1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437 438 439 440 441 442 443 444 445 446 447 448 449 450 451 452 453 454 455 456 457 458 459 460 461 462 463 464 465 466 467 468 469 470 471 472 473 474 475 476 477 478 479 480 481 482 483 484 485 486 487 488 489 490 491 492 493 494 495 496 497 498 499 500 501 502 503 504 505 506 507 508 509 510 511 512 513 514 515 516 517 518 519 520 521 522 523 524 525 526 527 528 529 530 531 532 533 534 535 536 537 538 539 540 541 542 543 544 545 546 547 548 549 550 551 552 553 554 555 556 557 558 559 560 561 562 563 564 565 566 567 568 569 570 571 572 573 574 575 576 577 578 579 580 581 582 583 584 585 586 587 588 589 590 591 592 593 594 595 596 597 598 599 600 601 602 603 604 605 606 607 608 609 610 611 612 613 614 615 616 617 618 619 620 621 622 623 624 625 626 627 628 629 630 631 632 633 634 635 636 637 638 639 640 641 642 643 644 645 646 647 648 649 650 651 652 653 654 655 656 657 658 659 660 661 662 663 664 665 666 667 668 669 670 671 672 673 674 675 676 677 678 679 680 681 682 683 684 685 686 687 688 689 690 691 692 693 694 695 696
// Copyright 2015-2016 Mozilla Foundation. See the COPYRIGHT
// file at the top-level directory of this distribution.
//
// Licensed under the Apache License, Version 2.0 <LICENSE-APACHE or
// https://www.apache.org/licenses/LICENSE-2.0> or the MIT license
// <LICENSE-MIT or https://opensource.org/licenses/MIT>, at your
// option. This file may not be copied, modified, or distributed
// except according to those terms.
#![doc(html_root_url = "https://docs.rs/charset/0.1.0")]
//! `charset` is a wrapper around [`encoding_rs`][1] that provides
//! (non-streaming) decoding for character encodings that occur in _email_ by
//! providing decoding for [UTF-7][2] in addition to the encodings defined by
//! the [Encoding Standard][3] (and provided by `encoding_rs`).
//!
//! _Note:_ Do _not_ use this crate for consuming _Web_ content. For security
//! reasons, consumers of Web content are [_prohibited_][4] from supporting
//! UTF-7. Use `encoding_rs` directly when consuming Web content.
//!
//! The set of encodings consisting of UTF-7 and the encodings defined in the
//! Encoding Standard is believed to be appropriate for consuming email,
//! because that's the set of encodings supported by [Thunderbird][5].
//! Furthermore, UTF-7 support is believed to be necessary based on the
//! experience of the Firefox OS email client. In fact, while the UTF-7
//! implementation in this crate is independent of Thunderbird's UTF-7
//! implementation, Thunderbird uses `encoding_rs` to decode the other
//! encodings. The set of _labels_/_aliases_ recognized by this crate matches
//! those recognized by Thunderbird.
//!
//! Known compatibility limitations (shared with Thunderbird and known from
//! Thunderbird bug reports):
//!
//! * JavaMail may use non-standard labels for legacy encodings such that
//! the labels aren't recognized by this crate even if the encodings
//! themselves would be supported.
//! * Some ancient Usenet posting in Chinese may not be decodable, because
//! this crate does not support HZ.
//! * Some emails sent in Chinese by Sun's email client for CDE on Solaris
//! around the turn of the millennium may not decodable, because this
//! crate does not support ISO-2022-CN.
//! * Some emails sent in Korean by IBM/Lotus Notes may not be decodable,
//! because this crate does not support ISO-2022-KR.
//!
//! This crate intentionally does not support encoding content into legacy
//! encodings. When sending email, _always_ use UTF-8. This is, just call
//! `.as_bytes()` on `&str` and label the content as `UTF-8`.
//!
//! [1]: https://crates.io/crates/encoding_rs/
//! [2]: https://tools.ietf.org/html/rfc2152
//! [3]: https://encoding.spec.whatwg.org/
//! [4]: https://html.spec.whatwg.org/#character-encodings
//! [5]: https://thunderbird.net/
//!
//! # Security considerations
//!
//! Again, this crate is for _email_. Please do _NOT_ use it for _Web_
//! content.
//!
//! Never try to perform any security analysis on the undecoded data in
//! ASCII-incompatible encodings and in UTF-7 in particular. Always decode
//! first and analyze after. UTF-7 allows even characters that don't have to
//! be represeted as base64 to be represented as base64. Also, for consistency
//! with Thunderbird, the UTF-7 decoder in this crate allows e.g. ASCII
//! controls to be represented without base64 encoding even when the spec
//! says they should be base64-encoded.
//!
//! This implementation is non-constant-time by design. An attacker who
//! can observe input length and the time it takes to decode it can make
//! guesses about relative proportions of characters from different ranges.
//! Guessing the proportion of ASCII vs. non-ASCII should be particularly
//! feasible.
extern crate base64;
extern crate encoding_rs;
use encoding_rs::CoderResult;
use encoding_rs::Encoding;
use encoding_rs::GB18030;
use encoding_rs::GBK;
use encoding_rs::UTF_16BE;
use std::borrow::Cow;
/// The UTF-7 encoding.
pub const UTF_7: Charset = Charset {
variant: VariantCharset::Utf7,
};
/// A character encoding suitable for decoding _email_.
///
/// This is either an encoding as defined in the [Encoding Standard][1]
/// or UTF-7 as defined in [RFC 2152][2].
///
/// [1]: https://encoding.spec.whatwg.org/
/// [2]: https://tools.ietf.org/html/rfc2152
///
/// Each `Charset` has one or more _labels_ that are used to identify
/// the `Charset` in protocol text. In MIME/IANA terminology, these are
/// called _names_ and _aliases_, but for consistency with the Encoding
/// Standard and the encoding_rs crate, they are called labels in this
/// crate. What this crate calls the _name_ (again, for consistency
/// with the Encoding Standard and the encoding_rs crate) is known as
/// _preferred name_ in MIME/IANA terminology.
///
/// Instances of `Charset` can be compared with `==`. `Charset` is
/// `Copy` and is meant to be passed by value.
///
/// _Note:_ It is wrong to use this for decoding Web content. Use
/// `encoding_rs::Encoding` instead!
#[derive(PartialEq, Debug, Copy, Clone, Hash)]
pub struct Charset {
variant: VariantCharset,
}
impl Charset {
/// Implements the
/// [_get an encoding_](https://encoding.spec.whatwg.org/#concept-encoding-get)
/// algorithm with the label "UTF-7" added to the set of labels recognized.
/// GBK is unified with gb18030, since they decode the same and `Charset`
/// only supports decoding.
///
/// If, after ASCII-lowercasing and removing leading and trailing
/// whitespace, the argument matches a label defined in the Encoding
/// Standard or "utf-7", `Some(Charset)` representing the corresponding
/// encoding is returned. If there is no match, `None` is returned.
///
/// This is the right method to use if the action upon the method returning
/// `None` is to use a fallback encoding (e.g. `WINDOWS_1252`) instead.
/// When the action upon the method returning `None` is not to proceed with
/// a fallback but to refuse processing, `for_label_no_replacement()` is more
/// appropriate.
///
/// The argument is of type `&[u8]` instead of `&str` to save callers
/// that are extracting the label from a non-UTF-8 protocol the trouble
/// of conversion to UTF-8. (If you have a `&str`, just call `.as_bytes()`
/// on it.)
#[inline]
pub fn for_label(label: &[u8]) -> Option<Charset> {
if let Some(encoding) = Encoding::for_label(label) {
Some(Charset::for_encoding(encoding))
} else if is_utf7_label(label) {
Some(UTF_7)
} else {
None
}
}
/// This method behaves the same as `for_label()`, except when `for_label()`
/// would return `Some(Charset::for_encoding(encoding_rs::REPLACEMENT))`,
/// this method returns `None` instead.
///
/// This method is useful in scenarios where a fatal error is required
/// upon invalid label, because in those cases the caller typically wishes
/// to treat the labels that map to the replacement encoding as fatal
/// errors, too.
///
/// It is not OK to use this method when the action upon the method returning
/// `None` is to use a fallback encoding (e.g. `WINDOWS_1252`) with `text/html`
/// email. In such a case, the `for_label()` method should be used instead in
/// order to avoid unsafe fallback for labels that `for_label()` maps to
/// `Some(REPLACEMENT)`. Such fallback might be safe, though not particularly
/// useful for `text/plain` email, though.
#[inline]
pub fn for_label_no_replacement(label: &[u8]) -> Option<Charset> {
if let Some(encoding) = Encoding::for_label_no_replacement(label) {
Some(Charset::for_encoding(encoding))
} else if is_utf7_label(label) {
Some(UTF_7)
} else {
None
}
}
/// Returns the `Charset` corresponding to an `&'static Encoding`.
///
/// `GBK` is unified with `GB18030`, since those two decode the same
/// and `Charset` only supports decoding.
#[inline]
pub fn for_encoding(encoding: &'static Encoding) -> Charset {
let enc = if encoding == GBK { GB18030 } else { encoding };
Charset {
variant: VariantCharset::Encoding(enc),
}
}
/// Performs non-incremental BOM sniffing.
///
/// The argument must either be a buffer representing the entire input
/// stream (non-streaming case) or a buffer representing at least the first
/// three bytes of the input stream (streaming case).
///
/// Returns `Some((Charset::for_encoding(encoding_rs::UTF_8), 3))`,
/// `Some((Charset::for_encoding(encoding_rs::UTF_16LE), 2))` or
/// `Some((Charset::for_encoding(encoding_rs::UTF_16BE), 2))` if the
/// argument starts with the UTF-8, UTF-16LE or UTF-16BE BOM or `None`
/// otherwise.
#[inline]
pub fn for_bom(buffer: &[u8]) -> Option<(Charset, usize)> {
if let Some((encoding, length)) = Encoding::for_bom(buffer) {
Some((Charset::for_encoding(encoding), length))
} else {
None
}
}
/// Returns the name of this encoding.
///
/// Mostly useful for debugging
pub fn name(self) -> &'static str {
match self.variant {
VariantCharset::Encoding(encoding) => encoding.name(),
VariantCharset::Utf7 => "UTF-7",
}
}
/// Checks whether the bytes 0x00...0x7F map exclusively to the characters
/// U+0000...U+007F and vice versa.
#[inline]
pub fn is_ascii_compatible(self) -> bool {
match self.variant {
VariantCharset::Encoding(encoding) => encoding.is_ascii_compatible(),
VariantCharset::Utf7 => false,
}
}
/// Decode complete input to `Cow<'a, str>` _with BOM sniffing_ and with
/// malformed sequences replaced with the REPLACEMENT CHARACTER when the
/// entire input is available as a single buffer (i.e. the end of the
/// buffer marks the end of the stream).
///
/// This method implements the (non-streaming version of) the
/// [_decode_](https://encoding.spec.whatwg.org/#decode) spec concept.
///
/// The second item in the returned tuple is the encoding that was actually
/// used (which may differ from this encoding thanks to BOM sniffing).
///
/// The third item in the returned tuple indicates whether there were
/// malformed sequences (that were replaced with the REPLACEMENT CHARACTER).
///
/// _Note:_ It is wrong to use this when the input buffer represents only
/// a segment of the input instead of the whole input.
///
/// # Panics
///
/// If the size calculation for a heap-allocated backing buffer overflows
/// `usize`.
#[inline]
pub fn decode<'a>(self, bytes: &'a [u8]) -> (Cow<'a, str>, Charset, bool) {
let (charset, without_bom) = match Charset::for_bom(bytes) {
Some((charset, bom_length)) => (charset, &bytes[bom_length..]),
None => (self, bytes),
};
let (cow, had_errors) = charset.decode_without_bom_handling(without_bom);
(cow, charset, had_errors)
}
/// Decode complete input to `Cow<'a, str>` _with BOM removal_ and with
/// malformed sequences replaced with the REPLACEMENT CHARACTER when the
/// entire input is available as a single buffer (i.e. the end of the
/// buffer marks the end of the stream).
///
/// When invoked on `UTF_8`, this method implements the (non-streaming
/// version of) the
/// [_UTF-8 decode_](https://encoding.spec.whatwg.org/#utf-8-decode) spec
/// concept.
///
/// The second item in the returned pair indicates whether there were
/// malformed sequences (that were replaced with the REPLACEMENT CHARACTER).
///
/// _Note:_ It is wrong to use this when the input buffer represents only
/// a segment of the input instead of the whole input.
///
/// # Panics
///
/// If the size calculation for a heap-allocated backing buffer overflows
/// `usize`.
#[inline]
pub fn decode_with_bom_removal<'a>(self, bytes: &'a [u8]) -> (Cow<'a, str>, bool) {
match self.variant {
VariantCharset::Encoding(encoding) => encoding.decode_with_bom_removal(bytes),
VariantCharset::Utf7 => decode_utf7(bytes),
}
}
/// Decode complete input to `Cow<'a, str>` _without BOM handling_ and
/// with malformed sequences replaced with the REPLACEMENT CHARACTER when
/// the entire input is available as a single buffer (i.e. the end of the
/// buffer marks the end of the stream).
///
/// When invoked on `UTF_8`, this method implements the (non-streaming
/// version of) the
/// [_UTF-8 decode without BOM_](https://encoding.spec.whatwg.org/#utf-8-decode-without-bom)
/// spec concept.
///
/// The second item in the returned pair indicates whether there were
/// malformed sequences (that were replaced with the REPLACEMENT CHARACTER).
///
/// _Note:_ It is wrong to use this when the input buffer represents only
/// a segment of the input instead of the whole input.
///
/// # Panics
///
/// If the size calculation for a heap-allocated backing buffer overflows
/// `usize`.
#[inline]
pub fn decode_without_bom_handling<'a>(self, bytes: &'a [u8]) -> (Cow<'a, str>, bool) {
match self.variant {
VariantCharset::Encoding(encoding) => encoding.decode_without_bom_handling(bytes),
VariantCharset::Utf7 => decode_utf7(bytes),
}
}
}
#[inline(never)]
fn is_utf7_label(label: &[u8]) -> bool {
let mut iter = label.into_iter();
// before
loop {
match iter.next() {
None => {
return false;
}
Some(&byte) => match byte {
0x09u8 | 0x0Au8 | 0x0Cu8 | 0x0Du8 | 0x20u8 => {
continue;
}
b'u' | b'U' => {
break;
}
_ => {
return false;
}
},
}
}
// inside
let tail = iter.as_slice();
if tail.len() < 4 {
return false;
}
match (tail[0] | 0x20, tail[1] | 0x20, tail[2], tail[3]) {
(b't', b'f', b'-', b'7') => {}
_ => {
return false;
}
}
iter = (&tail[4..]).into_iter();
// after
loop {
match iter.next() {
None => {
return true;
}
Some(&byte) => match byte {
0x09u8 | 0x0Au8 | 0x0Cu8 | 0x0Du8 | 0x20u8 => {
continue;
}
_ => {
return false;
}
},
}
}
}
#[inline]
fn utf7_ascii_up_to(bytes: &[u8]) -> usize {
for (i, &byte) in bytes.into_iter().enumerate() {
if byte == b'+' || byte >= 0x80 {
return i;
}
}
bytes.len()
}
#[inline]
fn utf7_base64_up_to(bytes: &[u8]) -> usize {
for (i, &byte) in bytes.into_iter().enumerate() {
match byte {
b'a'...b'z' | b'A'...b'Z' | b'0'...b'9' | b'+' | b'/' => {}
_ => {
return i;
}
}
}
bytes.len()
}
#[inline]
fn utf7_base64_decode(bytes: &[u8], string: &mut String) -> bool {
// The intermediate buffer should be long enough to fit a line
// of 80 base64 bytes and should also be a multiple of 3. This
// way, normal email lines will be handled in one go, but
// longer sequences won't get split between base64 groups of
// 4 input / 3 output bytes.
let mut decoder = UTF_16BE.new_decoder_without_bom_handling();
let mut buf = [0u8; 60];
let mut tail = bytes;
let mut had_errors = false;
let mut trailing_error = false;
loop {
let (last, mut cap) = if tail.len() <= 80 {
(true, tail.len())
} else {
(false, 80)
};
let len;
loop {
match base64::decode_config_slice(&tail[..cap], base64::STANDARD_NO_PAD, &mut buf[..]) {
Ok(l) => {
len = l;
break;
}
Err(_) => {
assert!(last);
had_errors = true;
trailing_error = true;
tail = &tail[..tail.len() - 1];
cap -= 1;
}
}
}
let mut total_read = 0;
loop {
let (result, read, err) = decoder.decode_to_string(&buf[total_read..len], string, last);
total_read += read;
had_errors |= err;
match result {
CoderResult::InputEmpty => {
if last {
if trailing_error {
string.push_str("\u{FFFD}");
}
return had_errors;
}
break;
}
CoderResult::OutputFull => {
let left = len - total_read;
let needed = decoder.max_utf8_buffer_length(left).unwrap();
string.reserve(needed);
}
}
}
tail = &tail[80..];
}
}
#[inline(never)]
fn decode_utf7<'a>(bytes: &'a [u8]) -> (Cow<'a, str>, bool) {
let up_to = utf7_ascii_up_to(bytes);
if up_to == bytes.len() {
let s: &str = unsafe { std::str::from_utf8_unchecked(bytes) };
return (Cow::Borrowed(s), false);
}
let mut had_errors = false;
let mut out = String::with_capacity(bytes.len());
out.push_str(unsafe { std::str::from_utf8_unchecked(&bytes[..up_to]) });
let mut tail = &bytes[up_to..];
loop {
// `tail[0]` is now either a plus sign or non-ASCII
let first = tail[0];
tail = &tail[1..];
if first == b'+' {
let up_to = utf7_base64_up_to(tail);
had_errors |= utf7_base64_decode(&tail[..up_to], &mut out);
if up_to == tail.len() {
if up_to == 0 {
// Plus sign didn't start a base64 run and also
// wasn't followed by a minus.
had_errors = true;
out.push_str("\u{FFFD}");
}
return (Cow::Owned(out), had_errors);
}
if up_to == 0 {
if tail[up_to] == b'-' {
// There was no base64 data between
// plus and minus, so we had the sequence
// meaning the plus sign itself.
out.push_str("+");
tail = &tail[up_to + 1..];
} else {
// Plus sign didn't start a base64 run and also
// wasn't followed by a minus.
had_errors = true;
out.push_str("\u{FFFD}");
}
} else if tail[up_to] == b'-' {
tail = &tail[up_to + 1..];
} else {
tail = &tail[up_to..];
}
} else {
had_errors = true;
out.push_str("\u{FFFD}");
}
let up_to = utf7_ascii_up_to(tail);
out.push_str(unsafe { std::str::from_utf8_unchecked(&tail[..up_to]) });
if up_to == tail.len() {
return (Cow::Owned(out), had_errors);
}
tail = &tail[up_to..];
}
}
#[derive(PartialEq, Debug, Copy, Clone, Hash)]
enum VariantCharset {
Utf7,
Encoding(&'static Encoding),
}
#[cfg(test)]
mod tests {
use super::*;
fn utf7_no_err(bytes: &[u8]) -> String {
let (cow, had_errors) = UTF_7.decode_without_bom_handling(bytes);
assert!(!had_errors);
cow.into()
}
fn utf7_err(bytes: &[u8]) -> String {
let (cow, had_errors) = UTF_7.decode_without_bom_handling(bytes);
assert!(had_errors);
cow.into()
}
// Any copyright to the test code below this comment is dedicated to the
// Public Domain. http://creativecommons.org/publicdomain/zero/1.0/
#[test]
fn test_for_label() {
assert_eq!(Charset::for_label(b" uTf-7\t "), Some(UTF_7));
assert_eq!(
Charset::for_label(b" uTf-8\t "),
Some(Charset::for_encoding(encoding_rs::UTF_8))
);
assert_eq!(
Charset::for_label(b" iSo-8859-1\t "),
Some(Charset::for_encoding(encoding_rs::WINDOWS_1252))
);
assert_eq!(
Charset::for_label(b" gb2312\t "),
Some(Charset::for_encoding(encoding_rs::GB18030))
);
assert_eq!(
Charset::for_label(b" ISO-2022-KR\t "),
Some(Charset::for_encoding(encoding_rs::REPLACEMENT))
);
assert_eq!(Charset::for_label(b"u"), None);
assert_eq!(Charset::for_label(b"ut"), None);
assert_eq!(Charset::for_label(b"utf"), None);
assert_eq!(Charset::for_label(b"utf-"), None);
}
#[test]
fn test_for_label_no_replacement() {
assert_eq!(
Charset::for_label_no_replacement(b" uTf-7\t "),
Some(UTF_7)
);
assert_eq!(
Charset::for_label_no_replacement(b" uTf-8\t "),
Some(Charset::for_encoding(encoding_rs::UTF_8))
);
assert_eq!(
Charset::for_label_no_replacement(b" iSo-8859-1\t "),
Some(Charset::for_encoding(encoding_rs::WINDOWS_1252))
);
assert_eq!(
Charset::for_label_no_replacement(b" Gb2312\t "),
Some(Charset::for_encoding(encoding_rs::GB18030))
);
assert_eq!(Charset::for_label_no_replacement(b" ISO-2022-KR\t "), None);
assert_eq!(Charset::for_label_no_replacement(b"u"), None);
assert_eq!(Charset::for_label_no_replacement(b"ut"), None);
assert_eq!(Charset::for_label_no_replacement(b"utf"), None);
assert_eq!(Charset::for_label_no_replacement(b"utf-"), None);
}
#[test]
fn test_for_label_and_name() {
assert_eq!(Charset::for_label(b" uTf-7\t ").unwrap().name(), "UTF-7");
assert_eq!(Charset::for_label(b" uTf-8\t ").unwrap().name(), "UTF-8");
assert_eq!(
Charset::for_label(b" Gb2312\t ").unwrap().name(),
"gb18030"
);
}
#[test]
fn test_utf7_decode() {
assert_eq!(utf7_no_err(b""), "");
assert_eq!(utf7_no_err(b"ab"), "ab");
assert_eq!(utf7_no_err(b"+-"), "+");
assert_eq!(utf7_no_err(b"a+-b"), "a+b");
assert_eq!(utf7_no_err(b"+ACs-"), "+");
assert_eq!(utf7_no_err(b"+AGEAKwBi-"), "a+b");
assert_eq!(utf7_no_err(b"+JgM-"), "\u{2603}");
assert_eq!(utf7_no_err(b"+JgM."), "\u{2603}.");
assert_eq!(utf7_no_err(b"+JgM "), "\u{2603} ");
assert_eq!(utf7_no_err(b"+JgM--"), "\u{2603}-");
assert_eq!(utf7_no_err(b"+JgM"), "\u{2603}");
assert_eq!(utf7_no_err(b"+JgMmAw-"), "\u{2603}\u{2603}");
assert_eq!(utf7_no_err(b"+JgMmAw."), "\u{2603}\u{2603}.");
assert_eq!(utf7_no_err(b"+JgMmAw "), "\u{2603}\u{2603} ");
assert_eq!(utf7_no_err(b"+JgMmAw--"), "\u{2603}\u{2603}-");
assert_eq!(utf7_no_err(b"+JgMmAw"), "\u{2603}\u{2603}");
assert_eq!(utf7_no_err(b"+2D3cqQ-"), "\u{1F4A9}");
assert_eq!(utf7_no_err(b"+2D3cqQ."), "\u{1F4A9}.");
assert_eq!(utf7_no_err(b"+2D3cqQ "), "\u{1F4A9} ");
assert_eq!(utf7_no_err(b"+2D3cqQ--"), "\u{1F4A9}-");
assert_eq!(utf7_no_err(b"+2D3cqQ"), "\u{1F4A9}");
assert_eq!(utf7_no_err(b"+JgPYPdyp2D3cqdg93KnYPdyp2D3cqdg93KnYPdyp2D3cqdg93KnYPdyp2D3cqdg93KnYPdyp2D3cqdg93KnYPdyp2D3cqdg93KnYPdyp2D3cqdg93KnYPdyp2D3cqdg93KnYPdyp2D3cqdg93KnYPdyp2D3cqdg93KnYPdyp2D3cqdg93KnYPdyp2D3cqdg93KnYPdyp2D3cqdg93KnYPdyp"), "\u{2603}\u{1F4A9}\u{1F4A9}\u{1F4A9}\u{1F4A9}\u{1F4A9}\u{1F4A9}\u{1F4A9}\u{1F4A9}\u{1F4A9}\u{1F4A9}\u{1F4A9}\u{1F4A9}\u{1F4A9}\u{1F4A9}\u{1F4A9}\u{1F4A9}\u{1F4A9}\u{1F4A9}\u{1F4A9}\u{1F4A9}\u{1F4A9}\u{1F4A9}\u{1F4A9}\u{1F4A9}\u{1F4A9}\u{1F4A9}\u{1F4A9}\u{1F4A9}\u{1F4A9}\u{1F4A9}\u{1F4A9}\u{1F4A9}\u{1F4A9}\u{1F4A9}\u{1F4A9}\u{1F4A9}\u{1F4A9}\u{1F4A9}\u{1F4A9}\u{1F4A9}");
assert_eq!(utf7_err(b"+"), "\u{FFFD}");
assert_eq!(utf7_err(b"+J-"), "\u{FFFD}");
assert_eq!(utf7_err(b"+Jg-"), "\u{FFFD}");
assert_eq!(utf7_err(b"+J"), "\u{FFFD}");
assert_eq!(utf7_err(b"+Jg"), "\u{FFFD}");
assert_eq!(utf7_err(b"+."), "\u{FFFD}.");
assert_eq!(utf7_err(b"+J."), "\u{FFFD}.");
assert_eq!(utf7_err(b"+Jg."), "\u{FFFD}.");
assert_eq!(utf7_err(b"+ "), "\u{FFFD} ");
assert_eq!(utf7_err(b"+J "), "\u{FFFD} ");
assert_eq!(utf7_err(b"+Jg "), "\u{FFFD} ");
assert_eq!(utf7_err(b"+JgMmA-"), "\u{2603}\u{FFFD}\u{FFFD}");
assert_eq!(utf7_err(b"+JgMmA"), "\u{2603}\u{FFFD}\u{FFFD}");
assert_eq!(utf7_err(b"+JgMmA."), "\u{2603}\u{FFFD}\u{FFFD}.");
assert_eq!(utf7_err(b"+JgMmA "), "\u{2603}\u{FFFD}\u{FFFD} ");
assert_eq!(utf7_err(b"+JgMm-"), "\u{2603}\u{FFFD}");
assert_eq!(utf7_err(b"+JgMm"), "\u{2603}\u{FFFD}");
assert_eq!(utf7_err(b"+JgMm."), "\u{2603}\u{FFFD}.");
assert_eq!(utf7_err(b"+JgMm "), "\u{2603}\u{FFFD} ");
assert_eq!(utf7_err(b"+2D3cq-"), "\u{FFFD}\u{FFFD}");
assert_eq!(utf7_err(b"+2D3cq"), "\u{FFFD}\u{FFFD}");
assert_eq!(utf7_err(b"+2D3cq."), "\u{FFFD}\u{FFFD}.");
assert_eq!(utf7_err(b"+2D3cq "), "\u{FFFD}\u{FFFD} ");
assert_eq!(utf7_err(b"+2D3c-"), "\u{FFFD}");
assert_eq!(utf7_err(b"+2D3c"), "\u{FFFD}");
assert_eq!(utf7_err(b"+2D3c."), "\u{FFFD}.");
assert_eq!(utf7_err(b"+2D3c "), "\u{FFFD} ");
assert_eq!(utf7_err(b"+2D3-"), "\u{FFFD}");
assert_eq!(utf7_err(b"+2D3"), "\u{FFFD}");
assert_eq!(utf7_err(b"+2D3."), "\u{FFFD}.");
assert_eq!(utf7_err(b"+2D3 "), "\u{FFFD} ");
assert_eq!(utf7_err(b"+2D-"), "\u{FFFD}");
assert_eq!(utf7_err(b"+2D"), "\u{FFFD}");
assert_eq!(utf7_err(b"+2D."), "\u{FFFD}.");
assert_eq!(utf7_err(b"+2D "), "\u{FFFD} ");
assert_eq!(utf7_err(b"+2-"), "\u{FFFD}");
assert_eq!(utf7_err(b"+2"), "\u{FFFD}");
assert_eq!(utf7_err(b"+2."), "\u{FFFD}.");
assert_eq!(utf7_err(b"+2 "), "\u{FFFD} ");
// Lone high surrogate
assert_eq!(utf7_err(b"+2D0-"), "\u{FFFD}");
assert_eq!(utf7_err(b"+2D0"), "\u{FFFD}");
assert_eq!(utf7_err(b"+2D0."), "\u{FFFD}.");
assert_eq!(utf7_err(b"+2D0 "), "\u{FFFD} ");
assert_eq!(utf7_err(b"+2D0AYQ-"), "\u{FFFD}a");
assert_eq!(utf7_err(b"+2D0AYQ"), "\u{FFFD}a");
assert_eq!(utf7_err(b"+2D0AYQ."), "\u{FFFD}a.");
assert_eq!(utf7_err(b"+2D0AYQ "), "\u{FFFD}a ");
assert_eq!(utf7_err(b"+2D3/QQ-"), "\u{FFFD}\u{FF41}");
assert_eq!(utf7_err(b"+2D3/QQ"), "\u{FFFD}\u{FF41}");
assert_eq!(utf7_err(b"+2D3/QQ."), "\u{FFFD}\u{FF41}.");
assert_eq!(utf7_err(b"+2D3/QQ "), "\u{FFFD}\u{FF41} ");
// Lone low surrogate
assert_eq!(utf7_err(b"+AGHcqQ-"), "a\u{FFFD}");
assert_eq!(utf7_err(b"+AGHcqQ"), "a\u{FFFD}");
assert_eq!(utf7_err(b"+AGHcqQ."), "a\u{FFFD}.");
assert_eq!(utf7_err(b"+AGHcqQ "), "a\u{FFFD} ");
}
}