Search
Friday, 28th of August 2020, 16:49:11 UTC
16:49:27
beach
I was referring to (setf (symbol-function...)
16:49:54
beach
SYMBOL-FUNCTION is basically obsolete.
16:49:55
phoe
also, (setf (symbol-function 'foo) nil) doesn't work
16:50:06
phoe
neither does SETF FDEFNIITION
16:50:35
shka_
there is also function inside slime for that
16:51:09
thetabit
that's good to know! lol I have been using on lisp for a lot of my studies, and symbol-function is used throughout the book
16:51:25
phoe
it will work, just, not everywhere
16:51:36
phoe
SETF functions are its sore spot
16:51:40
beach
It is best to always use FDEFINITION.
16:52:40
beach
Also the name SYMBOL-FUNCTION gives the impression that there is actually a function slot in the symbol.
16:53:04
shka_
slime-undefine-function to be precise
16:54:46
Josh_2
_death: hows your fork of Parmenides going?
16:54:54
phoe
beach: the same way PACKAGE-NAME gives the impression that there is actually a name slot in the package
16:55:16
phoe
the storage of both are implementation details.
16:55:31
beach
ACTION vanishes to have dinner.
16:55:44
phoe
I read PACKAGE-NAME as a mapping from a package object to a string of some sort, not as some sort of "this has a slot"
16:56:10
Bike
symbol-function is bad because it implies the function is a property of the symbol and not e.g. an environment
16:56:20
Bike
but mostly because fdefinition is beter.
16:56:35
shka_
heh, Bike hangs around beach a lot :P
16:56:57
_death
Josh_2: I submitted a pull request to your repo.. we can play programmer ping-pong improving it
16:57:24
_death
Josh_2: so after you improve it a little, send a pull request back
16:58:19
phoe
Bike: what name would be better? FIND-FUNCTION?
16:58:21
_death
Josh_2: the point is to have fun, so no pressure
16:59:04
phoe
I mean, we have FDEFINITION but FIND-CLASS and they have somewhat distinct semantics when unbinding stuff
16:59:11
phoe
I guess it's time to change the standardâ„¢
16:59:51
Josh_2
i'll probs do a little later tonight
17:00:04
shka_
phoe: we can push this into the limbo of "we are not talking about that part of the standard" where it can hang out with symbol-plists and progv
17:00:29
phoe
hey uh wait progv is actually pretty cool
17:00:38
Bike
two hardest things in computer science, etc
17:01:12
phoe
progv is the only real way you can bind arbitrary dynavars at runtime without using the compiler or evaluator
17:01:17
phoe
which *sometimes* is useful
17:01:25
phoe
(with a stress on "sometimes")
17:01:52
White_Flame
ACTION has used symbol-plists as well
17:02:00
phoe
but the possibility is there
17:02:03
_death
Josh_2: cool.. also feel free to discuss things in issues/todo file/pull request comments.. I guess irc too, though that has less value for archival purposes (and we may be busy at times)
18:38:30
mseddon
I wonder, is there a nice way in CL and emacs to do something equivalent to elisp's (declare (indent indent-spec))?
18:39:25
mseddon
https://www.gnu.org/software/emacs/manual/html_node/elisp/Indenting-Macros.html#Indenting-Macros
18:40:03
jackdaniel
you may hint slime about indentation from swank package (I don't remember the interface)
18:40:17
jackdaniel
trivial-indent is defined to make that easy I think
18:40:22
jackdaniel
(library trivial-indent that is)
18:40:28
mseddon
jackdaniel, aha thanks, I shall consult the relevant docs.
3:22:55
beach
Good morning everyone!
4:46:08
Lord_of_Life_
** NICK Lord_of_Life
Saturday, 29th of August 2020, 4:49:11 UTC