| |
| |
| |
|
Page: 1 2 3 4 5
Comments:
<0> wsmoak: ah, indeed, slime is using different streams, and implements its own REPL...
<1> beslyrus: just got back
<0> wsmoak: I don't know enough slime/swank. You should read the sources of swank and see if there's a switch to log the forms evaluated, otherwise, you could add this feature.
<1> rahul: re: futex overhead with GC - wouldn't you have a problem that on lock allocation you'd create a watched memory outside the lisp data heap to avoid the OS address and lisp address getting out of sync? Then you'd have to have a portable way of freeing the use of that memory address for reuse when the referring lock was collected. Otherwise you'd eventually run out of addresses to watch. Or am I missing something?
<2> do any open source CL implementations have something like environments present in ACL ?
<3> SBCL has the sb-cltl2 (or whatever it's called) macro
<3> err
<3> contrib
<2> hmm
<4> ieslick: hmm, you'd have to not move the lock while it's being accessed
<5> ls
<6> emacs-unicode-2-cvs-2006-11-12tar.bz2 scheme48.texinfo
<7> how come when i type - or + or * into slime, each on a line by itself, I get - back from the REPL
<7> oh wait
<8> clhs -
<9> http://www.lispworks.com/reference/HyperSpec/Body/a__.htm
<7> never mind
<7> thanks
<10> ******************** BUSH WANTS: $245billion more for Iraq. . . . for a country in Civil War. Imagine if the $350 billion already spent there, plus this extra $250 billion - $600+billion dollars, nearly $1Trillion dollars. Imagine if that was spent on the United States and important domestic issues. Bush is the ANTICHRIST
<11> how interesting.
<12> nerplexia: Bush is just a thug, who might be making a good chunk of that $1 T
<11> if only there were a militia with guts and ability to topple the government.
<12> Auris-: And lose themselves a take of the slice of $1T?
<10> bush/israel are gonna use tactical nuclear weapons
<10> to attack irans nuclear facilities
<10> and start WW3
<0> nerplexia: and the relationship with lisp is?
<10> llisper
<10> updated info from cnn.com
<10> WASHINGTON (AP) -- President Bush will send Congress a $2.9 trillion spending request Monday that seeks billions of dollars more to fight the Iraq war and tries to restrain the spiraling cost of the government's big health care programs.
<10> WASHINGTON (AP) -- President Bush will send Congress a $2.9 trillion spending request Monday that seeks billions of dollars more to fight the Iraq war and tries to restrain the spiraling cost of the government's big health care programs. - - - - http://www.cnn.com/2007/POLITICS/02/04/bush.budget.ap/index.html
<8> Auris-, llisper: thanks for feeding the troll. where would we have been without you?
<12> jsnell: I couldn't resist a swipe at my reasoning for all the daily dose of boring tv headlines. But I agree with you here, and its best to not do so.
<13> prxq: Still around?
<12> just did a "$ time clbuild world" now
<14> So, hmm, what's new and exciting in the world of Lisp?
<6> even the bots just left #lisp ...
<0> lisp's doomed, if the bots leave!
<6> hooray for lisp revival!
<15> luis: New in the world of lisp? Found the bug (operator error, oops) causing cliki to not be writable. That's about it for me.
<14> nyef: I was hoping for something related to NyefOS :)
<15> Sorry, not hacking on that right now.
<6> What are some popular ajax libraries that would be nice to have support for under hunchentoot?
<6> I've "done" prototype already
<14> Urfin: what have you done to prototype?
<6> err, I'm doing some simplistic ajax framework for hunchentoot that will however support different "backends"
<6> prototype serves as one of such backends
<14> Urfin: are you using parenscript?
<6> luis: not at the moment, although it looks like it can make javvascript less painful
<12> clbuild is a convenient script on Linux. $ time clbuild world returned in 15 mins. But I did not have darcs, so it may have errored a few commands.
<7> hey I just had a kind of nice idea, what if functions were differentiated from each other by a nonsensical three or four character suffix like sfow
<7> instead of by a prefix
<7> then you could use tab completion to get the function you want, so long as you know its base name
<6> and the point of that nonsensical suffix would be?
<7> serve the same purpose as a package name
<7> heck it wouldnt need to be nonsensical, it could just be the package name
<7> so you just type the name of the function and press tab, and let your editor worry about which package it's in
<15> Umm... WTF?
<15> Nevermind the whole "don't have to use package prefixes when they're for the current package" thing?
<15> And your argument now leading to limiting package names to four characters in length...
<7> just so it's shorter
<7> 2 character suffix actually would be enough in almost all cases
<6> maybe just make a smarter completion which DWYM?
<15> Yes, because there's a really good two-character mnemonic for "sb-debugger-internals". Or "nevermore-microengine-emulator".
<7> not a mnemonic
<7> just a distinguishing mark
<7> might be sb-debugger-internalsxq
<7> in case there were someone else who defined their own sb-debugger-internalswp
<15> ... And now you need a @$@# LOOKUP TABLE for parsing the mark back to the package name??
<7> no it just means that you never have to shadow anything
<7> you do need a lookup table but you already have that with tab-completion
<15> So, if I tab-complete "wp", it'll tell me that it refers to internal debugger code?
<7> no you might tab complete sb-debug
<15> But I don't need to tab-complete sb-debug, that's the package name.
<7> and it would ask you to choose sb-debugger-internalsxq or sb-debugger-internalswp
<7> ok then sb-debugg
<15> And I know the difference between the xq and wp variants... how?
<15> Which is which?
<15> And what do you do about uninterned symbols?
<7> I don't know, you are right that could be a problem
<7> the editor could solve it for you by telling you the package of names automatically, or you could just remember which one you want
<6> rtoym: this solution seems to be still looking for its problem
<15> brainly-green: Next scenario: For some deployments, I don't have editor integration into my Lisp system. -Now- how does this work?
<7> actually there was a problem I was thinking about
<13> Urfin: Heh.
<7> a situation where you would have one huge package with names that potentially conflict internally
<8> whatever the problem, it can't be serious enough to warrant this
<15> Urfin: Looking for its problem? No, it's looking for a -mercy killing-.
<7> ok well whatever, I think it would be useful for what I was thinking about, maybe it would not be necessary otherwise
<15> brainly-green: At which point you split the package?
<8> the solution to that problem is to use different meaningful names
<8> rather than using the same name + a meaningless suffix, which you seem to be suggesting
<7> manually splitting the package or requiring programmers to manually differentiate the names would not be adequate for this particular thing I'm thinking about, but never mind, maybe it is only for this one thing
<7> I am thinking about an online code database where many functions are contributed, maybe each coming from a different place--how do you coordinate the names to make them different
<7> automatically so you don't have to manually go in there and change things
<7> it wouldl be similar to what imageshack uses for naming its pictures
<7> and suffix rather than prefix would work with apropos
<6> pictures names don't have to be remembered by people
<8> I have no idea of how you envision making the code coming from different places conform to this scheme without manually changing it
<7> this would also be combined by the ability to select which functions you want to use from the package--you don't take the whole package
<7> in most cases you would only select 1 funtion of any given base name, independent of the suffix
<15> See, this is where you look at either abusing the package system, or using a custom reader/printer/package system overlay to manage things.
<7> the latter, nyef
<15> So, custom reader and printer?
<7> custom website
<7> anyway I'm not trying to sell this idea here
<16> problem solved! commence the hacking!
<7> right
<17> helo. what should i say to emacs to make it indenting according to Common Lisp standard? :) i remember there was a command, but cannot find it :(
<6> how's prefix not working with apropos?
<18> killerstorm: (slime-setup) should do the trick in ~/.emacs, iirc
<0> killerstorm: It is done automatically once you are in lisp or common-lisp mode (vs. emacs lisp mode).
<15> Didn't it have something to do with lisp-indent-function?
<6> I have (setq lisp-indent-function 'common-lisp-indent-function)
<7> urfin: is there a way to type the end of a function and have apropos give you the prefix?
<0> brainly-green: apropos always searches part of the name.
<15> (Hrm... lisp-indent-function -> lif -> life-instinct-function -> I should watch Serial Experiments Lain again.)
<17> thanks, Urfin, that was that command :)
<6> brainly-green: try (apropos "bind")
<15> brainly-green: Why don't you try it? (apropos "PLY")
<7> hmm well I mean something like tab completion only working backward
<6> use "fuzzy completion"
<6> also allows me to expand bind to destructuring-bind
<7> huh what's the key for that
<7> not C-c C-i
<6> (setq slime-complete-symbol-function 'slime-fuzzy-complete-symbol)
<7> what's the function name for it?
<6> and then just normal key you have for completion
<7> hmm interesting except when I type bind <tab> the first choice is bit-nand
<7> how do I change it back?
<7> what's the default function
<6> use the second choice (or rather the one you want)
<7> right yes it's useful, but I'm used to <tab> being end-completion
<7> maybe I will set that fuzzy completion to some other key but how do I change <tab> back to end-completion
<7> without restarting slime
<6> I think the default is 'slime-simple-complete-symbol
<6> (if not check the docs)
<7> thanks
<6> fuzzy completion has won me over after I tried it on "mvb"
<6> but maybe you are right in that it should be etended to pacckages too, say so that "ai:ins" wouls expand to asdf-install:install
<6> Heh, I'm writing plain text and feel the need to define a macro :) Guess it's time to read up on what emacs has to offer in this respect
<19> It can offer you abbrev-mode
<6> Adrinael: thanks, maybe that's the ticket
<20> night
<21> wow. It's almost as if I hadn't gone offline for a day.
<16> must be comforting to know that #lisp has not p***ed you by.
<0> hefner: in general both too wide and too long a function is sign you're not using enough abstractions.
<16> or too many method and cl*** names that are 30 characters long..
Return to
#lisp
or
Go to some related
logs:
#web
#javascript
#ubuntu
debian xserver-xorg ratpoison
#php
#openzaurus
KungFuDeathGrip
port centronix lpt basic
#fluxbox
h3xis
|
|