libera/commonlisp - IRC Chatlog

18:34:00 Lord_of_Life_ ** NICK Lord_of_Life

18:38:53 White_Flame and of course, you can destructuring-bind with &key as well

18:40:35 hexology ah, that's also helpful White_Flame

18:41:32 White_Flame and of coruse defmacro itself does handle (defmacro do-fancily (var (&key fanciness ...)) ...)

18:41:44 White_Flame since it destructures deeper than ordinary lambda lists

19:23:49 hexology White_Flame: i also did not know that about defmacro, that i think _really_ solves my problem in this case

19:47:49 _death to avoid nasty surprises, you should by default take care to evaluate the forms (passed as keyword arguments) in the order specified by the user and only once, if they are to be evaluated at all..

19:48:06 jcowan pjb: Alists take twice as much space but are normally treated as persistent data structures, whereas plists are normally mutated.

19:48:19 jcowan or no, try again

19:49:01 jcowan space is the same

19:52:22 dbotton So I have hit my first hard crash with sbcl and the following error : Heap exhausted during garbage collection: 16 bytes available, 48 requested.

19:52:41 dbotton it is not in my regular code. Any tips ?

19:52:45 dbotton :)

19:53:59 aeth start with more memory?

19:54:33 dbotton do you know off had how? (I am looking up now)

19:55:20 aeth I think it's `--dynamic-space-size`

19:55:22 aeth best guess from the manpage

19:56:46 aeth you can see what it defaults to with (sb-ext:dynamic-space-size)

19:57:03 dbotton that seems to have worked :) thanks

19:57:38 _death it's also possible that you're leaking memory (by holding references to stale data) or that you're consing like mad when you probably need not

19:58:22 aeth strangely, you specify it in megabytes at startup and see what is reserved in bytes

19:59:13 aeth it's apparently only 2^30 bytes (i.e. 1 GiB) by default on x86-64 so it should be easy to run out if you're working with a lot of data

19:59:41 aeth You couldn't even write a web browser with that

20:07:57 pjb jcowan: a-lists and p-lists take exactly the same space, and have exactly the same access times.

20:08:33 pjb (setf (cdr (assoc key a-list)) value) is a common idiom.

20:09:36 pjb Nowadays, there may be a slight difference due to cache effect, depending on where the conses are stored. But not depending on the structure.

20:10:28 jcowan Yes, I withdrew the remark about space. But while you obviously can mutate an alist or use a plist persistently, you generally don't.

20:32:58 dbotton _death the issue is from mgl-pax when make documentation

20:33:56 dbotton The manual keeps growing in size so, I assume from that

20:39:28 _death I see.. so aeth's suggestion is a good workaround

23:00:47 gin Is there the C equivalent of "foo\n" in Common Lisp?

23:01:31 aeth "foo~%" in FORMAT

23:02:02 aeth #.(format nil "foo~%")

23:10:11 gin aeth: ah! so I have to call (format nil ...) to do it. there is no way to specify a newline escape sequenc in a literal string then?

23:17:43 kakuhen #\Newline maybe? Not sure if I've ever wanted to do this over simply formatting a string

23:29:46 Bike gin: you just write an actual newline in the literal

23:30:06 Bike gin: if that doesn't suit you, the cl-interpol library provides alternate syntax that lets you use perl escape squences like \n

23:30:58 gin thx kakuhen, Bike !

23:56:48 gin common lisp does not have the alist-get (present in emacs lisp)?

23:58:09 masinter assoc?

23:59:59 gin looks like i have to do (cdr (assoc key alist)) to get the (alist-get key alist) behavior.

0:05:53 Bike gin: the alexandria library provides an assoc-value that does this.

1:29:05 dbotton anyone know a good way to find a random open port to listen on that is x-platform?

1:30:53 hayley usocket can pick one when creating a server on port 0, then you can query the port it chose.

1:35:08 dbotton i'll give a try

1:46:50 jcowan https://llthw.common-lisp.dev/ <-- Learn Lisp the Hard Way

1:48:31 dbotton hayley thanks seems to work for me. thanks

3:44:36 mathrick does Trivia have lambda-list matchers (ie. a drop-in replacement for DESTRUCTURING-BIND, where I can use stuff like &OPTIONAL and &KEY)?

3:45:14 mathrick I'm looking at its wiki, and so far I don't see an obvious way to do that

3:45:46 Bike not a trivia expert, but i expect that would need some real complicated semantics to handle lambda list variables that are subpatterns

3:46:33 mathrick I also don't see a way to replicate (destructuring-bind (head &rest tail) somelist)

3:47:05 Bike aren't there cons patterns?

3:47:59 mathrick I think so?

3:48:14 mathrick Bike: possibly, but I'm looking at Trivia to replace my spaghetti piles of D-B / M-V-B / LET, so it's kinda a letdown if it doesn't have a way to do what D-B already does...

3:49:07 mathrick I might go with metabang-bind, since it seems to have a syntax that's much more like the CL I already know, even though Trivia is the well-maintained community standard

3:53:50 mathrick oh, I think it has the LAMBDA-LIST pattern, which is that