How about the "AutoDecode" filename?

[nenge123]

	function decodeText(value, encoding) {
		encoding = encoding && encoding.trim().toLowerCase();
		if (encoding == "cp437") {
			return decodeCP437(value);
		} else {
			return autoDecode(value, encoding);
		}
	}
	function charByteLen(text){
		return new TextEncoder().encode(text).byteLength;
	}
	function autoDecode(value, encoding) {
        let u8txt = new TextDecoder('utf8').decode(value),
            u8len = charByteLen(u8txt);
        if (value.length != u8len) {
            let cptxt = decodeText(value, 'cp437'),
                cplen = charByteLen(cptxt);
            if (encoding&&encoding != 'cp437') {
                let mytxt = new TextDecoder(encoding).decode(value),
                    mylen = charByteLen(mytxt);
                if (mylen < u8len && mylen < cplen) return mytxt;
            }
            if (cplen < u8len) return cptxt;
            return 'unknow.' + u8txt.split('.').pop();
        }
        return u8txt;
    }

[nenge123]

i think is very ok~

[gildas-lormeau]

Well, when I run decodeText(new Uint8Array([36, 0]), "utf-16") with your code, it returns the string decoded in UTF-8, which is not correct.

See https://jsfiddle.net/d5sL4tqb/

[nenge123]

@gildas-lormeau

that nothing error!
{
filenameEncoding:'utf-16'
}
k=new Uint8Array([36, 0]);
kk= e=>new TextEncoder().encode(e).length;

u8 =new TextDecoder('utf-8').decode(k) //'$\x00'
u16 =new TextDecoder('utf-16').decode(k) //$

kk(u8) //length 2
kk(u16) // length1

1<2 return u16;

I think the cp437判断上存在逻辑错误
或者干脆去掉？“&& mylen < cplen”

[nenge123]

我的判断流程是，即便不是UTF8先检测，
再检测用户自定义编码
(或许可以 filenameEncoding:['gbk','utf-16']) 多编码检测
最后检测 cp437

[nenge123]

auto decode isn't decode any code.

it's make sure can Perfect decoding my setting encoding.

[gildas-lormeau]

Feel free to modify the implementation here https://jsfiddle.net/d5sL4tqb/ in order to fix the issues. BTW, decodeText is used in zip.js to decode 2 different types of data: filenames and comments. The returned value 'unknow.' + u8txt.split('.').pop() makes no sense for comments, unfortunately. Finally, the purpose of rawFilename and rawComment in zip.js is to allow you to do whatever you want in your implementation, for example, auto-detect encoding.

[nenge123]

the site can't visit.the cdn file out of china great wall

[gildas-lormeau]

Here is the code

const CP437 = "\0☺☻♥♦♣♠•◘○◙♂♀♪♫☼►◄↕‼¶§▬↨↑↓→←∟↔▲▼ !\"#$%&'()*+,-./0123456789:;<=>?@ABCDEFGHIJKLMNOPQRSTUVWXYZ[\\]^_`abcdefghijklmnopqrstuvwxyz{|}~⌂ÇüéâäàåçêëèïîìÄÅÉæÆôöòûùÿÖÜ¢£¥₧ƒáíóúñѪº¿⌐¬½¼¡«»░▒▓│┤╡╢╖╕╣║╗╝╜╛┐└┴┬├─┼╞╟╚╔╩╦╠═╬╧╨╤╥╙╘╒╓╫╪┘┌█▄▌▐▀αßΓπΣσµτΦΘΩδ∞φε∩≡±≥≤⌠⌡÷≈°∙·√ⁿ²■ ".split("");
const VALID_CP437 = CP437.length == 256;

function decodeCP437(stringValue) {
  if (VALID_CP437) {
    let result = "";
    for (let indexCharacter = 0; indexCharacter < stringValue.length; indexCharacter++) {
      result += CP437[stringValue[indexCharacter]];
    }
    return result;
  } else {
    return new TextDecoder().decode(stringValue);
  }
}

function decodeText(value, encoding) {
  encoding = encoding && encoding.trim().toLowerCase();
  if (encoding == "cp437") {
    return decodeCP437(value);
  } else {
    return autoDecode(value, encoding);
  }
}

function charByteLen(text) {
  return new TextEncoder().encode(text).byteLength;
}

function autoDecode(value, encoding) {
  let u8txt = new TextDecoder("utf8").decode(value),
    u8len = charByteLen(u8txt);
  if (value.length != u8len) {
    let cptxt = decodeText(value, "cp437"),
      cplen = charByteLen(cptxt);
    if (encoding) {
      let mytxt = new TextDecoder(encoding).decode(value),
        mylen = charByteLen(mytxt);
      if (mylen < u8len && mylen < cplen) return mytxt;
    }
    if (cplen < u8len) return cptxt;
    return "unknow." + u8txt.split(".").pop();
  }
  return u8txt;
}

const value = new Uint8Array([36, 0]);
const encoding = "utf-16";
console.log("expected: " + new TextDecoder(encoding).decode(value));
console.log("result: " + decodeText(value, encoding));