libera/#commonlisp - IRC Chatlog
Search
15:49:59
phoe
ldb: (ql:quickload :spell) or (ql:quickload :spell/simple), the latter occupies less RAM since both load the whole dictionary into memory but the latter does not have any meanings associated with words
15:51:04
phoe
beach: (spell:english-lookup "will") ;=> (#<SPELL::VERB {1010A76823}> #<SPELL::VERB {1010A76893}> #<SPELL::NOUN {1010A76903}>)
15:54:50
phoe
good, then the job is already sort of done - with most of it by you in the initial commit
16:09:01
mfiano
Maybe your mind thought that it was something phoe wrote and your PR'd, so you deleted it once it was merged? :D
17:16:41
beach
Hey, this is great! Loading the FASL takes only a second or so. So I can already incorporate the spell checker into Second Climacs.
17:17:18
beach
I didn't mean "incorporate". I meant, I can use the library to check the spelling of comments and plain text.
17:27:07
phoe
yes, compilation takes the longest because of creating and filling and possibly rehashing the hash table
17:27:25
phoe
I wonder if it could be made quicker by creating the hash table with an appropriate :hash-table-size, hmm
17:29:20
phoe
the paragraph API is probably not the best since it accepts a string; maybe you can optimize it by using the word lookup function and then doing the standard stuff: optimize for speed, inilining the accessor, blah blah
17:33:33
beach
Anyway, time to go fix dinner for my (admittedly small) family. I'll be back tomorrow.
19:26:11
jcowan
What is the effect of a top-level assignment to an unbound variable, such as (setf foo 32)? Is it effectively become dynamically bound?
21:24:36
phoe
AFAIK SBCL creates an "undefined" variable, which is global/lexical/shadowable and does not exhibit dynamic behavior
21:25:43
phoe
as in, on all implementations I can quickly check, (progn (setf foo 42) (defun bar () foo) (defun baz () (let ((foo 24)) (baz))) (bar)) ;=> 42
21:47:05
phoe
in practice everyone seems to fall back to what old lisps did, with also a warning from SBCL
0:14:53
nij-
Hi! Is it possible to handle a condition by popping out the choice panel to the user for N seconds, and close the panel to do something else if no choice is provided?
0:22:30
Bike
nij-: that would have to be a feature of the debugger. i do not know of any debuggers that do that, but you could write one.
0:34:26
nij-
Perhaps this is an XY problem. I want to let a lot of threads running in the background. But I find it annoying to have a debugger popping out everytime a condition is met.
0:43:11
Bike
having a debugger disappear while while i'm deciding what to do would probably just piss me off
2:25:32
hayley
stylewarning: For no good reason whatsoever, here is Coq compiled to CL <https://plaster.tymoon.eu/view/3637>
2:49:58
hayley
By doing really crappy compilation of the JSON output of the Coq Extraction module. I suspect Coalton would generate much better code (mostly around munging Lisp-1/2 and FFI'ing curried functions to Lisp uncurried functions).
2:50:40
hayley
As for why, because one course this semester involves "using static analysis tools" for cloud stuffs, for which they probably mean a linter.
2:51:48
hayley
Then I had the terrible thought last night to feed verified code to CL. Not that my compiler is sound, or even close to working for anything other than ADD.