libera/#commonlisp - IRC Chatlog
Search
14:09:41
Fade
if there's a space it doesn't exist in any context that I can interact with it using the standard emacs machinery.
14:11:39
jmercouris
it is a library used by circe and others to show like a prompt and history, like in the REPL
14:12:18
Fade
I'm not getting an error, I'm just not getting the command, so I'm not surehow to debug it.
14:19:31
Colleen
us: Unknown command. Possible matches: 8, time, set, say, mop, roll, get, search, login, grant,
14:23:11
Fade
hum. it only seems to be happening with the one system I'm trying to work on. my other systems behave normally. sly behaves normally prior to any nonefault systems being loaded.
15:37:24
Fade
well, for posterity, the answer is loading the system :clog will do some violence to your Sly REPL
16:29:22
Josh_2
Hi, I have these classes https://plaster.tymoon.eu/view/2486#2486 but when I try to make an instance of one-hundred I get the condition "IPN-PAYMENT-SUCCESS is a forward referenced class .. IPN-PAYMENT-SUCCESS is a direct superclass" how do I fix this?
17:34:41
Inline
there is a xlib:display-force-output *display* when replaced with xlib:display-finish-output it works as expected
17:42:28
Inline
ok, the only difference between the two seems to be this: both force any buffered output to be sent to the X server, but finish-output is the only one waiting till all requests are received and processed by the X server
17:45:32
Nilby
When I'm working with CLX interactively I often make it put an implicit display-finish-output after every repl comamnd.
18:20:07
Inline
the windows themselves are supposed to be empty, but the windows themselves are the bouncing ones
18:27:46
Inline
the bounce and shove bounce demos are not about things jumping inside windows, they are about the windows itself jumping around in some manner
18:28:21
Inline
mine head down to the bottom of my screen as if i have left a ball down on the floor
18:31:43
Inline
if i knew who the maintainer of CLX was i'd just post the examples (the corrections as a diff file) maybe
18:34:22
dieggsy
is there an easy way of calling some code inline at macro expansion time without another defmacro ?
18:37:23
Nilby
You can just call any function that's allready defined when the macro is being defined.
18:39:28
Nilby
I'm not sure if that's what you mean? but something like: (defmacro foo (x) `(print ,(sqrt x)))
18:41:06
dieggsy
Nilby: i think im basically asking for the macro-equivalent of a lambda. like an unnamed macro that i can call/expand inline
18:43:11
Josh_2
With ironclad a 'sha512 hmac' requires making a hmac-digest with my private key, then using hmac-digest with the octets I wish to digest?
18:53:24
flip214
Josh_2: I'm using (ironclad::hkdf-extract 'ironclad:sha256 secret string), what are you doing?
18:57:52
Nilby
dieggsy: If you have an example of what you want to do, it's likely someone here can come up with a nice way.
19:01:14
dieggsy
Nilby: it would literally be like an inline (mexpand (format nil "~a~b" "foo" "bar")) becomes "foobar" in the compiled code. As opposed to having to (defmacro thing () (format nil "~a~b" "foo" "bar")) and calling (thing). I can of course just write that mexpand macro myself, was just wondering if there's some clever built-in i don't know about
19:11:59
dieggsy
i'm not against writing a macro, just would be nice if we had a built-in inline form that does the same. of course, when you can write that yourself, no point baking in every little thing into the language lol
19:56:17
dieggsy
I'm doing something kind of weird and injecting a "git status" output into code as a string at macro expansion time
19:56:51
dieggsy
is that just chance, or is it somehow guaranteed that files are somehow compiled within the context of where they are in the file ssytem
20:41:50
dieggsy
it's quite weird that pathname-directory doesn't return a pathname - am I using the wrong function?
20:43:54
Josh_2
finally figured it out, turns out in one place you use a private key, the other you have to use a different key
22:15:37
pjb
waleee: I don't think there's a hard requirements for most /etc/ files. There's a standard sysexit code for such an occurence: EX_OSFILE
22:16:44
pjb
waleee: another consideration is that there's a configuration to specify what resource, and in what order, is used to perform various lookups, including DNS lookups.
22:18:17
pjb
waleee: version 2 of GNU glibc uses /etc/nsswitch.conf https://man7.org/linux/man-pages/man5/nsswitch.conf.5.html https://www.oreilly.com/library/view/linux-network-administrators/1565924002/ch06.html
22:19:52
waleee
pjb: it fails with "Failed to find the WRITE-DATE of /etc/hosts: No such file or directory"
22:25:25
pjb
gethostent() if you need to scan all the entries (for some value of "all", depending on the configuration, and what's available locally, I guess).
22:25:46
pjb
waleee: no, it's a posix function. It should be used by iolib/sockets instead of reading /etc/hosts.
22:34:26
waleee
wtf my email was registered on a ubuntu one-account (needed for launchpad bug-reports)
23:55:38
jcowan
That changes the definitions of the symbols t and nil, which is not the same as mutating either one (or any othr symbol)
23:57:38
pjb
Well, it is possible that an implementation uses the string passed to intern or make-symbol as string name, and if you mutate this string you could thus change the symbol name.
0:01:04
jcowan
Back when the pname was on the alist of a symbol, you could mutate it, but that wouldn't affect the inaccessible hash table that maps pnames to their symbols.
0:01:30
jasom
jcowan: what about (setf (get nil ...)...)? If the plist is part of the symbol, you're mutating the symbol, no?
0:04:02
jcowan
The story is about mutating symbol names. in Lisp 1.5, pnames were on the plist, but they were not Lisp objects (no strings yet), so you could only mutate a pname by assembly-language operations. And even if you mutated the plist of NIL to be T, it would still *work like* NIL in the sense of evaluating to itself and being the empty list. So typing in NIL would print out T, because that is the new pname of NIL.
0:05:38
pjb
jcowan: the story is foremost that modern implementations implement protections. The standard doesn't mandate them. It only forbids conforming programs to perform any retrospection on the CL stuff.