host.rs 9.5 KB

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960616263646566676869707172737475767778798081828384858687888990919293949596979899100101102103104105106107108109110111112113114115116117118119120121122123124125126127128129130131132133134135136137138139140141142143144145146147148149150151152153154155156157158159160161162163164165166167168169170171172173174175176177178179180181182183184185186187188189190191192193194195196197198199200201202203204205206207208209210211212213214215216217218219220221222223224225226227228229230231232233234235236237238239240241242243244245246247248249250251252253254255256257258259260261262263264265266267268269270271272273274275276277278279280281282283284285286287288289290291292293294295296297298299300301302303304305306307308309310311312313314315316317318319320321322323
  1. // Copyright 2013-2014 Simon Sapin.
  2. //
  3. // Licensed under the Apache License, Version 2.0 <LICENSE-APACHE or
  4. // http://www.apache.org/licenses/LICENSE-2.0> or the MIT license
  5. // <LICENSE-MIT or http://opensource.org/licenses/MIT>, at your
  6. // option. This file may not be copied, modified, or distributed
  7. // except according to those terms.
  8. use std::ascii::AsciiExt;
  9. use std::cmp;
  10. use std::fmt::{self, Formatter};
  11. use std::net::{Ipv4Addr, Ipv6Addr};
  12. use parser::{ParseResult, ParseError};
  13. use percent_encoding::{from_hex, percent_decode};
  14. use idna;
  15. /// The host name of an URL.
  16. #[derive(PartialEq, Eq, Clone, Debug, Hash, PartialOrd, Ord)]
  17. #[cfg_attr(feature="heap_size", derive(HeapSizeOf))]
  18. pub enum Host {
  19. /// A (DNS) domain name.
  20. Domain(String),
  21. /// A IPv4 address, represented by four sequences of up to three ASCII digits.
  22. Ipv4(Ipv4Addr),
  23. /// An IPv6 address, represented inside `[...]` square brackets
  24. /// so that `:` colon characters in the address are not ambiguous
  25. /// with the port number delimiter.
  26. Ipv6(Ipv6Addr),
  27. }
  28. impl Host {
  29. /// Parse a host: either an IPv6 address in [] square brackets, or a domain.
  30. ///
  31. /// Returns `Err` for an empty host, an invalid IPv6 address,
  32. /// or a or invalid non-ASCII domain.
  33. pub fn parse(input: &str) -> ParseResult<Host> {
  34. if input.len() == 0 {
  35. return Err(ParseError::EmptyHost)
  36. }
  37. if input.starts_with("[") {
  38. if !input.ends_with("]") {
  39. return Err(ParseError::InvalidIpv6Address)
  40. }
  41. return parse_ipv6addr(&input[1..input.len() - 1]).map(Host::Ipv6)
  42. }
  43. let decoded = percent_decode(input.as_bytes());
  44. let domain = String::from_utf8_lossy(&decoded);
  45. let domain = match idna::domain_to_ascii(&domain) {
  46. Ok(s) => s,
  47. Err(_) => return Err(ParseError::InvalidDomainCharacter)
  48. };
  49. if domain.find(&[
  50. '\0', '\t', '\n', '\r', ' ', '#', '%', '/', ':', '?', '@', '[', '\\', ']'
  51. ][..]).is_some() {
  52. return Err(ParseError::InvalidDomainCharacter)
  53. }
  54. match parse_ipv4addr(&domain[..]) {
  55. Ok(Some(ipv4addr)) => Ok(Host::Ipv4(ipv4addr)),
  56. Ok(None) => Ok(Host::Domain(domain.to_ascii_lowercase())),
  57. Err(e) => Err(e),
  58. }
  59. }
  60. /// Serialize the host as a string.
  61. ///
  62. /// A domain a returned as-is, an IPv6 address between [] square brackets.
  63. pub fn serialize(&self) -> String {
  64. self.to_string()
  65. }
  66. }
  67. impl fmt::Display for Host {
  68. fn fmt(&self, f: &mut Formatter) -> fmt::Result {
  69. match *self {
  70. Host::Domain(ref domain) => domain.fmt(f),
  71. Host::Ipv4(ref addr) => addr.fmt(f),
  72. Host::Ipv6(ref addr) => {
  73. try!(f.write_str("["));
  74. try!(write_ipv6(addr, f));
  75. f.write_str("]")
  76. }
  77. }
  78. }
  79. }
  80. fn write_ipv6(addr: &Ipv6Addr, f: &mut Formatter) -> fmt::Result {
  81. let segments = addr.segments();
  82. let (compress_start, compress_end) = longest_zero_sequence(&segments);
  83. let mut i = 0;
  84. while i < 8 {
  85. if i == compress_start {
  86. try!(f.write_str(":"));
  87. if i == 0 {
  88. try!(f.write_str(":"));
  89. }
  90. if compress_end < 8 {
  91. i = compress_end;
  92. } else {
  93. break;
  94. }
  95. }
  96. try!(write!(f, "{:x}", segments[i as usize]));
  97. if i < 7 {
  98. try!(f.write_str(":"));
  99. }
  100. i += 1;
  101. }
  102. Ok(())
  103. }
  104. fn longest_zero_sequence(pieces: &[u16; 8]) -> (isize, isize) {
  105. let mut longest = -1;
  106. let mut longest_length = -1;
  107. let mut start = -1;
  108. macro_rules! finish_sequence(
  109. ($end: expr) => {
  110. if start >= 0 {
  111. let length = $end - start;
  112. if length > longest_length {
  113. longest = start;
  114. longest_length = length;
  115. }
  116. }
  117. };
  118. );
  119. for i in 0..8 {
  120. if pieces[i as usize] == 0 {
  121. if start < 0 {
  122. start = i;
  123. }
  124. } else {
  125. finish_sequence!(i);
  126. start = -1;
  127. }
  128. }
  129. finish_sequence!(8);
  130. (longest, longest + longest_length)
  131. }
  132. fn parse_ipv4number(mut input: &str) -> ParseResult<u32> {
  133. let mut r = 10;
  134. if input.starts_with("0x") || input.starts_with("0X") {
  135. input = &input[2..];
  136. r = 16;
  137. } else if input.len() >= 2 && input.starts_with("0") {
  138. input = &input[1..];
  139. r = 8;
  140. }
  141. if input.is_empty() {
  142. return Ok(0);
  143. }
  144. if input.starts_with("+") {
  145. return Err(ParseError::InvalidIpv4Address)
  146. }
  147. match u32::from_str_radix(&input, r) {
  148. Ok(number) => Ok(number),
  149. Err(_) => Err(ParseError::InvalidIpv4Address),
  150. }
  151. }
  152. fn parse_ipv4addr(input: &str) -> ParseResult<Option<Ipv4Addr>> {
  153. let mut parts: Vec<&str> = input.split('.').collect();
  154. if parts.last() == Some(&"") {
  155. parts.pop();
  156. }
  157. if parts.len() > 4 {
  158. return Ok(None);
  159. }
  160. let mut numbers: Vec<u32> = Vec::new();
  161. for part in parts {
  162. if part == "" {
  163. return Ok(None);
  164. }
  165. if let Ok(n) = parse_ipv4number(part) {
  166. numbers.push(n);
  167. } else {
  168. return Ok(None);
  169. }
  170. }
  171. let mut ipv4 = numbers.pop().expect("a non-empty list of numbers");
  172. // Equivalent to: ipv4 >= 256 ** (4 − numbers.len())
  173. if ipv4 > u32::max_value() >> (8 * numbers.len() as u32) {
  174. return Err(ParseError::InvalidIpv4Address);
  175. }
  176. if numbers.iter().any(|x| *x > 255) {
  177. return Err(ParseError::InvalidIpv4Address);
  178. }
  179. for (counter, n) in numbers.iter().enumerate() {
  180. ipv4 += n << (8 * (3 - counter as u32))
  181. }
  182. Ok(Some(Ipv4Addr::from(ipv4)))
  183. }
  184. fn parse_ipv6addr(input: &str) -> ParseResult<Ipv6Addr> {
  185. let input = input.as_bytes();
  186. let len = input.len();
  187. let mut is_ip_v4 = false;
  188. let mut pieces = [0, 0, 0, 0, 0, 0, 0, 0];
  189. let mut piece_pointer = 0;
  190. let mut compress_pointer = None;
  191. let mut i = 0;
  192. if len < 2 {
  193. return Err(ParseError::InvalidIpv6Address)
  194. }
  195. if input[0] == b':' {
  196. if input[1] != b':' {
  197. return Err(ParseError::InvalidIpv6Address)
  198. }
  199. i = 2;
  200. piece_pointer = 1;
  201. compress_pointer = Some(1);
  202. }
  203. while i < len {
  204. if piece_pointer == 8 {
  205. return Err(ParseError::InvalidIpv6Address)
  206. }
  207. if input[i] == b':' {
  208. if compress_pointer.is_some() {
  209. return Err(ParseError::InvalidIpv6Address)
  210. }
  211. i += 1;
  212. piece_pointer += 1;
  213. compress_pointer = Some(piece_pointer);
  214. continue
  215. }
  216. let start = i;
  217. let end = cmp::min(len, start + 4);
  218. let mut value = 0u16;
  219. while i < end {
  220. match from_hex(input[i]) {
  221. Some(digit) => {
  222. value = value * 0x10 + digit as u16;
  223. i += 1;
  224. },
  225. None => break
  226. }
  227. }
  228. if i < len {
  229. match input[i] {
  230. b'.' => {
  231. if i == start {
  232. return Err(ParseError::InvalidIpv6Address)
  233. }
  234. i = start;
  235. is_ip_v4 = true;
  236. },
  237. b':' => {
  238. i += 1;
  239. if i == len {
  240. return Err(ParseError::InvalidIpv6Address)
  241. }
  242. },
  243. _ => return Err(ParseError::InvalidIpv6Address)
  244. }
  245. }
  246. if is_ip_v4 {
  247. break
  248. }
  249. pieces[piece_pointer] = value;
  250. piece_pointer += 1;
  251. }
  252. if is_ip_v4 {
  253. if piece_pointer > 6 {
  254. return Err(ParseError::InvalidIpv6Address)
  255. }
  256. let mut dots_seen = 0;
  257. while i < len {
  258. // FIXME: https://github.com/whatwg/url/commit/1c22aa119c354e0020117e02571cec53f7c01064
  259. let mut value = 0u16;
  260. while i < len {
  261. let digit = match input[i] {
  262. c @ b'0' ... b'9' => c - b'0',
  263. _ => break
  264. };
  265. value = value * 10 + digit as u16;
  266. if value == 0 || value > 255 {
  267. return Err(ParseError::InvalidIpv6Address)
  268. }
  269. }
  270. if dots_seen < 3 && !(i < len && input[i] == b'.') {
  271. return Err(ParseError::InvalidIpv6Address)
  272. }
  273. pieces[piece_pointer] = pieces[piece_pointer] * 0x100 + value;
  274. if dots_seen == 0 || dots_seen == 2 {
  275. piece_pointer += 1;
  276. }
  277. i += 1;
  278. if dots_seen == 3 && i < len {
  279. return Err(ParseError::InvalidIpv6Address)
  280. }
  281. dots_seen += 1;
  282. }
  283. }
  284. match compress_pointer {
  285. Some(compress_pointer) => {
  286. let mut swaps = piece_pointer - compress_pointer;
  287. piece_pointer = 7;
  288. while swaps > 0 {
  289. pieces[piece_pointer] = pieces[compress_pointer + swaps - 1];
  290. pieces[compress_pointer + swaps - 1] = 0;
  291. swaps -= 1;
  292. piece_pointer -= 1;
  293. }
  294. }
  295. _ => if piece_pointer != 8 {
  296. return Err(ParseError::InvalidIpv6Address)
  297. }
  298. }
  299. Ok(Ipv6Addr::new(pieces[0], pieces[1], pieces[2], pieces[3],
  300. pieces[4], pieces[5], pieces[6], pieces[7]))
  301. }