* Get the encoding of a buffer. * Checks the start, middle, and end of the buffer for characters that are unrecognized within UTF8 encoding. * History has shown that inspection at all three locations is necessary. * @returns Will be `null` if `buffer` was not provided. Otherwise will be either `'
(buffer: Uint8Array, opts?: EncodingOpts)
| 33 | * @returns Will be `null` if `buffer` was not provided. Otherwise will be either `'utf8'` or `'binary'` |
| 34 | */ |
| 35 | function getEncoding(buffer: Uint8Array, opts?: EncodingOpts): "utf8" | "binary" | null { |
| 36 | // Check |
| 37 | if (!buffer) return null; |
| 38 | |
| 39 | // Prepare |
| 40 | const textEncoding = "utf8"; |
| 41 | const binaryEncoding = "binary"; |
| 42 | const chunkLength = opts?.chunkLength ?? 24; |
| 43 | let chunkBegin = opts?.chunkBegin ?? 0; |
| 44 | |
| 45 | // Discover |
| 46 | if (opts?.chunkBegin === undefined) { |
| 47 | // Start |
| 48 | let encoding = getEncoding(buffer, { chunkLength, chunkBegin }); |
| 49 | if (encoding === textEncoding) { |
| 50 | // Middle |
| 51 | chunkBegin = Math.max(0, Math.floor(buffer.length / 2) - chunkLength); |
| 52 | encoding = getEncoding(buffer, { |
| 53 | chunkLength, |
| 54 | chunkBegin, |
| 55 | }); |
| 56 | if (encoding === textEncoding) { |
| 57 | // End |
| 58 | chunkBegin = Math.max(0, buffer.length - chunkLength); |
| 59 | encoding = getEncoding(buffer, { |
| 60 | chunkLength, |
| 61 | chunkBegin, |
| 62 | }); |
| 63 | } |
| 64 | } |
| 65 | |
| 66 | // Return |
| 67 | return encoding; |
| 68 | } else { |
| 69 | // Extract |
| 70 | chunkBegin = getChunkBegin(buffer, chunkBegin); |
| 71 | if (chunkBegin === -1) { |
| 72 | return binaryEncoding; |
| 73 | } |
| 74 | |
| 75 | const chunkEnd = getChunkEnd(buffer, Math.min(buffer.length, chunkBegin + chunkLength)); |
| 76 | |
| 77 | if (chunkEnd > buffer.length) { |
| 78 | return binaryEncoding; |
| 79 | } |
| 80 | |
| 81 | const contentChunkUTF8 = buffer.slice(chunkBegin, chunkEnd); |
| 82 | |
| 83 | // Detect encoding |
| 84 | for (let i = 0; i < contentChunkUTF8.length; ++i) { |
| 85 | const charCode = contentChunkUTF8[i]; |
| 86 | if (charCode === 65533 || charCode <= 8) { |
| 87 | // 8 and below are control characters (e.g. backspace, null, eof, etc.) |
| 88 | // 65533 is the unknown character |
| 89 | // console.log(charCode, contentChunkUTF8[i]) |
| 90 | return binaryEncoding; |
| 91 | } |
| 92 | } |
no test coverage detected