| |
| |
| |
|
Page: 1 2 3 4 5
Comments:
<0> that hashtable may be local to the statemachine, etc
<1> "c faq"? i could use a good laugh..
<2> http://www.plethora.net/~seebs/faqs/c-iaq.html#section-1
<3> attila_lendvai: That sounds maybe like it could work. So a hashtable lookup is less expensive than an FBOUNDP, (setf symbol-function <blah>), FMAKUNBOUND combination?
<4> that will depend on the implementation
<4> in sbcl, probably. in an implementation where the symbol-function is stored in a slot of the symbol, the opposite
<0> that may depend, but i think the code will be much easier to debug if/when the need comes
<5> erider pasted "jsnell can you have a look at this pls" at http://paste.lisp.org/display/23317
<4> no, that's too long. anything in particular that I should look at?
<6> jsnell: I pointed at the core and still no result. Also I have been working and the env and I still can't launch lisp without being in the bin
<7> comprehensibility? he's perfectly comprehensible to me: http://common-lisp.net/pipermail/clo-devel/2006-July/000308.html
<4> how did you point sbcl to the core? I see no --core in that paste
<6> jsnell: I changed the directory to point at the core
<4> what directory?
<6> home/erider/programming/sbcl/lib/sbcl/sbcl
<4> what do you mean by "you changed the directory point to the core"
<6> it was ~/sbcl/bin/sbcl
<0> kire: strange guy. i got bored on his mail, writes like a lawyer or something... :)
<4> oh, you mean inferior-lisp-program. why did you do that? it should point to the sbcl executable
<4> and you should do as I said earlier, and either ensure that the SBCL_HOME environment variable is set correctly also in emacs, or use the --core command line argument to sbcl
<6> jsnell: I was trying things out
<8> anyone know any resources on dynamic compilation and jits ?
<8> from a programming perspective
<8> doesnt have to be lisp specific
<4> erider: ok. I can see that giving you any more instructions is going to be a waste of time. have fun
<6> The SBCL_HOME is set to /home/erider/programming/sbcl/lib/sbcl
<6> jsnell: should it be /home/erider/programming/sbcl/bin/sbcl
<4> ok, I'll try one final time then. set inferior-lisp-program to point to the sbcl executable, and use the --core argument to tell sbcl where the core is
<9> katana__: it's not that different from compilation in general. What aspect are you intersted in?
<6> jsnell: I don't know how to do that
<8> beach eh the general aspects, i have a stream of "bytecode" thats currently interpreted
<8> where do i begin if i want to have dynamic compilation ?
<8> what kind of analysis do i need to do on the bytecode ?
<8> how do i translate blocks of bytecode to machine code ?
<6> jsnell: how do you set the --core argument?
<9> katana__: depends on the level of the bytecodes. Presumably, the bytecodes define some kind of virtual machine. You have to map that virtual machine to the real one.
<8> i ***ume you mean registers and instructions
<8> yes ok lets say i do this
<9> katana__: the instructions to generate could the same ones that the interpreter executes, except that you don't need to analyse the bytecodes at runtime.
<8> beach so this effectively saves me the cost of analysing bytecode at runtime which doesnt sound that dramatic
<8> where do the big speedups come from ?
<8> i ***ume you have some kind of analysis on a block of bytecode
<8> where you reorganize the code and possible combine operations together
<9> it could be dramatic, depending on how different the byte code virtual machine is from the real one.
<6> jsnell: ok I think I figured it out
<8> beach do you know any examples of lisp programs that do this ?
<9> katana__: on a modern machine, testing and branching can be expensive, and that's what happens when the bytecodes are analyzed.
<9> katana__: JIT compilation? No, I don't. Most Lisp systems either interpret the bytecodes or compile the Lisp code directly to machine code without going through the byte code.
<8> i meant something like an emulator for an architecture thats written in lisp
<8> and uses jit techniques
<9> yeah, I understood what you mean. And I don't know of any. But that doesn't mean the don't exist.
<9> *they
<10> Is there a way to have tab completion within an emacs lisp buffer like using tab in the slime repl?
<9> C-c TAB
<9> or C-c C-i if you don't want to change the position of your fingers too much
<10> M-/ seems to work pretty good thanx
<9> marcelino: they are very different
<10> i am not sure of the differences i want to be able to type long variable names the short way :)
<9> marcelino: try typing `(with-ou' and then hitting M-/ (nothing will happen) then try C-c C-i and see
<9> marcelino: or, if you use fuzzy completion, try typing `(m-v-b' and C-c C-i and see it expand to `(multiple-value-bind'.
<10> i'll have a try thanks for the tip
<9> marcelino: the problem with M-/ (which is useful in other cases) is that it doesn't know anything about Lisp
<11> hi
<12> Hi
<11> i would like to do applications in lisp, in windows and with a beutifull GUI. WHere i could read about that? (sorry for my english)
<8> depends what you mean by beautiful ui ;p
<9> "windows" as the MS operating system, or just GUI components?
<8> i wouldnt use lisp for end-user gui applications
<9> katana__: oh?
<11> windows as the OS yes, and beautifull, bonitas. :)
<9> jeremiah: I hear Lispworks is good for that.
<11> thx beach
<8> beach, meaning i cant afford to buy lispworks
<2> jeremiah: cl-opengl xD
<11> do you know an example of an application with a gui in Lispworks?
<11> cl-opengl?, mmm
<11> it sounds to render
<2> jeremiah: parse error
<9> jeremiah: I don't know of any, because I use SBCL on GNU/Linux, but others may.
<11> ok beach, you dont use guis?
<13> how can I set breakpoints and step with slime and sbcl/clisp
<9> jeremiah: yes, I do. I use McCLIM on SBCL
<11> oh ok
<11> thx for the info, i will investigate now
<8> you could try ltk aswell
<8> its "beautiful" on osx
<9> katana__: so you don't think http://common-lisp.net/project/gsharp/rapsoden-sjunger.png is appropriate?
<8> beach i do ;p
<8> somehow though, i dont think i'd like to ship end-user applications using mcclim
<9> why is that?
<12> use swing *gasp*
<8> because i get this feeling that they wouldnt sell very well
<10> does anyone know of lisp code dealing with process syncronization or dealing with deadlocks?
<2> how odd.. I have yet to see a happy response on here when I mention cl-opengl
<9> katana__: interesting you should mention that, because I was just contacted by a company that would like to help me develop Gsharp so that they can suggest it to their customers.
<8> congrats
<9> thanks
<11> allegro is good?
<9> so I hear, yes.
<14> katana: Do you think they'd care if you use McCLIM or not?
<12> jeremiah ; Technically? Or do you ask about the license?
<11> technically, i could buy the license, or is better the sbcl or a gnu version?
<9> jeremiah: most people here probably use SBCL.
<8> adeht depends on the application and the type of user
<8> dont have anything against mcclim
<15> heh. http://boinkor.net/misc/i-love-adwords.png
<16> jeremiah: http://lemonodor.com/archives/000583.html
<2> jeremiah: if you need to ask, the answer is probably "grab sbcl"
<12> jeremiah ; It depends on what you intend to do. But if you want to ship your software to customers, be aware that Allegro requires you to pay for each copy so shipped.
<11> oh, i see, i will use the sbcl then
<11> all the channel cant be in an error, hx friends
<8> sbcl on windows is too bleeding edge afaik
<11> ag, dosent work in windows?
<17> You know, there's only one real way to make SBCL/Win32 -not- be too bleeding edge.
<2> throw it away?
<2> (windows)
<17> No, USE IT!
<2> use windows?!?!?
<11> uf, sorry my english don let me to understand too much complex constructions
<8> yeah **** that ;p
<17> The more people who use SBCL/Win32, the more bugs will be found, the more bugs that are found, the more bugs can be fixed.
<11> <17> You know, there's only one real way to make SBCL/Win32 -not- be too bleeding edge.
<10> just curious to know if there exists lisp code dealing with cl***ic problems of synchronization and deadlocks (e.g. bounder-buffer problem, Monitor implementation using semaphores etc)
<11> nyef, what do you mean?
<2> nyef: maybe after I get a vmware license
<17> jeremiah: Almost nobody uses SBCL/Win32. If it had more users, it would get more developer attention.
<18> xarq: I don't think anybody here cares about your attitude towards Windows
<17> (Especially since we'd try and convert as many of the users into developers as possible.)
<11> i see, thx nyef
<17> I mean, I don't even use SBCL/Win32, and I did the initial port.
<2> cmm-: that's odd, I don't recall stating my attitude towards using windows
<11> oh, nyef, you are a genious then, nice to meet ypu
<11> i am only a novice
<18> xarq: I must be reading too much into excess punctuation, then :)
<2> cmm-: guess so
<11> so in windows i use lispworks and in linux sbcl
<2> jeremiah: try clisp on windows
<11> thx xarq, i will try
<12> wxcl for the gui part.
<11> ok, if i learn fats i will write my experience in real world applications in windows with beautifull guis in a blog
<11> in spanish of course
<2> fats?
<11> fast, sorry
<17> And here I was wondering what the filesystem had to do with anything...
<11> you must learn spanish to read my blog then
<16> i would love to see someone do a beautiful gui app in lisp. show me something like https://aerith.dev.java.net/ done in semi-portable lisp, and i'll be impressed.
<8> this looks like 100% opengl
<2> katana__: you're getting the idea now...
<11> i will try to do it, i think lisp will give me a difference over other windows programmers
<1> is there any reason to do one of these over the other: (member 'nil '(1 nil 2)) versus (memeber nil '(1 nil 2))? or are they exactly the same?
<1> minus the typo of "memeber" ofcouse
Return to
#lisp
or
Go to some related
logs:
#javascript
img2iso fc5 rpm
+gentoo install script +gui
#lisp
#math
#debian
how to install kplayer in kubuntu
div inside body
RPC: failed to contact portmap (errno -5) knoppix
linux grum md raid howto
|
|
