@bobfrankston/iflow-direct 0.1.62 → 0.1.64
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/imap-protocol.js +41 -4
- package/package.json +1 -1
package/imap-protocol.js
CHANGED
|
@@ -344,6 +344,40 @@ function unquote(s) {
|
|
|
344
344
|
return s.slice(1, -1).replace(/\\(.)/g, "$1");
|
|
345
345
|
return s;
|
|
346
346
|
}
|
|
347
|
+
/** Build a TextDecoder for an encoded-word charset label, trying progressively
|
|
348
|
+
* looser readings of the label before giving up on UTF-8.
|
|
349
|
+
*
|
|
350
|
+
* TextDecoder throws RangeError on a label it doesn't know, and the caller's
|
|
351
|
+
* catch then returned the RAW encoded text — so ONE unrecognized label put
|
|
352
|
+
* `London_Startup_Is_Selling_Permanent=2C_Painless=2C_=22Stick-On=22` in the
|
|
353
|
+
* user's subject line (Bob 2026-08-02, NNSquad via mailman). Labels seen in
|
|
354
|
+
* the wild that TextDecoder rejects:
|
|
355
|
+
* `en_US.UTF-8` — a POSIX LOCALE (language_TERRITORY.CODESET), not a
|
|
356
|
+
* charset; the real charset is the codeset after the dot.
|
|
357
|
+
* `utf-8*en` — RFC 2231 language suffix appended to the charset.
|
|
358
|
+
* `utf8` — no hyphen.
|
|
359
|
+
* The raw label is tried FIRST so legitimate dotted names (`ansi_x3.4-1968`)
|
|
360
|
+
* are never mangled by the locale rule. Last resort is UTF-8, which decodes
|
|
361
|
+
* leniently (U+FFFD for bad bytes) rather than throwing — a couple of
|
|
362
|
+
* replacement characters beat a subject full of quoted-printable. */
|
|
363
|
+
function decoderFor(label) {
|
|
364
|
+
const raw = label.trim().toLowerCase();
|
|
365
|
+
const candidates = [raw];
|
|
366
|
+
const noLang = raw.split("*")[0]; // RFC 2231 suffix
|
|
367
|
+
if (noLang !== raw)
|
|
368
|
+
candidates.push(noLang);
|
|
369
|
+
const locale = noLang.match(/^[a-z]{1,8}(?:_[a-z0-9]{1,8})?\.(.+)$/);
|
|
370
|
+
if (locale)
|
|
371
|
+
candidates.push(locale[1]); // POSIX codeset
|
|
372
|
+
candidates.push(noLang.replace(/^utf[_]?8$/, "utf-8"));
|
|
373
|
+
for (const cs of candidates) {
|
|
374
|
+
try {
|
|
375
|
+
return new TextDecoder(cs);
|
|
376
|
+
}
|
|
377
|
+
catch { /* unknown label — try the next reading */ }
|
|
378
|
+
}
|
|
379
|
+
return new TextDecoder("utf-8");
|
|
380
|
+
}
|
|
347
381
|
/** Decode IMAP encoded-word (=?charset?encoding?text?=) */
|
|
348
382
|
function decodeImapString(s) {
|
|
349
383
|
if (!s)
|
|
@@ -352,13 +386,13 @@ function decodeImapString(s) {
|
|
|
352
386
|
const unfolded = s.replace(/\?=\s+=\?/g, "?==?");
|
|
353
387
|
return unfolded.replace(/=\?([^?]+)\?([BQ])\?([^?]+)\?=/gi, (_match, charset, encoding, text) => {
|
|
354
388
|
try {
|
|
355
|
-
const
|
|
389
|
+
const decoder = decoderFor(charset);
|
|
356
390
|
if (encoding.toUpperCase() === "B") {
|
|
357
391
|
const raw = atob(text);
|
|
358
392
|
const bytes = new Uint8Array(raw.length);
|
|
359
393
|
for (let j = 0; j < raw.length; j++)
|
|
360
394
|
bytes[j] = raw.charCodeAt(j);
|
|
361
|
-
return
|
|
395
|
+
return decoder.decode(bytes);
|
|
362
396
|
}
|
|
363
397
|
else {
|
|
364
398
|
// Quoted-printable: collect bytes then decode with charset
|
|
@@ -375,11 +409,14 @@ function decodeImapString(s) {
|
|
|
375
409
|
i++;
|
|
376
410
|
}
|
|
377
411
|
}
|
|
378
|
-
return
|
|
412
|
+
return decoder.decode(new Uint8Array(bytes));
|
|
379
413
|
}
|
|
380
414
|
}
|
|
381
415
|
catch {
|
|
382
|
-
|
|
416
|
+
// Only reachable now for malformed base64 (atob throws). Strip the
|
|
417
|
+
// Q-encoding's underscore-for-space at least, so the fallback isn't
|
|
418
|
+
// more wrong than it has to be.
|
|
419
|
+
return encoding.toUpperCase() === "Q" ? text.replace(/_/g, " ") : text;
|
|
383
420
|
}
|
|
384
421
|
});
|
|
385
422
|
}
|