IRC channel logs

***chris is now known as Guest2590

***Guest2590 is now known as chrislck

<lilyp>sneek: later tell dthompson yeah the joke kinda is that the inner lambda has to roll the main loop – which fwiw would work if your game/engine had a global main-loop:iter

<sneek>Will do.

<lilyp>sneek later tell dthompson I don't think either Tsukundere or Chickadee have something like that though

<sneek>Okay.

<andrzejku>hello

<andrzejku>I am trying to understand macros

<andrzejku>do you have maybe something better than official Guile Reference?

<andrzejku>generally I get and idea

<unmatched-paren>andrzejku: I found this helpful to understand why we use define-syntax instead of define-macro: https://en.wikipedia.org/wiki/Hygienic_macro#The_hygiene_problem

<andrzejku> http://paste.debian.net/1248708/

<andrzejku>there are usually something with (if condition) in examples

<andrzejku>but I want to put a several expressions

<andrzejku>in sequence

<unmatched-paren>one moment

<andrzejku>unmatched-paren: ahh thsi hygiene I read it is about to not capture accidentally outer variables

<andrzejku>I acutally have problem with defining

<unmatched-paren> https://paste.sr.ht/~unmatched-paren/19d8a59b83479a2e8e75247fd845cbf9c81a4178

<andrzejku>also should I always use ...

<andrzejku>as I understood (begin exp ...) will execute all expressions given as the parameters

<unmatched-paren>if you want to be able to do (foo [ARG1] [ARG2] [ARG3]...) then use ..., if you want to do (foo (lambda () [ARG1] [ARG2] [ARG3])) then just accept a lambda argument

<unmatched-paren>although if you're going to accept a lambda you should probably just use a function

<unmatched-paren>(define (smth name body) (display name) (newline) (body))

<unmatched-paren>works just as well

<andrzejku>unmatched-paren: shit thanks for your time

<andrzejku>unmatched-paren: hmm I still got an error

<andrzejku> http://paste.debian.net/1248711/

<andrzejku>I tried to use your function

<unmatched-paren>source expression failed to match any pattern in form (syntax-rules () ((_ name body ...) (display name) (newline) body ...))

<andrzejku>maybe I have to import some module

<unmatched-paren>that's not my function :)

<andrzejku>unmatched-paren: that one https://paste.sr.ht/~unmatched-paren/19d8a59b83479a2e8e75247fd845cbf9c81a4178

<unmatched-paren>Ah, you meant my macro

<andrzejku>yes

<andrzejku>macro

<unmatched-paren>Hmm, that's strange. I'm pretty sure that's the right syntax

<andrzejku>I think too

<andrzejku>I am running the file using guile filename.scm

<unmatched-paren>btw, there's a shorthand for define-syntax/syntax-rules with one pattern

<andrzejku>I found that Guile is not good in errors and maybe some modules must be imported

<unmatched-paren>(define-syntax-rule (smth name body ...) (display name) (newline) body ...)

<unmatched-paren>no, there's no module imports needed here

<andrzejku>unmatched-paren: ya this I saw

<andrzejku>I wanted to improve it later

<unmatched-paren>but "guile is not good in errors" is pretty accurate :)

<andrzejku>unmatched-paren: ok the standard macro when

<andrzejku>works

<unmatched-paren>wdym "standard macro"?

<andrzejku> https://www.gnu.org/software/guile/manual/guile.html#Defining-Macros

<unmatched-paren>ah

<andrzejku>unmatched-paren: it looks for me that we can not put (define-syntax-rule (smth name body ...) (display name))

<andrzejku>the rest of arguments sequentially

<andrzejku>just display is accepted

<unmatched-paren>strange

<andrzejku>oh maybe it must me a lambda?

<unmatched-paren>i guess you could do (begin body ...)

<unmatched-paren>instead of body ...

<unmatched-paren>try that

<andrzejku>I tried already

<unmatched-paren>hmm

<andrzejku>not working

<andrzejku>(define-syntax-rule (smth name body ...) (display name) (begin body ...))

<unmatched-paren>Ah, i think I know the problem

<unmatched-paren>you can only do (syntax-rules ((...) (...)))

<unmatched-paren>you can only have two forms inside the body of the rule

<unmatched-paren>so (syntax-rules ((...) (...) (...) (...))) as we're doing won't worj

<unmatched-paren>work

<unmatched-paren>you need to wrap the body itself in a begin form

<unmatched-paren>and then you don't need the begin around the body ... either

<andrzejku>unmatched-paren: what's begin actually?

<unmatched-paren>begin just executes each of its forms, one after the other

<unmatched-paren>it's like { ... } in C-like language

<unmatched-paren>languages

<andrzejku>ya I pack the second rule argument in (lambda () () ()) but it needs additional () to be executed

<andrzejku>aa ok

<andrzejku>get it

<unmatched-paren>s/executes/evaluates/, mustn't forget the correct terminology :)

<unmatched-paren>if i did the lispers would come after me O_o

<andrzejku>unmatched-paren: thank you it works

<andrzejku>just one more question

<unmatched-paren>okay :)

<andrzejku>body and ... it is the same like the list of expr expr expr

<andrzejku>and expr is the function

<andrzejku>which can be executed

<unmatched-paren>i don't quite follow you

<unmatched-paren>(define-syntax-rules (foo expr ...) (begin expr ...))

<unmatched-paren>the exprs don't have to be functions

<andrzejku>I mean we write (begin expr ...)

<unmatched-paren>in fact they usually aren't

<andrzejku>but I can actually pass (+ 3 2)

<unmatched-paren>the exprs are actually just lists

<unmatched-paren>(list '+ 3 2)

<andrzejku>I read that it will be evaluated as 5

<unmatched-paren>and so is the "template"

<andrzejku>and then we have (begin 5 ...)

<unmatched-paren>(list 'begin 'expr '...')

<andrzejku>begin 5)

<unmatched-paren>the two are spliced together using guile's syntax-rules algorithm

<andrzejku>ohh

<unmatched-paren>and then treated as literal syntax

<unmatched-paren>guile syntax is just lists

<unmatched-paren>this is illustrated by the traditional macro style, define-macro:

<unmatched-paren>(define-macro (foo bar) ...)

<unmatched-paren>now `bar` is bound to a syntax object

<unmatched-paren>which is either a list or an atom like 3, "hello", or 'foo

<unmatched-paren>and a macro is supposed to return a syntax object itself

<unmatched-paren>(define-macro (foo bar) `(display ,foo))

<unmatched-paren>(define-macro (foo bar) `(display ,bar))

<unmatched-paren>sorry, mistake :)

<andrzejku>unmatched-paren: ahh that's enough hard to understand

<unmatched-paren>so define-macro uses quote/unquote to construct actual syntax objects

<unmatched-paren>but it suffers from the hygiene problem

<andrzejku>unmatched-paren: is it explained in any book?

<andrzejku>I started with structure and interpretation

<andrzejku>but just wanted to write a macro

<andrzejku>hehe

<unmatched-paren>andrzejku: https://spritely.institute/static/papers/scheme-primer.html#scheme-extensibility

<andrzejku>unmatched-paren: thank you

<unmatched-paren>there's probably a few books, but it's a simple enough concept

<unmatched-paren>you don't need a book to understand it

<andrzejku>ok

<andrzejku>unmatched-paren: by the way I got a little schemer and The Scheme Programming Language

<andrzejku>looks more than enough :P

<dsmith-work>Happy Friday, Guilers!!

<dthompson>happy friday

<sneek>Welcome back dthompson, you have 2 messages!

<sneek>dthompson, lilyp says: yeah the joke kinda is that the inner lambda has to roll the main loop – which fwiw would work if your game/engine had a global main-loop:iter

<sneek>dthompson, lilyp says: I don't think either Tsukundere or Chickadee have something like that though

<dthompson>lilyp: yeah chickadee just uses coroutines instead. in the context of a coroutine a call-with-cursor procedure would work as expected.

*spk121 just can't autotools any more

***daviid` is now known as daviid

IRC channel logs

2022-07-29.log