca32: not everybody knows haskell
misc_ ok
luca32: but you can write doc and test
pjcj or hack on parrot 14:15
skew luca32: with embed_haskell you'll have to learn it eventually if you want to know all of perl6 :)
luca32 what's the state of parrot 14:18
whish it was 2003
i believe parrot is a too mature project right now 14:19
it would be interesting to start instead a parrot.NET
port
gaal|work misc_, the problem still happens for me on my (SMP) work machine, but i'd be glad if you could confirm this.
skew I'm not sure if that's really in the standard, though Larry Wall promised inline Haskell while Perl 5 was being developed: www.dcs.gla.ac.uk/~partain/haskerl/wall-1.html 14:20
gaal|work look at the date :) 14:21
misc_ gaal|work: i still have problem with cpan install Test::TAP::HTMLMatrix, even if i managed to make it work as simple user 14:23
no, maybe i didn't managed to use it as simple user :/
gaal|work i wonder if that might be a bug in Module::Build :)
misc_ i am starting from scratch on a nother server 14:25
gaal|work oh, crap - now it happens even w/o the UPLOAD command thing. 14:27
misc_ gaal|work: still doesn't work, give the same type of error : 14:29
t/comprehensive....Operation `eq': no method found,
gaal|work did installing the Model module work? 14:35
misc_ Test::Model ?
gaal|work test tap model
misc_ Test::TAP::Model doesn't install on the two server i tested 14:36
gaal|work oh, in that case the htmlmatrix for sure won't work.
why doesn't it install thoguh?
misc_ gaal|work: pastebin.com/281282 14:38
gaal|work misc_, just so i'm sure we're on the same subtest here, could you please run from the command line: 14:40
prove -vl t/comprehensive.t
you have to be at the base dir of Test::TAP::Model for this. 14:41
misc_ test 22 fails
pastebin.com/281284 14:42
gaal|work: it is a amd64 server without smp this time, but i doubt it change something
gaal|work no, this bit i don't think has anything to wiht smp.
(btw, 22 is passing; 23 is dying.) 14:43
oh 14:44
weird
i'm getting it too now
hmm
well, that makes it easier to debug at least :)
hcchien seen autrijus 14:50
jabbot hcchien: autrijus was seen 1 days 12 minutes 21 seconds ago
Juerd Let's put sigils at the *end*! 15:21
It solves all the sigil-part-of-the-name problems, and it's retro-ish (think BASIC) 15:22
:P
lightstep some scheme systems have those 15:23
jabbot pugs - 2853 - * Added more documentation to Net::IRC. 15:34
Juerd I'm seeing double 15:42
I like svnbot's mentioning of the username
I like jabbot's using a normal privmsg
Can that be combined into one bot?
(And reduce waste of screen estate, s/ --//) 15:43
iblech Juerd: Bots should always use NOTICE, not PRIVMSG, according to RFC 1459 (the IRC RFC)
jabbot pugs - 2854 - Added more comments to svnbot.p6. 15:44
iblech (But if we want to change, it's a matter of changing one "notice" to "privmsg".)
lightstep haskell.org/hawiki/LambdaBot_2fNotices 15:45
Juerd I know the RFC
lightstep this page contains solutions
Juerd And prefer in this case to radically ignore it.
Oh, and I like that svnbot.p6 is more realtimeish
iblech: I think it's wise to follow the de facto standard rather than the 12 year old official one 15:46
iblech: Note too that bots have become more user-ish over time
iblech: While at the time rfc1459 was written, bots were for channel management only, not for providing information 15:47
iblech Juerd: FWIW, RFC 2812 ("Internet Relay Chat: Client Protocol") was released April 2000, and says the same. 15:48
Juerd: But I don't have a very strong opinion on this. 15:49
Juerd Also, it is said to be like this to avoid infinite loops
Which is a problem only if the automated message was in reply to an IRC event 15:50
iblech True.
lightstep there could be a cycle (perhaps svn listening to irc?) 15:51
Juerd lightstep: I think we have to rule out that possibility for sanity's sake. 15:52
iblech Or a bot which commits something when it sees certain svn events.
But as you say, Juerd, sanity :)
?quit Moving svnbot6 to a 24/7-online machine 16:01
Juerd autrijus: You asked for registration of my slides or talk 16:02
autrijus: I'm afraid the only thing available is a photograph: amsterdam.pm.org/meetings/reports/2...1a955.html 16:03
:)
I believe that the page shown was the fifth
gaal|work a bot that commits a fix every time someone reports a bug? 16:08
Juerd Now that would come in handy. 16:09
You write it
gaal|work easy. someone just spec 'use DWIM' first. 16:10
Juerd It's going to be a builtin, not a module 16:11
dwim();
jabbot pugs - 2855 - Oops! Because of a too early return(), $ 16:14
Juerd Yay for parrot 0.2.0
robkinyon hey gaal - did you make that fix for cygwin we talked about? 17:01
ingy hola 17:13
gaal|work: I got your email. Adding your test to the suite which I am currently refactoring. 17:15
jabbot pugs - 2856 - Net::IRC will now discover the bot's use 18:24
iblech BTW, lambdafolks, it'd be cool to be able to connect() to IPv6-hosts, too 18:32
FYI, if svnbot6 goes crazy, please stop it by MSGing "?quit" 18:34
jabbot pugs - 2857 - * compile sub calls with no arguments to 18:44
theorbtwo Allo, all! 18:52
Juerd Hi 18:53
luca32 autrijus is here? 20:16
are y here?
is there someone who has news about the parrot hack event in Vienna?
how many people have already booked? 20:17
pjcj I don't think it's generally open. It seems to be sort of semi official at Leo's house. But I might be completely wrong. 20:18
luca32 it's open
there is also a website about it 20:19
it's only necessary to subscribe
who's leo?
pjcj OK, then I am completely wrong.
Leo is the release manager for parrot.
luca32 i have not decided yet
so i would like to know if i am still in time
leo is austrian? 20:20
pjcj yes
where is the website?
luca32 i have to take a loojk at my gmail account
There will be a Parrot/PUGS Hack-a-thon at the Austrian Perl Workshop, which 20:23
> takes place on 9th and 10th June in Vienna, Austria.
conferences.yapceurope.org/apw2005 20:24
pjcj Ah yes. In that case I refer the right honourable gentleman to my previous answer. 20:25
But maybe you should contact the conference organisers. 20:26
luca32 pjcj are u involved in parrot/pugs?
pjcj I mostly lurk on irc and mailing lists. 20:27
luca32 :P
k
are u austrian?
pjcj Wating for my supply of appropriately shaped tuits.
No, I'm English.
luca32 k
must code something 20:28
i am going
see ya
by the way
pjcj tchuss
luca32 i am a phd in computer science
right now i work for a multinational consulting firm
which business is MS oriented
pjcj I suggest looking into parrot.
luca32 but my roots are in unix systems
pjcj No need to be coy ;-)
luca32 during my spare time i enjoy doing some real engineering stuff 20:29
and not IT
:P
pjcj parrot 0.2.0 was just released today
electronic engineering?
luca32 now i am coding a plug in for lightwave that does mesh unfolding
pjcj or the real real stuff
luca32 hope to release soon it
after i finish this i would like to be involved in some other challenging project 20:30
i was thinking about a .net port of parrto
parrot
pjcj as in parrot running on .net or the other way around?
luca32 parrot running on .net 20:31
pjcj interesting - are you on the parrot mailing list?
luca32 maybe someone could be interested in stuff like this
yeah
but never posted
pjcj there were some people who were into .net, but I don't remember them posting recently
(or their names, I'm afraid) 20:32
actually, they were mono people, but it's all the same stuff, right? ;-) 20:33
luca32 more or less 20:34
pjcj I'm currently chasing a bug in XS code, but I'm about to give it up for the night.
luca32 k 20:35
pjcj good luck with parrot, if you decide to work on it
luca32 k
but i won't approach it before 1 and a half months 20:36
now i am begging to lurk
pjcj I'm sure it'll still be there ...
luca32 sure 20:37
good night
pjcj night
Corion reads the backlog and wonders ... 21:07
theorbtwo Allo, Corion. 21:08
Corion Hi theorbtwo ! Back from the UK already?
theorbtwo Yep. 21:09
Got in this morning.
Touchdown was around... 12:30, local.
We're sick, though. 21:10
Corion theorbtwo: Ah, so you've been back some time already.
Ugh - why that? The british cuisine?
theorbtwo The London weather had more to do with it, I think, but some of that too. 21:11
crysflame what is "boiled frog treatment"? 21:12
Corion Oh - so you've brought a cold back ... I thought you had visited castaways sister? Or is she in London?
crysflame er, hello #perl6
theorbtwo Two of castaway's sisters live in greater london, but we didn't see either of them.
We went to a london.pm social, and met broquaint, and did a lot of walking about the city. 21:14
We almost stopped in at the work of one of her sisters, but callled and found out that she's at castaway's parents for the weekend. 21:15
clkao waves at theorbtwo 21:16
theorbtwo Allo, clkao.
clkao don't think we actually met in the social :P 21:18
Corion Ah - I didn't know she had two - I only knew of the one that recently married. But you met a lot of people, that's nice! 21:19
Corion goes to bed now.
Get well soon theorbtwo (and castaway too! )
theorbtwo Hmm, I don't remember... 21:21
There were a lot of people at the social, though, and I'm not very good at being social with large groups of strangers.
...and it was very noisy; I had problems hearing. 21:22
But I had lots of fun anyway.
Oh, you aren't in AUTHORS. What's your real name, clkao? 21:23
clkao i write svk
mugwump ah, is that the literal translation of "kao chialiang"? :) 21:27
clkao umm :p 21:28
mugwump theorbtwo: just call the hotsprings operator on him; pugs.kwiki.org/?PugsOperators 21:29
stevan: around? 21:32
autrijus How are you lambdacamels! 21:39
mugwump morning autrijus ! Up early/late?
autrijus early.
mugwump nice
autrijus I'm dayshifting to a sleep-at-8pm, wake-at-5pm cycle
er.
s/5pm/5am/
mugwump that must be close to sunrise 21:40
autrijus yup 21:41
gaal hey all
lightstep 9 hrs is lotsa fun
gaal ingy: thanks :)
robkinyon: ping
clkao actually i'm in pugs AUTHORS.. i wonder why 21:43
elmex hmpf 21:54
autrijus clkao: maybe because you reported three bugs, all of which resolved? :) 21:57
clkao oh did i
i was only complaining 21:58
:)
howdy autrijus. is svk.elixus.org slow for you?
autrijus bearably so
i.e. I'm fine
clkao i found i was added at r32
Juerd autrijus: Seen the "slide"? 21:59
autrijus Juerd: no. url?
Juerd 18:06 < Juerd> autrijus: I'm afraid the only thing available is a photograph:
amsterdam.pm.org/meetings/reports/2...1a955.html
autrijus heh cool! 22:01
Juerd++
ok, so, release plan for 6.2.3.
I propose we wait until this Sunday, that is 6 days from now.
because I want to get rx// and object support in :)
and the tree had not been stabilised enough to make an easy snapshot release. 22:02
autrijus checks regressions
hm. my impression was wrong, we only have 6 or so regressions 22:03
mugwump object support!
autrijus let me take a look to see if I can fix them quickly 22:04
mugwump there is a plan for this? I saw the closure state `hack'
autrijus mugwump: there's a plan for this, based on your Class.hs 22:05
but I want rx// support first (because it's easier) :) 22:06
mugwump via PGE?
autrijus aye
mugwump awesome
autrijus what's more, rx// support will be there even if you don't build against libparrot! 22:07
autrijus cackles evilly
autrijus hacks that in
mugwump pulling another IMC?
autrijus no, running external bin/parrot
;)
mugwump oh, evil 22:08
obra hello 22:28
23:08 hawkalooogie is now known as hawkaloogie
jabbot pugs - 2858 - * New, improved, parrot-based PGE 23:24
stevan mugwump: pong 23:27
autrijus rx// support rsn.
stevan autrijus: hello :) 23:28
rx// requires parrot though right?
obra hey autrijus
autrijus yo obra 23:30
stevan: an installed parrot, not libparrot
it will be faster (no system() overhead) if you link against libparrot.
stevan ok 23:31
stevan goes to download/install parrot
mugwump hey stevan, I was just reading your message re: .isa on built-ins and thought that perhaps it would be a good idea to prototype the behaviour
stevan mugwump: what are you thinking? write some tests? 23:32
or prototype it in perl5?
mugwump well, we've got 'objects' of sorts, implemented by closures in Pugs, right?
stevan mugwump: and I hacked some inside-out invocant based object stuff last night too 23:33
mugwump coo, is that in pugs?
stevan ext/Tree-Simple/
since we dont have proper namespaces or bless, it is a little delicate
basically new() returns a random string, which is the hash key to an %INSTANCES hash inside the package 23:34
it does work though, and even for recursive data-types like Trees 23:35
mugwump is the trick used in svn.openfoundry.org/pugs/examples/cashiers.p6 not ok? 23:38
oh, it's using hash subscript rather than .
stevan autrijus: are there instructions for building libparrot?
mugwump: either one is fine with me, I did the inside out object things as a proof of concept 23:39
mugwump sure. I'd prefer an interface that doesn't require me to call 'methods' with <> though :) 23:40
stevan mugwump: yes it makes it more future-proof too
although I didn't figure out inheritance with it yet
it would have to be pretty manual 23:41
mugwump For these purposes it doesn't matter
stevan ok 23:42
autrijus stevan: set the PUGS_EMBED env to parrot
stevan: maybe write it up on Makefile.PL or README or somewhere
mugwump right, well what I'll start on is translating the Class.hs into Tree::Simple-style OO ... to at least have a prototype metamodel in Perl 6 23:43
stevan mugwump: I have not looked at Class.hs
stevan goes to look right now 23:44
autrijus: I will document my parrot experience somewhere
probably README
mugwump: src/Pugs/Class.hs?
mugwump nods 23:45
Juerd autrijus: I had to "make clean" in order for the procedure to pick up that I wanted parrot embedded 23:47
autrijus Juerd: nod
mugwump: ooh! that will r0x0r 23:48
mugwump ok, how does ext/Perl/MetaClass.pm sound for a location? :)
stevan mugwump: how about ext/Perl-MetaClass/lib/Perl/MetaClass.pm 23:49
then we can add tests and such
mugwump: Class.hs goes /way/ over my head
but I will help out any way I can
tests are what I am good at :) 23:50
I can grok how to use things, and how they should be used, but the underlying theory is not my thing
mugwump this is basically the schema for those .meta objects mentioned in S12 23:51
stevan actually the further down i get, the more I understand :) 23:53
mugwump excellent!
how does the MMD know that the '$self' you return is a Tree::Simple object in Tree::Simple? 23:55
stevan it doesnt
thats the fragile part
we have no proper namespaces
mugwump right, so these methods are all clobbering each other for now
stevan potentially 23:56
mugwump ok, so what about possibilities with 'magic types' ... 23:57
mugwump looks for MMD typing tests
stevan what do you mean 'magic types'?
blessed ?
mugwump: I think MMD would work in some cases 23:58
mugwump ie, returning some 'unusual' or non-sensical type that is used for MMD
(on the invocant)
stevan multi sub foo ($self: Int $id)
multi sub foo ($self: Str $id)
that should DWIM 23:59
mugwump sure, but sort of like:
multi sub foo ($self: Str $id where exists %INSTANCES<id>)
stevan mugwump: $self actually is the key into %INSTANCES

Powered by Cro and the Raku® Programming Language.

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

Thank you!