SVK users: search.cpan.org/dist/SVN-Mirror/ | paste: sial.org/pbot/perl6 | pugs.blogs.com | pugscode.org | pugs.kwiki.org | www.treehugger.com/files/th_images/paradigm.jpg
Set by audreyt on 17 August 2006.
00:01 araujo joined
araujo hello :-) 00:01
markstos_ hello 00:02
araujo hi markstos_
araujo just read a bit about perl6 and was interested to take a look at what the community has so far :-) 00:03
markstos_ araujo: You can try it out right here in the IRC channel: 00:04
?eval say "hello world!"
evalbot_12700 OUTPUT[hello world! ] Bool::True
araujo oh nice :-)
markstos_ Yeah. I just used to post an OO problem I'm having: 00:05
?eval class Foo; has %!PARAM is rw; submethod BUILD (%param) { %!PARAM = %param; }; Foo.new( a => 'b' );
evalbot_12700 Error: Undeclared variable: "%param"
markstos_ You can see that some things look different!
araujo well, yeah, i kind of remember a few things from perl5 ....
i got quite a time without using it 00:06
markstos_ Here's a nice starting point for documentation: feather.perl6.nl/syn/
lambdabot Title: Official Perl6 Documentation
araujo Ive been some kind of interested, since pugs is being written on Haskell
araujo is a lambda haskell folk
markstos_ Ah.
araujo :-} 00:07
oh, yay, we got lambdabot here too!
markstos_ There's definitely still Haskell programming let to do if you are interested. :)
araujo Thanks markstos_
i will check the docs
markstos_, sounds cool to me :-)
TreyHarris markstos_: well, to start with, evalbot requires that you use class Foo { }, because it's not a compilation unit
markstos_ And I'm qualified to give you a commit bit if you find something you want to hack on...
araujo isn't a perl 'monger' though 00:08
cool markstos_ , thanks!
markstos_ TreyHarris: Ah. I still get the error with that {} syntax, though. 00:09
araujo wouldn't mind to read the code and report typos , small problems though
markstos_ araujo: Great, thanks! Your help is appreciated. I know Perl but not Haskell, so I help on other parts of the project. It takes all kinds! 00:10
Juerd Seeing police cars is frightening if you're doing something illegal, even if you're quite certain they're after someone else :) 00:13
araujo markstos_, Cool!
markstos_, it sounds like an interesting project indeed.
It's good to see two programming languages community working together :-) 00:14
markstos_ araujo: yes! I'm not a Haskell programmer, but I'm big fan and user of darcs, and have appreciated what I've learned about Haskel far 00:15
Perl6 has been fun for me to learn about. I've been blogging about some of recent experiences here:
use.perl.org/~markjugg/journal/
lambdabot Title: Journal of markjugg (792)
markstos_ Juerd: Have you been watching DVDs on Linux again ?
Juerd markstos_: Heh, no. I've been taking pictures of highway road signs.
TreyHarris markstos_: well, another thing is you need a slurpy hash there. otherwise you need to pass in an actual hash. so submethod (*%params) 00:16
markstos_ And that's illegal?
Juerd It's kind of illegal to have 0 velocity on a motorway :)
araujo markstos_, nice , i check it
Juerd Can't take good pictures while driving.
markstos_ TreyHarris: Thanks. I had begun to figure out I need the * from another example, but i hadn't figured out why yet.
araujo I used to program on perl. But i have quite a time without touching it (since 2003 around)
Juerd araujo: Don't worry; Perl 6 is sufficiently different in internals that it wouldn't help you anyway. 00:17
araujo: Only the outside is clearly coming from Perl 5
araujo I have read a bit about this new perl6 , and it practically seems like a 'new' language from the look.
Juerd It is
araujo Juerd, oh, yeah, so my first impressions are sort of true then :-) 00:18
Juerd It's a new language, that borrows from many existing languages, but mostly Perl 5.
araujo Cool!
Juerd Many of the ideas you find in Perl 5 are still there. But the syntax is very different in many ways. Still, though, it looks like Perl.
markstos_ araujo: I suspect that even some influences of Haskell have snuck into it...
araujo yeah, i get it.
markstos_, really? ... such as? 00:19
markstos_ I actually findit more different than the same.
araujo: Oh, now you've called me out.
Juerd markstos_: Compare it to Python or Ruby, and you'll find that Perl 6 is much more like Perl 5 than those languages are.
markstos_ I have to recall what I what I thinking of. :)
I just recall Audrey Tang blogging about some things she liked about Haskell, but I'm not for certain how the use of Haskell has actually influenced the project. 00:21
araujo haha
00:22 eden_c joined
araujo it's ok, i probably got too many questions because i _need_ to start reading about this stuff :-) 00:22
Juerd araujo: Ask them :)
araujo Juerd, :-)
araujo bookmarks markstos_ links to check when he gets back from dinner 00:23
off to eat now, thanks markstos_ Juerd
00:24 eden_c left
Juerd Bon apetit 00:25
markstos_ TreyHarris: The slurpy seems to help, but leads to a new error I don't understand. When I try examine the hash I popular, a casting error is reported: 00:31
class Foo { has %!PARAM is rw; submethod BUILD (*%param) { %!PARAM = *%param; } }; Foo.new( a => 'b' ); say %!PARAM.perl;
?eval class Foo { has %!PARAM is rw; submethod BUILD (*%param) { %!PARAM = *%param; } }; Foo.new( a => 'b' ); say %!PARAM.perl;
evalbot_12700 Error: Cannot cast into Hash: VRef <Sub:0x895ea48>
markstos_ I'm working on CGI.pm for Perl6, and all I want to do is have args passed to new() availabe later through param()... 00:32
It seemed so easy at the outset... 00:33
00:36 Limbic_Region joined
Limbic_Region anyone know where I can get a decent nmake.exe without downloading the toolkit? 00:37
00:40 mako132_ joined
markstos_ I solved my problem. I was confused. I was calling %!PARAM outside of a class. 00:47
It would have been nice if the error message was more along those lines, though: "%!PARAM is declared as an attribute of the Foo class, but is being referenced outside of it. "
00:57 neoesque joined
svnbot6 r12701 | markstos++ | Allow Initializing the CGI object with your hash of params 01:01
01:03 Limbic_Region joined 01:19 diakopter joined, Limbic_Region joined 01:21 justatheory joined, diakopter joined
Limbic_Region attempts to compile pugs using nmake.exe as he refuses to install AND REGISTER the toolkit with M$ 01:29
Juerd It must surely suck to be forced to use Windows 01:30
Limbic_Region actually, I am not forced to use windows at home (where I am now) 01:31
in fact, my desktop is dual boot - this is my wife's shiny new laptop
Juerd Oh. Why do you prefer Windows on it?
Limbic_Region I don't - it makes things easier for my wife 01:32
I haven't had a chance to partition the drive and install linux on the laptop - it is brandy new
Juerd ssh to your desktop ;)
Limbic_Region desktop is off-line ATM - need to install wireless card in it 01:33
Juerd Modern linux installers can resize partitions for you
Limbic_Region long story
Juerd Ah, wireless. Wireless sucks for nonlaptops
Limbic_Region Juerd - yes I know
and to the second point, it wasn't originally that way - the desktop was hardwired to the wireless router
Juerd It's hard to tell you something you don't know already :)
Limbic_Region but unfortunately I needed to move the wifi router to a different floor in the house so the laptop could get reception everywhere - leaving my desktop SOL 01:34
though I am not really a winblows hater
Juerd You say "M$" and "Winblows". You may not hate it, but I do sense some annoyance :) 01:35
01:37 Bit-Man joined
Limbic_Region sure - but I have yet to find an OS I am 100% happy with 01:37
it's not that I don't believe Microsoft can put out quality software, I just believe they choose not to most of the time 01:38
it's not that I don't think Bill is an uber smart guy, humanatarian, and a great business man - I just think many of those business practices suck 01:39
what does "==> Refreshing specs from svn.perl.org/perl6/doc/trunk/design/syn..." do in the perl Makefile.PL process?
Limbic_Region will brb 01:42
Juerd I'm never 100% happy with any piece of software. But my Kubuntu desktop gets a lot closer than any Windows installation I've seen so far.
01:46 thomason joined 02:04 dolmans joined 02:06 Limbic_Region joined
Limbic_Region yay - pugs successfully built 02:12
cool beaners and moldy chicken
02:19 onsen joined 02:20 marksto1 joined
svnbot6 r12702 | dolmans++ | add eval_dies_ok. 02:21
02:22 LimbicRegion joined, dada joined
dada www.djdada.dl.am RnB/HipHop Songs to Download & Musicclips! check it out ;) 02:22
svnbot6 r12703 | markstos++ | Add link to Documention.pod from syn_index.html, in hopes 02:24
r12703 | markstos++ | that Agentz will add it to the HTMLization system.
02:25 dada left
svnbot6 r12704 | markstos++ | Add mention of kwid and Documentation.pod (S26) to Differences.pod 02:30
r12705 | dolmans++ | use eval_dies_ok to ensure user-defined non-assoc infix cannot be used chaining.
r12706 | markstos++ | Clarify on syn_index that Documentation spec is "DRAFT" status. 02:32
r12706 | markstos++ | (As opposed to the official-but-still-changing nature of the other specs...)
03:03 kanru joined 03:25 discordja left 03:34 hikozaemon joined
svnbot6 r12707 | markstos++ | simple typo fix 03:36
03:46 mako132_ joined 03:51 markstos_ joined
araujo back 04:05
Perl 6 gains the lazy evaluation of lists that has been a feature of some functional programming languages such as Haskell 04:06
:-)
04:14 justatheory joined
TimToady araujo: did anyone send you a commit bit yet? 04:26
04:26 mako132_ joined 04:27 mdiep_ joined
araujo TimToady, commit bit? ... 04:29
no , i am starting my learning process :-) 04:30
TimToady that's no excuse. 04:32
all of us are learning.
araujo :-) 04:33
TreyHarris TimToady: did you see my question earlier? what happens when you have executable statements inside a class declaration?
araujo TimToady, well, i would be more tha happy to help whenever i can definitely
TimToady araujo _at_ gentoo?
araujo TimToady, yes sir 04:34
TimToady TreyHarris: they execute once at class construction time. 04:35
TreyHarris ok, but RHS of accessor declarations are automatically run as if they were inside BUILD? 04:36
04:37 Limbic_Region joined
TimToady sorry, flakey connection 04:39
Limbic_Region just spent the last 3 hours diagnosing a flakey connection 04:40
turned out, all I had to do was release and renew the dhcp setting on my wireless router
TimToady well, I'm in a remote location that just happens to have wifi, sort of 04:41
\^H
yes, pseudo assign to attrib is suppoed to happen at BUILD gtime 04:42
TreyHarris ok. 04:43
TimToady A12 explains better than S12, probably. 04:45
araujo: still waiting for openfoundry to log me in... 04:46
04:46 kanru joined
TimToady araujo: commit bit is on its way. 04:49
traditional first checkin is to add your name to AUTHORS 04:50
and then you're emotionally committed. :)
checkin philosophy is to ask for forgiveness rather than permission.. :) 04:51
have fun! 04:52
04:57 Khisanth joined
araujo TimToady, oh, that was ... fast :-P 05:01
Thanks TimToady ! , i'll do my best
TimToady seemed kinda slow to me. :)
araujo hah, well, considering i just joined this channel ;-)
TimToady lambdafolk generally seem to catch on about as fast as anyone... 05:02
araujo has pointed several times he is not any perl monger too :-)
TimToady we consider differences to be features around here.
obra You made the mistake of suggesting that you're at least passingly familiar with Haskell
araujo haha nice one 05:03
obra ;)
Welcome
araujo Hi obra , and thanks!
TimToady sure, I've learned Haskell several times already...
araujo ya, ive been coding Haskell from quite a time now :-)
obra Haskell, is, sadly, over my head.
araujo haha, i sometimes feel like that about perl ;-) 05:04
obra Around perl6ish things, I mostly just suggest things people might want to hack on.
Limbic_Region equates trying to learn Haskell with trying to stuff the square in the round hole - it just don't fit
TimToady which timezone are you in, araujo?
Limbic_Region will get it eventually though
araujo TimToady, VET
araujo is from venezuela
Limbic_Region those edges are bound to wear down eventually
05:05 kanru joined
araujo :-) 05:05
Limbic_Region seen audreyt 05:06
?seen audreyt
lambdabot audreyt is in #perl6 and #haskell. I last heard audreyt speak 12 hours, 46 minutes and 57 seconds ago.
Limbic_Region that's the last time I heard her speak too 05:07
TimToady the other lambdafolk are scattered about the world, so around the world, so you can usually get advice here most times o of the day o rnight. 05:08
araujo nice :-) 05:09
TimToady sorry, relocated to better reception now, so not typing blind anymore...
araujo hopes to have some perl mongers around willing to answer questions from time to time :-) 05:10
TimToady the place is infested with those... 05:13
araujo haha
Limbic_Region araujo - in the very unlikely event there isn't a perl monger here that can answer your question - #perl and PerlMonks are great alternatives
araujo good
TimToady though in both of those places most people will be more familiar with Perl 5 than Perl 6. 05:14
Limbic_Region ah yes, nice augmentation TimToady 05:15
if there isn't a perl 6 monger here that can answer your question - the synopses dev.perl.org/perl6 or the mailing list perl6.language are great alternatives
lambdabot Title: Perl 6 - perl6
araujo TimToady, i see 05:17
aufrank (slightly) OT: does anyone know where I can download an archive of the info files for the prl 5.8.8?
Limbic_Region wonders what info files are 05:19
aufrank www.gnu.org/software/texinfo/ 05:22
lambdabot Title: Texinfo - GNU Documentation System - GNU Project - Free Software Foundation (FSF ...
Limbic_Region aufrank - oh, those info pages - as in alternatives to man pages 05:23
aufrank yeah, those :)
Limbic_Region would assume there is a pod2info translator out there and generating your own with a little File::Find magic shouldn't be too difficult
or even man2info 05:24
sorry I couldn't be of more help
aufrank np
Limbic_Region did you try googling? 05:25
www.stanford.edu/services/pubsw/pac...linfo.html
lambdabot Title: IT Services: Perl info 5.6
Limbic_Region of course - that's for 5.6 but it looks like they are out there if you dig 05:26
aufrank digs 05:28
Limbic_Region pod2info is on CPAN 05:35
or at least so claims perlamiga 05:36
05:38 agentzh joined
agentzh Limbic_Region: pugs' Makefile.PL invokes docs/Perl6/Spec/update and downloads the latest version of synopses via LWP. 05:39
my smartlinks.pl also invokes the ``update'' script to sync with the synopses repos. 05:40
markstos: should we offer links to syn drafts on syn_index.html? i really doubt it. 05:43
Limbic_Region agentzh - yeah, the firewall on this laptop doesn't work the same as my desktop so I didn't realize it was silently hanging the LWP request
Limbic_Region fixed it though - thanks
agentzh LR: you can set the HTTP_proxy env?
you know, i'm also behind a firewall.
(two firewalls, actually) 05:44
and an HTTP proxy is just fine.
oh, i just remember your brand new laptop at home has no web connection at all. :) 05:45
Limbic_Region no - I am connected on that laptop 05:46
agentzh markstos: there're already serveral syn drafts under docs/Perl6/Spec, but my smartlinks skip them all since they're not official yet.
LR: good to hear that. :)
Limbic_Region firewall not allowing the LWP connection - nothing to do with proxy
agentzh oh, dear... 05:47
Limbic_Region windows sw based firewalls tend to limit what programs have access to the internet not just what ports are open and where you are connecting to
so because the firewall hadn't been told perl.exe was ok - it just silently hung
agentzh ah, i see. the win firewall. hehe. 05:48
so it's trivial to fix.
i was thinking you had fixed Makefile.PL or something else in the pugs tree.
Limbic_Region oh, sorry - just getting the new laptop set up with everything I want (on this partition anyway) 05:50
agentzh ah 05:51
<markstos_> agentz: S29 currently cuts off after the "ref" entry in the HTML version. I'm not sure why. feather.perl6.nl/syn/S29.html
lambdabot Title: S29
agentzh markstos: looking now.
<svnbot6> r12687 | audreyt++ | * eval 'UnicodeHere' now works again. Thanks to agentzh++ for noticing. 05:54
perlbot What kind of idiot karmas himself? Your kind of idiot!
agentzh audreyt: you mean the failing test in t/syntax/comments.t?
perlbot: shut up!
ayrnieu ayrnieu-- 05:55
agentzh marstos: it meems that S29 contains Pod errors. fixing now. 05:57
05:59 avarab is now known as avar
agentzh podchecker has told me that Functions.pod has 8 pod syntax errors. oh, my God. 05:59
05:59 kattana_ joined
agentzh and Documentation.pod also has 4 pod syntax errors. 06:00
svnbot6 r12708 | agentz++ | [docs/Perl6/Spec] 06:13
r12708 | agentz++ | - fixed all the POD syntax errors and warnings reported
r12708 | agentz++ | by the excellent podchecker script.
06:13 kanru joined
agentzh sorry, it should have been "[docs/Perl6/Spec/Function.pod]" in the svn log. :) 06:15
could anyone tell me the middle name of Ingy? 06:17
svnbot6 r12709 | agentz++ | [docs/Perl6/Spec/CPAN.pod]
r12709 | agentz++ | - fixed all the POD syntax errors and warnings reported
r12709 | agentz++ | by the excellent podchecker script.
agentzh d?t? 06:20
aufrank dĆøt
or maybe dƶt
I can never remember 06:21
audreyt see AUTHORS :)
but I don't think it's his middle name... "dot net" is the last name
it's a preposition, part of his last name
like the German "von", French "de", 06:22
or Dutch "van"
ayrnieu or the American ""
audreyt lol
agentzh thanks, already fixed in Documentation.pod. :) 06:25
svnbot6 r12710 | agentz++ | [docs/Perl6/Spec/Documentation.pod]
r12710 | agentz++ | - fixed the Pod syntax errors and warnings which i'm able
r12710 | agentz++ | to fix.
audreyt actually, the american equivalent would be "\cH"
svnbot6 r12710 | agentz++ | - changed the file encoding to UTF-8.
06:26 zgh joined
agentzh audreyt: do you think if we should put drafts into feather.perl6.nl/spec ? 06:26
i'm doubting.
audreyt sure, as long as it's clearly marked as such
agentzh audreyt: okay, thanks for the answer.
audreyt "unofficial draft written by members of the Perl community but not approved by the Perl 6 Design Team." 06:27
is the wording ingy used in Perl6::Bible
agentzh ok
audreyt: i hope there will be 6.2.13 and 6.2.14 to fix basic issues in pugs. :) 06:36
audreyt: i'm just hoping pugs to be somewhat ``trusty'' or sort of ``stable'' before 6.28.0. 06:37
06:38 Daveman joined
audreyt well... the problem is, a stable implementation of december 2005's perl6 06:41
is kinda of not very useful :)
agentzh hmm 06:42
audreyt in particular: the object model and capture-based calling convention
really needs to land before we can stabilize anything
gaal moose.
audreyt currently pugs does not even support user-defined &each and &zip 06:43
and introspection is completely not there
agentzh yeah
gaal itrospection would be fun to write once vv is there
audreyt today I think I'll fix the regression caused by the Var refactoring
(which has delayed my time working on vv by a week, but is probably worth it by the speed and internal clarity win alone :)) 06:44
agentzh audreyt: i'll be happy if you're keeping fixing failing tests in t/
audreyt and then go back to Capture callconv
gaal I want to finish the \() stuff from last saturday
audreyt agentzh: right, that's what I'm doing. but "regressions"
i.e. tests that used to pass in the previous release but now fails 06:45
agentzh that's great. :)
audreyt would take precedence over new tests
agentzh i think we just need more audreys here. :)
audreyt or more hackathons :) 06:47
audreyt needs to get the Brazil hackathon figured out...
agentzh audreyt: it's not a complaint. that's merely a wish. :)
YAPC::SA?
audreyt yeah
or hey, maybe you can fly over to Taipei with xinming and other .cn foles 06:48
agentzh that'll be wonderful. :)
audreyt and we can have a hackathon here
which would be far less expensive than flying to Brazil
agentzh indeed!
audreyt I hear that September onward you can get a visa with much more ease.
cmarcelo anyone mentioned Brazil?
agentzh audreyt: saddly the school will start in September. :/ 06:49
audreyt failing that, I can also come to YAPC::Beijing if there will still be one
cmarcelo: yeah. I plan to fly to Brazil directly after OOPSLA
cmarcelo audreyt: I tested i'm my linux and in feather, s/StrMap/Hash/ for both AST/Internals.hs and interning, no segfaults.. do you have any other segfault case with Hash that I can try reproduce? 06:50
audreyt so that should give me a couple weeks of hacking time
agentzh audreyt: sweet. :)
audreyt cmarcelo: no, not at this moment, my macbook is in repair
cmarcelo: but I kinda like StrMap better than Hash now
as we can now put embedded nulls in hash keys too
?eval my %hash{"foo\0bar"} = 1; %hash.keys.[0].chars 06:51
evalbot_12710 7
audreyt and having hash ordered by default with minimal speed loss (<5%) is I think a win...
gaal is that ordered by insertion or collation? 06:52
audreyt by collation.
autosorted hash, that is
gaal so the 5% is insert time yes?
audreyt I believe so. lookup time should be of no dicernable difference 06:53
agentzh gaal, audreyt: does feather smoke pugs periodically? and where's the resulting tests.yml then?
audreyt I don't think it does anymore
eric256 used to run the smoke loop there 06:54
agentzh ohhhhh
used to == no longer
audreyt aye
agentzh gaal: any comments? 06:55
cmarcelo audreyt: re StrMap: thats fine so. that Hash segfaults were the last big problem (i remember) in pugs' judy..
agentzh or any suggestions?
gaal agentzh: well, I vote for fixing that :) but no tuits to do it myself
audreyt agentzh: running the smoke loop requires a bit of itmeout detection
agentzh gaal: is it possible to use the data on the smoke server?
audreyt: true.
gaal agentzh: not really, it's html produce, not raw TAP data 06:56
also it's on iblech++'s server, not feather
audreyt (brb)
gaal audreyt: I don't understand Prim dispatch. look:
agentzh gaal: is it possible to let the user also upload .yml?
gaal ?eval \(42)
evalbot_12710 "--- !hs/VInt \n- 42\n"
audreyt ?eval \(1,2)
evalbot_12710 Error: No compatible subroutine found: "&circumfix:\( )"
gaal this is good, it means my capture crutch works 06:57
that is bad, it meas it doesn't
Prim:1996
audreyt gaal: at this moment you need to intercept it
gaal nothing I put there seems to work
audreyt it's too late to do it there
gaal hmm.
audreyt you need to do like &each
and &goto and &assuming
and ignore the prim dispatching for now
i.e. grep for "bloody"in Pugs.Eval 06:58
and do it like that.
gaal Eval's too late too, isn't it?
audreyt it's not
in Eval the args were yet unevaluated
gaal or, you mean my capture is parsed, just not dispatched?
audreyt so you can do whatever you want
gaal *correctly
audreyt aye
to wit:
pugs> :d \(1: 2,3, x=>4) 06:59
Noop;
App &circumfix:\( ) (1:
2,
3,
Syn "named" {"x";
4})
which is the Correct OPParse.
gaal ack
audreyt so nix the Prim line
gaal but xlating it to feed in the evaluator is pretty bogus, isn't it?
audreyt and swap Eval out? I'll be back in a bit
it is, but we can unbogify it once the Exp supports Capt node
gaal did't we do that already? EE? 07:00
ah, no :(
audreyt no.
gaal because of circularity
audreyt right.
gaal which can be fixed on 6.5
(committed in my branch) 07:01
audreyt which is having build problems on win32 at this moment
gaal puts on the broken record
I will not buy this win32, it is scratched
audreyt please have patience, I beseech thee. :)
gaal greps for 'bloody' 07:03
audreyt will bbiab 07:04
07:06 kanru joined
agentzh gaal: if there's no autosmoke on feather, then i'll feel less motivated to add smoke results to the HTML pages generated by smartlinks.pl. :( 07:13
even if it's easy.
gaal how can I help you get autosmokes? 07:14
agentzh i just want an autoupdated tests.yml or something like that.
gaal there are two approaches to getting this done:
07:15 drbean joined
gaal 1. hack audreyt's autobuild and add 'make smoke' 07:15
2. copy her autobuild and modify it
#2 is wasteful, a little, because it means an extra pugs build
but #1 means evalbot suffers
so I say go for #2, just be sure it's niced
agentzh timeout issue is also important here. 07:16
gaal you mean for hanging tests?
agentzh aal:
eys
yes
so i would rather reuse others' smoke results instead to do it myself.
gaal agentzh: I don't know if anybody still runs autosmokes. 07:17
agentzh i mean manual smoke.
gaal we used to do it -- nothingmuch, eric, me
but really feather is the perfect place to run it if you ask me
okay, wanna add hang detection to the harness maybe? 07:18
agentzh i think so.
but i don't have to tuits to add that myself since i've never run smoke on my machine.
*the 07:19
gaal "make smoke". :-)
agentzh yes, i know the command. i just don't want to, since it's terribly time consuming. :)
gaal it's not *that* bad
say 45 minutes
depending on your hardware, of course. 07:20
agentzh gaal: if you can find anyone who likes to setup auto smoke on feather, i'll add smoke result support to smartlinks.pl in less than 1 hour. 07:21
gaal just add it :)
agentzh i have no data
gaal what do you need right now? data?
agentzh do you have a tests.yml?
gaal ah, larry and I made you some
scroll back a bit 07:22
agentzh url?
gaal sec
wall.org/~larry/tests.yml
agentzh thanks
gaal sure
agentzh gaal: one last thing, is this url auto updating? 07:23
gaal nope! but that's orthogonal :)
agentzh okay
gaal audreyt: data Capt ... c_feeds :: [Feed a] 07:27
multiple feeds are semicolon based, yes?
so current parsing can't construct them anyway, and for now I should ignore them?
s/based/separated 07:28
TimToady ==> and <== also theoretically make feeds
gaal TimToady: ACK, but not in current pugs :) 07:29
TimToady but yes, zip(@a;@b;@c) is a zip of 3 feeds
07:29 drbean joined
gaal also they construct one feed each, right? 07:29
07:30 larsen joined
TimToady yes, but the multiple feeds get flattened unless bound to @@ 07:30
gaal which Syn is this?
(gluttony, surely)
TimToady 9 I think
gaal thank you
TreyHarris how do you write zip with :shortest? i can't make the syntax work 07:37
ayrnieu TreyHarris - how did you try? 07:38
07:40 elmex joined
svnbot6 r12711 | agentz++ | [util/smartlinks.pl] 07:40
r12711 | agentz++ | - changed the confusing var name $links to $linktree in
r12711 | agentz++ | the Perl 5 source.
agentzh there's a bug in smartlinks.pl which results in incomplete HTML outputs. i'll try to fix it first. :/ 07:41
markstos: thanks for reporting this.
TreyHarris ayrnieu: oh, i went to the source, it's unimpl. but thanks 07:42
07:45 iblechbot joined
ayrnieu TreyHarris - oh, I was just curious. 07:45
07:47 pina joined, pina left
gaal is \(1;;2) a syntax error? 07:54
(empty feed)
agentzh oh, there seems to be a bug in File::Temp or Pod::Simple. well, not sure. but now i've fixed smartlinks.pl by avoiding temp files. 07:55
TimToady I don't know why it should be a syntax error. 07:56
gaal TimToady: okay 07:57
I won't make it an evaluation error either then :)(
08:02 marmic joined
svnbot6 r12712 | agentz++ | [util/smartlinks.pl] 08:05
r12712 | agentz++ | - fixed the HTML truncating bug. it was due to
r12712 | agentz++ | non-flushed output buffer.
r12712 | agentz++ | - removed the dependency File::Temp by using an
r12712 | agentz++ | undocumented feature of Pod::Simple::HTML.
gaal heh, agentzh++
agentzh :)
Juerd: i've just installed YAML::Syck and also upgraded YAML on feather. 08:09
svnbot6 r12713 | audreyt++ | * remove t\rules\rules.t as it's superceded by the spec-based tests in t\p6regex
agentzh markstos: S29 (and also other synopses) is now complete on feather: feather.perl6.nl/syn/S29.html 08:13
lambdabot Title: S29
08:18 foo\ joined 08:19 prefiks joined
agentzh markstos: smartlinks pointing to syn drafts won't work. this is not a bug, but an intentional feature. or we can discuss it later. 08:20
svnbot6 r12714 | agentz++ | [util/smartlinks.pl] 09:10
r12714 | agentz++ | - added very detailed Pod documentation which explains
r12714 | agentz++ | the internal algorithms and data structures we're
r12714 | agentz++ | using. so others can understand and/or hack on the
r12714 | agentz++ | source code more easily.
r12715 | agentz++ | [util/smartlinks.pl]
r12715 | agentz++ | - removed 'mkdir "tmp"' which was legacy.
r12716 | agentz++ | [util/smartlinks.pl] 09:13
r12716 | agentz++ | - added one "FIXME" comment.
agentzh supper & 09:15
09:15 agentzh left 09:46 elmex joined 09:56 cookys joined 10:21 chris2 joined 10:37 DaGo joined 10:38 buetow joined 11:01 dolmans joined 11:06 agentzh joined
agentzh TimToady: oh, sorry, i didn't download tests.yml and now it's not available at wall.org/~larry/tests.yml. 11:13
gaal agentzh: I'll send you mine. sec
agentzh gaal++
gaal perlcabal.org/~gaal/tmp/tests.yml 11:15
agentzh gaal: should i use the OO interface provided by Test::TAP::Model or directly access the data structure in .yml? 11:17
gaal TTM is better
(TTM is a refectoring of yaml_harness)
agentzh okay 11:18
11:32 zgh joined, soisoisoi joined 11:37 zgh joined 11:42 Nop0x90 joined 11:43 zgh joined 11:45 kane-xs joined
TimToady woops, is there now. 11:45
zzz &
11:48 zgh joined 11:56 zgh joined
svnbot6 r12717 | polettix++ | [t/builtins/strings/quotemeta.t] 11:57
r12717 | polettix++ | Worked around the lack of a Config.pm module.
r12717 | polettix++ | Marked absence with a todo.
12:03 zgh joined, Aankhen`` joined, markstos_ joined 12:05 jferrero joined 12:09 foo\ joined
kane-xs seen audreyt 12:11
12:17 norageek2 joined 12:19 zgh joined 12:20 marmic joined 12:25 norageek2 left 12:26 zgh joined 12:36 zgh joined 12:41 zgh joined 12:45 drbean joined 12:48 zgh joined 12:59 zgh joined
pasteling "markstos_" at 12.176.101.89 pasted "bad mojo: What happens when you declare variable with 'my' that has the same name as an arg" (13 lines, 430B) at sial.org/pbot/19274 13:02
audreyt markstos_: indeed, that should raise a warning
or an error
but probably a warning about redeclaration 13:03
markstos_ audreyt: Thanks!
audreyt: So it should warn, but work as expected, with the arg getting a value ?
audreyt it's conceptually same as
sub f { my %p; my %p; }
markstos_ That's what I thought should happen.
audreyt which is the same as
sub f {my %p; %p} 13:04
which is not what happened in perl5
i.e. in perl6 redecl of the same var in the same storage class
is a warning but noop
markstos_ It was one of those frustrating experiences where I made a reduced test case, but the test case passed..... until I found extra "my" in my code.
audreyt yeah. a warning should definitely happen
markstos_ Thanks!
13:04 zgh joined
audreyt np :) write a test that demonstrates this shadowing? 13:04
markstos_ Here's anothe case that could use a better error-- calling an attribute outside of a class: 13:05
?eval class Foo { }; say %!PARAM.perl;
evalbot_12717 Error: Cannot cast into Hash: VRef <Sub:0x895ea48>
audreyt aye
markstos_ It took me a while to figure that out one, too. :) 13:06
audreyt: Test case for issue #1 coming up. Care where I file it?
audreyt t/var/my.t seems logical 13:07
markstos_ coming up.
audreyt moose!
markstos_ Is moose the new cool ?
audreyt or the new hot... or both 13:09
13:12 zgh joined
markstos_ audreyt: What about this second case I reported. Do you agree it should have a better warning? I could write a test for it, too. :) 13:13
svnbot6 r12718 | markstos++ | A new failing test case for my. audreyt confirmed it should work, but emit a warning.
audreyt well, it's hard to test the presence of a warning 13:16
maybe it should be a flat-out error
markstos_ ...until Test::Warn is ported. :)
audreyt so let's make it that, and you can test that it would be a compile time error :) 13:17
markstos_ I think a syntax error for the second case would be reasonable.
audreyt test for that then.
13:17 zgh joined
markstos_ I use eval_ok() for that ? 13:17
audreyt sure
markstos_ coming up.
gaal hey audreyt. got some Eval code but it's not finished... 13:18
some coercions I'm not sure about, too 13:19
audreyt does it compile? :)
gaal nope :)
audreyt check it in anyway? :)
gaal let me scp up a patch
perlcabal.org/~gaal/capt.0.patch 13:21
audreyt 404 13:23
gaal oops perlcabal.org/~gaal/tmp/capt.0.patch
I wonder if there's a better way to manage this double-accum pattern, this is the second time I'm encountering it 13:24
13:24 zgh joined
gaal the first two args to argsFeed are accumulators, the third is the actual argstack, here used only with one semi-separated-section (we need a name for that!) 13:25
the first accum is the total Feed list, the second is the currently constructed Feed, Maybed because \(1;;2) is legal. 13:26
and it's in Eval because I'm assuming this should reduce.
audreyt that's correct 13:27
13:27 onsen joined
clkao mmmm onsen 13:28
onsen hello :) 13:29
svnbot6 r12719 | markstos++ | New test: calling a trait outside of a class should be a syntax error. 13:31
r12719 | markstos++ | ( I /think/ I got the syntax right for testing this... )
13:32 zgh joined, Revision17 joined
audreyt gaal: looking at the patch now 13:36
gaal cool 13:37
audreyt (sorry, on a semi-defunct win32 work env)
I had to download patch ;)
gaal shall I paste the errors?
heh, sorry :)
audreyt no, I got it now 13:38
agentzh gaal: i've implemented that feature.
audreyt compiling
gaal /a.{6}/++ 13:39
audreyt it seems to compile just fine
after a one line change that is
TimToady btw, not expecting a fix soon, but my Test.pm.yml makes my pugs segv, even if complete tree is regened.
audreyt ! 13:40
you mean you cannot run smoke anymore?
TimToady I can run the part of it that doesn't use Test.pm :/
audreyt heh
13:40 nothingmuch joined
gaal TimToady: can you regen the .yml itself? rm blib6/lib/Test.pm.yml ; make blib6/lib/Test.pm.yml 13:40
13:40 zgh joined
audreyt howlong has it been segfaulting? 13:40
TimToady since sometime yesterday 13:41
I'll try the specific regen, but certainly a full make won't work.
gaal there was a strictness fix in hssyck, right?
audreyt in drift yaml hat is
reverting that
actually, I don't even have a full svk mirror yet :/ 13:42
gaal: can you revert it? should be indriftyaml
gaal sure
audreyt gaal++
TimToady complains about unexpected " at line 298 13:43
audreyt gaal: you don't need separate IValue and Pure instances
instance ICoercible P ValCapt where asStr _ = return (cast "<capt>") -- XXX
is sufficient 13:44
I'll commit in a bit
gaal TimToady: line 298 of Test.pm?
TimToady yes, print "not " unless $cond; 13:45
print "not " unless $cond;
apparently from some previous line 13:46
audreyt weird, I can't quite duplicate that
13:46 f0rth joined
audreyt and it works once you remove .yml? 13:46
oh wait, it must be unrelated, if it's a parse error
TimToady I don't believe the error message shows with a full make, but a partial .yml would explain the segv... 13:47
gaal um, I'll hold off with the revert patch then...
maybe we need a checksum to protect against this then. 13:48
TimToady ./pugs -c ext/Test/lib/Test.pm works 13:49
fails if I put -CParse-YAML in
gaal asYAML [ver, desc, check dump, dump] where dump = asYAML ast 13:50
13:50 zgh joined
gaal hmm, ok, let me ci the revert. 13:50
TimToady: please try w/12720 13:52
svnbot6 r12720 | gaal++ | * revert audreyt++'s r12647, to see if it fixes TimToady++'s Test.pm.yml segfault
gaal brb
TimToady compiling 13:53
agentzh an HTML-related question: <td><pre>...</pre></td> results in unbearable big table cells. how can i get rid of it? 13:54
i've tried many methods, but saddly they don't work. 13:55
TimToady now I get pugs: user error (no tag found)
agentzh TimToady: please regenerate your prelude.pm.yml 13:56
gaal rm the yml file and regen it manually again?
agentzh i've ever got this error too.
Juerd agentzh: Don't use <pre>.
agentzh Juerd: then what should i use?
gaal "regen it again" is superfluously reduntant, and also repeats itself
agentzh that's code line, you know. :)
TimToady same parse error, unexp "
Juerd agentzh: Just expand \n to <br>, and whitespace to &nbsp; yourself. 13:57
TimToady looking at prelude now
agentzh Juerd: and also set a proper code font?
Juerd agentzh: Obviously
gaal agentzh: "<code>"
Juerd agentzh: Don't change every horizontal whitespace to nbsp, or you'll have the same effect as pre.
agentzh Juerd: when using <code>? 13:58
Juerd agentzh: In leading whitespace, substitute every odd, and in other whitespace every even.
agentzh Juerd: ahh
gaal audreyt: doing that change you suggested with IValue etc. didn't work 13:59
looking forward to your commit ;)
Juerd e.g. s/^( )/"&nbsp; " x (length($1) / 2)/gem; s/ / &nbsp;/g
Or something like that.
(untested)
agentzh Juerd: thank you. will use that. :)
TimToady blew away both .pm.yml files, same problem 14:00
Juerd Won't work, delemiter error.
gaal /^( *)/
Juerd Eh, right
TimToady biab & 14:03
gaal looks like a 6.4 issue, Test.pm worksforme 14:04
(but then what triggered it?)
wolverian alternatively, use xhtml2 and its <l> element for each code line. 14:05
(no, I'm not being entirely serious. I wish, though..)
gaal ah, oops, against trunk Code.hs is still included. beh 14:07
wolverian I suppose <l> wouldn't help, in fact. it doesn't create block structure in the dom. oh well. 14:08
audreyt waits for svk to fully catch up the repo history... 14:09
gaal weird, touching Val and recompiling didn't fix it. 14:10
SamB rm the .hi instead?
gaal SamB: the module was recompiled 14:11
but the proposed fix didn't work
audreyt anymoose... finally got it to sync fully 14:12
committing
gaal yay
audreyt gaal: I handled the argsFeed case incorrectly it seems 14:17
the feeds only recognizes invocants
gaal could certianly have been a bug of mine...
svnbot6 r12721 | audreyt++ | Gaal's \(...) Capture-literal patch, hacked slightly so it compiles.
audreyt argsFeed fAcc _ [] = return fAcc
please fix that line appropriately
dinner, bbiab 14:18
araujo morning 14:28
markstos_ Good morning araujo. 14:29
araujo morning markstos_ :-)
gaal yay, it works (for positionals. fixing nameds.) 14:31
audreyt gaal: it didn't compile as you were missing three fromVal 14:32
that converts oldval to newval
I think that's all
TimToady: "svn di ext" shows nothing?:)
gaal and bugs in argsFeed now :) turns out this is bogus: let (VStr key) = castV key'
but what's in key'? or shall I fromVal it too? 14:33
TimToady audreyt: no output
gaal (I was sure it _was_ a VStr by :D....)
audreyt Syn "named [VStr key, ] <-
would do
14:33 mako132_ joined
audreyt TimToady: I can't seem to duplicate this on feather nor win32 here... 14:34
rm blib6/lib/Test.pm.yml ; make blib6/lib/Test.pm.yml 14:35
does this also fail?
TimToady yes, unexp " line 298 14:37
14:37 kanru joined
audreyt does "svn di" show anything? 14:37
araujo markstos_, i am in (with a commit bit!) now :-)
gaal well, it's better but the reduction's still a bit off. committing...
TimToady nothing
not surprising, considered I blew it all away and redownloaded 14:38
audreyt extremely puzzling, then 14:39
markstos_ araujo: Great! You must have found someone else to give one out then.
TimToady *whistles* 14:40
araujo markstos_, blame TimToady ;-)
hi TimToady !
audreyt so if you do have a correct blib6/lib/Test.pm.yml
TimToady good morning, hope you got more sleep than I did last night...
audreyt such as can be downloaded from perlcabal.org/~audreyt/tmp/Test.pm.yml
markstos_ araujo: do you have a first task in mind to commit then ?
audreyt put into blib6/lib/ and "touch"ed
does it make smoke work? 14:41
araujo TimToady, sort of :-)
markstos_, not really .. well i was thinking to update the AUTHORS file
i just got up, and i need to find my way around all this code first :-) 14:42
14:42 lichtkind joined
lichtkind hi PerlJam 14:43
can anybody explain why on $*IN there is alsways an = = ? and not a = 14:44
gaal captures landed
ayrnieu lichtkind - one of them is an operator.
lichtkind ayrnieu thanks but what operator coul it be if its also = 14:45
integral "$a = =$*IN", the first = is &infix:<=> for assignment, the second = is &prefix:<=> for fetching a value from the iterator 14:46
TimToady audreyt: still a segv
svnbot6 r12722 | gaal++ | * Capture construction, supporting invocants, positional and 14:49
r12722 | gaal++ | (possibly recurring) named args:
r12722 | gaal++ | pugs> \(1: 2, :a<b> :a<b>)
r12722 | gaal++ | VPure (CaptMeth
r12722 | gaal++ | { c_invocant = VPure (IFinite 1)
r12722 | gaal++ | , c_feeds = [ MkFeed
r12722 | gaal++ | { f_positionals = [VPure (IFinite 2)]
r12722 | gaal++ | , f_nameds = fromList
r12722 | gaal++ | [ ("a", [VPure (MkStr "b"), VPure (MkStr "b") ])]
r12722 | gaal++ | } 14:50
r12722 | gaal++ | ]
r12722 | gaal++ | })
r12722 | gaal++ | This doesn't seem to reduce everything correctly yet, e.g.
r12722 | gaal++ | \(1: ("a"~"b") => "c") breaks.
TimToady audreyt: in fact, the .pm.yml I generate is identical to the one you sent me
audreyt hm. I'm puzzled
the segv happens after the parse error
or before? 14:51
lichtkind thanks integral thyt where one of the missing peaces for my perl6 talk
audreyt isn't quite sure what's the error sequence
gaal: that is entirely correct;
"a"~"b"=>c
is not a named param
in fact, neither is "a"=>"c".
only a=>"c" is
integral lichtkind: hmm?
gaal how do you programmatically select a named param then? 14:52
audreyt [,]...
would be the only way
lichtkind integral ?
audreyt which I think is a feature, actually :)
markstos_ It looks like $SIG{__WARN__} is not spec'ed for Perl6? Does something replace it? (yet?)
integral lichtkind: I didn't understand what you meant
gaal I'm not sure I agree :)
audreyt gaal: consider that positionals can also be passed by name 14:53
gaal: it is quite nice to be able to trnanslate those to positionals on compile time
and invalidate bogus calls without requiring runtime dispatch...
"a b c" => 'x'
14:53 diakopter joined
lichtkind integral your answer were one of the last missing sentences of my perl6 talk in 2 weeks in frankfurt 14:54
audreyt would be an example; you cannot have $::("a b c") declared as a param
TimToady maybe I should just try a reboot. could have a corrupted memory image persisting or something.
integral lichtkind: oh!
gaal okay, okay. so say I want a mona^H^H^H^Hset of combinators for captures -- at the least, paste inplace and paste add-semisection
TimToady biab &
audreyt TimToady: if you're on win32, that might actually fix things...
TimToady :P
audreyt TimToady: but I'm not sure I understand the sequence 14:55
when you "make blib6/lib/Test.pm.yml"
it's generated, but then segfaulted?
and when you actually run the tests, it fails with parse error?
lichtkind integral have to go im in icafe .. thanks
by @ all 14:56
audreyt gaal: it's [,]...
gaal \([,] $capta ; $captb) ? \([,] $capta, $captb)
lichtkind your grat
great
audreyt \([,] =$capta, =$captb)
is the currently specced way
gaal what's =$capt?
audreyt it's a shorthand for
$$capt, @$capt, %$capt
gaal ooh 14:57
audreyt when occuring after [,]
(bbiab)
15:06 agentzh joined
agentzh gaal, Juerd: <code> with interweaved spaces and `&nbsp;` results in exactly what i want. thank you! 15:19
gaal boings 15:24
TimToady reboot didn't help, so if something is corrupt, it's on the disk. 15:26
will now upgrade 6.4.1 to 6.4.2 (available via yum) 15:27
15:27 Limbic_Region joined
TimToady (pity they don't have 6.5 yet...) 15:27
gaal TimToady: there's a binary build of 6.5 you can fetch 15:28
svnbot6 r12723 | markstos++ | Add a cross reference to help people looking for docs on references to learn about Captures.
r12724 | gaal++ | * Now that Captures are handled in Eval, revert the changes in Prim
r12724 | gaal++ | from r12474 that were neither helpful nor pretty.
gaal I put one under my home dir 15:29
all I needed was readline 4
TimToady yeah, I know, but that's work. :)
audreyt feather is also 6.4.1 15:31
so I really doubt it's ghcvers problem
Limbic_Region morning audreyt
or rather - evening your time
TimToady but if my ghc is corrupt somehow, upgrade will fix
audreyt greetings
that is true 15:32
gaal TimToady's box was hax0red by ruby zealots? 15:33
svnbot6 r12725 | gaal++ | * Add note about Capture FAQ
r12726 | agentz++ | [util/smartlinks.pl]
r12726 | agentz++ | - implemented the smoke result rendering feature
r12726 | agentz++ | requested by gaal++
r12726 | agentz++ | you can now use the --test-res option to add smoke
r12726 | agentz++ | results into the HTML:
r12726 | agentz++ | $ util/smartlinks.pl --test-res=tests.yml t/*/*.t t/*/*/*.t
gaal whee!
audreyt wheeeeee!
gaal agentzh++ agentzh++ agentzh++ :)
audreyt wow this is quickly turning into something most awesome
Limbic_Region also likes the way 6.5 compiles - much better look and feel 15:34
agentzh gaal: not yet add --test-res option to feather's script, since there's no autoupdating tests.yml. :/
gaal now to set up autosmokes on feather again :)
agentzh i'm feeling quite comfortable with the smoke-result-rendered HTML-fied Synopses. :) 15:35
they look very nice~~~
gaal demo?
oh, never mind, I can gen one myself :)
agentzh here's a demo:
svn.berlios.de/svnroot/repos/unisim...n/S02.html
lambdabot Title: S02
gaal even better :P
agentzh :) 15:36
gaal lol pugs' commit rss feed doesn't escape html, so the commit log that said :a<b> threw the rest of the message into bold
agentzh skipped tests are currently not marked at all.
and TODO tests are marked as normal tests.
(i'm using actual_ok of Test::TAP::Model::Subtest) 15:37
gaal *very* pretty :)
agentzh gaal: thanks!
Limbic_Region audreyt - still on Win32?
Juerd agentzh: "5 āˆš, 3 Ɨ" in the clickable links?
agentzh Juerd: pardon?
Juerd agentzh: test result stats 15:38
Per snippet :)
agentzh ah, you mean a small summary, right?
Juerd So you don't have to click-open everything to see what failed
Yes
audreyt Limbic_Region: yes. I figured out the nested atomicity problem
Juerd s/to see what failed/to look for failed tests/
agentzh Hmm, that's easy...
audreyt and found a glaring bug in one of the examples
Limbic_Region: fixing it now
agentzh Juerd: thanks for suggesting this. :)
Juerd agentzh: Thanks for creating this :) 15:39
agentzh hehe
Limbic_Region yay - that cleans up 6 more examples
or rather - potentially, there may be other problems 15:40
Limbic_Region wonders if he was successful in getting eric256 to come back and fix his examples/games/wizard.pl
svnbot6 r12727 | gaal++ | * make pass and fail marks a litle more prominent
audreyt it can be reduced to "sub f {} my &g; sub h {} 1;" 15:41
?eval sub f {} my &g; sub h {} 1;
evalbot_12726 (no output)
audreyt praises easy access to the stack trace
gaal the haskell stack trace?
audreyt yeah +RTS -xc 15:42
gaal after make fastprof yes?
audreyt just prof 15:43
gaal er yes. I need to buy some RAM before I can do that :-)
agentzh gaal: personally i don't quite like bold marks. :) 15:44
Limbic_Region it is amazing what 6 years will do WRT prices
Juerd It's amazing they still profit.
agentzh gaal: but i'm fine with that. :) 15:45
gaal agentzh: ah, I didn't mean to start an argument about visualization.. the green just seemed a little faint for me...
Limbic_Region Juerd - I think I have a hunch as to why that is - at least with WinTel PCs
Juerd Limbic_Region: Which is? No overproduction, ever?
Limbic_Region I have spent two days cleaning up crap trial ware software they install
agentzh gaal: okay. :)
Limbic_Region I am thinking they have SW sponsors
Juerd I see chips for sale now, that were manufactured in 2004. That must be a terrible loss.
Limbic_Region: Hm, seriously? 15:46
Limbic_Region yes
Juerd I sell Wintel boxes with lots of open source software :)
Limbic_Region in fact, there is at least one lawsuit over it
in any case, this laptop (my wife's) is much bigger and badder than my desktop (which was top of the line when I bought it) and cost literrally 1/5 the price 15:47
Juerd Yea
When I think of my first laptop...
My current laptop is expensive, but costs the same as that entry level laptop back then. 15:48
gaal hey, where are captures actually parsed ruleVal shouldn't be enough... 15:49
Limbic_Region well, the nice thing is that I should now be able to do a little hacking outside of work
TimToady agentzh: allow me to also say that I'm exceedingly impressed.
gaal aha! that's for the noncircumfix case, and indeed falls short: 15:50
?eval \42
evalbot_12727 \42
agentzh TimToady: oh, i'm also impressed to hear that. ;-)
gaal fixing :)
Limbic_Region agentzh++
Juerd Why do people in American movies stress what they say by announcing the number of words?
agentzh Limbic_Region: :)
Juerd "One word: ...", "Three words: ..."
To me, this is extremely silly :)
gaal crude error detection protocol?
Juerd gaal: Heh 15:51
TimToady It might be related to the game of charades.
Juerd That's in bad American comedies. "Two words: ..." "That's three words"
TimToady but culturally I think it's a form of golf
Juerd TimToady: Adding a checksum makes it *longer*
TimToady "I can express that idea in fewer words than you can..." 15:52
Juerd Heh
TimToady it's meta-info, so doesn't count. :)
Limbic_Region Juerd, after thinking about it for 30 seconds, I think it has to do with our need for instant gratification
15:52 dolmen joined
svnbot6 r12728 | markstos++ | add commented-out links to FAQ from new docs index. 15:52
r12728 | markstos++ | Needs some help (from agentz?) to actually make the pod2html conversion happen.
Limbic_Region when we see a movie review - we don't want two read a couple of paragraphs
we want
Juerd Limbic_Region: Whoa. Explain that, please :)
Ah
Limbic_Region it can be summed up in 5 words ...
go see this movie today
etc
Juerd In a movie review, maybe. But I'm talking about the movies themselves...
TimToady there's a one-word metaform from Mr Rogers: "Can you say ___?" 15:53
Limbic_Region Juerd - I think it stems from that
Juerd I see
Limbic_Region but again, that's only 30 seconds of thinking about it
agentzh markstos: looking
Juerd Limbic_Region: That's 28 more than I have
I'm lazy. I don't think about such things; I ask on IRC :P
Limbic_Region heh
well, I like to think of myself as not being as bad as my fellow americans 15:54
TimToady It's a form of intellectual snobbery to be able to sum something up in very few words, I think. But you have to call your shots, like in pool.
Limbic_Region but I really have a hard time adjusting to the relaxed culture of the Philippines when I visit
Limbic_Region agrees with that too
My time is too valuable to waste explaining something to you so hear it is in N words 15:55
Juerd is often able to sum up very complex questions in one word
"Why?"
gaal \(my $x) # allowed?
audreyt sure
my$x is allows whwnever $x is allowed in exprs
Juerd Oh, gaal does that too :)
gaal so ruleCapture should c <- parseTerm, basically, recursively 15:56
TimToady the upgrade seems to have fixed my problem. so must have been disk corruption, I guess, unless 6.4.2 fixed it somehow...
audreyt not likely, seeing feather has 6.4.1 that works...
though the word "disk corruption" sounds alarming
Juerd Oh, right, I was upgrading feather. Thanks for reminding
gaal maybe 6.4.2 fixed the disk corruption
Juerd 584 upgraded, 15 newly installed, 3 to remove and 3 not upgraded.
Need to get 343MB of archives.
After unpacking 18.0MB of additional disk space will be used.
audreyt groans
gaal the Glorious Glasgow etc.
TimToady well, maybe it was just a cosmic ray. 15:57
Limbic_Region TimToady - in your cursory investigation of all the examples that were failing to parse, were most (if not all) the golf examples due to white space issues? 15:58
16:00 xinming joined
agentzh markstos: you meant the link pointed to S26.html? 16:00
markstos_ agentz: I'd like that created, but I also just added some hopeful links to the FAQ PODs 16:04
agentzh url?
for example?
markstos_ one sec.
The PODs are here: 16:05
svn.openfoundry.org/pugs/docs/Perl6/FAQ/
lambdabot Title: Revision 12728: /docs/Perl6/FAQ
agentzh markstos: ah, i see. thanks.
markstos_ I guessed they would have HTML counter parts at FAQ/Capture.html, and FAQ/FUD.html
agentzh L<FAQ::Capture>? 16:06
markstos_ The HTML versions will make them a lot more visible.
agentzh markstos: just like L<S02/...>
markstos_ agentzh: Why not use the use full module name, which would be more normal: Perl6::FAQ::Capture.
agentzh i'm fine with that. :) 16:07
markstos_ But S02 corresponds to Perl6::Spec::Syntax.
Great.
agentzh just use the syntax, and i'll look into that later.
markstos_ Thanks.
agentzh i'm more motivated to fix things when they're broken. 16:08
so feel free to leave links broken. don't comment them out.
because i'm lazy. ;)
gaal is it okay to fold \42 to &circumfix:<\\( )> inside Parser?
to put it another way, is \42 precisely \(42), for all values of 42? 16:11
markstos_ agentz. Ok. :)
agentzh markstos: btw, i'm not using pod2html. i'm using podhtm.pl living in the Pugs repos (util/podhtm.pl) :) 16:12
audreyt gaal: no. 16:13
svnbot6 r12729 | markstos++ | Go ahead and link directly to .pod FAQ files, until HTMLization happens.
gaal aw.
audreyt gaal: \42 is prefix:\ on 42, \(42) is circumfix \( )
prec is diff
markstos_ hmm. can't connect to apache on feather now. feather.perl6.nl/syn/ 16:14
markstos_ should really do the dishes now.
gaal audreyt: okay, so ruleCapture can be retired completely yes?
markstos_ wifi in the kitchen is dangerous.
audreyt aye
gaal good
TimToady Limbic_Region: yes, they were mostly place where P6 requires whitespace to keep postfixes straight from infixes or statement blocks. 16:16
Limbic_Region ok - that I think I can fix
on my own that is
TimToady if($x){.say} is illegal two ways.
16:17 zgh joined
Limbic_Region listens intently 16:17
while enjoying my breakfast that is
TimToady if($x) parses as a sub call.
Juerd Oh my
TimToady and if($x){.say} parses as a hash subscript on the sub call
Juerd Some people won't like this :)
TimToady gee, I've never heard that sentiment before. :)
Juerd Impressive :) 16:18
agentzh Juerd: feather's website is down?
Juerd I personally am glad with these changes. More horizontal whitespace would be great for many programmers.
agentzh: Could be. It's in the middle of a dist-upgrade
agentzh okay. ;)
Juerd agentzh: I never check if things work while it's upgrading. I check afterwards :)
TimToady I was actually kind of delighted to see how much clarity it forced on the golf examples. :)
Limbic_Region agentzh - yep, it's down not just you 16:19
Juerd TimToady: Don't worry. People WILL find other ways of golfing :)=
agentzh LR: i hear that. :)
Limbic_Region well, my plan was just to whitespace everything until it started working and then remove them one by one to see when they started failing
but perhaps that isn't such a good idea
audreyt on my first perlgolf tournament, I asked whether non-ascii characters count as one character... 16:20
Juerd Who's going to YAPC::Europe btw?
audreyt and perl6 golfers will certainly take advantage that the nonascii << counts as one char
svnbot6 r12730 | gaal++ | * fix \42. still orz: my $x; \$x
r12730 | gaal++ | * retire old Parser code for Captures
gaal I don't quite see why toVV' is entered in \my $x... 16:21
still looking.
TimToady audreyt: did you see my spec changes to "is" yesterday?
Limbic_Region I did
TimToady++
that one actually made sense to me
which is a scary thought
Juerd Nobody going to YAPC here? :(
miyagawa_ Juerd: i will. 16:22
audreyt Limbic_Region: your brain is being larrified
Juerd Yay, I won't be the only one from here there then :)
audreyt TimToady: I noticed, and looks sane
TimToady et tu, audrey
gaal audreyt: if you have a profbuild handy, could you stacktrace \my $x please? r12730 16:23
Juerd Je pense que c'est "toi" :)
audreyt gaal: sure. a sec 16:24
Juerd Doesn't feather have that?
audreyt is wrangling with inlinePerformSTM
Limbic_Region audreyt - well, that may be but more likely it means I am once again paying enough attention to do more than just ask annoying questions that are answered in the synopses and complain about problems without offering to help fix them
audreyt - that's the problem I reported? 16:25
16:25 Psyche^ joined
audreyt Limbic_Region: nested atomically? yes 16:25
gaal I rememeber inlinePerformIO, but what's the STM counterpart?
audreyt ditto with atomic 16:26
bbiab again
gaal: <Pugs.AST.Internals.toVV',Pugs.Eval.argsFeed,Pugs.Eval.reduceApp,Pugs.Eval.reduc
e,Pugs.Eval.evaluate,Pugs.Eval.evaluateMain,Pugs.Run.runEnv,Pugs.runProgramWith,
Pugs.doRun,Pugs.run',Pugs.run,Pugs.Run.runWithArgs,Pugs.mainWith,Pugs.pugsMain,P
ugs.CAF><Pugs.AST.Internals.toVV',Pugs.Eval.argsFeed,Pugs.Eval.reduceApp,Pugs.Ev
al.reduce,Pugs.Eval.evaluate,Pugs.Eval.evaluateMain,Pugs.Run.runEnv,Pugs.runProg
ramWith,Pugs.doRun,Pugs.run',Pugs.run,Pugs.Run.runWithArgs,Pugs.mainWith,Pugs.pu
gaal heh
audreyt gsMain,Pugs.CAF>pugs-prof: don't know how to toVV': VUndef
gaal thanks
Juerd: have what, profpugs? don't think so 16:27
Juerd Why not? :)
gaal smokepugs is a sweeter fruit and we don't have that :)
agentzh Juerd: stat info added: svn.berlios.de/svnroot/repos/unisim...n/S02.html 16:29
lambdabot Title: S02
16:30 Psyche^ is now known as Patterner
Juerd agentzh: Yay 16:32
agentzh :)
Juerd agentzh: s/--/&mdash;/ :)
agentzh okay, fixing
gaal /g :) 16:33
um no actually
not on code.
svnbot6 r12731 | agentz++ | [util/smartlinks.pl]
r12731 | agentz++ | - added test stat to links' names:
r12731 | agentz++ | 7 ?\226?\136?\154, 1 ?\195?\151
r12731 | agentz++ | which was suggested by Juerd++
gaal ah, it's just on the stats. ok ok
agentzh committing 16:34
Limbic_Region heh - agentzh, when I loaded the S02.html link and minimized it - I read the S02 as 502 and thought there was an http error 16:35
agentzh LR: (?)
Limbic_Region oh nothing - just eyes playing tricks on me 16:36
agentzh okay. :)
svnbot6 r12732 | agentz++ | [util/smartlinks.pl] 16:38
r12732 | agentz++ | - on stats info: s/--/&mdash;/ per Juerd++
agentzh the demo page is also updated. :) 16:39
so much for tonight. gotta run to bed... 16:40
night, all~~~ &
16:40 agentzh left
audreyt Juerd: apache is down for upgrading, right? 16:41
Juerd audreyt: Yes. Due to a design mistake in Debian, it's down longer than necessary.
It could come up any moment
Setting up apache (1.3.34-4) ... 16:44
Starting apache 1.3 web server....
markstos_ it's up now. 16:47
svnbot6 r12733 | gaal++ | * '\my $moose' now works.
gaal Signatures up next, I suppose 16:50
svnbot6 r12734 | agentz++ | [util/smartlinks.pl]
r12734 | agentz++ | - turn off stats info when smoke data is not available.
Juerd Can roles initialize?
i.e. Can Blah do something on "$foo does Blah;"? 16:51
For web programming, it'd be nice to have "$param<file> does Upload" and "$param<url> does ParamArray", or something alike. 16:52
audreyt runtime does?
it would trigger BUILD, I think.
TimToady the BUILD in the role 16:54
don't want to reBUILD the existing object
audreyt aye
was just too lazy to type down the whole line :) 16:55
TimToady I was just channeling you. :)
audreyt et tu...
TimToady et three
markstos_ linguists... 16:57
gaal audreyt: parsing :(), plan of antler. is circumfix:<:( )> enough? what parses typeconstraints? 16:58
ooh!
?eval Int
evalbot_12734 ::Int 16:59
gaal that's a start :)
audreyt need prolly to switch to another parser for :()
gaal parameterize makeCircumOp? 17:00
Limbic_Region wow, that STM problem must be a doozy 17:01
or perhaps I am just too used to - found the bug......fixed
audreyt it's very complicated :) 17:02
note that it only occurs when &f is used in parsing without a content
i.e.
sub f {} my &g := &f; sub h {} 1;
breaks bug
sub f {} my &g ::= &f; sub h {} 1;
works
reason is that the content of &g is consulted during parsing 17:03
but it has no content (yet)
so it goes to create a new storage cell for it
but we are currently already in the parser, which is another atomic transaction, and so it triggered explosion
clkao what's ::= ?
and is there :::= ? 17:04
audreyt clkao: BEGIN time binding
no.
gaal BEING { := }
audreyt LIGHTNESS { := }
clkao ok
gaal AND NOTHINGNESS
Juerd TimToady, audreyt: thanks.
audreyt np :) 17:06
Juerd my $foo, $bar := *%param; will be GREAT for web programming :) 17:07
And if coercion can be controlled, my @foo, $bar, Upload $baz :) 17:08
(Or are parens needed with my?)
TimToady yes
Juerd my (@foo, $bar, Upload $baz) := *%param; then?
Combined with "where", this is absolutely wonderful. 17:09
TimToady MAIN binding works the same way
TimToady needs to wander off &
gaal ok looks like end of cycles for me for today, I'm off to a bike ride 17:10
have fun y'all
Juerd Uhhuh. I'm looking for something that doesn't require a named sub, though: PHP-like programming. (No matter how much I dislike it, I must admit that for smaller projects it's much easier to both create and maintain)
markstos_ bikerides++
audreyt gaal: have fun :) 17:13
gaal thanks! & 17:15
TimToady oh, it's [,]%param now, now *%param 17:16
*not
diakopter [,]not
17:16 xinming joined
TimToady [,]now 17:17
*not
:)
markstos_ Is that the new slurp notation for subroutine params? 17:18
s/slurp/slurpy/
audreyt subroutine arguments.
params still use *.
(params are in signatures)
17:19 justatheory joined
markstos_ still ponders the difference quietly. 17:19
audreyt sub f ($param) {}
markstos_ likes examples. :)
audreyt f($arg) 17:20
makes some sense? :)
markstos_ Yes!
I had always thought the two as interchangable concepts, at least in Perl5, which doesn't really use signatures the same way. 17:21
Well, when that part my code breaks after a pugs upgrade, I'll know what to do...
wolverian looppeel
er, ww 17:22
markstos_ hunts and gathers some lunch. &
audreyt Limbic_Region: also "nmake ghci" doesn't yet work on win32 17:23
so my repl loop is 120 seconds
as opposed to 12
which significantly affects my productivity :)
Limbic_Region: but it's fixed all the same... 17:25
17:26 mako132_ joined
audreyt and I really need to fix ghci for win32 tomorrow if I'm to actually do hacking here... 17:27
but sleep first. :)
Limbic_Region: please confirm that the fix worksforyou now when you get back.
svnbot6 r12735 | audreyt++ | * examples/vmethods/time.pl: fix one typo.
r12736 | audreyt++ | * uninitialized 'my &f' no longer performs a newTVar, so the dreaded 'nested atomically' bug won't be trigered when its params are queried for parsing.
17:32 iblechbot joined
Limbic_Region audreyt++ 17:34
Limbic_Region checks
diakopter is exploring ways to make trypugs (webpugs) work so that one user can multicast their session to others. 17:39
Limbic_Region audreyt++ # down to 18 failing examples 17:45
17:51 mako132_ joined
Juerd my %foo of Bar | Baz | Array of (Bar | Baz), possible? 17:56
audreyt a smoke report of r12737 would be lovely 17:57
sadly I can't stay up to finish the smoke loop now
see you tomorrow :)
Juerd Bye. Sleep well.
svnbot6 r12737 | audreyt++ | * do not hijack \3 to mean \(3,) just now; first, it probably means \(3:); second, VRef's cast instance to newVal is not written, and so it'd break a lot of existing code. maybe fix that tomorrow...
diakopter is there anyone here who wants to talk about trypugs/webpugs ? 18:06
markstos_ diakopter: You could bounce some ideas off me. 18:08
I'm a professional website developer by day.
diakopter ok. i was one for a while also. 18:09
I've selected anyterm as a starting point for the javascript portion 18:10
markstos_ I'm not familiar with anyterm, but Ok. 18:11
diakopter server side, i'm starting with the webgui framework to handle all the http requests, since webgui contains an asynchronous daemon that can host all the pugs virtual sessions. 18:12
markstos_ unrelated: The Perl6 FAQ appears to have been written in POD originally, but is not in the pugs tree with the other FAQs. Anyone know the story? It would be nice to to merge it in and have them in one place, but I don't want to reverse engineer the POD.
dev.perl.org/perl6/faq.html
lambdabot Title: Perl6 FAQ - perl6
markstos_ dakopter: OK. I'm not familiar with WebGUI either, but I have done some AJAX development. 18:13
diakopter WebGUI was featured on perlcast a few months ago. it's a large GPL'd perl web app framework 18:14
it has never garnered viral attention 18:15
markstos_ diakopter: Yes I heard a bit about it at YAPC::NA.
svnbot6 r12738 | markstos++ | Add link to general FAQ to Feather's doc page.
markstos_ I'm used to the design of a small framework that is extensible throug plugins. Namely, CGI::Application. 18:16
diakopter I wrote several k LoC in it, including the ajax dashboard/portal application
ah yes
markstos_ But I suspect WebGUI would work fine here.
diakopter: So are you asking for a peer review of your general approach before you go further? 18:17
diakopter yes, sort of.
yes :)
overall architecture, featureset
all the implementation details are still up in the air
I want to make sure I'm not restricting the features any by any early wrong design choices 18:18
markstos_ diakopter: What do at work for programming projects is to create a "web blueprint". ...
basically, create some static HTML pages that look like the "screens" of the web app, and some notes about the logic to connect them.
It usually doesn't take long relative to the programming, and is much quicker to refine if there are "big picture" issues. 18:19
diakopter ok. this has two screens - an entry page that lists other people's shared sessions to "watch", and the "create my own session" link to go to the actual pseudo terminal
in my mind so far
markstos_ Ok. 18:20
Would you allow anonymoose, or unshared sessions?
diakopter I doubt it. pugs sessions take too much ram 18:21
too much potential for abuse
errr
yes on the unshared sessions
markstos_ :)
Would there be a max number of sessions supported at once?
diakopter i was going to enforce email address validation
yeah, defined by the app admin
markstos_ Ok.
diakopter later of course an autoadjusting algorithm could be added
markstos_ email address validation? You mean, you can't just use it casually without signing up ? 18:22
diakopter yeah, because of the abuse potential
do you have other ideas to deal with that problem?
markstos_ Wouldn't running pugs in a safe mode, like the eval bot, handle that ?
diakopter I was figuring a 2-3 minute timeout before each session is killed
hmm 18:23
18:23 aufrank joined
markstos_ How you keep each user connected to the same context, so when they type line 3 and press return, it's in the same lexical context as line 2 ? 18:23
diakopter you mean having only one active pugs session, shared among all the clients?
markstos_ I think the evalbot runs pugs in a "safe" way, so that file system interactions and other "dangerous" actions are disabled. 18:24
diakopter right but that doesn't address state persistence 18:25
markstos_ You are right about that.
Have you tried contacting the tryruby maintainer? tryruby.hobix.com/
lambdabot Title: try ruby! (in your browser)
diakopter haven't yet. 18:26
markstos_ If you could see their code and "port" it, that might be the easiest.
They must have answered the persistency issue, too.
diakopter yeah. although I think the structured tutorial is very limited. I was imagining a pugs environment where entire modules could be pasted in, loaded.
markstos_ You might also discover a good bit just by poking at the JavaScript it's running.
"entire modules" doesn't sound like a beginner trying the language. I'm not sure it's necessary to support that. 18:27
diakopter hmm; ok
point taken
another option is to persist the command history and spawn a new pugs session for each request, recreating the state 18:28
markstos_ Neat. On try ruby, watch what happens if you type "help" or "2 + 8".
It's like an interactive tutorial. The web app captures certain inputs and provides helpful feedback.
diakopter yeah I know; i've gone through it 18:29
markstos_ I'm just discovering it. :) Impressive.
diakopter the server returns the terminal response along with the html and javascript code to eval
anyterm does it similarly. 18:30
and anyterm is already gpl 18:32
markstos_ Good. It sounds like you are on the right track.
Great.
diakopter it would be neat to develop a structured tutorial
I suppose it could be integrated with a testing framework, to see if the state outcome matches what the tutorial step is demanding, to allow for timtowtdi 18:33
markstos_ Possibly. 18:34
You might also take an existing tutorial and integrate it in. For example, here's a Perl6 OO cookbook: 18:35
cog.cognitivity.com/perl6/
lambdabot Title: Perl6 Object Oriented Cookbook (v0.2.1)
markstos_ It may be too advanced to be a good match for this project, though.
diakopter so I was going to have WebGUI's daemon ("spectre") handle the pugs session management, and the WebGUI wobject (application) be a mere wrapper front-end to that time-sharing server.
hunh. hadn't seen that. 4 years old. 18:36
markstos_ Really? I just found it on Google a moment ago. I haven't read through it.
diakopter copyright 2002. i assume that means ago 18:37
age
18:37 lightstep joined
diakopter any lurkers have anything to add? please interject 18:37
markstos_ Yep. Some pages say 2003, but still.. 18:38
However, for the random pages I choose, the syntax was still accurate.
18:40 SubStack joined
Juerd Silly question... How do you flatten an arrayref? :) 18:46
That is... array :) 18:47
.foo returns an array. Is it:
*.foo
@{.foo}
lambdabot Unknown command, try @list
Juerd @ .foo
markstos_ @( .foo ) ?
lambdabot Maybe you meant: . v
markstos_ ?eval my $aref = [<a b c>]; say @( $aref ); 18:48
evalbot_12738 OUTPUT[a b c ] Bool::True
18:48 xerox joined
markstos_ ?eval my $aref = [<a b c>]; say @( $aref ).perl; 18:48
evalbot_12738 OUTPUT[["a", "b", "c"] ] Bool::True 18:49
18:50 diakopter_ joined 18:52 justatheory joined
diakopter_ markstos_ - any other thoughts? 18:59
brb
18:59 diakopter joined 19:00 rashakil joined
markstos_ diakopter: Not for now. I appreciate your efforts! 19:01
19:04 chris2 joined
Juerd ?eval [ <a b c> ].kv 19:12
evalbot_12738 (0, "a", 1, "b", 2, "c")
Juerd ?eval [ <a b c> ].kv.reverse
evalbot_12738 ("c", 2, "b", 1, "a", 0)
Juerd ?eval my %foo = [ <a b c> ].kv.reverse
evalbot_12738 {("a" => 0), ("b" => 1), ("c" => 2)}
Juerd Lovely!
diakopter ?eval [ <a..z> ].kv.reverse 19:14
evalbot_12738 ("a..z", 0)
svnbot6 r12739 | markstos++ | Overhaul's feather's index and "syn" index pages to make them a much more
r12739 | markstos++ | comprehensive Perl6 resource.
markstos_ Lovely? Why does the zero appear at the beginning rather than the end when you treat a three element array as a hash ? 19:15
diakopter apparently I haven't a clue
markstos_ takes snack break & 19:17
Juerd diakopter:
?eval [ 'a'..'z' ].kv.reverse 19:18
evalbot_12739 ("z", 25, "y", 24, "x", 23, "w", 22, "v", 21, "u", 20, "t", 19, "s", 18, "r", 17, "q", 16, "p", 15, "o", 14, "n", 13, "m", 12, "l", 11, "k", 10, "j", 9, "i", 8, "h", 7, "g", 6, "f", 5, "e", 4, "d", 3, "c", 2, "b", 1, "a", 0)
Juerd diakopter: <> is the new qw()
markstos_: Hashes are unordered
Ignore that pugs ignores this :) 19:19
tnx.nl/5294IJOR # HTTP::Headers for Perl 6 (lacks convenience methods) 19:20
diakopter each pugs instance uses 8MB ram 19:27
initially. 19:28
executing 2**112412 causes it to take up 3 more MB ram 19:30
hmm; I guess it have two modes 19:32
session persistence mode for those who validate their email addresses, and state-recreation mode for those who don't.
it could have two modes, I mean
so on a 512 MB jail, apache/mod_perl/webgui will take about 110MB ram plus 10MB ram per apache child. 19:33
so, assume 5 children 19:34
44 available empty pugs sessions
not allowing for swapping.
that would be more than plenty, I think. 19:35
not accounting for whatever ram the system is using. assume 200MB
okay, so 19 pugses 19:36
diakopter ceases spamming the channel 19:37
(for now)
19:37 chris2_ joined
markstos_ diakopter: Sounds like reasonable math. 19:38
diakopter hmm; and a 512MB jail from prgmr.com is.... 19:40
well, the site was up yesterday. 19:42
19:48 chris2_ is now known as chris2
svnbot6 r12740 | markstos++ | fix broken <li> tags, which must have been broken in kwiki, too, since they were copy/pasted from there... 19:50
19:56 araujo joined
diakopter finally gets a local WebGUI dev environment set up. 19:59
markstos_ Perl6 blogging about references vs. Capture objects: use.perl.org/~markjugg/journal/30762
diakopter prepares to hack spectre to host pugs sessions
lambdabot Title: Journal of markjugg (792)
gaal audreyt: re: Signatures needing alternate parsers - so do Captures! makeCircumOp+parseNoParenParamList doesn't permit \(1; 2) 20:12
markstos_ diakopter: I just got this result on tryruby. It provides a hit about how it works: "YourĀ sessionĀ hasĀ beenĀ closed,Ā eitherĀ dueĀ toĀ inactivityĀ orĀ aĀ bitĀ ofĀ codeĀ whichĀ ranĀ tooĀ long.Ā RefreshĀ theĀ pageĀ andĀ youĀ canĀ beginĀ aĀ newĀ session.Ā " 20:15
Juerd ?eval Object.new 20:25
evalbot_12740 Object.new()
diakopter markstos_: ah, okay 20:33
20:35 elmex joined
diakopter I'm going to posit a guess that he's doing state reconstruction via each's session history 20:35
markstos_ diakopter: I tried to confirm if you could even refer to a previous line from the next line, but I don't know enough ruby to even declar a simple variable. :) 20:37
diakopter ah; it doesn't like 2**500000 20:39
also: "If you happen to leave or refresh the page, your session will still be here for unless it is left inactive for ten minutes." 20:40
[sic]
Juerd 2;0 juerd@feather:~$ pugs -e'my $foo = [ 1..5 ]; say $foo.perl' 20:41
[1, 2, 3, 4, 5]
Eek
Wrong channel
20:44 Limbic_Region joined 20:53 chris2 joined 20:59 justatheory joined
svnbot6 r12741 | markstos++ | Add some practical examples of life without references to Differences.pod 21:09
Juerd tnx.nl/5296LSEC 21:11
lambdabot Title: TNX
Juerd I wonder if it will compile. I haven't had the guts to try it, yet.
markstos_ Juerd: Why not try? 21:14
Juerd Tried. 21:15
Needed to change all @foo ==> map { .. } to non-pipe form
I thought pipes worked
Also, I thought hyper @foo>>.keys worked, but it doesn't 21:16
Now I have something that compiles.
Exciting
?eval class Foo { has %foo; method bar { %foo } } Foo.new.bar 21:17
evalbot_12741 Error: Undeclared variable: "%foo"
Juerd ?eval class Foo { has %.foo; method bar { %.foo } } Foo.new.bar 21:18
evalbot_12741 {}
Juerd Hm
?eval class Foo { has %!foo; method bar { %!foo } } Foo.new.bar
evalbot_12741 {}
Juerd Hm, ok
Thought you could leave out the !
markstos_ I guess not.
diakopter markstos_: I think I'll just do the state-recreating commandHistory-persisting version first :) 21:19
markstos_ diakopter: Sounds good! 21:20
Here's my own problem. I have an hash element that is an array that I want to append to, but I get a syntax error with this:
my %h; my $b = 5; %h<a> = @( @$b, @( %h<a> ));
?eval my %h; my $b = 5; %h<a> = @( @$b, @( %h<a> ));
evalbot_12741 Error: unexpected "(" expecting "::"
markstos_ Hmm, I think I just tried to hard to re-invent 'push'. I found an alternate solution. 21:24
Juerd I give up. 21:31
Pugs isn't at all complete enough to do this. 21:32
On %!headers.keys it calls &keys in its own class, instead of Hash's
And I don't like moving around syntax all the time. That's nice if you have lots of time, but unfortunately, I don't. 21:33
Juerd postpones trying Perl 6 again
Do we have a historic graph or table of how many tests pass? 21:37
Or perhaps a chart, for the nitpicky :) 21:38
21:46 diakopter joined
TreyHarris Juerd: no, I asked the other day and apparently the problem is that only HTML is saved, not TAP output 21:47
In any case, audreyt's said that at this point she's just dealing with regression as she tries to replace the object model and implement Captures correctly
so tests that used to fail, she's fixing to make pass, but new tests she's allowing to stay failing
but fixing the object model and the captures is what's required to make feeds, hypermethodcalls, etc., work right, so it is an indirect answer to you, if not a direct one 21:48
21:49 markstos_ joined
Juerd Ah, that might explain why objects work so badly now 21:49
markstos_ Objects are working OK for me... but perhaps I'm not working them as hard as you are... 21:51
TreyHarris i've found workarounds for almost everything, but i'll grant it's annoying to have to workaround when Perl 6 has such nice syntax that's not working
Juerd ?eval class Foo { has %.bar; method keys { die "shouldn't be called" }; method quux { %.bar.keys } } Foo.new.quux 21:52
evalbot_12741 Error: shouldn't be called
Juerd ?eval class Foo { has %.bar; method xkeys { die "shouldn't be called" }; method quux { %.bar.keys } } Foo.new.quux
evalbot_12741 ()
Juerd Method resolution is wrong.
markstos_ Doh.
Juerd Foo has nothing to do with %.bar
But Pugs does think it does 21:53
markstos_ ?eval print if 'hello world' ~~ m/hello/;
evalbot_12741 OUTPUT[] Bool::True
markstos_ print if 'hello world' ~~ rx/hello/;
?eval print if 'hello world' ~~ m/hello/;
evalbot_12741 OUTPUT[] Bool::True
markstos_ That little test doesn't work here. It complains that parrot is not installed. Do I really have to install for Regexp support with pugs ?
Juerd In boolean context, rx// and m// are the same.
markstos_: AFAIK, yes. Otherwise you only get Perl 5 regexes. 21:54
And then only if you have Perl 5 embedded.
markstos_ So that was the shortcut to get regular expressions working so quickly...
Juerd Shortcut?
markstos_ As opposed to implementing them in Haskell. 21:55
Juerd These are important techniques.
markstos_ embedding parrot, you mean ?
Juerd Implementing them in Haskell is, in my opinion, madness, if we already have them elsewhere.
TreyHarris ?eval class Foo { has %.bar; method keys { die "shouldn't be called" }; method quux { %(%.bar).keys } }; Foo.new.quux
evalbot_12741 Error: shouldn't be called
markstos_ As a user, it's just an annoyance at the moment, because I have to something else to install now.
Juerd Implementing rules in Parrot is (I think) much harder than implementing them in Haskell, so I wouldn't call it a shortcut.
markstos_ I see your perspective about how it's a good technical decision. 21:56
It just conflicts with convenience at the moment. :)
Juerd Pugs isn't really built for users just yet :)
markstos_ I've been noticing...
I like it anyway. :)
Juerd I like it, but every time I try to use it, it's just a frustrating experience. 21:57
My brain already speaks Perl 6, and has a hard time accepting that much of it isn't implemented yet. 21:58
merlyn my brain still speaks perl4 as enhanced by perl5
TreyHarris lol 21:59
Juerd Things like @foo.kv.reverse come entirely naturally to me, before I try if they work.
In Perl 5 it'd be { my $i = 0; $bar{$_} = $i++ for @foo }
Much more complex.
Another thing that I like is @foo ==> map { ... } ==> map { ... }, but Pugs doesn't do that yet 22:00
Or @foo>>.keys
Err, @foo>>.key
merlyn $bar{0..$#foo} = @foo
@bar{0..$#foo} = @foo
lambdabot Unknown command, try @list
merlyn that's how I'd do that
Juerd merlyn: @bar{@foo} = 0..$#foo
merlyn oh, right
backwardsin 22:01
Juerd Good idea. I didn't think of it.
markstos_ Juerd: Perhaps I should just get an account on feather. From this page, it looks like parrot 0.4.5 is broken on FreeBSD, and thus I assume 0.4.6 wouldn't work any better.
www.freshports.org/lang/parrot
lambdabot Title: FreshPorts -- lang/parrot
Juerd markstos_: Perhaps you should indeed. Do you know how to request one?
TreyHarris but you hit the nail on the head: the tension with Perl has always been the most brilliant code is rejected by the people that matter (e.g., managers, learning programmers) because it's too opaque. i think in Perl 6 a lot more of the brilliant is also obvious 22:02
markstos_ checks feather website
Juerd I like it that people call that single page a website 22:03
markstos_ The instructions say to let Juerd know.
Juerd Oooohhh!
markstos_ You looked at the website, didn't you ?
Juerd Someone made it much better.
markstos_ grins...
Juerd I hadn't seen the "site" in this form yet
markstos_ That was me, a few hours ago.
TreyHarris how did you update feather without having an account on it? 22:04
markstos_ I'm L33t?
Juerd Hehe
markstos_ subversion and cron.
Juerd Long live wiki style delevopment :)
markstos_: Thanks man :)
markstos_ Juerd: about adding more pages:
I'm confused that syn/index.html is stored as syn_index.html in svn. 22:05
I expected directories work work as a 1-to-1 mapping with svn and the feather directory.
Juerd markstos_: Please request your account by sending me an email with your real name, and preferred username
markstos_: This all got changed recently, and I haven't yet seen how they did it 22:06
syn symlinks to somewhere
syn/index.html symlinks back to syn_index
I have no idea why, but there's probably a good reason.
markstos_ to ~agentz/syn, I htink.
It was probably more convenient to agentz. Anyway, other directories can created normally under there ? 22:07
Juerd Very probably.
22:07 Oswyn joined
markstos_ Ok. I expect I'll continue to work on this site some of I have time. 22:08
Juerd markstos_: Please use normal quotes, not backticks 22:14
svnbot6 r12742 | juerd++ | s/autrijus/audreyt/
markstos_ Where are backticks used? in the list of links to kwiki ? 22:15
svnbot6 r12743 | juerd++ | s:g/``|''/"/
22:15 ajs_ joined 22:19 justatheory joined
Juerd markstos_: In quotes. I don't know if you did that, or someone else. 22:22
markstos_ That doesn't sound like something I would do, but it's possible I copy/pasted some HTML from kwiki that was strange that like. 22:23
BTW, How I can view a patch with svn ? I tried svn diff -r123, but that didn't work as I expected. 22:24
Nevermind, I figured that out:
svn diff -r 122:123 22:25
It's simpler with darcs, just darcs diff -p 'human patch name'
No number lookups. :)
svnbot6 r12744 | juerd++ | Random website improvements 22:30
markstos_ waits patiently for feather login 22:32
Juerd Sorry, forgot to re-enable my mail beep 22:35
markstos_ Beep! 22:36
Juerd Your password is markstos. Please change it immediately.
markstos_ updated. 22:37
Thanks!
Juerd You're welcome, in both senses 22:38
Please create a ~/.qmail if you're not going to read email on feather 22:39
And please create a ~/.plan describing your intentions with the machine
Possibly hidden from the outside world by adding a ~/.nofinger
22:40 diakopter joined 22:41 jferrero joined 23:00 diakopter_ joined 23:01 diakopter joined 23:06 mako132_ joined 23:07 diakopter_ joined, xdg joined
svnbot6 r12745 | juerd++ | google box 23:07
r12746 | juerd++ | typofix 23:10
r12747 | juerd++ | It's better on one line :)
r12748 | juerd++ | html improvements 23:19
araujo didn't know compiling pugs would take so long 23:21
Juerd It used to take much longer even 23:23
araujo oh ok, so is it a normal behaviour Juerd ? 23:25
i was thinking it was probably my box
sometimes some packages are compiled very slowly
Juerd It could *also* be your box. 23:30
What are the specs of your box?
araujo amd64 , 1.8ghz , 512ram 23:31
araujo needs to get more memory
Big haskell applications tend to compile too slow :-( 23:32
ready! 23:34
Juerd 512 MB is just enough for Pugs, but you'd better not be running X and firefox at the same time :) 23:37
512 MB is a bit on the low end for an amd64 1.8 - was this a conscious choice? 23:38
23:58 dolmans 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!