Camelia, the Perl 6 bug

IRC log for #parrot, 2008-02-06

Parrot | source cross referenced

| Channels | #parrot index | Today | | Search | Google Search | Plain-Text | summary

All times shown according to UTC.

Time Nick Message
16:58 ilbot2 joined #parrot
16:58 Topic for #parrotis now  #parrot Parrot 0.5.2 Released | http://parrotcode.org/ | see http://www.parrotcode.org/misc/parrotsketch-logs/ for logs
16:58 Topic for #parrotis now  #parrot Parrot 0.5.2 Released | http://parrotcode.org/ | see http://www.parrotcode.org/misc/parrotsketch-logs/ for logs
16:59 moritz I know got logs online here: http://irclog.perlgeek.de/parrot/today
17:02 moritz s/\bk//
17:18 Theory joined #parrot
17:27 sjansen joined #parrot
17:29 sjansen Is there any way to get piper to shut up about the channel being logged? I think telling everyone who wanders in is noble, but I'm now _fully_ informed.
17:29 particle we're working on it
17:30 particle right now we don't have the source. we're trying to get it in an svn repo, so we can modify at will.
17:33 avar ..or use the bot #perl6 uses
17:33 moritz avar: I just fired it up to log here as well
17:35 avar let's ban Piper !
17:37 particle don't overreact. piper's not any worse than freenode's nickserv & freenode-connect
17:42 diakopter for some, joining a channel is done more often than connecting to a network, so in those cases, it is worse.  That, and magnet/piper don't do authentication (like nickserv can).
17:44 * diakopter welcomes ilbot2
17:44 diakopter purl, meet ilbot2
17:44 purl diakopter: sorry...
17:44 diakopter purl, love on ilbot2
17:44 purl diakopter: i'm not following you...
17:44 diakopter purl, figures.
17:44 purl diakopter: excuse me?
17:50 paq joined #parrot
17:59 cognominal_ joined #parrot
18:16 Coke as soon as we get piper source, we can hack on it. (I would like to use something that lets us keep the logs going where they're going.)
18:18 * Coke wonders if you could write a sudoku solver using just perl6 rules.
18:19 moritz Coke: I've seen a mini (4x4) sudoku solver in p5 regexes
18:19 particle i don't see why not
18:19 moritz it's a nice backtracking problem, you just have to formulate it in a way that it fits into regexes
18:27 zaphod joined #parrot
18:40 alvar joined #parrot
18:43 zev Coke: http://perl.plover.com/NPC/NPC-3SAT.html and http://perl.plover.com/NPC/NPC-3COL.html
18:54 ruoso_ joined #parrot
19:00 Limbic_Region joined #parrot
19:00 Limbic_Region Juerd ping
19:00 Juerd pong
19:01 Limbic_Region did something just happen with feather?
19:01 Juerd nafaik. What kind of something?
19:01 Limbic_Region I just tried to logon to irc chat and was prompted for my username/password
19:01 Tene Like a song and dance number.
19:01 Juerd Limbic_Region: Errh
19:02 Juerd Limbic_Region: It's supposed to do that - just once, though.
19:02 Limbic_Region I gave it the same info 3 times and it said "authorization failed - either these are the wrong credentials or your browser doesn't know how to provide them"
19:02 Juerd Hm
19:02 Juerd MITM?
19:02 purl MITM is Man-in-the-Middle
19:02 Limbic_Region and then
19:02 Juerd [Wed Feb 06 19:59:26 2008] [error] [client 127.0.0.1] AuthExtern MyShadowPwauth
19:02 Juerd [/usr/local/bin/pwauth]: Failed (51) for user limbic
19:03 Juerd [Wed Feb 06 19:59:26 2008] [error] [client 127.0.0.1] user limbic: authenticatio
19:03 Juerd n failure for "/cgi-bin/cgiirc/irc.cgi": Password Mismatch
19:03 Juerd They're actual failures on feather
19:03 Juerd Any idea why it thinks you're at 127.0.0.1?
19:03 Juerd ssh tunnel?
19:04 Limbic_Region joined #parrot
19:04 Juerd 20:03 <@Juerd> Any idea why it thinks you're at 127.0.0.1?
19:04 Juerd 20:03 <@Juerd> ssh tunnel?
19:04 Limbic_Region this is odd
19:04 Limbic_Region I was chatting away and I got asked for my credentials again
19:04 Limbic_Region but if I close the page, and go back - I don't get prompted at all
19:04 Limbic_Region and no, I have no idea why it thinks that
19:05 Limbic_Region I am coming in from behind a corporate proxy
19:05 Juerd Your browser caches credentials
19:05 Juerd So if you have ANY window open, it keeps them
19:05 Juerd What port are you using?
19:05 Limbic_Region except that it is the same credentials that feather said is wrong
19:05 Juerd Oh, https of course.
19:05 Juerd sslh proxies that on feather, hence the 127.1
19:06 Juerd Well, at this point I don't know what's wrong
19:06 Juerd Could be your keyboard, in theory :)
19:06 Juerd Or anything else in the chain
19:06 Limbic_Region nor do I - but it is incredibly odd.  I give the same password 3 times and it says it is wrong.  I reload and it takes me straight to IRC as though the password was right
19:06 Juerd Uhhuh, I agree it is strange.
19:06 Limbic_Region and then while I am chatting - it pops up the authentication window again
19:07 Juerd I'm not going to look into it now, though. If it happens more, please ping me again.
19:07 Limbic_Region whatever - seems to be working fine now
19:07 Juerd I'm blaming your browser for now :D
19:07 Juerd Easiest for myself
19:07 Limbic_Region joined #parrot
19:18 jisom joined #parrot
19:19 Coke http://groups.google.com/group/perl.cpan.tes​ters/browse_thread/thread/586052a2ba760877/9​d1ead86320ee9d3?hl=en&amp;q=parrot+-site%3Ap​erl.org+-site%3Aparrotcode.org+-site%3Acpan.​org+-perl.perl6+-perl.cvs#9d1ead86320ee9d3 :: why is this just showing up now?
19:19 shorten Coke's url is at http://xrl.us/bfrrk
19:22 particle coke: that's likely before Makefile.PL existed
19:22 particle and we should probably remove old parrot distros from cpan
19:38 paq joined #parrot
19:38 Coke particle: +1 # I've asked about that before.
19:39 Coke before 1.0, we can get away with just the last 2 releases, I think.
19:46 kj oh this looks interesting (reading the title): http://listeningtoreason.blogspot.com/2008/​02/net-dlr-makes-building-new-language.html
19:46 shorten kj's url is at http://xrl.us/bfrtc
19:47 kj let's check if they can beat the usability of PCT :-)
19:48 moritz looks like they stole some PCT ideas ;)
19:50 kj looks like it. Of course, maybe they thought of it before... the idea of a universal set of AST nodes was conceived years a go
19:50 kj for parrot at least
19:51 zaphod one could say that LISP thought of it in the 60s
19:51 kj heh :-)
19:51 kj also for new languages?
19:51 zaphod such as?
19:52 kj sorry, I misunderstand
19:52 zaphod I was wondering what you meant by "also for new languages?"....maybe I'm misunderstanding :)
19:53 kj what they say on this site is that you only have to construct the AST, and you're done. That's the approach with PCT as well. What do you mean with LISP in that context?
19:53 zaphod you can (almost) look at LISP as simply a serialization of an AST
19:53 kj oohok. But LISP was actually designed to be a language by itself, no?
19:54 kj it's way before my time :-)
19:55 zaphod LISP was an implementation of the lambda calculus.
19:55 zaphod But the syntax for code is the same as the syntax for the data structures.  So a LISP program is just a LISP data structure and ends up being a tree
19:56 kj ah i see
19:56 teknomunk joined #parrot
19:58 paq joined #parrot
19:58 zaphod An interesting read, and pertinent to parrot probably is Olin Shivers' PhD dissertation.  He works with Scheme and a representation of Scheme programs in Continuation Passing Style.
19:59 zaphod http://citeseer.ist.psu.edu/630499.html :: not a bad read
20:01 TimToady phone
20:15 * diakopter points TimToady toward http://www.codeproject.com/KB/recipes/Irony.aspx
20:20 kj I'm happy to write some article on using PCT to implement a compiler, but I'd like to give a complete example including break and return statements
20:20 zaphod return statements are not currently supported in PCT AFAIK
20:21 kj yeah i know. that's why i mentioned it ;-)
20:23 zaphod Although the small compiler I wrote doesn't have break or return it might be good for an intro article since it is so small and simple
20:25 zaphod and I would be willing to help with the article :P
20:27 kj why don't you write it? :-)
20:29 Tene I've been thinking about writing some documentation...
20:30 Coke documentation++
20:30 Coke documentation++
20:30 Coke documentation++
20:30 kj Tene++ # THINKING about documentation is the first step :-)
20:30 kj any particular topic?
20:30 Coke knowing is half the battle?
20:30 kj heh :-)
20:31 zaphod kj: hmmmm...what kind of article were you thinking of
20:31 Tene language implementation, PCT stuff, everything I wish I had had a good document about when I started playing with parrot.
20:31 kj zaphod: using the PCT to implement a non-trivial language
20:31 kj we've seen abc
20:31 kj it's kinda trivial
20:32 kj (i think)
20:32 Tene lolcode is a bit better, but isn't finished yet.
20:32 zaphod what would be non-trivial?
20:32 Tene Still needs some things.
20:32 kj Tene: I wrote some docs on PCT in docs/pct/
20:32 kj Tene: if you have a look at those, and see what's missing in your opinion
20:32 Tene Hmm.  I should read those to avoid duplication.
20:32 kj then you could add the next step?
20:33 Tene Probably.
20:33 purl Really? Probably? Are you Certain it's not certain? Are you sure it's unsure? I think you need to look harder.
20:33 Tene Hmm... I need to figure out what to work on with particle tonight.
20:33 kj zaphod: non-trivial, well, Lua came to mind when I was typing. NOt sure if it's too non-trivial :-)
20:34 Tene Hopefully I can really attack cardinal next week.
20:34 zaphod yeah...non-trivial quick starts steping into the bounds of book instead of article :D
20:34 Tene I'm off work for most of next week.
20:34 kj Tene: cardinal is slow like **** currently
20:34 Tene I noticed.
20:34 kj it has to do with overlapping subrule I think
20:34 Tene I think I can clean that out a bit.
20:35 Tene I have some fixes to the AST that let assignment to variables work, but I haven't made a clean patch yet.
20:35 Tene one of the things was setting 'is lvalue' on infix:= in the grammar, iirc
20:35 kj zaphod: that's true, but Lua is quite straightforward in most parts. Some complex parts are lexicals for instance, but also the "leaves" of the AST: the behavior of values
20:36 kj (which can be implemented in PMCs btw)
20:36 kj zaphod: it could be a series of articles
20:37 kj with exercises "left to the reader" :-P
20:37 kj (which are given in the next episode)
20:38 zaphod sounds interesting and not too hard (I hope)...I'll give it some thought.  I am just getting into this, and would like to do more, but I start working on my master's work in march, so time will be a little short
20:38 kj ah ok. I'll be getting a Real Job myself too soon.
20:39 kj I sent a patch to fperrad, the maintainer of the lua implementation; the patch contains the start of a PCT based impl.
20:40 kj anyway, I really like to do it if a complete implementation can be provided... if the hard parts are left out, there's not much point in doing  it
20:41 kj tene: one thing that could use a manual is using the optable in a PGE based grammar
20:43 zaphod Sounds like an interesting languge for the articles would be one with interesting variables (to bring in PMCs), standard control structures (to mention most of the PAST::Op things), scopes/subs/closures (PAST::Block) and maybe even something that could have some simple but noticable optimizations done to the PAST (to show a new stage in the compiler)
20:43 kj (of course, copying an optable from another language is one solution, but you're bound to encounter some error messages if you make the slightest errror)
20:43 Coke Whee: http://www.perlfoundation.org/who_s_who
20:44 kj zaphod: yes. different articles: one about pmcs, one about operators, one about PAST nodes and using "make" command
20:45 kj not sure about optimzations... I don't have a clue how PAST could be optimized
20:46 kj except for the obvious register loading/storing of vars and such
20:46 particle congrats, coke!
20:46 zaphod I got it in my small compiler to provide tailcall (a functional language without tailcall is a short running program)
20:47 kj a tailcall eh. you emitted that :inline?
20:47 zaphod yeah...with a hack because PAST doesn't like the inline doing its own .return
20:49 kj afk for a bit
21:06 Coke particle: danke.
21:09 alvar joined #parrot
21:19 davidfetter joined #parrot
21:21 Coke ... crap, I'm going to have YA email address, aren't i.
21:22 svnbotlt r25555 | kjs++ | trunk:
21:22 svnbotlt : [docs] add a document about optables in pge-based grammars. this is a stub.
21:22 svnbotlt diff: http://perlsix.org/svn/parrot/revision/?rev=25555
21:34 svnbotlt r25556 | kjs++ | trunk:
21:34 svnbotlt : [docs] add a bit to optable parsing guide. this is as much as i can do right now. more later (if nobody beats me to it)
21:34 svnbotlt diff: http://perlsix.org/svn/parrot/revision/?rev=25556
21:35 integral joined #parrot
21:38 shamu hi folks
21:38 davidfetter hi
21:38 shamu someone just committed something that causes 'perl Configure.pl' to fail -- Checking MANIFEST...No such file: languages/plua/config/makefiles/root.in
21:38 shamu No such file: languages/plua/lua.pir
21:41 kjs_ joined #parrot
21:41 Coke kj: bad dog.
21:42 kjs_ hey! I was trying to add documentation
21:42 kjs_ uhm. let's see. what's the easiest way to do this
21:43 Coke let me fix it.
21:43 Coke (that's the easiest. =-)
21:45 kj damn. something went wrong here too... now my file is gone :-(
21:45 shamu adding documentation caused the manifest to be wrong?
21:45 spinclad local change leaked into repo?
21:46 kj yea
21:46 kj shamu: i created a patch earlier today
21:46 kj which made svn think i added these file
21:46 kj so when regenerating manifest, they were added too
21:46 kj i forgot about thta
21:46 shamu ah ok thansk
21:47 Coke if you've svn added, but not committed, those changes show up.
21:47 kj exactly
21:47 Coke which is SUPPOSED to prevent issues where someone does the add, updates the manifest missing those files, and commits.
21:47 Coke MANIFEST MUST DIE.
21:47 shamu ah ok
21:48 kj shit. now my file is gone completely. Anybody has a copy of docs/pct/pct_optable_guide.pod?
21:48 shamu which commit was that?
21:48 Coke kj: MANIFEST updated.
21:48 kj Coke: thanks
21:48 zaphod kj: I seem to
21:49 particle kj: revert it
21:49 kj I guess it's time to get a clean parrot again..
21:49 Coke kj: ... just svn up that one file?
21:50 Coke (er, what particle said.)
21:50 kj i added that one file yes
21:50 Coke "svn revert <missing filename>
21:50 kj ah ok
21:50 kj (wasn't sure what you meant)
21:51 shamu thanks, it works now.
21:51 kj yay! it's back
21:51 kj Coke: thanks again.
21:52 svnbotlt r25557 | coke++ | trunk:
21:52 svnbotlt : back out some inadvertent additions to the MANIFEST.
21:52 svnbotlt diff: http://perlsix.org/svn/parrot/revision/?rev=25557
21:52 Coke np.
21:56 zaphod left #parrot
22:17 jonathan evening @all
22:17 Tene Yes, it is!
22:22 sjansen A mighty fine afternoon.
22:28 Andy joined #parrot
22:36 * Tene has an idea.
22:37 * moritz has lots of ideas, most of them are crap ;)
22:39 jonathan Last night I had an idea. "Let's hack on Rakudo before I sleep!" Then I went to bed and lay awake for two hours working out how to implement More Stuff.
22:40 Tene It might be pretty straightforward to write a graphic app to explore parse trees and PAST trees.
22:41 jonathan Tene: Parrot has SDL bindings OR just get the dump of it and parse it and use it in whatever you like.
22:41 Tene Right.
22:42 Tene SDL wouldn't be much fun.
22:42 jonathan I've never used it; I haven't written any meaningful GUI app in years.
22:42 Tene I just need a simple gtk or tk treeview
22:44 Alias_ Tene: Install PPI::Tester
22:44 Alias_ There is a precedent for this sort of thing
22:45 Alias_ I wrote that for PPI debugging, it was mostly simple to write and incredibly useful
22:46 Alias_ Cut per-bug fix time down from an hour to 15 mins
22:46 * Tene bbiab
22:47 DarkWolf84 joined #parrot
22:52 gryphon joined #parrot
23:09 wknight8111 joined #parrot
23:40 kid51 joined #parrot
23:40 Andy joined #parrot

| Channels | #parrot index | Today | | Search | Google Search | Plain-Text | summary

Parrot | source cross referenced