about summary refs log tree commit diff stats
diff options
context:
space:
mode:
-rw-r--r--README.md45
1 files changed, 40 insertions, 5 deletions
diff --git a/README.md b/README.md index 1d5e86b..85a5994 100644 --- a/README.md +++ b/README.md
@@ -1,13 +1,48 @@
1# Chicanery: subtle, opinionated improvements to R7RS in CHICKEN 1# Chicanery: subtle, opinionated improvements to R7RS Scheme (in CHICKEN)
2 2
3`chicanery` is my attempt at a prelude module for CHICKEN scheme. It imports and re-exports all `r7rs` modules, using `utf8` where possible, and it corrects a few (in my mind) rough spots in the language, to whit: 3While I was reading the [R7RS
4specification](https://standards.scheme.org/official/r7rs.pdf), I was a little
5annoyed by how broken up the standard library seemed to be. While the most
6egregious example is maybe `(scheme case-lambda)`, which exports ... only
7`case-lambda`, the other libraries like `(scheme write)` for `display`, `(scheme
8cxr)` for functions like `caddr` but not `cddr`, and the rest didn't really seem
9logical to me. Plus, in CHICKEN scheme, the one I usually use, `utf8` is an egg
10you need to install separately...
4 11
5- `map`, `for-each`, and `append` are now generalized over lists, vectors, strings, and bytevectors (where appropriate). The list versions of these procedures have been named `list-map`, `list-for-each`, and `list-append` respectively. 12To ameliorate these minor warts in what I otherwise consider to be an excellent
6- `ref` and `copy` have been defined as generalized functions over the above data types. 13language, `chicanery` was born. It's kind of like a prelude, I suppose? It
14imports all `r7rs` modules and re-exports their identifiers, and makes sure the
15implementation is Unicode-aware. It also includes a few extras ;)
16
17## Chicanery extras
18
19I also thought it was strange that `map`, `for-each`, and `append` apply to
20lists only, while `vector-map` and `string-for-each`, exist, for example.
21`chicanery` prefixes the default `map`, `for-each`, and `append` functions with
22`list-`, and redefines those identifiers to functions that are generic over
23Scheme's base collection types (lists, strings, vectors, bytevectors where
24applicable). I didn't make the functions fully-generic to keep them efficient
25and because I don't know how to do that, but you can still use `vector-append`
26or `list-map` if you know what type you need and want speed.
27
28In a similar vein, I've also added generic functions `ref` and `copy` that
29dispatch to `string-ref`, `list-copy`, and the like.
30
31Other extras include
32
33- `(atom? x)` determines whether `x` is an atom (i.e., not a pair or null)
34- `(read-port)`, `(read-port port)` reads a port until hitting end-of-file (IDK
35 why this isn't in R7RS!), in chunks of `read-port-chunk-size`
36- `(defined? x)` returns whether the symbol `x` is bound to a variable
37- `(with-input-from-string str thunk)` calls `thunk` with `str` bound as the current-input-port.
38- `(with-output-to-string thunk)` calls `thunk` and returns a string of the output.
39- `(displayed x)`, `(->string x)` returns `x` as a string (via `display`)
40- `(written x)` returns `x` as a string (via `write`)
41- `(print x ...)` displays `x ...` followed by a newline
7 42
8## Todo 43## Todo
9 44
10- Make sure `set!` works in, like, `(set! (ref 5 '(1 2 3 4 5)) 10)` 45- Support multiple scheme implementations. I tried doing this (see the [multiple-impls](https://git.acdw.net/chicanery/?h=multiple-impls) branch), but it kept not working in weird ways, plus it was like whack-a-mole with all the different schemes and just exhausting.
11 46
12## License 47## License
13 48