freenode/#lisp - IRC Chatlog
Search
13:56:16
makomo
beach: i think you have an extra space before the closing paren in the output of your example in si-units.pdf
13:57:19
beach
Normally, I would delegate that kind of work to my coauthor, but none has been named for this book yet.
15:13:02
Xach
When looking for patterns in streams from arbitrary sources (not necessarily well-encoded, or possibly with control codes, etc) I've often thought it would be better to convert the pattern to binary, rather than convert the whole stream all the time according to some encoding. And then only convert the potentially-small data of interest to character data. But I've only tried it on a small scale so I'm not sure
15:13:43
Xach
It's on my mind today because of a renewed interest in something Expect-like in sbcl, where the program output often has a lot of junk in it.
15:18:34
Xach
Younder: in the past i've used protocol context. for example, http headers have line terminators and colons that mean special things.
21:53:45
fiddlerwoaroof
If anyone wants a docker image with sbcl built to work with CFFI:STATIC-IMAGE-OP, here's a Dockerfile: https://fwoar.co/pastebin/9b30249f564662b48b1612759c2328204fe4bb04.nil.html
21:54:58
fiddlerwoaroof
This will essentially let you generate a binary that has no dependencies on any shared objects loaded by CFFI
21:58:59
fiddlerwoaroof
That one's slightly wrong, this one's better: https://fwoar.co/pastebin/7893627c963cdbd4b63de76ef72240dd22bb58e9.nil.html
23:35:44
fiddlerwoaroof
The reason I did all this work is that CFFI generates runtime dependencies on shared objects that make it hard to transfer a dumped lisp image to someone else's system
23:36:08
fiddlerwoaroof
(osicat is the worst offender here, ime, because it generates a so as part of its build in the ASDF cache directories)
23:37:58
fiddlerwoaroof
(you asked me about it a couple months ago, but things have been kinda crazy around here)