freenode/#lisp - IRC Chatlog
Search
16:53:29
phf
hello, does anybody know if you can make capi's browser-pane (lispworks) display html explicitly, without doing (..-navigate url)?
16:54:13
phoe_
phf: best ask on LispWorks's forums - they'll be able to answer CAPI questions much quicker
16:54:48
phoe_
they have a community forum specialized for LW, where Freenode's much more specialized in general standard CL questions and those related to free/open implementations
16:55:56
phf
phoe_: i like to test the limit of #lisp's question answering abilities. fwiw i get roughly the same quality of answers for free/open implementations as i do for commercial ones, that is to say "low"
16:56:52
beach
phf: Most questions about free implementations are answered to the best ability of the participants.
16:58:26
beach
phf: Contrary to common belief, #lisp is not a "Lisp support channel". It is a forum for people who use and develop Common Lisp software, so its main purpose is not to answer questions about implementations.
17:00:08
beach
phf: What you said is quite insulting to many people who spend a lot of time trying to help out whenever someone needs it.
17:03:08
phf
beach: well, i used to spend time on #lisp answering questions, and being a professional lisp developer i would suspect that my answers were worthwhile. i find that the s/n has dropped significantly, but that's ~my~ impressions, and that's why i don't lurk here. i periodically drop by to ask questions though, which is selfish of me, but i also want to see if there's perhaps some change in the landscape
17:04:30
phf
beach: i appreciate all the work ~you~ and others that i can name specifically did for #lisp and lisp community though. i also apologize for baiting. an answer "go elsewhere" is mildly annoying though, because i know for a fact there are people who do professional lispworks dev lurking here
17:06:38
phf
Bike: i agree, i usually wait though until i find the answer to then also answer my own question in case anyone was curious, and/or reading the logs
17:07:21
jackdaniel
paid software comes with paid support, free software has to depend on people donating their free time
17:12:08
phf
Bike: well, i'm not talking about ~quality of support~ here, but rather the s/n. it would be entirely silly of me to come to #lisp and expect correct answer!
17:14:45
phf
well, in that case i will then just add that if the only thing you have in your browser widget is a url browsing function, then adding a local webserver and hosting ad hoc pages is one solution, but i'd like it to be that of last resort..
17:22:33
drmeister
I was looking at the Google paper on energy efficiency of different languages paper again.
17:23:04
drmeister
The guy who wrote the Lisp (using SBCL) examples took a lot of pains to improve performance.
17:23:05
drmeister
https://github.com/greensoftwarelab/Energy-Languages/blob/master/Lisp/mandelbrot/mandelbrot.lisp
17:27:11
drmeister
Kudos to the programmer and... is this what needs to be done to get the 'ultimate' in performance?
17:28:37
drmeister
Lisp is the fastest, most efficient dynamic language - much, much better than Python.
17:32:49
phoe_
drmeister: the "ultimate" in performance is basically either writing as close to compiler-optimized routines as you can or writing as close to assembly as you can
17:38:29
Harag
I have never really used trace but tonight I need to get to grips with it, I started (tried in repl and slime emacs menus) a trace on a function called by an a hunchentoot page but I am not seeing any output of the trace anywhere. What would "block" trace output to repl or slime trace buffer?
17:42:46
phoe_
drmeister: http://benchmarksgame.alioth.debian.org/u64q/program.php?test=mandelbrot&lang=sbcl&id=3 is (almost) completely portable though and it's twice as slow as the version you linked.
17:43:24
phoe_
Harag: can you DECLAIM NOTINLINE the function that you are trying to trace and recompile the callers of that function?
17:43:52
phf
Harag: you want to look into *inferior-lisp* buffer, that's where the trace is most likely being output
17:48:02
phf
Harag: your hunchentoot handlers are running in own threads, where various output streams are ~not~ bound to the swank ones, so you want to look at the buffer where the actual sbcl process output goes
17:56:43
Harag
that trace out put is a real let down two lines and the out put of the function if I want to see whats going on inside do I have to trace every one of the functions used inside it?
17:59:55
Harag
_death: yeah that would be a longer term solution I suppose, thanx will book mark and investigate
19:36:38
phf
found a mac specific way to throw a webkit window up with lispworks and render an arbitrary string http://paste.lisp.org/+7OHR i distilled it from examples/objc/web-kit.lisp. good enough for me
19:49:53
_death
may be cute, but converting to a string in order to convert it to a bit vector is not what I'd consider good code.. note dlowe's smily
19:52:33
didi
Bike: I actually have something like (format t "~{ ... ~}" (loop for i below 7 collect (if (logbitp i datum) 'true 'false)))
20:05:03
Bike
didi: i mean, you could avoid some consing, but probably the i/o will dominate the time
20:47:54
vtomole
How do you return a lists of lists all at once? (return-at-once '((1 apple) (2 orange) (3 pear))) => (1 apple) (2 orange) (3 pear)
20:52:00
vtomole
What if they are in a list? (values '((1 apple) (2 orange) (3 pear))) != (1 apple) (2 orange) (3 pear) Do I need to write a function that recusively returns values?
20:58:37
Bike
i'm a bit worried about the concept confusion behind "remove the outermost parentheses" is all.
20:59:39
Bike
there is a limit on the number of arguments to functions, but there's also a limit on the number of values you can return, and it's probably not any higher.
21:01:07
vtomole
There is a library that takes s-expr in order ( run (1 apple) (2 pear) (3 orange)) But I want to able to do (return-at-once '((1 apple) (2 orange) (3 pear))) to converit it what the run function wants
21:01:56
White_Flame
that looks like a macro. if it was a function it would be (run '(1 apple) '(2 pear) '(3 orange))
21:02:32
Bike
if it is a function, you can use the elements of a list as arguments to a function more easily
21:04:29
White_Flame
then if you want to generate parameters to call it, that generator needs to be a macro as well
22:03:32
iqubic
I know I should't be trying to compare CLOS to Java or C++, but this has been nagging me for a bit now.
22:08:47
_death
iqubic: http://www.gigamonkeys.com/book/object-reorientation-classes.html see footnote 12
22:12:21
iqubic
_death: What happens if a class inherits two, or more, different slot specifiers with the same name (One from each superclass)?
22:14:18
mfiano
structs do a couple things classes cannot, but moreso the other way around. Generally, stay away from structs unless you need arrays with named accessor functions for example
22:16:44
pjb
iqubic: the only standard way is to use (eval `(defclass ,new-name (,@new-superclasses) ,@new-slots))
22:40:29
emaczen
http://paste.lisp.org/display/358268 -- can anyone show me how to get a commonlisp array from *bytes*
22:43:26
emaczen
pillton: I've tried different combinations of forms before evaluating (cffi:foreign-array-to-lisp ...)
22:57:19
pillton
In all honesty, you are doing it the wrong way. Use IPC or implement some higher level C functions that save you from all of the defcstruct madness. You will go insane if you keep following this path.
23:07:05
pillton
I'm saying that your current approach is very brittle because it relies on the layout of C structures which you are not in control of.
23:10:07
emaczen
I was actually trying to use ABCL, java opencv bindings but I could never get it configured
23:56:30
sizur
you guys can do everything easy, so i assumed there would be a simple cli command for that already :)
23:56:59
sizur
but i understand since it's probably easy to invode adhoc command for that, the cli tool was never needed
23:59:30
jmercouris
sizur: elisp is missing a lot of things, you'll have better luck in #emacs for elisp specific questions
1:00:15
jmercouris
I'm trying to run the example here: https://trac.clozure.com/ccl/wiki/CocoaBridge and instead of drawing a red window, I just get the following: https://imgur.com/a/64OxP
1:07:27
jmercouris
It's interesting because it updates the "Window" section of the menubar, but the window is NOT visible, even when using "Bring all to front"
2:16:14
jmercouris
I'm looking at it now, seems pretty good, but just wondering if there was something else, I like reading from multiple resources usually
2:41:25
emaczen
Or sorry, you should pass the code you want called in the function argument to #'gui:execute-in-gui
2:42:34
emaczen
jmercouris: if you are looking at GUI work with commonlisp, ABCL is really easy to use Java's swing libraries
2:43:18
emaczen
I've just tinkered around with ABCL for other reasons but it is as easy as (setf frame (new 'Jframe))
2:48:52
jmercouris
emaczen: Thank you for the advice, but I already use EQL for gui, just toying around with cffi for some other ideas
2:52:02
jmercouris
you can install it and check out the examples section: https://gitlab.com/eql/EQL5
2:52:31
iqubic
I thought it was a function to check if two symbols pointed to the same memory location.
3:42:21
turkja
Does anyone know any good examples of GUI apps written for ABCL? I mean some serious app, not just "hello java".