| |
| |
| |
|
Page: 1 2 3 4 5 6 7
Comments:
<0> Heh :)
<1> A good book.
<1> I also like "The Mind's I"
<2> I wrote that down for some reason. Maybe I should figure out why.
<1> And "Metamagical Themas"
<2> I just get randomly fascinated by Godel's theorems and their equivalents in other areas
<2> and pondering if they're related to P?=NP, and other areas that I don't really have any formal education in
<1> Recursion is fun.
<2> Strange loops are too.
<1> And fractals.
<2> I was just making a note of something on one page in which he explains that human intelligence must be composed of rules... and metarules... and metametarules... and so on, infinitely.
<2> Meh. Fractals are kinda mind-boggling. I spent too much time staring at the mandelbrot trying to see "Where does the order come from?!?!?! AAAAAAAAAAAAAAAUGH!"
<1> Order appears in the strangest places.
<2> Yeah.
<2> Mmmmm, stochastic resonance...
<1> Sounds good. :)
<3> morning
<1> Morning.
<2> I should be going to bed in the next two hours or so.
<2> If I can sleep.
<1> Yeah, same here.
<4> well at this point with respect to FC4 x86_64 it appears that cmucl 19b is the last release that works and that the sbcl is AMD specific.
<4> so with the sbcl situation being a new port and the problem w CMUCL (in 'save-lisp') being a well known one, there is nothing to report.
<5> new port?
<6> morning
<7> So, I don't suppose anyone wants to help me get started with clg for a bit. I have the latest CVS version, compiled and is capable of running the "hello world" program included... Now I'm just wondering how to get it to load in the glade file, and then let me poke it.
<5> Xophe: regarding the horror in low.lisp which you mentioned a couple of weeks ago, using typecase instead of case on (cl***oid-name ...) seems to work fine
<6> well, that's some good news :-)
<6> looking at your stepper patch, one thing that I'm trying to eradicate (slowly) is the use of "type" as a variable when it doesn't mean a lisp type
<6> I prefer "kind" if possible
<5> thanks. I'll change that
<6> (looks nice, though)
<8> I intend to learn the Lisp programming language on my own, on my free time, and using worthy internet lisp-learning resources. Can anyone ***ist me by giving me some links to good resources?
<9> minion, tell luakagon about clhs
<10> luakagon: direct your attention towards clhs: To look up a symbol in the HyperSpec, try saying "clhs symbol". For more information on the HyperSpec see http://www.cliki.net/CLHS .
<9> minion, tell luakagon about learning
<10> Sorry, I couldn't find anything in the database for ``learning''.
<9> Err.
<11> minion: tell Luakagon about pcl-book
<10> Luakagon: look at pcl-book: "Practical Common Lisp", an introduction to Common Lisp by Peter Seibel, available at http://www.gigamonkeys.com/book/ and in dead-tree form from Apress (as of 11 April 2005).
<8> Gracias piso and minion.
<12> OT, but what is it with people who drink coffee with the equivalent of about 15 teaspoons of sugar?
<13> rkrush: They don't like coffee, but they like the energy boost?
<9> I have coworkers who respond to my complaints that the coffee tastes bad that coffee is SUPPOSED to taste bad, and there is something wrong with it if it does not.
<9> They would probably go into shock if they ever tasted real coffee.
<13> *sipping a cup of infusion-brewed freshly ground french-roast*
<9> The free "coffee" served at my employers is some sort of serfdom-accentuating device thought up by some thoroughly sadistic enemy of mankind.
<9> Fortunately, tomorrow I will resign.
<13> Oh? Interesting new job, or just feel a need to change?
<9> I've felt the need from the first day I worked here, but now I get some sort of sensible alternative.
<9> Whether it's actually interesting remains to be seen. .-)
<9> I certainly hope so.
<13> Any place I'll have heard of?
<14> trying it out is not sufficient, but absolutely necessary
<9> No, it is small and only just come into existence.
<13> Ah. What sort of thing will you be doing?
<9> Nerdiness related to programming, I think.
<13> Heh. Never would have guessed. ;)
<9> I hope more technical work and less soul-****ing meetings and debate.
<9> Debate is ok if only it goes under the heading of technical and not soul-****ing.
<15> you ain't gonna work for no soul-****ing jerk
<5> Xophe: do you have time for a typesystem question?
<16> jsnell: go ahead
<5> I noticed that the evaluator patch still has (deftype compiled-function () 'function). I tried redefining that as '(and function (not sb-eval:interpreted-function)), but apparently standard CL types shouldn't unparse to composite types like that
<5> which places do I need to change to properly define the type?
<16> oo, that's tricky
<16> possibly the easiest way around that is to add a special case in the intersection :unparse type method
<16> I think there might be a similar special case for the KEYWORD type
<16> the trickiness comes from the fact that an object can change from being not typep compiled-function to being typep compiled-function without changing its identity, so compiled-function is definitely not a cl***oid-like type
<16> so I think special-casing the unparse is the right answer
<5> I'll try that. thanks!
<17> I am using slime, and occasionally the print-eval-debug? loop will stop. I will be typing several malformed lisp expressions, and they show up, but the lisp environment stops
<17> i kill slime and restart, but perhaps this is familiar to somebody
<18> zithowa: shouldn't happen.. what does the *inferior-lisp* buffer say when it locks up like that?
<16> jsnell: could you give me a quick state-of-play summary on the status of (a) Unicode / alien / filesystem etc and (b) the clos-typechecking stuff?
<5> the unicode-alien stuff has been committed, with the changes we discussed (add :element-type), except that I didn't add the ascii-string alien-type (people can just define it themselves if it's really needed)
<5> it's possible that the current implementation where the alien / pathname external formats are controlled by the same variable will prove to be insufficient
<16> so the pathname external format is :default for filesystem functions within sbcl itself?
<5> right
<16> ok
<5> and it's conceivable that someone could want say utf-8 for aliens in general, but iso-8859-1 for the pathnames
<17> antifuchs: the last line in *inferior-lisp* is NIL preceded by ;;Swang started at port: 32795
<16> jsnell: sure, yes
<5> but I thought it'd be better to get some real-world experience with it before over-engineering
<16> that's OK; let's see how many complaints we get now about people not being able to read their mp3 library directories :-)
<17> antifuchs: i may have been wrong to say emacs feezes. I can still operate emacs, but the slime buffer behaves incorrectly
<16> "The archaic character U+10341 GOTHIC LETTER NINETY was given the numeric value 90." -- go, go, Unicode 5.0 :-)
<17> this usually happens along with a 'commencing garbage collection' message in the minibuffer
<5> zithowa: is that on older version of cmucl? if so, upgrade to 19c
<16> perhaps more interestingly, Unicode 5.0 supports stable case-folding, which probably makes char-equal less painful
<5> I don't think I've done anything about clos typechecking after we last discussed it
<16> fine
<17> $cmucl --version outputs CMU Common Lisp CVS release-19a 19a-release-20040728 + minimal debian patches
<5> so http://jsnell.iki.fi/tmp/clos-typechecking-3.diff is probably what I have
<17> aha, not 19c
<5> for solving the existing test failure, I think we can't use setq detection in the walker as I thought previously
<16> oh? Why not?
<16> (that was what was in cmucl briefly before it got removed again when people observed some unclear symptom with mcclim...)
<5> since the walker might encounter the setq after the slot-value
<16> oh, it only walks once? Hm
<5> at least that's how I thought it worked
<16> you're probably right. The solution is then to rewrite slot value calls to a compiler macro form, have the walker return a symbol-macrolet magic binding if it's found a setq, and let the compiler macro function check for that in its environment?
<17> jsnell: do you know much about CMUCL on Debian? CLiki indicates something is broken right now
<5> zithowa: sorry, I use neither of those
<16> making the expansion something like (symbol-macrolet ((%setq-p% 'yes)) (maybe-pv-optimized-set-slot-value arg 'x 3) (setq arg 17) ...) for (defmethod foo ((arg bar)) (setf (slot-value arg 'x) 3) (setq arg 17) ...)
<17> jsnell: :). i am just learning lisp, so I really do not have a preference for any implementation, i just want Slime to work :(
<16> jsnell: while I was away I also thought a little bit about PCL and threadsafety
<16> I think I can make generic function calls threadsafe given a lock in each gf
<16> (threadsafety in the presence of cl*** redefinition is really quite hard)
<5> that solution for the slot-value issue seems reasonable. I'll have a look at it later this week, unless you get around to it before
<5> zithowa: as I said, upgrading to 19c should fix the slime problem
<16> I'm unlikely to have time for non-trival hacking before the weekend
<16> I might make SBCL Unicode 5.0 compliant, if that's as simple as copying over the new CharacterData.txt file :-)
<5> ah, judging by the nick you're back at work. how was the vacation?
<17> jsnell: thanks
<16> lovely. (Until the flight home...)
<16> well this is all very nice
<16> climacs does marvellous things with a file named .lisp
<19> good lord
<16> (given a sufficiently new sbcl)
<16> hello, Xach
<20> Xof: Define "marvellous things"
<16> well, it can load and save it, for starters
<16> but also I am marvelling at all the work that therp has clearly done
<20> How mature is climacs?
<16> it's a couple of years old
<20> Usable for daily editing?
<20> years? damn
<20> I thought it was more like weeks
<20> then again, I only hear about it here
<21> tanuki_: Climacs is from november 2004 or thereabouts. It's mature as in "it doesn't crash and it is generally reliable", but it is not mature as in "it does everything an Emacs should do."
<21> It has some neat features, but it's almost useless for anything else than editing plain files or hacking Common Lisp code.
<20> Athas: Eh. I meant it as something closer to the latter
<21> Patches are welcome!
<16> did I say therp? I meant Athas
<16> Athas: while you're here, (make-instance '<hit tab> puts me into the debugger
<22> is polling bad ?
<23> sometimes
<22> Zhivago: for an IM client is it bad to check the file descriptors for new messages ?
<22> its only 1 connection and i only want to check for new ims like every 1 or 5 ms
<16> yes
<22> yes what ?
<16> yes, it's bad
<22> wtf
<16> (why are you asking here?)
<22> what are my other options
<22> its a programming channel thats why
<16> it's a lisp channel. (You want something like select() to call back when data is present)
Return to
#lisp
or
Go to some related
logs:
#css
#web
ex-gfpics
1146 mysql temporary table doesn't exist
#css
#ubuntu
#gcc
#osdev
kUbuntu edit system menu
slackware libdbus
|
|