exifreader 4.39.0 → 4.39.1
This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
- package/dist/exif-reader.js +1 -1
- package/dist/exif-reader.js.map +1 -1
- package/package.json +1 -1
- package/src/tag-decoder.js +40 -1
package/package.json
CHANGED
package/src/tag-decoder.js
CHANGED
|
@@ -6,6 +6,23 @@ import TextDecoder from './text-decoder.js';
|
|
|
6
6
|
|
|
7
7
|
const TAG_HEADER_SIZE = 5;
|
|
8
8
|
|
|
9
|
+
// Windows-1252 (CP1252) mapping for the 0x80–0x9F range, which is where it
|
|
10
|
+
// differs from ISO-8859-1. Bytes outside this range map identically in the
|
|
11
|
+
// two encodings. Five positions (0x81, 0x8D, 0x8F, 0x90, 0x9D) are unassigned
|
|
12
|
+
// in CP1252; we keep their raw byte value so the data is still
|
|
13
|
+
// representable. We do not delegate to TextDecoder('windows-1252') because
|
|
14
|
+
// Node.js's built-in implementation regressed in 20.18.3 / 22.13.0 / 23.4.0
|
|
15
|
+
// and returns the raw bytes for 0x80–0x9F instead of the WHATWG-specified
|
|
16
|
+
// codepoints (nodejs/node#60888, still tracked under nodejs/node#61041).
|
|
17
|
+
const WINDOWS_1252_C1_MAP = {
|
|
18
|
+
0x80: 0x20ac, 0x82: 0x201a, 0x83: 0x0192, 0x84: 0x201e, 0x85: 0x2026,
|
|
19
|
+
0x86: 0x2020, 0x87: 0x2021, 0x88: 0x02c6, 0x89: 0x2030, 0x8a: 0x0160,
|
|
20
|
+
0x8b: 0x2039, 0x8c: 0x0152, 0x8e: 0x017d,
|
|
21
|
+
0x91: 0x2018, 0x92: 0x2019, 0x93: 0x201c, 0x94: 0x201d, 0x95: 0x2022,
|
|
22
|
+
0x96: 0x2013, 0x97: 0x2014, 0x98: 0x02dc, 0x99: 0x2122, 0x9a: 0x0161,
|
|
23
|
+
0x9b: 0x203a, 0x9c: 0x0153, 0x9e: 0x017e, 0x9f: 0x0178,
|
|
24
|
+
};
|
|
25
|
+
|
|
9
26
|
export default {
|
|
10
27
|
decode,
|
|
11
28
|
TAG_HEADER_SIZE
|
|
@@ -25,7 +42,29 @@ function decode(encoding, tagValue) {
|
|
|
25
42
|
}
|
|
26
43
|
|
|
27
44
|
const stringValue = tagValue.map((charCode) => String.fromCharCode(charCode)).join('');
|
|
28
|
-
|
|
45
|
+
try {
|
|
46
|
+
return decodeURIComponent(escape(stringValue));
|
|
47
|
+
} catch (error) {
|
|
48
|
+
// The bytes are not valid UTF-8. Many real-world IPTC blocks have no
|
|
49
|
+
// Coded Character Set tag but use Windows-1252 (en/em dashes,
|
|
50
|
+
// ellipsis, smart quotes, etc.). Remap the bytes through CP1252 so
|
|
51
|
+
// those characters render correctly instead of staying as C1 control
|
|
52
|
+
// characters.
|
|
53
|
+
return decodeWindows1252(tagValue);
|
|
54
|
+
}
|
|
55
|
+
}
|
|
56
|
+
|
|
57
|
+
function decodeWindows1252(tagValue) {
|
|
58
|
+
const isDataView = tagValue instanceof DataView;
|
|
59
|
+
const length = isDataView ? tagValue.byteLength : tagValue.length;
|
|
60
|
+
const getByte = isDataView ? (i) => tagValue.getUint8(i) : (i) => tagValue[i];
|
|
61
|
+
const codePoints = new Array(length);
|
|
62
|
+
for (let i = 0; i < length; i++) {
|
|
63
|
+
const byte = getByte(i);
|
|
64
|
+
const mapped = WINDOWS_1252_C1_MAP[byte];
|
|
65
|
+
codePoints[i] = mapped !== undefined ? mapped : byte;
|
|
66
|
+
}
|
|
67
|
+
return String.fromCharCode.apply(null, codePoints);
|
|
29
68
|
}
|
|
30
69
|
|
|
31
70
|
function decodeAsciiValue(asciiValue) {
|