freenode/#lisp - IRC Chatlog
Search
12:32:47
eigenhombre
Hi, is there some magic required to load libraries requiring FFI in SBCL? When I "(ql:quickload :cl-charms)" I get "Error while trying to load definition for system cl-charms frompathname/Users/eigenhombre/quicklisp/dists/quicklisp/software/cl-charms-20181210-git/cl-charms.asd: COMPILE-FILE-ERROR while compiling #<CL-SOURCE-FILE "cffi" "src"
12:32:47
eigenhombre
"cffi-sbcl"> [Condition of type ASDF/FIND-SYSTEM:LOAD-SYSTEM-DEFINITION-ERROR]"
12:33:42
eigenhombre
(Sorry, IRC and Common Lisp newbie). This happens on OS X whether I install SBCL with Homebrew or from source. Other quicklisp library installs work fine.
12:55:30
ck_
I am on osx as well, can load cl-charms without problems through quicklisp. It compiles some c sources with clang.
12:56:02
ck_
is it possible that that part didn't work for you? Also, try (ql:quickload :cl-charms :verbose T) to see more information
16:31:38
pjb
Well, require with a single argument is implementation dependent; in that case, I don't see how getting different results when evaluating it at compilation-time or at run-time makes any difference. If you want conforming results, pass it the path to load!
17:04:46
Harag
I submitted a project to quicklisp, and it passed build but since then there have been some bug fixes, will a fresh pull be done before the ql release or do I need to submit an update request in the issues?
17:28:30
munksgaard[m]
Is anyone here associated with lisp.se? All links I can find seem to download immediately instead of opening in the browser: (fragment :421d7a1490788a2c1c1785dff7c479ddb501a0ad97b7bbf399f7abae69abc0f9)
17:38:14
rmg1917
Is there a way to configure SLIME so that M-x slime opens a REPL buffer in a window on the bottom of the screen, instead of the entire right half of the screen? I'm reading through slime.el and I can't figure out exactly where the new window is created and selected (not super familiar with ELisp or emacs). Is it possible to configure this? Which
17:41:45
Shinmera
No, but you can just change the layout so it's at the bottom. C-x 1 C-x 2 C-x o C-x b repl RET
17:43:25
Shinmera
The way new windows pop up is An Emacs Thing™, and it can get a lot more annoying than that, I'm afraid.
17:45:12
Shinmera
Emacs' new window... algorithm (?) is so terrible I don't know if it's sadistic on purpose or just incompetent.
17:48:33
Shinmera
Can't generate docs, you gotta write them, sorry. Otherwise entire swaths of people would be out of a job already!
17:50:12
Harag
Shinmera: ok so what is the compromise .. what generator gives you the basics that you can flesh out?
17:51:35
Shinmera
I don't know what you mean or want, so I can't tell you. I use Staple to aggregate documentation into a page to publish.
18:14:40
Harag
Shinmera: ...urg.. tried Staple ...my doc strings made sense in the context of the code but without the context (surrounding code) they become nonsensical...
19:07:08
minion
The URL https://gitlab.common-lisp.net/users/sign_in?secret=56862810 will be valid until 19:15 UTC.
20:01:12
munksgaard[m]
I'm trying to use `(ironclad:byte-array-to-hex-string #(44 242 77 186 95))` but ironclad complains that I'm not supplying a `(vector (unsigned-byte 8))`. How do I coerce the array correctly?
20:02:57
munksgaard[m]
Or rather, how do I convert the array from `(simple-vector 32)` to `(vector (unsigned-byte 8))`
20:04:16
Bike
simple-vector = (vector t), i.e. a vector of objects. a (vector (unsigned-byte 8)) is specialized to only deal with bytes, so it's stored more efficiently.
20:20:43
whartung
is there a technique where you have an elaborate macro (or macro system), and within the scope of the macro, there are DEFUNs that are scoped, but outside of the macro, they are not.
20:25:55
Bike
almost, but labels also allows recursive definitions, and for earlier definitions to refer to later ones
20:46:57
whartung
well, for me, I wanted a “kinda sorta DSL”, so I wanted capabilities only within the macro, and I didn’t want to have to code walk the thing to hunt them down. and it’s “nice”, in this case, to have a set of functions that don’t pollute the global name space since they’re not applicable outside of the macro.
20:49:29
_death
well, if you mean local definitions in the expansion, that's ok.. though sometimes it's a good idea to also provide global definitions that error.. I thought you might be talking about definitions to help in computing the expansion
20:52:58
whartung
yea, basically you (should) be able to write normal CL within the sections of the macro, but I also want some “special functions”, that are part of the DSL, to be avaiable to those routines as well.
20:57:34
whartung
honestlly, thinkikng about it, not sure how that works in my case. I guess I expand the macro in to a function, and FLET the defitnitions as part of the expansion.
21:02:13
Bike
don't remember what you're thanking me for but i'm confident whatever i did was perfectly great
23:34:56
eigenhombre
ck_: thanks for the suggestions earlier. I tried installing llvm / clang via Homebrew but that did not seem to address the CFFI problem. Using your 'verbose' suggestion provides the error "... Lock on package SB-ALIEN violated when interning SHARED-OBJECT-FILE while in package CFFI-SYS."