Finch/std/json.fch
Nexoniarz ea6d2a57e5 a standard library in Finch: path, time, proc, json
import json   import path   import time   import proc  -- found in std/ after the program's own files,
so a program's own json.fch always wins.

path: join, base, dir, ext, stem, parts, clean, is_abs.
time: now, millis, sleep, date, local, format, days_in_month (clock_gettime, strftime through C import).
proc: run, output (an error with the command's output when it fails), env, pid, cwd.
json: parse (errors with line and column), stringify, pretty, new_*, get/at/set/push/has/names and
str_or/int_or/float_or/bool_or; \u escapes with surrogate pairs.

The C importer now leaves a plain (non-const) char * parameter a ptr[char], a buffer C writes into;
only const char * is text going in. str(ptr) and str([]u8) are in the self-hosted compiler.

boot.sh: 40 passed, 0 failed.
2026-10-25 20:58:26 +01:00

518 lines
13 KiB
Text

// JSON: read it, build it, write it.
//
// import json
//
// doc := try json.parse("{\"name\": \"Finch\", \"tags\": [\"fast\", \"small\"], \"stars\": 5}")
// print(doc.get("name").str_or("")) Finch
// print(doc.get("tags").at(1).str_or("")) small
// print(doc.get("stars").int_or(0)) 5
// print(doc.get("missing").is_null()) true (a missing key is null, so chains never fail)
//
// out := json.new_object()
// out.set("ok", json.new_bool(true))
// list := json.new_list()
// list.push(json.new_number(1.5))
// out.set("values", list)
// print(json.stringify(out)) {"ok":true,"values":[1.5]}
// print(json.pretty(out)) the same, over several lines
enum Kind { Null, Bool, Number, Text, List, Object }
// One JSON value. `kind` says which of the other fields it uses: flag for Bool, num for Number,
// text for Text, items for List, and items with keys (the same length, in the same order) for Object.
struct Json {
Kind kind
bool flag
float num
str text
[]Json items
[]str keys
}
fn new_null() -> Json {
return Json(kind: Kind.Null)
}
fn new_bool(bool b) -> Json {
return Json(kind: Kind.Bool, flag: b)
}
fn new_number(float n) -> Json {
return Json(kind: Kind.Number, num: n)
}
fn new_text(str s) -> Json {
return Json(kind: Kind.Text, text: s)
}
fn new_list() -> Json {
return Json(kind: Kind.List)
}
fn new_object() -> Json {
return Json(kind: Kind.Object)
}
// ---------- looking inside ----------
fn Json.is_null() -> bool {
return self.kind == Kind.Null
}
// the number of elements of a list, of keys of an object, of bytes of a text; 0 for the rest
fn Json.len() -> int {
return match self.kind {
List, Object -> self.items.len
Text -> self.text.len
else -> 0
}
}
fn Json.has(str key) -> bool {
return self.kind == Kind.Object && self.keys.contains(key)
}
// the value of a key of an object; null if there is none (or this is not an object)
fn Json.get(str key) -> Json {
if self.kind != Kind.Object {
return new_null()
}
i := self.keys.find(key)
if i < 0 {
return new_null()
}
return self.items[i]
}
// the element of a list; null if there is none
fn Json.at(int i) -> Json {
if self.kind != Kind.List || i < 0 || i >= self.items.len {
return new_null()
}
return self.items[i]
}
// the value if it is of that kind, else what was given
fn Json.str_or(str fallback) -> str {
if self.kind == Kind.Text {
return self.text
}
return fallback
}
fn Json.float_or(float fallback) -> float {
if self.kind == Kind.Number {
return self.num
}
return fallback
}
fn Json.int_or(int fallback) -> int {
if self.kind == Kind.Number {
return int(self.num)
}
return fallback
}
fn Json.bool_or(bool fallback) -> bool {
if self.kind == Kind.Bool {
return self.flag
}
return fallback
}
// the keys of an object, in the order they were written
fn Json.names() -> []str {
return self.keys
}
// ---------- building ----------
// add to the end of a list
fn Json.push(Json v) {
self.items.push(v)
}
// give a key a value in an object; a key that is already there is replaced
fn Json.set(str key, Json v) {
i := self.keys.find(key)
if i >= 0 {
self.items[i] = v
return
}
self.keys.push(key)
self.items.push(v)
}
// ---------- writing ----------
fn quote(str s) -> str {
[]str parts = ["\""]
for c in s {
n := int(c)
if c == '"' {
parts.push("\\\"")
} else if c == '\\' {
parts.push("\\\\")
} else if c == '\n' {
parts.push("\\n")
} else if c == '\r' {
parts.push("\\r")
} else if c == '\t' {
parts.push("\\t")
} else if n < 32 {
parts.push("\\u00" + hexdigit(n / 16) + hexdigit(n % 16))
} else {
parts.push(str(c))
}
}
parts.push("\"")
return parts.join("")
}
fn hexdigit(int n) -> str {
digits := "0123456789abcdef"
return str(digits[n])
}
fn number_text(float n) -> str {
if n == float(int(n)) && n > -1000000000000000.0 && n < 1000000000000000.0 {
return str(int(n))
}
return str(n)
}
struct Writer {
[]str out
}
fn write(Json j, str indent, str pad, ptr[Writer] w) {
nl := ""
inner := ""
sep := ","
colon := ":"
if indent.len > 0 {
nl = "\n"
inner = pad + indent
colon = ": "
}
match j.kind {
Null -> w.out.push("null")
Bool -> {
if j.flag {
w.out.push("true")
} else {
w.out.push("false")
}
}
Number -> w.out.push(number_text(j.num))
Text -> w.out.push(quote(j.text))
List -> {
if j.items.len == 0 {
w.out.push("[]")
return
}
w.out.push("[" + nl)
for i, item in j.items {
if i > 0 {
w.out.push(sep + nl)
}
w.out.push(inner)
write(item, indent, inner, w)
}
w.out.push(nl + pad + "]")
}
Object -> {
if j.items.len == 0 {
w.out.push("{}")
return
}
w.out.push("{" + nl)
for i, item in j.items {
if i > 0 {
w.out.push(sep + nl)
}
w.out.push(inner + quote(j.keys[i]) + colon)
write(item, indent, inner, w)
}
w.out.push(nl + pad + "}")
}
}
}
// the value as text on one line
fn stringify(Json j) -> str {
w := Writer()
write(j, "", "", addr(w))
return w.out.join("")
}
// the value as text over several lines, indented by two spaces
fn pretty(Json j) -> str {
w := Writer()
write(j, " ", "", addr(w))
return w.out.join("")
}
// ---------- reading ----------
struct Reader {
str src
int i
}
// the line and column of a position, for a message
fn where(ptr[Reader] r) -> str {
line := 1
col := 1
for k in 0..r.i {
if k < r.src.len && r.src[k] == '\n' {
line += 1
col = 1
} else {
col += 1
}
}
return "line " + str(line) + ", column " + str(col)
}
fn peek(ptr[Reader] r) -> char {
if r.i < r.src.len {
return r.src[r.i]
}
return char(0)
}
fn skip_space(ptr[Reader] r) {
while r.i < r.src.len && (r.src[r.i] == ' ' || r.src[r.i] == '\n' || r.src[r.i] == '\t' || r.src[r.i] == '\r') {
r.i += 1
}
}
fn unexpected(ptr[Reader] r, str wanted) -> str {
if r.i >= r.src.len {
return "json: the text ends where " + wanted + " should be (" + where(r) + ")"
}
return "json: found '" + str(r.src[r.i]) + "' where " + wanted + " should be (" + where(r) + ")"
}
// does `word` come next?
fn comes(ptr[Reader] r, str word) -> bool {
return r.src.len - r.i >= word.len && r.src.sub(r.i, r.i + word.len) == word
}
fn expect_word(ptr[Reader] r, str word) -> ! {
if !comes(r, word) {
return error(unexpected(r, "'" + word + "'"))
}
r.i += word.len
}
fn hex_value(char c) -> int {
if c >= '0' && c <= '9' {
return int(c) - int('0')
}
if c >= 'a' && c <= 'f' {
return int(c) - int('a') + 10
}
if c >= 'A' && c <= 'F' {
return int(c) - int('A') + 10
}
return -1
}
// a code point as UTF-8
fn utf8(int cp) -> str {
[]u8 b
if cp < 128 {
b.push(u8(cp))
} else if cp < 2048 {
b.push(u8(192 + cp / 64))
b.push(u8(128 + cp % 64))
} else if cp < 65536 {
b.push(u8(224 + cp / 4096))
b.push(u8(128 + (cp / 64) % 64))
b.push(u8(128 + cp % 64))
} else {
b.push(u8(240 + cp / 262144))
b.push(u8(128 + (cp / 4096) % 64))
b.push(u8(128 + (cp / 64) % 64))
b.push(u8(128 + cp % 64))
}
return str(b)
}
fn read_hex4(ptr[Reader] r) -> int! {
n := 0
for k in 0..4 {
if r.i >= r.src.len || hex_value(r.src[r.i]) < 0 {
return error(unexpected(r, "a hexadecimal digit"))
}
n = n * 16 + hex_value(r.src[r.i])
r.i += 1
}
return n
}
fn read_string(ptr[Reader] r) -> str! {
r.i += 1 // the opening quote
[]str parts
start := r.i
while true {
if r.i >= r.src.len {
return error("json: a text is never closed (" + where(r) + ")")
}
c := r.src[r.i]
if c == '"' {
parts.push(r.src.sub(start, r.i))
r.i += 1
return parts.join("")
}
if int(c) < 32 {
return error("json: a line break or control character inside a text (" + where(r) + ")")
}
if c != '\\' {
r.i += 1
continue
}
parts.push(r.src.sub(start, r.i))
r.i += 1
e := peek(r)
r.i += 1
if e == 'n' {
parts.push("\n")
} else if e == 't' {
parts.push("\t")
} else if e == 'r' {
parts.push("\r")
} else if e == 'b' {
parts.push(utf8(8))
} else if e == 'f' {
parts.push(utf8(12))
} else if e == '"' || e == '\\' || e == '/' {
parts.push(str(e))
} else if e == 'u' {
cp := try read_hex4(r)
// a pair of 😀 stands for one character above U+FFFF
if cp >= 55296 && cp < 56320 && comes(r, "\\u") {
r.i += 2
low := try read_hex4(r)
cp = 65536 + (cp - 55296) * 1024 + (low - 56320)
}
parts.push(utf8(cp))
} else {
return error("json: unknown escape \\" + str(e) + " in a text (" + where(r) + ")")
}
start = r.i
}
}
fn read_number(ptr[Reader] r) -> Json! {
start := r.i
if peek(r) == '-' {
r.i += 1
}
while r.i < r.src.len && ((r.src[r.i] >= '0' && r.src[r.i] <= '9') || r.src[r.i] == '.' || r.src[r.i] == 'e' || r.src[r.i] == 'E' || r.src[r.i] == '+' || r.src[r.i] == '-') {
r.i += 1
}
n := float(r.src.sub(start, r.i)) or {
r.i = start
return error(unexpected(r, "a number"))
}
return new_number(n)
}
fn read_value(ptr[Reader] r) -> Json! {
skip_space(r)
c := peek(r)
if r.i >= r.src.len {
return error(unexpected(r, "a value"))
}
if c == '{' {
obj := new_object()
r.i += 1
skip_space(r)
if peek(r) == '}' {
r.i += 1
return obj
}
while true {
skip_space(r)
if peek(r) != '"' {
return error(unexpected(r, "a key in quotes"))
}
key := try read_string(r)
skip_space(r)
if peek(r) != ':' {
return error(unexpected(r, "':'"))
}
r.i += 1
v := try read_value(r)
obj.set(key, v)
skip_space(r)
if peek(r) == ',' {
r.i += 1
continue
}
if peek(r) == '}' {
r.i += 1
return obj
}
return error(unexpected(r, "',' or '}'"))
}
}
if c == '[' {
list := new_list()
r.i += 1
skip_space(r)
if peek(r) == ']' {
r.i += 1
return list
}
while true {
v := try read_value(r)
list.push(v)
skip_space(r)
if peek(r) == ',' {
r.i += 1
continue
}
if peek(r) == ']' {
r.i += 1
return list
}
return error(unexpected(r, "',' or ']'"))
}
}
if c == '"' {
s := try read_string(r)
return new_text(s)
}
if c == 't' {
try expect_word(r, "true")
return new_bool(true)
}
if c == 'f' {
try expect_word(r, "false")
return new_bool(false)
}
if c == 'n' {
try expect_word(r, "null")
return new_null()
}
if c == '-' || (c >= '0' && c <= '9') {
return try read_number(r)
}
return error(unexpected(r, "a value"))
}
// the JSON in `text`; an error with the line and column where it went wrong
fn parse(str text) -> Json! {
r := Reader(src: text)
v := try read_value(addr(r))
skip_space(addr(r))
if r.i < r.src.len {
return error(unexpected(addr(r), "the end of the text"))
}
return v
}