domain/rdata/rfc1035/txt.rs
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 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437 438 439 440 441 442 443 444 445 446 447 448 449 450 451 452 453 454 455 456 457 458 459 460 461 462 463 464 465 466 467 468 469 470 471 472 473 474 475 476 477 478 479 480 481 482 483 484 485 486 487 488 489 490 491 492 493 494 495 496 497 498 499 500 501 502 503 504 505 506 507 508 509 510 511 512 513 514 515 516 517 518 519 520 521 522 523 524 525 526 527 528 529 530 531 532 533 534 535 536 537 538 539 540 541 542 543 544 545 546 547 548 549 550 551 552 553 554 555 556 557 558 559 560 561 562 563 564 565 566 567 568 569 570 571 572 573 574 575 576 577 578 579 580 581 582 583 584 585 586 587 588 589 590 591 592 593 594 595 596 597 598 599 600 601 602 603 604 605 606 607 608 609 610 611 612 613 614 615 616 617 618 619 620 621 622 623 624 625 626 627 628 629 630 631 632 633 634 635 636 637 638 639 640 641 642 643 644 645 646 647 648 649 650 651 652 653 654 655 656 657 658 659 660 661 662 663 664 665 666 667 668 669 670 671 672 673 674 675 676 677 678 679 680 681 682 683 684 685 686 687 688 689 690 691 692 693 694 695 696 697 698 699 700 701 702 703 704 705 706 707 708 709 710 711 712 713 714 715 716 717 718 719 720 721 722 723 724 725 726 727 728 729 730 731 732 733 734 735 736 737 738 739 740 741 742 743 744 745 746 747 748 749 750 751 752 753 754 755 756 757 758 759 760 761 762 763 764 765 766 767 768 769 770 771 772 773 774 775 776 777 778 779 780 781 782 783 784 785 786 787 788 789 790 791 792 793 794 795 796 797 798 799 800 801 802 803 804 805 806 807 808 809 810 811 812 813 814 815 816 817 818 819 820 821 822 823 824 825 826 827 828 829 830 831 832 833 834 835 836 837 838 839 840 841 842 843 844 845 846 847 848 849 850 851 852 853 854 855 856 857 858 859 860 861 862 863 864 865 866 867 868 869 870 871 872 873 874 875 876 877 878 879 880 881 882 883 884 885 886 887 888 889 890 891 892 893 894 895 896 897 898 899 900 901 902 903 904 905 906 907 908 909 910 911 912 913 914 915 916 917 918 919 920 921 922 923 924 925 926 927 928 929 930 931 932 933 934 935 936 937 938 939 940 941 942 943 944 945 946 947 948 949 950 951 952 953 954 955 956 957 958 959 960 961 962 963 964 965 966 967 968 969 970 971 972 973 974 975 976 977 978 979 980 981 982 983 984 985 986 987 988 989 990 991 992 993 994 995 996 997 998 999 1000 1001 1002 1003 1004 1005 1006 1007 1008 1009 1010 1011 1012 1013 1014 1015 1016 1017 1018 1019 1020 1021 1022 1023 1024 1025 1026 1027 1028 1029 1030 1031 1032 1033 1034 1035 1036 1037 1038 1039 1040 1041 1042 1043 1044 1045 1046 1047 1048 1049 1050 1051 1052 1053 1054 1055 1056 1057 1058 1059 1060 1061 1062 1063 1064 1065 1066 1067 1068 1069 1070 1071 1072 1073 1074 1075 1076 1077 1078 1079 1080 1081 1082 1083 1084 1085 1086 1087 1088 1089 1090 1091 1092 1093 1094 1095 1096 1097 1098 1099 1100 1101 1102 1103 1104 1105 1106 1107 1108 1109 1110 1111 1112 1113 1114 1115 1116 1117 1118 1119 1120 1121 1122
//! Record data for the TXT record.
//!
//! This is a private module. It’s content is re-exported by the parent.
use crate::base::charstr::CharStr;
#[cfg(feature = "serde")]
use crate::base::charstr::DeserializeCharStrSeed;
use crate::base::cmp::CanonicalOrd;
use crate::base::iana::Rtype;
use crate::base::rdata::{
ComposeRecordData, LongRecordData, ParseRecordData, RecordData,
};
use crate::base::scan::Scanner;
#[cfg(feature = "serde")]
use crate::base::scan::Symbol;
use crate::base::wire::{Composer, FormError, ParseError};
#[cfg(feature = "bytes")]
use bytes::BytesMut;
use core::cmp::Ordering;
use core::convert::{Infallible, TryFrom};
use core::{fmt, hash, mem, str};
use octseq::builder::{
infallible, EmptyBuilder, FreezeBuilder, FromBuilder, OctetsBuilder,
ShortBuf,
};
use octseq::octets::{Octets, OctetsFrom, OctetsInto};
use octseq::parse::Parser;
#[cfg(feature = "serde")]
use octseq::serde::{DeserializeOctets, SerializeOctets};
//------------ Txt ----------------------------------------------------------
/// TXT record data.
///
/// TXT records hold descriptive text. While it may appear as a single text,
/// it internally consists of a sequence of one or more
/// [character strings][CharStr]. The type holds this sequence in its encoded
/// form, i.e., each character string is at most 255 octets long and preceded
/// by an octet with its length.
///
/// The type provides means to iterate over these strings, either as
/// [`CharStr`s][CharStr] via [`iter_charstrs`][Self::iter_charstrs] or
/// as plain octets slices via [`iter`][Self::iter]. There is a short cut for
/// the most common case of there being exactly one character string in
/// [`as_flat_slice`][Self::as_flat_slice]. Finally, the two methods
/// [`text`][Self::text] and [`try_text`][Self::try_text] allow combining the
/// content into one single octets sequence.
///
/// The TXT record type is defined in [RFC 1035, section 3.3.14].
///
/// # Presentation format
///
/// TXT record data appears in zone files as the white-space delimited
/// sequence of its constituent [character strings][CharStr]. This means that
/// if these strings are not quoted, each “word” results in a character string
/// of its own. Thus, the quoted form of the character string’s presentation
/// format is preferred.
///
/// # `Display`
///
/// The `Display` implementation prints the sequence of character strings in
/// their quoted presentation format separated by a single space.
///
/// # Serde support
///
/// When the `serde` feature is enabled, the type supports serialization and
/// deserialization. The format differs for human readable and compact
/// serialization formats.
///
/// For human-readable formats, the type serializes into a newtype `Txt`
/// wrapping a sequence of serialized [`CharStr`]s. The deserializer supports
/// a non-canonical form as a single string instead of the sequence. In this
/// case the string is broken up into chunks of 255 octets if it is longer.
/// However, not all format implementations support alternative
/// deserialization based on the encountered type. In particular,
/// _serde-json_ doesn’t, so it will only accept sequences.
///
/// For compact formats, the type serializes as a newtype `Txt` that contains
/// a byte array of the wire format representation of the content.
///
/// [RFC 1035, section 3.3.14]: https://tools.ietf.org/html/rfc1035#section-3.3.14
#[derive(Clone)]
#[repr(transparent)]
pub struct Txt<Octs: ?Sized>(Octs);
impl Txt<()> {
/// The rtype of this record data type.
pub(crate) const RTYPE: Rtype = Rtype::TXT;
}
impl<Octs: FromBuilder> Txt<Octs> {
/// Creates a new Txt record from a single slice.
///
/// If the slice is longer than 255 octets, it will be broken up into
/// multiple character strings where all but the last string will be
/// 255 octets long.
///
/// If the slice is longer than 65,535 octets or longer than what fits
/// into the octets type used, an error is returned.
pub fn build_from_slice(text: &[u8]) -> Result<Self, TxtAppendError>
where
<Octs as FromBuilder>::Builder:
EmptyBuilder + AsRef<[u8]> + AsMut<[u8]>,
{
let mut builder = TxtBuilder::<Octs::Builder>::new();
builder.append_slice(text)?;
builder.finish().map_err(Into::into)
}
}
impl<Octs> Txt<Octs> {
/// Creates new TXT record data from its encoded content.
///
/// The `octets` sequence most contain correctly encoded TXT record
/// data. That is, it must contain a sequence of at least one character
/// string of at most 255 octets each preceded by a length octet. An
/// empty sequence is not allowed.
///
/// Returns an error if `octets` does not contain correctly encoded TXT
/// record data.
pub fn from_octets(octets: Octs) -> Result<Self, TxtError>
where
Octs: AsRef<[u8]>,
{
Txt::check_slice(octets.as_ref())?;
Ok(unsafe { Txt::from_octets_unchecked(octets) })
}
/// Creates new TXT record data without checking.
///
/// # Safety
///
/// The passed octets must contain correctly encoded TXT record data.
/// See [`from_octets][Self::from_octets] for the required content.
unsafe fn from_octets_unchecked(octets: Octs) -> Self {
Txt(octets)
}
}
impl Txt<[u8]> {
/// Creates new TXT record data on an octets slice.
///
/// The slice must contain correctly encoded TXT record data,
/// that is a sequence of encoded character strings. See
pub fn from_slice(slice: &[u8]) -> Result<&Self, TxtError> {
Txt::check_slice(slice)?;
Ok(unsafe { Txt::from_slice_unchecked(slice) })
}
/// Creates new TXT record data on an octets slice without checking.
///
/// # Safety
///
/// The passed octets must contain correctly encoded TXT record data.
/// See [`from_octets][Self::from_octets] for the required content.
unsafe fn from_slice_unchecked(slice: &[u8]) -> &Self {
// SAFETY: Txt has repr(transparent)
mem::transmute(slice)
}
/// Checks that a slice contains correctly encoded TXT data.
fn check_slice(mut slice: &[u8]) -> Result<(), TxtError> {
if slice.is_empty() {
return Err(TxtError(TxtErrorInner::Empty))
}
LongRecordData::check_len(slice.len())?;
while let Some(&len) = slice.first() {
let len = usize::from(len);
if slice.len() <= len {
return Err(TxtError(TxtErrorInner::ShortInput));
}
slice = &slice[len + 1..];
}
Ok(())
}
}
impl<Octs> Txt<Octs> {
/// Parses TXT record data from the beginning of a parser.
pub fn parse<'a, Src: Octets<Range<'a> = Octs> + ?Sized>(
parser: &mut Parser<'a, Src>,
) -> Result<Self, ParseError>
where
Octs: AsRef<[u8]>,
{
let len = parser.remaining();
let text = parser.parse_octets(len)?;
let mut tmp = Parser::from_ref(text.as_ref());
while tmp.remaining() != 0 {
CharStr::skip(&mut tmp)?
}
Ok(Txt(text))
}
/// Scans TXT record data.
pub fn scan<S: Scanner<Octets = Octs>>(
scanner: &mut S,
) -> Result<Self, S::Error> {
scanner.scan_charstr_entry().map(Txt)
}
}
impl<Octs: AsRef<[u8]> + ?Sized> Txt<Octs> {
/// Returns an iterator over the character strings as slices.
///
/// The returned iterator will always return at least one octets slice.
pub fn iter(&self) -> TxtIter {
TxtIter(self.iter_charstrs())
}
/// Returns an iterator over the character strings.
///
/// The returned iterator will always return at least one octets slice.
pub fn iter_charstrs(&self) -> TxtCharStrIter {
TxtCharStrIter(Parser::from_ref(self.0.as_ref()))
}
/// Returns the content if it consists of a single character string.
pub fn as_flat_slice(&self) -> Option<&[u8]> {
if usize::from(self.0.as_ref()[0]) == self.0.as_ref().len() - 1 {
Some(&self.0.as_ref()[1..])
} else {
None
}
}
/// Returns the length of the TXT record data.
///
/// Note that this is the length of the encoded record data and therefore
/// never the length of the text, not even if there is only a single
/// character string – it is still preceded by a length octet.
///
/// Note further that TXT record data is not allowed to be empty, so there
/// is no `is_empty` method.
#[allow(clippy::len_without_is_empty)]
pub fn len(&self) -> usize {
self.0.as_ref().len()
}
/// Returns the text content.
///
/// The method appends the content of each character string to a newly
/// created octets builder. It does not add any delimiters between the
/// character string.
///
/// If your octets builder is not space limited, you can use
/// [`text`][Self::text] instead.
pub fn try_text<T: FromBuilder>(
&self,
) -> Result<T, <<T as FromBuilder>::Builder as OctetsBuilder>::AppendError>
where
<T as FromBuilder>::Builder: EmptyBuilder,
{
// Capacity will be a few bytes too much. Probably better than
// re-allocating.
let mut res = T::Builder::with_capacity(self.len());
for item in self.iter() {
res.append_slice(item)?;
}
Ok(res.freeze())
}
/// Returns the text content.
///
/// The method appends the content of each character string to a newly
/// created octets builder. It does not add any delimiters between the
/// character string.
///
/// This method is only available for octets builder types that are not
/// space limited. You can use [`try_text`][Self::try_text] with all
/// builder types.
pub fn text<T: FromBuilder>(&self) -> T
where
<T as FromBuilder>::Builder: EmptyBuilder,
<<T as FromBuilder>::Builder as OctetsBuilder>::AppendError:
Into<Infallible>,
{
infallible(self.try_text())
}
}
impl<SrcOcts> Txt<SrcOcts> {
/// Converts the octets type.
///
/// This is used by the macros that create enum types.
pub(in crate::rdata) fn convert_octets<Target: OctetsFrom<SrcOcts>>(
self,
) -> Result<Txt<Target>, Target::Error> {
Ok(Txt(self.0.try_octets_into()?))
}
/// Flattens the contents.
///
/// This is used by the macros that create enum types.
pub(in crate::rdata) fn flatten<Octs: OctetsFrom<SrcOcts>>(
self,
) -> Result<Txt<Octs>, Octs::Error> {
self.convert_octets()
}
}
//--- OctetsFrom
impl<Octs, SrcOcts> OctetsFrom<Txt<SrcOcts>> for Txt<Octs>
where
Octs: OctetsFrom<SrcOcts>,
{
type Error = Octs::Error;
fn try_octets_from(source: Txt<SrcOcts>) -> Result<Self, Self::Error> {
Octs::try_octets_from(source.0).map(Self)
}
}
//--- IntoIterator
impl<'a, Octs: AsRef<[u8]>> IntoIterator for &'a Txt<Octs> {
type Item = &'a [u8];
type IntoIter = TxtIter<'a>;
fn into_iter(self) -> Self::IntoIter {
self.iter()
}
}
//--- PartialEq and Eq
impl<Octs, Other> PartialEq<Txt<Other>> for Txt<Octs>
where
Octs: AsRef<[u8]>,
Other: AsRef<[u8]>,
{
fn eq(&self, other: &Txt<Other>) -> bool {
self.0.as_ref().eq(other.0.as_ref())
}
}
impl<Octs: AsRef<[u8]>> Eq for Txt<Octs> {}
//--- PartialOrd, CanonicalOrd, and Ord
impl<Octs, Other> PartialOrd<Txt<Other>> for Txt<Octs>
where
Octs: AsRef<[u8]>,
Other: AsRef<[u8]>,
{
fn partial_cmp(&self, other: &Txt<Other>) -> Option<Ordering> {
self.0.as_ref().partial_cmp(other.0.as_ref())
}
}
impl<Octs, Other> CanonicalOrd<Txt<Other>> for Txt<Octs>
where
Octs: AsRef<[u8]>,
Other: AsRef<[u8]>,
{
fn canonical_cmp(&self, other: &Txt<Other>) -> Ordering {
self.0.as_ref().cmp(other.0.as_ref())
}
}
impl<Octs: AsRef<[u8]>> Ord for Txt<Octs> {
fn cmp(&self, other: &Self) -> Ordering {
self.0.as_ref().cmp(other.0.as_ref())
}
}
//--- Hash
impl<Octs: AsRef<[u8]>> hash::Hash for Txt<Octs> {
fn hash<H: hash::Hasher>(&self, state: &mut H) {
self.0.as_ref().hash(state)
}
}
//--- RecordData, ParseRecordData, ComposeRecordData
impl<Octs> RecordData for Txt<Octs> {
fn rtype(&self) -> Rtype {
Txt::RTYPE
}
}
impl<'a, Octs> ParseRecordData<'a, Octs> for Txt<Octs::Range<'a>>
where
Octs: Octets + ?Sized,
{
fn parse_rdata(
rtype: Rtype,
parser: &mut Parser<'a, Octs>,
) -> Result<Option<Self>, ParseError> {
if rtype == Txt::RTYPE {
Self::parse(parser).map(Some)
} else {
Ok(None)
}
}
}
impl<Octs: AsRef<[u8]>> ComposeRecordData for Txt<Octs> {
fn rdlen(&self, _compress: bool) -> Option<u16> {
Some(u16::try_from(self.0.as_ref().len()).expect("long TXT rdata"))
}
fn compose_rdata<Target: Composer + ?Sized>(
&self,
target: &mut Target,
) -> Result<(), Target::AppendError> {
target.append_slice(self.0.as_ref())
}
fn compose_canonical_rdata<Target: Composer + ?Sized>(
&self,
target: &mut Target,
) -> Result<(), Target::AppendError> {
self.compose_rdata(target)
}
}
//--- Display
impl<Octs: AsRef<[u8]>> fmt::Display for Txt<Octs> {
fn fmt(&self, f: &mut fmt::Formatter) -> fmt::Result {
let mut first = true;
for slice in self.iter_charstrs() {
if !first {
f.write_str(" ")?;
}
else {
first = false;
}
write!(f, "{}", slice.display_quoted())?;
}
Ok(())
}
}
//--- Debug
impl<Octs: AsRef<[u8]>> fmt::Debug for Txt<Octs> {
fn fmt(&self, f: &mut fmt::Formatter) -> fmt::Result {
f.write_str("Txt(")?;
fmt::Display::fmt(self, f)?;
f.write_str(")")
}
}
//--- Serialize and Deserialize
#[cfg(feature = "serde")]
impl<Octs> serde::Serialize for Txt<Octs>
where
Octs: AsRef<[u8]> + SerializeOctets,
{
fn serialize<S: serde::Serializer>(
&self,
serializer: S,
) -> Result<S::Ok, S::Error> {
use serde::ser::SerializeSeq;
struct TxtSeq<'a, Octs>(&'a Txt<Octs>);
impl<'a, Octs> serde::Serialize for TxtSeq<'a, Octs>
where
Octs: AsRef<[u8]> + SerializeOctets,
{
fn serialize<S: serde::Serializer>(
&self,
serializer: S,
) -> Result<S::Ok, S::Error> {
let mut serializer = serializer.serialize_seq(None)?;
for item in self.0.iter_charstrs() {
serializer.serialize_element(item)?;
}
serializer.end()
}
}
if serializer.is_human_readable() {
serializer.serialize_newtype_struct("Txt", &TxtSeq(self))
}
else {
serializer.serialize_newtype_struct(
"Txt",
&self.0.as_serialized_octets(),
)
}
}
}
#[cfg(feature = "serde")]
impl<'de, Octs> serde::Deserialize<'de> for Txt<Octs>
where
Octs: FromBuilder + DeserializeOctets<'de>,
<Octs as FromBuilder>::Builder: EmptyBuilder + AsRef<[u8]> + AsMut<[u8]>,
{
fn deserialize<D: serde::Deserializer<'de>>(
deserializer: D,
) -> Result<Self, D::Error> {
use core::marker::PhantomData;
struct NewtypeVisitor<T>(PhantomData<T>);
impl<'de, Octs> serde::de::Visitor<'de> for NewtypeVisitor<Octs>
where
Octs: FromBuilder + DeserializeOctets<'de>,
<Octs as FromBuilder>::Builder:
OctetsBuilder + EmptyBuilder + AsRef<[u8]> + AsMut<[u8]>,
{
type Value = Txt<Octs>;
fn expecting(&self, f: &mut fmt::Formatter) -> fmt::Result {
f.write_str("TXT record data")
}
fn visit_newtype_struct<D: serde::Deserializer<'de>>(
self,
deserializer: D,
) -> Result<Self::Value, D::Error> {
if deserializer.is_human_readable() {
deserializer.deserialize_seq(ReadableVisitor(PhantomData))
} else {
Octs::deserialize_with_visitor(
deserializer,
CompactVisitor(Octs::visitor()),
)
}
}
}
struct ReadableVisitor<Octs>(PhantomData<Octs>);
impl<'de, Octs> serde::de::Visitor<'de> for ReadableVisitor<Octs>
where
Octs: FromBuilder,
<Octs as FromBuilder>::Builder:
OctetsBuilder + EmptyBuilder + AsRef<[u8]> + AsMut<[u8]>,
{
type Value = Txt<Octs>;
fn expecting(&self, f: &mut fmt::Formatter) -> fmt::Result {
f.write_str("TXT record data")
}
fn visit_str<E: serde::de::Error>(
self,
v: &str,
) -> Result<Self::Value, E> {
// This is a non-canonical serialization. We accept strings
// of any length and break them down into chunks.
let mut builder =
TxtBuilder::<<Octs as FromBuilder>::Builder>::new();
let mut chars = v.chars();
while let Some(ch) =
Symbol::from_chars(&mut chars).map_err(E::custom)?
{
builder
.append_u8(ch.into_octet().map_err(E::custom)?)
.map_err(E::custom)?;
}
builder.finish().map_err(E::custom)
}
fn visit_seq<A: serde::de::SeqAccess<'de>>(
self,
mut seq: A,
) -> Result<Self::Value, A::Error> {
let mut builder = <Octs as FromBuilder>::Builder::empty();
while seq.next_element_seed(
DeserializeCharStrSeed::new(&mut builder)
)?.is_some() {
LongRecordData::check_len(
builder.as_ref().len()
).map_err(serde::de::Error::custom)?;
}
if builder.as_ref().is_empty() {
builder.append_slice(b"\0").map_err(|_| {
serde::de::Error::custom(ShortBuf)
})?;
}
Ok(Txt(builder.freeze()))
}
}
struct CompactVisitor<'de, T: DeserializeOctets<'de>>(T::Visitor);
impl<'de, Octs> serde::de::Visitor<'de> for CompactVisitor<'de, Octs>
where
Octs: FromBuilder + DeserializeOctets<'de>,
<Octs as FromBuilder>::Builder:
OctetsBuilder + EmptyBuilder + AsRef<[u8]> + AsMut<[u8]>,
{
type Value = Txt<Octs>;
fn expecting(&self, f: &mut fmt::Formatter) -> fmt::Result {
f.write_str("TXT record data")
}
fn visit_borrowed_bytes<E: serde::de::Error>(
self,
value: &'de [u8],
) -> Result<Self::Value, E> {
self.0.visit_borrowed_bytes(value).and_then(|octets| {
Txt::from_octets(octets).map_err(E::custom)
})
}
#[cfg(feature = "std")]
fn visit_byte_buf<E: serde::de::Error>(
self,
value: std::vec::Vec<u8>,
) -> Result<Self::Value, E> {
self.0.visit_byte_buf(value).and_then(|octets| {
Txt::from_octets(octets).map_err(E::custom)
})
}
}
deserializer.deserialize_newtype_struct(
"Txt", NewtypeVisitor(PhantomData)
)
}
}
//------------ TxtCharStrIter ------------------------------------------------
/// An iterator over the character strings of a Txt record.
#[derive(Clone)]
pub struct TxtCharStrIter<'a>(Parser<'a, [u8]>);
impl<'a> Iterator for TxtCharStrIter<'a> {
type Item = &'a CharStr<[u8]>;
fn next(&mut self) -> Option<Self::Item> {
if self.0.remaining() == 0 {
None
} else {
Some(CharStr::parse_slice(&mut self.0).unwrap())
}
}
}
//------------ TxtIter -------------------------------------------------------
/// An iterator over the character strings of a Txt record.
#[derive(Clone)]
pub struct TxtIter<'a>(TxtCharStrIter<'a>);
impl<'a> Iterator for TxtIter<'a> {
type Item = &'a [u8];
fn next(&mut self) -> Option<Self::Item> {
self.0.next().map(CharStr::as_slice)
}
}
//------------ TxtBuilder ---------------------------------------------------
/// Iteratively build TXT record data.
///
/// This type allows building TXT record data by starting with empty data
/// and appending either complete character strings or slices of data.
#[derive(Clone, Debug)]
pub struct TxtBuilder<Builder> {
/// The underlying builder.
builder: Builder,
/// The index of the start of the current char string.
///
/// If this is `None`, there currently is no char string being worked on.
start: Option<usize>,
}
impl<Builder: OctetsBuilder + EmptyBuilder> TxtBuilder<Builder> {
/// Creates a new, empty TXT builder.
#[must_use]
pub fn new() -> Self {
TxtBuilder {
builder: Builder::empty(),
start: None,
}
}
}
#[cfg(feature = "bytes")]
impl TxtBuilder<BytesMut> {
/// Creates a new, empty TXT builder using `BytesMut`.
pub fn new_bytes() -> Self {
Self::new()
}
}
impl<Builder: OctetsBuilder + AsRef<[u8]> + AsMut<[u8]>> TxtBuilder<Builder> {
/// Tries appending a slice.
///
/// Errors out if either appending the slice would result in exceeding the
/// record data length limit or the underlying builder runs out of space.
fn builder_append_slice(
&mut self, slice: &[u8]
) -> Result<(), TxtAppendError> {
LongRecordData::check_append_len(
self.builder.as_ref().len(), slice.len()
)?;
self.builder.append_slice(slice)?;
Ok(())
}
/// Appends a slice to the builder.
///
/// The method breaks up the slice into individual octets strings if
/// necessary. If a previous call has started a new octets string, it
/// fills this one up first before creating a new one. Thus, by using
/// this method only, the resulting TXT record data will consist of
/// character strings where all but the last one are 255 octets long.
///
/// You can force a character string break by calling
/// [`close_charstr`][Self::close_charstr].
///
/// The method will return an error if appending the slice would result
/// in exceeding the record data length limit or the underlying builder
/// runs out of space. In this case, the method may have appended some
/// data already. I.e., you should consider the builder corrupt if the
/// method returns an error.
pub fn append_slice(
&mut self, mut slice: &[u8]
) -> Result<(), TxtAppendError> {
if let Some(start) = self.start {
let left = 255 - (self.builder.as_ref().len() - (start + 1));
if slice.len() < left {
self.builder_append_slice(slice)?;
return Ok(());
}
let (append, left) = slice.split_at(left);
self.builder_append_slice(append)?;
self.builder.as_mut()[start] = 255;
slice = left;
}
for chunk in slice.chunks(255) {
// Remember offset of this incomplete chunk
self.start = if chunk.len() == 255 {
None
} else {
Some(self.builder.as_ref().len())
};
self.builder_append_slice(&[chunk.len() as u8])?;
self.builder_append_slice(chunk)?;
}
Ok(())
}
/// Appends a single octet.
///
/// This method calls [`append_slice`][Self::append_slice], so all the
/// caveats described there apply.
pub fn append_u8(&mut self, ch: u8) -> Result<(), TxtAppendError> {
self.append_slice(&[ch])
}
/// Appends a complete character string.
///
/// If a character string had previously been started by a call to
/// [`append_slice`][Self::append_slice], this string is closed before
/// appending the provided character string.
///
/// The method will return an error if appending the slice would result
/// in exceeding the record data length limit or the underlying builder
/// runs out of space. In this case, the method may have appended some
/// data already. I.e., you should consider the builder corrupt if the
/// method returns an error.
pub fn append_charstr<Octs: AsRef<[u8]> + ?Sized>(
&mut self, s: &CharStr<Octs>
) -> Result<(), TxtAppendError> {
self.close_charstr();
LongRecordData::check_append_len(
self.builder.as_ref().len(),
usize::from(s.compose_len())
)?;
s.compose(&mut self.builder)?;
Ok(())
}
/// Ends a character string.
///
/// If a previous call to [`append_slice`][Self::append_slice] started a
/// new character string, a call to this method will close it.
pub fn close_charstr(&mut self) {
if let Some(start) = self.start {
let last_slice_len = self.builder.as_ref().len() - (start + 1);
self.builder.as_mut()[start] = last_slice_len as u8;
self.start = None;
}
}
/// Finishes the builder and returns TXT record data.
///
/// If the builder is empty, appends an empty character string before
/// returning. If that fails because the builder does not have enough
/// space, returns an error.
pub fn finish(mut self) -> Result<Txt<Builder::Octets>, TxtAppendError>
where
Builder: FreezeBuilder,
{
self.close_charstr();
if self.builder.as_ref().is_empty() {
self.builder.append_slice(b"\0")?;
}
Ok(Txt(self.builder.freeze()))
}
}
impl<Builder: OctetsBuilder + EmptyBuilder> Default for TxtBuilder<Builder> {
fn default() -> Self {
Self::new()
}
}
//============ Error Types ===================================================
//------------ TxtError ------------------------------------------------------
/// An octets sequence does not form valid TXT record data.
#[derive(Clone, Copy, Debug)]
pub struct TxtError(TxtErrorInner);
#[derive(Clone, Copy, Debug)]
enum TxtErrorInner {
Empty,
Long(LongRecordData),
ShortInput,
}
impl TxtError {
#[must_use]
pub fn as_str(self) -> &'static str {
match self.0 {
TxtErrorInner::Empty => "empty TXT record",
TxtErrorInner::Long(err) => err.as_str(),
TxtErrorInner::ShortInput => "short input",
}
}
}
impl From<LongRecordData> for TxtError {
fn from(err: LongRecordData) -> TxtError {
TxtError(TxtErrorInner::Long(err))
}
}
impl From<TxtError> for FormError {
fn from(err: TxtError) -> FormError {
FormError::new(err.as_str())
}
}
impl fmt::Display for TxtError {
fn fmt(&self, f: &mut fmt::Formatter) -> fmt::Result {
f.write_str(self.as_str())
}
}
//------------ TxtAppendError ------------------------------------------------
/// An error occurred while append to TXT record data.
#[derive(Clone, Copy, Debug)]
pub enum TxtAppendError {
/// Appending would have caused the record data to be too long.
LongRecordData,
/// The octets builder did not have enough space.
ShortBuf
}
impl TxtAppendError {
/// Returns a static string with the error reason.
#[must_use]
pub fn as_str(self) -> &'static str {
match self {
TxtAppendError::LongRecordData => "record data too long",
TxtAppendError::ShortBuf => "buffer size exceeded"
}
}
}
impl From<LongRecordData> for TxtAppendError {
fn from(_: LongRecordData) -> TxtAppendError {
TxtAppendError::LongRecordData
}
}
impl<T: Into<ShortBuf>> From<T> for TxtAppendError {
fn from(_: T) -> TxtAppendError {
TxtAppendError::ShortBuf
}
}
impl fmt::Display for TxtAppendError {
fn fmt(&self, f: &mut fmt::Formatter) -> fmt::Result {
f.write_str(self.as_str())
}
}
//============ Testing =======================================================
#[cfg(test)]
#[cfg(all(feature = "std", feature = "bytes"))]
mod test {
use super::*;
use crate::base::rdata::test::{
test_compose_parse, test_rdlen, test_scan,
};
use std::vec::Vec;
#[test]
#[allow(clippy::redundant_closure)] // lifetimes ...
fn txt_compose_parse_scan() {
let rdata = Txt::from_octets(b"\x03foo\x03bar".as_ref()).unwrap();
test_rdlen(&rdata);
test_compose_parse(&rdata, |parser| Txt::parse(parser));
test_scan(&["foo", "bar"], Txt::scan, &rdata);
}
#[test]
fn txt_from_slice() {
assert!(Txt::from_octets(b"").is_err());
let short = b"01234";
let txt: Txt<Vec<u8>> = Txt::build_from_slice(short).unwrap();
assert_eq!(Some(&short[..]), txt.as_flat_slice());
assert_eq!(short.to_vec(), txt.text::<Vec<u8>>());
// One full slice
let full = short.repeat(51);
let txt: Txt<Vec<u8>> = Txt::build_from_slice(&full).unwrap();
assert_eq!(Some(&full[..]), txt.as_flat_slice());
assert_eq!(full.to_vec(), txt.text::<Vec<u8>>());
// Two slices: 255, 5
let long = short.repeat(52);
let txt: Txt<Vec<u8>> = Txt::build_from_slice(&long).unwrap();
assert_eq!(None, txt.as_flat_slice());
assert_eq!(long.to_vec(), txt.text::<Vec<u8>>());
// Partial
let mut builder: TxtBuilder<Vec<u8>> = TxtBuilder::new();
for chunk in long.chunks(9) {
builder.append_slice(chunk).unwrap();
}
let txt = builder.finish().unwrap();
assert_eq!(None, txt.as_flat_slice());
assert_eq!(long.to_vec(), txt.text::<Vec<u8>>());
// Empty
let builder: TxtBuilder<Vec<u8>> = TxtBuilder::new();
let txt = builder.finish().unwrap();
assert_eq!(Some(b"".as_ref()), txt.as_flat_slice());
// Empty
let mut builder: TxtBuilder<Vec<u8>> = TxtBuilder::new();
builder.append_slice(b"").unwrap();
let txt = builder.finish().unwrap();
assert_eq!(Some(b"".as_ref()), txt.as_flat_slice());
// Invalid
let mut parser = Parser::from_static(b"\x01");
assert!(Txt::parse(&mut parser).is_err());
// Too long
let mut builder: TxtBuilder<Vec<u8>> = TxtBuilder::new();
assert!(builder
.append_slice(&b"\x00".repeat(u16::MAX as usize))
.is_err());
// Incremental, reserve space for offsets
let mut builder: TxtBuilder<Vec<u8>> = TxtBuilder::new();
assert!(builder
.append_slice(&b"\x00".repeat(u16::MAX as usize - 512))
.is_ok());
assert!(builder.append_slice(&b"\x00".repeat(512)).is_err());
}
#[test]
fn txt_canonical_compare() {
let data = [
"mailru-verification: 14505c6eb222c847",
"yandex-verification: 6059b187e78de544",
"v=spf1 include:_spf.protonmail.ch ~all",
"swisssign-check=CF0JHMTlTDNoES3rrknIRggocffSwqmzMb9X8YbjzK",
"google-site-\
verification=aq9zJnp3H3bNE0Y4D4rH5I5Dhj8VMaLYx0uQ7Rozfgg",
"ahrefs-site-verification_\
4bdac6bbaa81e0d591d7c0f3ef238905c0521b69bf3d74e64d3775bc\
b2743afd",
"brave-ledger-verification=\
66a7f27fb99949cc0c564ab98efcc58ea1bac3e97eb557c782ab2d44b\
49aefd7",
];
let records = data
.iter()
.map(|e| {
let mut builder = TxtBuilder::<Vec<u8>>::new();
builder.append_slice(e.as_bytes()).unwrap();
builder.finish().unwrap()
})
.collect::<Vec<_>>();
// The canonical sort must sort by TXT labels which are prefixed by
// length byte first.
let mut sorted = records.clone();
sorted.sort_by(|a, b| a.canonical_cmp(b));
for (a, b) in records.iter().zip(sorted.iter()) {
assert_eq!(a, b);
}
}
#[test]
fn txt_strings_eq() {
let records = [["foo", "bar"], ["foob", "ar"], ["foo", "bar"]];
let records = records
.iter()
.map(|strings| {
let mut builder = TxtBuilder::<Vec<u8>>::new();
for string in strings {
builder
.append_charstr(
CharStr::from_slice(string.as_bytes()).unwrap(),
)
.unwrap();
}
builder.finish().unwrap()
})
.collect::<Vec<_>>();
assert_ne!(records[0], records[1]);
assert_eq!(records[0], records[2]);
}
#[cfg(all(feature = "serde", feature = "std"))]
#[test]
fn txt_ser_de() {
use serde_test::{assert_tokens, Configure, Token};
let txt = Txt::from_octets(Vec::from(b"\x03foo".as_ref())).unwrap();
assert_tokens(
&txt.clone().compact(),
&[
Token::NewtypeStruct { name: "Txt" },
Token::ByteBuf(b"\x03foo"),
],
);
assert_tokens(
&txt.readable(),
&[
Token::NewtypeStruct { name: "Txt" },
Token::Seq { len: None },
Token::NewtypeStruct { name: "CharStr" },
Token::BorrowedStr("foo"),
Token::SeqEnd,
],
);
let txt = Txt::from_octets(
Vec::from(b"\x03foo\x04\\bar".as_ref())
).unwrap();
assert_tokens(
&txt.clone().compact(),
&[
Token::NewtypeStruct { name: "Txt" },
Token::ByteBuf(b"\x03foo\x04\\bar"),
],
);
assert_tokens(
&txt.readable(),
&[
Token::NewtypeStruct { name: "Txt" },
Token::Seq { len: None },
Token::NewtypeStruct { name: "CharStr" },
Token::BorrowedStr("foo"),
Token::NewtypeStruct { name: "CharStr" },
Token::BorrowedStr("\\\\bar"),
Token::SeqEnd,
],
);
}
#[cfg(all(feature = "serde", feature = "std"))]
#[test]
fn txt_de_str() {
use serde_test::{assert_de_tokens, Configure, Token};
assert_de_tokens(
&Txt::from_octets(Vec::from(b"\x03foo".as_ref()))
.unwrap()
.readable(),
&[
Token::NewtypeStruct { name: "Txt" },
Token::BorrowedStr("foo"),
],
);
}
#[test]
fn txt_display() {
fn cmp(input: &[u8], output: &str) {
assert_eq!(
format!("{}", Txt::from_octets(input).unwrap()),
output
);
}
cmp(b"\x03foo", "\"foo\"");
cmp(b"\x03foo\x03bar", "\"foo\" \"bar\"");
cmp(b"\x03fo\"\x04bar ", "\"fo\\\"\" \"bar \"");
// I don’t think we need more escaping tests since the impl defers
// to CharStr::display_quoted which is tested ...
}
}