freenode/#lisp - IRC Chatlog
Search
15:27:38
jmercouris
I get a 0: (CFFI::FL-ERROR "Unable to load any of the alternatives:~% ~S" ("libmysqlclient.20.dylib" "libmysqlclient.dylib"))
15:28:03
Bike
i believe it does whatever the implementation does, and the implementation probably defers to the operating system
15:28:33
Bike
https://common-lisp.net/project/cffi/manual/html_node/_002aforeign_002dlibrary_002ddirectories_002a.html
15:33:39
rpg
I have been having trouble with slime-selector recently -- none of the characters I type as selections matches against slime-selector-methods. E.g.: "No method for character: ?\R" -- any idea why this would be happening? I'm not sure what is going on, but it seems like my selector characters are being UPCASED, then matched against slime-selector-methods. Any idea what could be doing this upcasing?
15:40:48
Ukari
i want to export some methods named 'map' and 'do' which would accept special classtype argument in my package, but there has already been 'map' and 'do' function in package :cl
15:45:38
scymtym
for CL:MAP, you could also look into user-extensible sequences, but not many implementations support them
15:52:12
lieven
Ukari: basically, when you type "map" in your code, the reader converts it to "MAP" and that's the name of the symbol defined in the CL package. by writing |map| you instruct the reader to make a symbol with as name "map"
15:56:08
rpg
Looks like read-key-sequence should give me the lower-case character, but somehow it seems not.
15:57:14
Ukari
thanks you, lieven, thought i might change the function name instead of use |map| as a method name which would be export due to it seems a bit strange to be wrap with ||
15:59:41
rpg
*In the context of slime-selector* read-key-sequence is upcasing. Anyone have a clue about what could be doing this? some rogue dynamically-scoped variable?
16:03:24
rpg
Should this be worrying me: Warning: Bug in minibuffer-inactive-mode: it forgets to call `run-mode-hooks'
16:07:00
rpg
<_death> that's the usual advice, but I have 2042 lines of emacs-lisp in init.el and 240 in slime-config.el. I have been using emacs for almost 40 years. I know you are trying to be helpful, but that is almost like saying "start by scraping your machine, and then put all the binaries back one by one."
16:07:47
rpg
For anything involving interactive debugging, the time cost of this strategy is simply staggering.
16:18:27
ggole
And describe-buffer-case-table, which might be interesting to run in *slime-selector*'s minibuffer thingy
16:22:48
rpg
ggole: this looks ... odd: "⇧A uppercase, matches a " and "A lowercase, matches A" -- any idea what the "matches" means?
16:25:42
ggole
Hmm. Well, in scratch it says a lowercase, matches A and A uppercase, matches a, so it looks like entries in the upcase and downcase parts of the table
16:26:21
rpg
ggole: I think that's just a stupid way of printing. If you look at the first of the two strings, they show capital A as "A with shift."
16:27:44
rpg
I just tried shift R to see if somehow I was getting case-inversion behavior, but not so -- it looks like somehow read-key-sequence in slime's minibuffer is upcasing. But not in the buffer itself. I suspect some state leakage.
16:29:41
ggole
I would probably try temporarily ripping out some slime contribs to see if that changes things
16:32:08
rpg
ggole: This started when I started working with Python -- I'm wondering if some of the python emacs config gunk is leaking state.
16:34:51
rpg
That's why I was wondering if the warning about minibuffer-inactive-mode not running its hooks was a problem.
16:38:05
rpg
ggole: (abc ABC) as expected, but the only way I know how to do this is by jamming elisp into a call to message in a redefined slime-selector, which means I'm not sure that this reflects what happens in the read-key-sequence context.
16:39:24
jcowan
An array restricted to a subtype of number doesn't even necessarily have an implementation default initial value: it's perfectly cromulent for it to contain arbitrary junk
17:33:10
rpg
ACTION finds that debugging init.el by bisection is even worse than you expect if you have (package-initialize)
17:57:25
rpg
(key-description "r") returns "R". I have no idea why I am (seemingly) the only SLIME user who has this problem.
18:18:49
asarch
AI learns to play snake using Genetic Algorithm and Deep learning: https://www.youtube.com/watch?v=3bhP7zulFfY
18:22:25
asarch
Intelligence, emotion, character, personality. We are far from create the perfect automata, right? :-(
19:43:30
gendl
I'm having issues with "File not found" both in CCL and Allegro, where I've made a Windows symbolic link to a directory with mklink /d
19:44:51
gendl
the directories and files under the symlink show up fine in cygwin bash shell or Windows cmd prompt, but after a couple levels deep, they do not show up with (probe-file ...) or (directory ...) in CCL or Allegro.
19:45:17
TMA
gendl: does dir list <JUNCTION> or <SYMBOLIC LINK> (not sure of the exact spelling used)
19:46:39
TMA
gendl: I have never had problems with the latter + sbcl. I have used it even for the fasl cache without any problem
19:47:20
gendl
another possibly complicating factor is that the target of the symlink is under a drive ("m:") which is mounted from a Linux system using sshfs-win
19:48:02
gendl
but it doesn't seem that the sshfs mounting is a problem in and of itself, because the files do show if using the actual "m:/..." pathname.
19:48:33
gendl
I guess i can try making a similar symbolic link setup with a directory from the local C: drive and see if I get similar problems...
19:51:01
gendl
so the difference is - the machine which is not working is a Server version of Windows (server 2016), and the drive is mounted with sshfs instead of Virtualbox shared drive.
19:52:18
gendl
i suppose it could also be some weird permissions issues -- but with the same user in cygwin bash shell, everything shows up. So it's kind of bizarre. And one of the last showstoppers to the super cool build/CI system I'm trying to get running here...
19:53:09
gendl
everything in place, everything ready to go, then this weird kink shows up when trying to load actual files in Lisp...
20:55:04
gendl
Update: it's not just a Lisp problem. Windows file Explorer and cmd shell also have trouble seeing files a level or two down from the <SYMLINKD> directory. Only cygwin can see everything.
20:57:25
gendl
So this is some kind of issue with the sshfs, the windows <SYMLINKD>, or the combination of the two. Using the M: drive directly (the one mounted with sshfs) works fine. Anyway not a Lisp issue so I'll dig into it on the windows end of things and stop yammering on about it here... sorry for the noise..
21:45:15
pjb
Clearly, Microsoft doesn't know what belongs to kernels and what belongs to userspace. GUI inside the kernel, FS in user space, PFFFTHTH!
21:50:15
Aruseus
perhaps they wanted to show they could be innovative and implement the fs on top of the gui
21:52:15
pjb
Aruseus: probably they were innovative in a meeting such as https://www.youtube.com/watch?v=BKorP55Aqvg
21:58:10
Aruseus
pjb, are you can expert? I need you to implement copy-on-write using context menus and buttons.
21:59:26
pjb
Notice, there's an good expertly answer: http://img1.joyreactor.cc/pics/post/full/песочница-Не-мое-кто-в-теме-тот-поймет-эксперт-2734907.jpeg
22:03:03
pjb
https://www.youtube.com/watch?v=BKorP55Aqvg&list=PLKlTnphWMK_pbfHE4VilQg3GjVuXJdo5I&index=1
23:40:33
pjb
(loop … do (multiple-value-bind (foo bar) (multiple-value-list whatever) …)) might be more efficient indeed.
1:50:48
jcowan
Cygwin treats .lnk files as if they were really symlinks, but most Windows programs don't understand them at all.
1:53:39
jasom
I seem to recall someone saying that NTFS support symlinks, but it not being exposed at the OS layer.