220 lines
7.5 KiB
JavaScript
220 lines
7.5 KiB
JavaScript
/**
|
|
* Using strings in Ethereum (or any security-basd system) requires
|
|
* additional care. These utilities attempt to mitigate some of the
|
|
* safety issues as well as provide the ability to recover and analyse
|
|
* strings.
|
|
*
|
|
* @_subsection api/utils:Strings and UTF-8 [about-strings]
|
|
*/
|
|
import { getBytes } from "./data.js";
|
|
import { assertArgument, assertNormalize } from "./errors.js";
|
|
function errorFunc(reason, offset, bytes, output, badCodepoint) {
|
|
assertArgument(false, `invalid codepoint at offset ${offset}; ${reason}`, "bytes", bytes);
|
|
}
|
|
function ignoreFunc(reason, offset, bytes, output, badCodepoint) {
|
|
// If there is an invalid prefix (including stray continuation), skip any additional continuation bytes
|
|
if (reason === "BAD_PREFIX" || reason === "UNEXPECTED_CONTINUE") {
|
|
let i = 0;
|
|
for (let o = offset + 1; o < bytes.length; o++) {
|
|
if (bytes[o] >> 6 !== 0x02) {
|
|
break;
|
|
}
|
|
i++;
|
|
}
|
|
return i;
|
|
}
|
|
// This byte runs us past the end of the string, so just jump to the end
|
|
// (but the first byte was read already read and therefore skipped)
|
|
if (reason === "OVERRUN") {
|
|
return bytes.length - offset - 1;
|
|
}
|
|
// Nothing to skip
|
|
return 0;
|
|
}
|
|
function replaceFunc(reason, offset, bytes, output, badCodepoint) {
|
|
// Overlong representations are otherwise "valid" code points; just non-deistingtished
|
|
if (reason === "OVERLONG") {
|
|
assertArgument(typeof (badCodepoint) === "number", "invalid bad code point for replacement", "badCodepoint", badCodepoint);
|
|
output.push(badCodepoint);
|
|
return 0;
|
|
}
|
|
// Put the replacement character into the output
|
|
output.push(0xfffd);
|
|
// Otherwise, process as if ignoring errors
|
|
return ignoreFunc(reason, offset, bytes, output, badCodepoint);
|
|
}
|
|
/**
|
|
* A handful of popular, built-in UTF-8 error handling strategies.
|
|
*
|
|
* **``"error"``** - throws on ANY illegal UTF-8 sequence or
|
|
* non-canonical (overlong) codepoints (this is the default)
|
|
*
|
|
* **``"ignore"``** - silently drops any illegal UTF-8 sequence
|
|
* and accepts non-canonical (overlong) codepoints
|
|
*
|
|
* **``"replace"``** - replace any illegal UTF-8 sequence with the
|
|
* UTF-8 replacement character (i.e. `\ufffd`) and accepts
|
|
* non-canonical (overlong) codepoints
|
|
*
|
|
* @returns: Record<"error" | "ignore" | "replace", Utf8ErrorFunc>
|
|
*/
|
|
export const Utf8ErrorFuncs = Object.freeze({
|
|
error: errorFunc,
|
|
ignore: ignoreFunc,
|
|
replace: replaceFunc
|
|
});
|
|
// http://stackoverflow.com/questions/13356493/decode-utf-8-with-javascript#13691499
|
|
function getUtf8CodePoints(_bytes, onError) {
|
|
if (onError == null) {
|
|
onError = Utf8ErrorFuncs.error;
|
|
}
|
|
const bytes = getBytes(_bytes, "bytes");
|
|
const result = [];
|
|
let i = 0;
|
|
// Invalid bytes are ignored
|
|
while (i < bytes.length) {
|
|
const c = bytes[i++];
|
|
// 0xxx xxxx
|
|
if (c >> 7 === 0) {
|
|
result.push(c);
|
|
continue;
|
|
}
|
|
// Multibyte; how many bytes left for this character?
|
|
let extraLength = null;
|
|
let overlongMask = null;
|
|
// 110x xxxx 10xx xxxx
|
|
if ((c & 0xe0) === 0xc0) {
|
|
extraLength = 1;
|
|
overlongMask = 0x7f;
|
|
// 1110 xxxx 10xx xxxx 10xx xxxx
|
|
}
|
|
else if ((c & 0xf0) === 0xe0) {
|
|
extraLength = 2;
|
|
overlongMask = 0x7ff;
|
|
// 1111 0xxx 10xx xxxx 10xx xxxx 10xx xxxx
|
|
}
|
|
else if ((c & 0xf8) === 0xf0) {
|
|
extraLength = 3;
|
|
overlongMask = 0xffff;
|
|
}
|
|
else {
|
|
if ((c & 0xc0) === 0x80) {
|
|
i += onError("UNEXPECTED_CONTINUE", i - 1, bytes, result);
|
|
}
|
|
else {
|
|
i += onError("BAD_PREFIX", i - 1, bytes, result);
|
|
}
|
|
continue;
|
|
}
|
|
// Do we have enough bytes in our data?
|
|
if (i - 1 + extraLength >= bytes.length) {
|
|
i += onError("OVERRUN", i - 1, bytes, result);
|
|
continue;
|
|
}
|
|
// Remove the length prefix from the char
|
|
let res = c & ((1 << (8 - extraLength - 1)) - 1);
|
|
for (let j = 0; j < extraLength; j++) {
|
|
let nextChar = bytes[i];
|
|
// Invalid continuation byte
|
|
if ((nextChar & 0xc0) != 0x80) {
|
|
i += onError("MISSING_CONTINUE", i, bytes, result);
|
|
res = null;
|
|
break;
|
|
}
|
|
;
|
|
res = (res << 6) | (nextChar & 0x3f);
|
|
i++;
|
|
}
|
|
// See above loop for invalid continuation byte
|
|
if (res === null) {
|
|
continue;
|
|
}
|
|
// Maximum code point
|
|
if (res > 0x10ffff) {
|
|
i += onError("OUT_OF_RANGE", i - 1 - extraLength, bytes, result, res);
|
|
continue;
|
|
}
|
|
// Reserved for UTF-16 surrogate halves
|
|
if (res >= 0xd800 && res <= 0xdfff) {
|
|
i += onError("UTF16_SURROGATE", i - 1 - extraLength, bytes, result, res);
|
|
continue;
|
|
}
|
|
// Check for overlong sequences (more bytes than needed)
|
|
if (res <= overlongMask) {
|
|
i += onError("OVERLONG", i - 1 - extraLength, bytes, result, res);
|
|
continue;
|
|
}
|
|
result.push(res);
|
|
}
|
|
return result;
|
|
}
|
|
// http://stackoverflow.com/questions/18729405/how-to-convert-utf8-string-to-byte-array
|
|
/**
|
|
* Returns the UTF-8 byte representation of %%str%%.
|
|
*
|
|
* If %%form%% is specified, the string is normalized.
|
|
*/
|
|
export function toUtf8Bytes(str, form) {
|
|
if (form != null) {
|
|
assertNormalize(form);
|
|
str = str.normalize(form);
|
|
}
|
|
let result = [];
|
|
for (let i = 0; i < str.length; i++) {
|
|
const c = str.charCodeAt(i);
|
|
if (c < 0x80) {
|
|
result.push(c);
|
|
}
|
|
else if (c < 0x800) {
|
|
result.push((c >> 6) | 0xc0);
|
|
result.push((c & 0x3f) | 0x80);
|
|
}
|
|
else if ((c & 0xfc00) == 0xd800) {
|
|
i++;
|
|
const c2 = str.charCodeAt(i);
|
|
assertArgument(i < str.length && ((c2 & 0xfc00) === 0xdc00), "invalid surrogate pair", "str", str);
|
|
// Surrogate Pair
|
|
const pair = 0x10000 + ((c & 0x03ff) << 10) + (c2 & 0x03ff);
|
|
result.push((pair >> 18) | 0xf0);
|
|
result.push(((pair >> 12) & 0x3f) | 0x80);
|
|
result.push(((pair >> 6) & 0x3f) | 0x80);
|
|
result.push((pair & 0x3f) | 0x80);
|
|
}
|
|
else {
|
|
result.push((c >> 12) | 0xe0);
|
|
result.push(((c >> 6) & 0x3f) | 0x80);
|
|
result.push((c & 0x3f) | 0x80);
|
|
}
|
|
}
|
|
return new Uint8Array(result);
|
|
}
|
|
;
|
|
//export
|
|
function _toUtf8String(codePoints) {
|
|
return codePoints.map((codePoint) => {
|
|
if (codePoint <= 0xffff) {
|
|
return String.fromCharCode(codePoint);
|
|
}
|
|
codePoint -= 0x10000;
|
|
return String.fromCharCode((((codePoint >> 10) & 0x3ff) + 0xd800), ((codePoint & 0x3ff) + 0xdc00));
|
|
}).join("");
|
|
}
|
|
/**
|
|
* Returns the string represented by the UTF-8 data %%bytes%%.
|
|
*
|
|
* When %%onError%% function is specified, it is called on UTF-8
|
|
* errors allowing recovery using the [[Utf8ErrorFunc]] API.
|
|
* (default: [error](Utf8ErrorFuncs))
|
|
*/
|
|
export function toUtf8String(bytes, onError) {
|
|
return _toUtf8String(getUtf8CodePoints(bytes, onError));
|
|
}
|
|
/**
|
|
* Returns the UTF-8 code-points for %%str%%.
|
|
*
|
|
* If %%form%% is specified, the string is normalized.
|
|
*/
|
|
export function toUtf8CodePoints(str, form) {
|
|
return getUtf8CodePoints(toUtf8Bytes(str, form));
|
|
}
|
|
//# sourceMappingURL=utf8.js.map
|