freenode/#lisp - IRC Chatlog
Search
23:22:21
aeth
onion: It has been a while, but going through parts of the book, iirc, it looks like the book generally agrees with the consensus of this channel, except when the author promotes his own libraries (can't really fault the author for that).
23:25:35
jeosol_
@bike: I think I may have spotted the error, not sure. I checked the following and it returns nil : (eql (intern (format nil "GET-CELL-P~A" 0)) (intern (format nil "GET-CELL-~A" :PO)))
23:26:18
jeosol_
The form on the right site is now I defined the get-cell-p0 function initially (form on right side of eql), and then I am calling it in a different function as the form on the left side of the eql symbol
23:27:02
jeosol_
There are some reasons I am calling it that way, but I have to change this it seems.
0:50:23
jeosol_
what is the better way to create function name at runtime. For get-cell-p0 function I am doing something like: (intern (format nil "GET-CELL-P~A" 0), I need to create other corresponding functions like get-cell-p1, get-cell-p2, etc
0:54:31
Bike
well, usually you don't do that. if you do do that, something like that is fine, though you should use an explicit package and prefer find-symbol.
0:59:43
jeosol_
These functions are defined by a macro when I start the application. Then depending on the argument to another function, I have to pick one of get-cell-p1, get-cell-p2, etc
1:02:02
jeosol_
@bike, what I mean is that, I need to call get-cell-p0, or get-cell-p1, as follows (funcall (intern (format nil "GET-CELL-P~A" 0)) I J K) -- this if for get-cell-p0. The get-cell-? functions have been defined earlier by a macro and take three
1:16:20
jeosol_
@bike, that sounds imilar to what I do already: (LOOP FOR POINT ACROSS #(0 1 2 3) COLLECT (FUNCALL (INTERN (FORMAT NIL "GET-CELL-P~A" POINT)) I J K))
1:18:59
Bike
you could have (defvar *get-cell-functions* (vector ...)) and then just (funcall (aref *get-cell-functions* point) i j k), for example
1:28:57
Bike
or just have like (defun get-cell-p (point i j k) (case point ...)) if they're similar enough
1:35:14
jeosol_
I adjusted the macro to use the first format. I just packed the functions into the defvar. they have similar syntax. basically, the functions are extracting vertices from a polygon
4:28:40
fouric
onion: defadvice is an emacs thing, right? you install a function to be run before/after another one?
4:29:13
onion
*can defadvice done with plain CLOS :around etc.? i know it is part of lispworks and emacs. and i suppose LW wouldnt have it if it could be done with plain CLOS.
4:29:56
beach
onion: You may have to explain what defadvice does, since it is not standard Common Lisp.
4:30:44
pierpa
beach, it allows to have some CLOS functionality over non-generic functions. Meant for debugging purposes.
4:32:04
pierpa
for example, it allows to patch functions which we don't have the source of, with pre/post methods, and the like
4:34:45
beach
onion: So what is your question? Do you want to implement defadvice in systems that don't have it?
4:35:36
pierpa
here's the LW docs: http://www.lispworks.com/documentation/lw445/LWRM/html/lwref-268.htm#pgfId-1187582
5:31:09
blisp[m]
Voidengineer here. aka cryptonarauder aka evilangel. Ive seen some shit. Dont ask lol
5:32:13
blisp[m]
Been dying to get back on here. Got some great ideas. Starting with a block list processor.
6:19:54
beach
Here is a nice little project for someone who needs practice with CLOS: Design a protocol for input methods, and implement a few common ones. I am particularly interested in latin-1 and the telex input method for Vietnamese. Such a library would be immensely useful. I would use it in Second Climacs, and in various CLIM applications. And jmercouris can use it in the Next browser.
6:21:41
beach
It is not entirely trivial, because it must be possible to obtain effects such as the underlining of "preliminary" characters as Emacs does, and to handle erasing correctly.
6:49:56
shrdlu68
You don't realize what a nice community CL has until you go into other languages' channels.
6:51:11
shrdlu68
I've never had a question go unanswered here, people try to at least acknowledge questions/comments.
6:52:35
onion
beach: hmmmm, i've been thinking of something like that for a while, input methods/touch keyboard. and it came up again recently, i will check out CLIM-TOS =)
7:02:40
beach
onion: Since jackdaniel started doing that, several other people are now helping out as well. I am very pleased with the way things are going.
7:04:30
onion
that is great to hear. CLIM-TOS i notice is an implementation from allegro? i notice theres a few, are they all wrong licenses or too old or not portable ?
7:05:24
jackdaniel
well, actually franz/clim2 was made available, clim-tos is fork to make it work on sbcl and ccl
7:09:11
beach
onion: We started McCLIM way before the commercial CLIM implementations were made available.
7:09:20
jackdaniel
http://hellsgate.pl/files/9f02a348 – application on the left "mimics" pane hierarchy on the right
7:10:44
onion
oh nice! going to update my sources and then poke around TOS version and see what i can learn. i would like to work with japanese and right-to-left hebrew and other text stuff, it seems, and i am not sure that emacs is the place for it
7:12:01
onion
(my first project was gnustep to which i draw a lot of parallels with mcclim . touch-enabled is where i am looking toward)
7:13:39
onion
yeah GNUstep implements OPENSTEP for the most part, where is no Cocoa Touch on iPhones .
7:14:22
onion
beach: ohh, in clim3 the standard or implementation ? would it be silly to ask if 2 and 3 are both useful seperately.. ? hmm
7:14:50
beach
CLIM3 is a specification. CLIMatis is the implementation I worked on in parallel with the specification.
8:28:02
jackdaniel
yet another pic (this one is actually an accident - a bug propagated surrounding-output-with-border "upwards"): https://pbs.twimg.com/media/DYjqsfiX0AEgSKx.jpg
8:35:29
jackdaniel
hm, after a thought - it is not a bug, it is just it "covered" rectangle drawn before it (because there is a pane name it top-left corner and "dot" in bottom-right) - both surrounded by a border
8:47:55
beach
I am guessing that tools like that are going to turn out to be invaluable for debugging pane hierarchies. At the moment I just use Clouseau on the application frame, but that is quite a tedious technique.
8:49:59
jackdaniel
I've already found some bugs thanks to that (because it is easier to "see" the problem than deduce it)
8:50:49
jackdaniel
person with enough time could hack it into something for prototyping clim application layouts too I suppose
9:03:51
blisp[m]
Lol that reminds me of a quote from Pineapple Express. Have you ever seen the back of a dollar bill man well have you ever seen the back of a dollar bill on weed
9:05:02
Shinmera
SBCL, CLISP, and LispWorks also work, but aren't available as apps, as far as I know.
9:07:29
blisp[m]
Im just curious is there are any FreeBSD users in here you know why there's no clisp in ports
9:09:16
blisp[m]
Well honestly no real reason just that I started with it and I was surprised to see that it wasn't there
9:13:29
blisp[m]
Beach, what's the status of your list OS is it working well enough to provide any significant advantage over a non list OS especially for doing tasks that lisp was made for? And lastly what's the chances that it would support something like opencl?
9:14:19
blisp[m]
Are you still wanted to be super purist I would just write my own opencl implementation in lisp but I don't like to reinvent the wheel all the time
9:16:46
blisp[m]
Weird so I thought mezzano was yours. Steven looks like it's under active development
9:17:25
blisp[m]
I remember you telling me many years ago that you were riding a common list operating system I was asking about that
9:59:14
pjb
blisp[m]: clisp runs on FreeBSD. That the FreeBSD people don't care for making a clisp package is their problem. You can just fetch the sources of clisp and compile it on any system you care to use.
10:19:08
Pierpa
Afaiu, LW sells a runtime for Android and then a regular LW can deliver apps for it. It's not a development system which runs on android.
10:45:16
Younder
Ubuntu is boring, but well maintained and reasonably secure. I like Qubes OS, but the lack of NVIDIA support is unacceptable to me.