freenode/#lisp - IRC Chatlog
Search
12:40:09
hajovonta
unfortunately, I don't know. This is the blog post the data is from: http://blog.quicklisp.org/2015/08/july-2015-download-stats.html
12:40:44
hajovonta
the dates are for the blog posts, not for the exact months, so there is a little shift, it's a quick glance only
12:47:28
hjudt_
from a library i get the warning message "bare references to struct types are deprecated. please use (:pointer (:struct ...)) or (:struct ...) instead. what are "bare references to struct types", what do they look like?
12:48:42
Shinmera
CFFI type specifications that use the struct name symbol directly rather than (:struct foo)
14:42:41
Xach_
Shinmera: http://report.quicklisp.org/2017-12-21/failure-report/qt-libs.html -- getting this today
14:50:19
mfiano
There is a rather large bug with ASDF that came to realization after a bug report I submitted to a project recently. I hope it doesn't take SBCL a few years to update for a fix once one is available.
14:50:52
Shinmera
Xach_: There were changes between commonqt (qt), qtools, and qt-libs. Are all of them up-to-date?
14:52:17
Shinmera
The changes were related to ASDF in the sense that I finally decided to incorporate some stuff into CommonQt to get rid of gross hacks in qt-libs.
14:53:49
rumbler31
er, I remember that qt-libs concerned itself with installing native libs, and also for providing that the native libs get shipped with lisp applications, right?
14:56:40
rumbler31
are the changes you're talking about w.r.t. qt-libs and asdf concerned with using asdf to ship the native libraries with applications?
14:58:05
Shinmera
CommonQt's system definition includes a step that builds libcommonqt.so automatically on linux systems. Doing so requires qmake, gcc, and a bunch of foreign libraries. Qt-libs should prevent that, so there was a series of ugly hacks to change the system on the fly to stop it from doing that, and to also jump in between to reroute other loading mechanisms
14:58:29
Shinmera
Now CommonQt includes the capabilities I needed on its own, plus a new system qt+libs that uses qt-libs to deliver the binaries.
15:01:20
Xach_
Shinmera: Most things build now, thanks. qsci is still complaining: http://report.quicklisp.org/2017-12-21/failure-report/qt-libs.html#qsci
17:05:52
dim
I'm not sure how common it is around here (I think not much), I am now publishing my docs for pgloader at http://pgloader.readthedocs.io/en/latest/ ; using a Python based tooling --- that's the end-user docs, that said
17:06:37
sebboh
jasom: help! The free code your provided which I used for a local utility stopped working after I apt-get upgrade'd!
17:09:24
beach
KZiemian: Just to tell you what we think you should be doing with CLUS? OK, I'll give it some thought.
17:10:37
KZiemian
beach: you know, I work better and faster when someone ask: hey when it will be read?
17:13:55
KZiemian
pbj: read and understand source code of ccldoc enough that we can rewrite it to suits CLUS
17:15:44
KZiemian
other people in CLUS project works on diffrent things, you must ask them for more information
17:19:45
KZiemian
pbj: we also want make it possible to load not one page but bunch of hyperlinked pages
17:48:54
rpg
sebboh: I just checked -- the latest CFFI has this fixed. So if you get CFFI from git instead of from Quicklisp, you'll be fine.
17:57:41
asarch
Can you create a cons cell pointing its both CAR and CDR pointers to itself? If yes, how?
17:59:18
asarch
And if yes, can you mix them (its CAR pointer points to its CDR pointer and viceversa)?
18:00:30
pjb
asarch: (setf *print-circle* t) (let ((k (cons nil nil))) (setf (car k) k (cdr k) k)) #| --> #1=(#1# . #1#) |#
18:04:52
dmiles
like: (let ((c (cons () ()))(loc (make-locative))) (setf (car c) loc) (setf (cdr c) loc))
18:08:49
asarch
I just was re-reading my logs from the channel when I asked about: (cons '(x y z) 'w)
18:44:06
jasom
rpg: I changed it to a defsystem-depends-on and :cffi-grovel-file instead of cffi:grovel-file which seems the "right way"
18:46:01
jasom
rpg: https://github.com/jasom/clinenoise/commit/c3d3f89f0fe876b2549e28aa1c67a5aa5f6d6b69#diff-13f67aea30a79b6c1016c707babd1dec This change is the "right way" correct? Was the old version actually wrong or just frowned upon?
18:46:12
rpg
jasom: The change in https://github.com/jasom/clinenoise/commit/c3d3f89f0fe876b2549e28aa1c67a5aa5f6d6b69 yes, looks like the right way.
18:46:54
rpg
The old version wasn't actually wrong (or at least detected as such), but the new one is better because if cffi-grovel changes then ASDF will properly detect that fact -- it wouldn't before.
18:48:04
rpg
I was just confused because before I was looking at the use of MAKE-INSTANCE on OPERATION problem, which was fixed.
18:50:48
jasom
I think the latest CFFI in QL doesn't have that issue either, so perhaps sebboh needs to update his quicklisp dist
18:51:58
rpg
jasom: No, the EVAL-WHEN would not have caused an ASDF failure. But your change makes things potentially better.
19:23:30
sebboh
jasom, there's an unused variable warning in the C grovel stuff. Not important to me, only mentioning it because I'm already here. /home/hobbes/.cache/common-lisp/sbcl-1.4.2.debian-linux-x64/home/hobbes/quicklisp/local-projects/clinenoise/termios-grovel__grovel.c:13:7: warning: unused variable ‘autotype_tmp’ [-Wunused-variable]
19:30:07
emaczen
If slime and SBCL have trouble finding my sources what should I do? I think I've had this problem before and for some odd reason it keeps reoccuring...
19:31:18
jmercouris
Shinmera: If I run "ls -la ~/" for example, it doesn't show the contents of my home dir
19:31:27
jmercouris
Shinmera: These docs lol? https://gitlab.common-lisp.net/asdf/asdf/blob/master/uiop/README.md
19:35:16
Bike
if you just tried to give it "ls -la ~/" that's not going to work, it's not a shell, yeah?
19:37:08
Bike
i don't have an sbcl installed from source on this system, so i can't tell you the directory tree off the top of my head
19:38:14
jasom
jmercouris: contrast with (uiop:run-program `("ls" "-la" ,(uiop:native-namestring (user-homedir-pathname))) :output :string)