| 1 |
use core::num::Wrapping; |
| 2 |
use core::slice; |
| 3 |
use core::fmt::Write; |
| 4 |
#[cfg(not(feature = "std"))] |
| 5 |
use alloc::vec::Vec; |
| 6 |
use crate::error::{LhaError, LhaResult}; |
| 7 |
use crate::stub_io::Read; |
| 8 |
use crate::crc::Crc16; |
| 9 |
use super::*; |
| 10 |
|
| 11 |
/// Raw identifiers of extra headers. |
| 12 |
pub mod ext { |
| 13 |
/// The "Common" header's CRC-16 field will always be reset to 0 in the parsed header data. |
| 14 |
/// This is the necessary condition to verify header's checksum. |
| 15 |
pub const EXT_HEADER_COMMON: u8 = 0x00; |
| 16 |
pub const EXT_HEADER_FILENAME: u8 = 0x01; |
| 17 |
pub const EXT_HEADER_PATH: u8 = 0x02; |
| 18 |
pub const EXT_HEADER_MULTI_DISC: u8 = 0x39; |
| 19 |
pub const EXT_HEADER_COMMENT: u8 = 0x3F; |
| 20 |
pub const EXT_HEADER_MSDOS_ATTRS: u8 = 0x40; |
| 21 |
pub const EXT_HEADER_MSDOS_TIME: u8 = 0x41; |
| 22 |
pub const EXT_HEADER_MSDOS_SIZE: u8 = 0x42; |
| 23 |
pub const EXT_HEADER_UNIX_PERM: u8 = 0x50; |
| 24 |
pub const EXT_HEADER_UNIX_UIDGID: u8 = 0x51; |
| 25 |
pub const EXT_HEADER_UNIX_GROUP: u8 = 0x52; |
| 26 |
pub const EXT_HEADER_UNIX_OWNER: u8 = 0x53; |
| 27 |
pub const EXT_HEADER_UNIX_TIME: u8 = 0x54; |
| 28 |
pub const EXT_HEADER_OS9: u8 = 0xCC; |
| 29 |
pub const EXT_HEADER_EXT_ATTRS: u8 = 0x7F; |
| 30 |
} |
| 31 |
|
| 32 |
use ext::*; |
| 33 |
/// An iterator through extra headers, yielding the headers' raw content excluding |
| 34 |
/// the next header length field. |
| 35 |
pub struct ExtraHeaderIter<'a> { |
| 36 |
data: &'a [u8], |
| 37 |
header_length: u32, |
| 38 |
header_len32: bool |
| 39 |
} |
| 40 |
|
| 41 |
impl<'a> Iterator for ExtraHeaderIter<'a> { |
| 42 |
type Item = &'a [u8]; |
| 43 |
|
| 44 |
fn next(&mut self) -> Option<Self::Item> { |
| 45 |
let header_length = self.header_length as usize; |
| 46 |
if header_length == 0 { |
| 47 |
return None |
| 48 |
} |
| 49 |
let counter_size = if self.header_len32 { 4 } else { 2 }; |
| 50 |
let (res, data) = self.data.split_at(header_length); |
| 51 |
let (res, len) = res.split_at(header_length - counter_size); |
| 52 |
let len = if self.header_len32 { |
| 53 |
read_u32(len).unwrap() |
| 54 |
} |
| 55 |
else { |
| 56 |
read_u16(len).unwrap() as u32 |
| 57 |
}; |
| 58 |
self.header_length = len; |
| 59 |
self.data = data; |
| 60 |
Some(res) |
| 61 |
} |
| 62 |
} |
| 63 |
|
| 64 |
#[derive(Clone, Copy, Debug, Default)] |
| 65 |
#[repr(C)] |
| 66 |
#[repr(packed)] |
| 67 |
struct LhaRawBaseHeader { |
| 68 |
compression: [u8;5], |
| 69 |
compressed_size: [u8;4], |
| 70 |
original_size: [u8;4], |
| 71 |
last_modified: [u8;4], |
| 72 |
msdos_attrs: u8, |
| 73 |
lha_level: u8 |
| 74 |
} |
| 75 |
|
| 76 |
struct Parser<'a, R> { |
| 77 |
rd: &'a mut R, |
| 78 |
crc: Crc16, |
| 79 |
csum: Wrapping<u8>, |
| 80 |
len: usize |
| 81 |
} |
| 82 |
|
| 83 |
impl<R: Read> Parser<'_, R> { |
| 84 |
// NOTE: does not update wrapping sum |
| 85 |
fn read_u8_or_none(&mut self) -> LhaResult<Option<u8>, R> { |
| 86 |
let mut byte = 0u8; |
| 87 |
if 0 == self.rd.read_all(slice::from_mut(&mut byte)).map_err(LhaError::Io)? { |
| 88 |
return Ok(None) |
| 89 |
} |
| 90 |
// self.rd.by_ref().bytes().next().transpose().map(|mb| |
| 91 |
// mb.map(|byte| { |
| 92 |
self.update_checksums_no_wrapping_sum(slice::from_ref(&byte)); |
| 93 |
// byte |
| 94 |
// }) |
| 95 |
// ) |
| 96 |
Ok(Some(byte)) |
| 97 |
} |
| 98 |
|
| 99 |
fn read_u8(&mut self) -> LhaResult<u8, R> { |
| 100 |
let mut byte: u8 = 0; |
| 101 |
self.read_exact(slice::from_mut(&mut byte))?; |
| 102 |
Ok(byte) |
| 103 |
} |
| 104 |
|
| 105 |
fn read_u16(&mut self) -> LhaResult<u16, R> { |
| 106 |
let mut buf = [0u8;2]; |
| 107 |
self.read_exact(&mut buf)?; |
| 108 |
Ok(u16::from_le_bytes(buf)) |
| 109 |
} |
| 110 |
|
| 111 |
fn read_u32(&mut self) -> LhaResult<u32, R> { |
| 112 |
let mut buf = [0u8;4]; |
| 113 |
self.read_exact(&mut buf)?; |
| 114 |
Ok(u32::from_le_bytes(buf)) |
| 115 |
} |
| 116 |
|
| 117 |
fn read_exact(&mut self, buf: &mut [u8]) -> LhaResult<(), R> { |
| 118 |
self.rd.read_exact(buf).map_err(LhaError::Io)?; |
| 119 |
self.update_checksums(buf); |
| 120 |
Ok(()) |
| 121 |
} |
| 122 |
|
| 123 |
fn read_limit(&mut self, limit: usize) -> LhaResult<Box<[u8]>, R> { |
| 124 |
let mut buf = Vec::with_capacity(limit); |
| 125 |
self.read_limit_no_checksums(limit, &mut buf)?; |
| 126 |
self.update_checksums(&buf); |
| 127 |
Ok(buf.into_boxed_slice()) |
| 128 |
} |
| 129 |
|
| 130 |
fn update_checksums(&mut self, buf: &[u8]) { |
| 131 |
self.update_checksums_no_wrapping_sum(buf); |
| 132 |
self.csum = wrapping_csum(self.csum, buf); |
| 133 |
} |
| 134 |
|
| 135 |
fn update_checksums_no_wrapping_sum(&mut self, buf: &[u8]) { |
| 136 |
self.len += buf.len(); |
| 137 |
self.crc.digest(buf); |
| 138 |
} |
| 139 |
|
| 140 |
fn read_limit_no_checksums(&mut self, limit: usize, buf: &mut Vec<u8>) -> LhaResult<(), R> { |
| 141 |
buf.try_reserve_exact(limit).map_err(|_| LhaError::HeaderParse("memory allocation failed"))?; |
| 142 |
// TODO: use BorrowedBuf once stabilized |
| 143 |
let spare = unsafe { core::mem::transmute::<_, &mut [u8]>(&mut buf.spare_capacity_mut()[..limit]) }; |
| 144 |
self.rd.read_exact(spare).map_err(LhaError::Io)?; |
| 145 |
unsafe { buf.set_len(buf.len() + limit); } |
| 146 |
// if self.rd.by_ref().take(limit as u64).read_to_end(buf)? != limit { |
| 147 |
// return Err(LhaError::HeaderParse("file is too short")) |
| 148 |
// } |
| 149 |
Ok(()) |
| 150 |
} |
| 151 |
} |
| 152 |
|
| 153 |
impl LhaHeader { |
| 154 |
/// Attempt to parse the LHA header. Return `Ok(Some(LhaHeader))` on success. Return `Ok(None)` |
| 155 |
/// if the end of archive marker (a `0` byte) was encountered. |
| 156 |
/// |
| 157 |
/// The method validates all length and checksum fields of the header, but does not parse extra |
| 158 |
/// headers except: |
| 159 |
/// * The ["Common"][EXT_HEADER_COMMON] header for validating the header's CRC-16 checksum. |
| 160 |
/// * The ["MS-DOS Attributes"][EXT_HEADER_MSDOS_ATTRS] header for reading MS-DOS attributes. |
| 161 |
/// * The ["MS-DOS Size"][EXT_HEADER_MSDOS_SIZE] header for reading 64-bit file size. |
| 162 |
/// |
| 163 |
/// All extra data is available as raw bytes and extra headers can be iterated with [LhaHeader::iter_extra]. |
| 164 |
/// |
| 165 |
/// Instance methods can be further called on the parsed `LhaHeader` struct to attempt to parse the |
| 166 |
/// name and path of the file or other file's meta-data. |
| 167 |
/// |
| 168 |
/// # Errors |
| 169 |
/// Returns an error from the underlying reading operations or because a malformed header was encountered. |
| 170 |
pub fn read<R: Read>(rd: &mut R) -> LhaResult<Option<LhaHeader>, R> { |
| 171 |
let mut parser = Parser { |
| 172 |
rd, |
| 173 |
crc: Crc16::default(), |
| 174 |
csum: Wrapping(0), |
| 175 |
len: 0 |
| 176 |
}; |
| 177 |
let header_len = match parser.read_u8_or_none()? { |
| 178 |
Some(0)|None => return Ok(None), |
| 179 |
Some(len) => len |
| 180 |
}; |
| 181 |
let csum = parser.read_u8()?; |
| 182 |
// reset wrapping checksum which should not include the first 2 bytes |
| 183 |
parser.csum = Wrapping(0); |
| 184 |
|
| 185 |
let mut raw_header = LhaRawBaseHeader::default(); |
| 186 |
parser.read_exact(unsafe { |
| 187 |
// safe because LhaRawBaseHeader is packed and contains only byte type members |
| 188 |
struct_slice_mut(&mut raw_header) |
| 189 |
})?; |
| 190 |
if raw_header.lha_level > 3 { |
| 191 |
return Err(LhaError::HeaderParse("unknown header level")) |
| 192 |
} |
| 193 |
|
| 194 |
// read filename if level 0 or 1 |
| 195 |
let filename = if raw_header.lha_level < 2 { |
| 196 |
let filename_len = parser.read_u8()? as usize; |
| 197 |
if (header_len as usize) < parser.len + filename_len { |
| 198 |
return Err(LhaError::HeaderParse("wrong header size")) |
| 199 |
} |
| 200 |
parser.read_limit(filename_len)? |
| 201 |
} |
| 202 |
else { |
| 203 |
Box::new([]) |
| 204 |
}; |
| 205 |
|
| 206 |
// file CRC-16 |
| 207 |
let file_crc = parser.read_u16()?; |
| 208 |
|
| 209 |
// OS-TYPE |
| 210 |
let mut os_type = 0; |
| 211 |
if raw_header.lha_level > 0 { |
| 212 |
os_type = parser.read_u8()?; |
| 213 |
} |
| 214 |
|
| 215 |
// extended area, only 0 and 1 level |
| 216 |
let mut extended_area: Box<[u8]> = Box::new([]); |
| 217 |
if raw_header.lha_level < 2 { |
| 218 |
let mut min_len = parser.len; |
| 219 |
if raw_header.lha_level == 0 { |
| 220 |
min_len -= 2; // no extra headers |
| 221 |
} |
| 222 |
if (header_len as usize) < min_len { |
| 223 |
return Err(LhaError::HeaderParse("wrong header size")) |
| 224 |
} |
| 225 |
let mut extended_len = (header_len as usize) - min_len; |
| 226 |
if extended_len != 0 && raw_header.lha_level == 0 { |
| 227 |
// get os_type from level 0 extended area |
| 228 |
extended_len -= 1; |
| 229 |
os_type = parser.read_u8()?; |
| 230 |
} |
| 231 |
if extended_len != 0 { |
| 232 |
extended_area = parser.read_limit(extended_len)?; |
| 233 |
} |
| 234 |
}; |
| 235 |
|
| 236 |
// extra headers |
| 237 |
let mut long_header_len: u32 = 0; // a long header length found in level >= 2 |
| 238 |
let mut first_header_len: u32 = 0; |
| 239 |
let mut extra_headers = Vec::new(); |
| 240 |
// establish the first extra header length and the long header length |
| 241 |
match raw_header.lha_level { |
| 242 |
1 => { |
| 243 |
first_header_len = parser.read_u16()? as u32; |
| 244 |
} |
| 245 |
2 => { |
| 246 |
long_header_len = u16::from_le_bytes([header_len, csum]) as u32; |
| 247 |
first_header_len = parser.read_u16()? as u32; |
| 248 |
} |
| 249 |
3 => { |
| 250 |
long_header_len = parser.read_u32()?; |
| 251 |
first_header_len = parser.read_u32()?; |
| 252 |
if header_len != 4 || csum != 0 { |
| 253 |
return Err(LhaError::HeaderParse("invalid header")) |
| 254 |
} |
| 255 |
} |
| 256 |
_ => {} |
| 257 |
} |
| 258 |
|
| 259 |
// validate level 0 and 1 header checksum |
| 260 |
if raw_header.lha_level < 2 { |
| 261 |
if csum != parser.csum.0 { |
| 262 |
return Err(LhaError::HeaderParse("invalid header level checksum")) |
| 263 |
} |
| 264 |
} |
| 265 |
else if long_header_len < parser.len as u32 + first_header_len { |
| 266 |
return Err(LhaError::HeaderParse("wrong header size")) |
| 267 |
} |
| 268 |
|
| 269 |
let mut msdos_attrs = MsDosAttrs::from_bits_retain(raw_header.msdos_attrs as u16); |
| 270 |
let mut original_size = u32::from_le_bytes(raw_header.original_size) as u64; |
| 271 |
let mut compressed_size = u32::from_le_bytes(raw_header.compressed_size) as u64; |
| 272 |
let mut header_crc: Option<u16> = None; |
| 273 |
// read extra headers |
| 274 |
let min_header_len = if raw_header.lha_level == 3 { 5 } else { 3 }; |
| 275 |
let mut extra_header_len = first_header_len as usize; |
| 276 |
while extra_header_len != 0 { |
| 277 |
if extra_header_len < min_header_len { |
| 278 |
return Err(LhaError::HeaderParse("wrong extra header size")) |
| 279 |
} |
| 280 |
// check long header length (level 2, 3) |
| 281 |
if long_header_len != 0 { |
| 282 |
if (long_header_len as usize) < parser.len + extra_header_len - 2 { |
| 283 |
return Err(LhaError::HeaderParse("wrong header size")) |
| 284 |
} |
| 285 |
} |
| 286 |
else if compressed_size < (extra_headers.len() + extra_header_len) as u64 { |
| 287 |
// otherwise check skip size (level 1) |
| 288 |
return Err(LhaError::HeaderParse("wrong header size")) |
| 289 |
} |
| 290 |
parser.read_limit_no_checksums(extra_header_len, &mut extra_headers)?; |
| 291 |
let start = extra_headers.len() - extra_header_len; |
| 292 |
let header = &mut extra_headers[start..]; |
| 293 |
match header { |
| 294 |
// we need to extract the CRC-16 from header and clear it in order to calculate checksum |
| 295 |
[EXT_HEADER_COMMON, data @ ..] => { |
| 296 |
if header_crc.is_some() { |
| 297 |
return Err(LhaError::HeaderParse("double common CRC-16 header")) |
| 298 |
} |
| 299 |
if let Some(crc) = data.get_mut(0..2) { |
| 300 |
header_crc = read_u16(crc); |
| 301 |
for p in crc.iter_mut() { |
| 302 |
*p = 0; |
| 303 |
} |
| 304 |
} |
| 305 |
} |
| 306 |
[EXT_HEADER_MSDOS_ATTRS, data @ ..]| |
| 307 |
[EXT_HEADER_EXT_ATTRS, data @ ..] if data.len() >= 2 => { |
| 308 |
if let Some(attrs) = read_u16(&data[0..2]) { |
| 309 |
msdos_attrs = MsDosAttrs::from_bits_retain(attrs); |
| 310 |
} |
| 311 |
} |
| 312 |
[EXT_HEADER_MSDOS_SIZE, data @ ..] if raw_header.lha_level >= 2 && data.len() >= 16 => { |
| 313 |
if let (Some(compr), Some(orig)) = (read_u64(&data[0..8]), read_u64(&data[8..16])) { |
| 314 |
compressed_size = compr; |
| 315 |
original_size = orig; |
| 316 |
} |
| 317 |
} |
| 318 |
_ => {} |
| 319 |
} |
| 320 |
parser.update_checksums_no_wrapping_sum(header); |
| 321 |
extra_header_len = if raw_header.lha_level == 3 { |
| 322 |
read_u32(&header[header.len() - 4..]).unwrap() as usize |
| 323 |
} |
| 324 |
else { |
| 325 |
read_u16(&header[header.len() - 2..]).unwrap() as usize |
| 326 |
} |
| 327 |
} |
| 328 |
|
| 329 |
// validate long header length |
| 330 |
if long_header_len != 0 && |
| 331 |
long_header_len != parser.len as u32 |
| 332 |
{ |
| 333 |
if raw_header.lha_level == 2 && long_header_len == parser.len as u32 + 1 |
| 334 |
{ |
| 335 |
// read padding byte |
| 336 |
parser.read_u8()?; |
| 337 |
} |
| 338 |
else if raw_header.lha_level == 2 && long_header_len + 2 != parser.len as u32 { |
| 339 |
// some packers (Osk) don't include self in the header length |
| 340 |
return Err(LhaError::HeaderParse("wrong length of headers")) |
| 341 |
} |
| 342 |
} |
| 343 |
|
| 344 |
// validate headers CRC |
| 345 |
if let Some(crc) = header_crc { |
| 346 |
if crc != parser.crc.sum16() { |
| 347 |
return Err(LhaError::HeaderParse("wrong header CRC-16 checksum")) |
| 348 |
} |
| 349 |
} |
| 350 |
|
| 351 |
// adjust compressed size for level 1 |
| 352 |
if raw_header.lha_level == 1 { |
| 353 |
if extra_headers.len() as u64 > compressed_size { |
| 354 |
return Err(LhaError::HeaderParse("wrong length of skip size")) |
| 355 |
} |
| 356 |
compressed_size -= extra_headers.len() as u64; |
| 357 |
} |
| 358 |
|
| 359 |
let compression = raw_header.compression; |
| 360 |
let last_modified = u32::from_le_bytes(raw_header.last_modified); |
| 361 |
let extra_headers = extra_headers.into_boxed_slice(); |
| 362 |
|
| 363 |
Ok(Some(LhaHeader { |
| 364 |
level: raw_header.lha_level, |
| 365 |
compression, |
| 366 |
compressed_size, |
| 367 |
original_size, |
| 368 |
filename, |
| 369 |
os_type, |
| 370 |
msdos_attrs, |
| 371 |
last_modified, |
| 372 |
file_crc, |
| 373 |
extended_area, |
| 374 |
first_header_len, |
| 375 |
extra_headers |
| 376 |
})) |
| 377 |
} |
| 378 |
|
| 379 |
/// Return an iterator that will iterate through extra headers, yielding the headers' raw |
| 380 |
/// data, excluding the next header length field. |
| 381 |
/// |
| 382 |
/// # Note |
| 383 |
/// Each iterated raw header will have at least the size of 1 byte containing the header identifier. |
| 384 |
pub fn iter_extra(&self) -> ExtraHeaderIter<'_> { |
| 385 |
ExtraHeaderIter { |
| 386 |
data: &self.extra_headers, |
| 387 |
header_length: self.first_header_len, |
| 388 |
header_len32: self.level == 3 |
| 389 |
} |
| 390 |
} |
| 391 |
} |
| 392 |
|
| 393 |
fn read_u16(slice: &[u8]) -> Option<u16> { |
| 394 |
match slice { |
| 395 |
&[lo, hi] => Some(u16::from_le_bytes([lo, hi])), |
| 396 |
_ => None |
| 397 |
} |
| 398 |
} |
| 399 |
|
| 400 |
pub(super) fn read_u32(slice: &[u8]) -> Option<u32> { |
| 401 |
match slice { |
| 402 |
&[b0, b1, b2, b3] => Some(u32::from_le_bytes([b0, b1, b2, b3])), |
| 403 |
_ => None |
| 404 |
} |
| 405 |
} |
| 406 |
|
| 407 |
pub(super) fn read_u64(slice: &[u8]) -> Option<u64> { |
| 408 |
match slice { |
| 409 |
&[b0, b1, b2, b3, b4, b5, b6, b7] => Some(u64::from_le_bytes([b0, b1, b2, b3, b4, b5, b6, b7])), |
| 410 |
_ => None |
| 411 |
} |
| 412 |
} |
| 413 |
|
| 414 |
fn wrapping_csum(init: Wrapping<u8>, data: &[u8]) -> Wrapping<u8> { |
| 415 |
let sum: Wrapping<u8> = data.iter().copied().map(Wrapping).sum(); |
| 416 |
sum + init |
| 417 |
} |
| 418 |
|
| 419 |
pub(super) fn split_data_at_nil_or_end(data: &[u8]) -> (&[u8], Option<&[u8]>) { |
| 420 |
match memchr::memchr(0, data) { |
| 421 |
Some(index) => (&data[0..index], Some(&data[index + 1..data.len()])), |
| 422 |
None => (data, None) |
| 423 |
} |
| 424 |
} |
| 425 |
|
| 426 |
#[cfg(feature = "std")] |
| 427 |
pub(super) fn parse_pathname(data: &[u8], path: &mut PathBuf) { |
| 428 |
path.reserve(data.len()); |
| 429 |
// split by all possible path separators |
| 430 |
for part in data.split(|&c| c == 0xFF || c == b'/' || c == b'\\') { |
| 431 |
match part { |
| 432 |
b"."|b".."|[] => {} // ignore malicious and empty paths |
| 433 |
name => path.push(parse_str_nilterm(name, false, false).as_ref()) |
| 434 |
} |
| 435 |
} |
| 436 |
} |
| 437 |
|
| 438 |
pub(super) fn parse_pathname_to_str(data: &[u8], path: &mut String) { |
| 439 |
path.reserve(data.len()); |
| 440 |
// split by all possible path separators |
| 441 |
for part in data.split(|&c| c == 0xFF || c == b'/' || c == b'\\') { |
| 442 |
match part { |
| 443 |
b"."|b".."|[] => {} // ignore malicious and empty paths |
| 444 |
name => { |
| 445 |
if !path.is_empty() { |
| 446 |
path.push('/'); |
| 447 |
} |
| 448 |
path.push_str(parse_str_nilterm(name, false, false).as_ref()) |
| 449 |
} |
| 450 |
} |
| 451 |
} |
| 452 |
} |
| 453 |
|
| 454 |
#[cfg(feature = "std")] |
| 455 |
#[inline(always)] |
| 456 |
fn is_separator(c: char) -> bool { |
| 457 |
std::path::is_separator(c) |
| 458 |
} |
| 459 |
|
| 460 |
#[cfg(not(feature = "std"))] |
| 461 |
fn is_separator(c: char) -> bool { |
| 462 |
c == '/' || c == '\\' |
| 463 |
} |
| 464 |
|
| 465 |
pub(super) fn parse_str_nilterm( |
| 466 |
data: &[u8], nilterm: bool, ignore_sep: bool |
| 467 |
) -> Cow<str> |
| 468 |
{ |
| 469 |
if let Some(index) = data.iter().position(|&c| |
| 470 |
!(0x20..0x7f).contains(&c) || |
| 471 |
(!ignore_sep && is_separator(c as char)) |
| 472 |
) |
| 473 |
{ |
| 474 |
let mut out = String::with_capacity(data.len()*3); |
| 475 |
let (head, rest) = data.split_at(index); |
| 476 |
out.push_str(unsafe { // safe because head was validated |
| 477 |
core::str::from_utf8_unchecked(head) |
| 478 |
}); |
| 479 |
for byte in rest.iter() { |
| 480 |
match byte { |
| 481 |
0 if nilterm => break, |
| 482 |
0x00..=0x1f| |
| 483 |
0x7f..=0xff => { |
| 484 |
write!(out, "%{:02x}", byte).unwrap(); |
| 485 |
} |
| 486 |
&ch => { |
| 487 |
let c = ch as char; |
| 488 |
if !ignore_sep && is_separator(c) { |
| 489 |
out.push('_'); |
| 490 |
} |
| 491 |
else { |
| 492 |
out.push(c); |
| 493 |
} |
| 494 |
} |
| 495 |
} |
| 496 |
} |
| 497 |
Cow::Owned(out) |
| 498 |
} |
| 499 |
else { |
| 500 |
unsafe { // safe because data was validated |
| 501 |
Cow::Borrowed(core::str::from_utf8_unchecked(data)) |
| 502 |
} |
| 503 |
} |
| 504 |
} |
| 505 |
|
| 506 |
/// # Safety |
| 507 |
/// This function can be used safely only with packed structs that solely consist of |
| 508 |
/// `u8` or array of `u8` primitives. |
| 509 |
unsafe fn struct_slice_mut<T: Copy>(obj: &mut T) -> &mut [u8] { |
| 510 |
let len = core::mem::size_of::<T>() / core::mem::size_of::<u8>(); |
| 511 |
core::slice::from_raw_parts_mut(obj as *mut T as *mut u8, len) |
| 512 |
} |
| 513 |
|
| 514 |
#[cfg(feature = "std")] |
| 515 |
#[cfg(test)] |
| 516 |
mod tests { |
| 517 |
use super::*; |
| 518 |
use std::path::MAIN_SEPARATOR; |
| 519 |
|
| 520 |
fn parse_filename(data: &[u8]) -> Cow<str> { |
| 521 |
parse_str_nilterm(data, false, false) |
| 522 |
} |
| 523 |
|
| 524 |
#[test] |
| 525 |
fn split_data_at_nil_or_end_works() { |
| 526 |
assert_eq!((&b"Foo"[..], None), split_data_at_nil_or_end(b"Foo")); |
| 527 |
assert_eq!((&b"Foo"[..], Some(&b"Bar"[..])), split_data_at_nil_or_end(b"Foo\x00Bar")); |
| 528 |
assert_eq!((&[][..], Some(&b"Bar"[..])), split_data_at_nil_or_end(b"\x00Bar")); |
| 529 |
} |
| 530 |
|
| 531 |
#[test] |
| 532 |
fn path_parser_works() { |
| 533 |
assert_eq!("", parse_filename(b"")); |
| 534 |
assert_eq!("Hello World!", parse_filename(b"Hello World!")); |
| 535 |
if std::path::is_separator('/') { |
| 536 |
assert_eq!("_Hello_World_", parse_filename(b"/Hello/World/")); |
| 537 |
} |
| 538 |
if std::path::is_separator('\\') { |
| 539 |
assert_eq!("_Hello_World_", parse_filename(br"\Hello\World\")); |
| 540 |
} |
| 541 |
assert_eq!("Hello%00World%7f", parse_filename(b"Hello\x00World\x7f")); |
| 542 |
assert_eq!("Hello%01World%ff", parse_filename(b"Hello\x01World\xff")); |
| 543 |
assert_eq!("Hello", parse_str_nilterm(b"Hello\x00World\xff", true, false)); |
| 544 |
if std::path::is_separator('/') { |
| 545 |
assert_eq!("He_llo", parse_str_nilterm(b"He/llo\x00World\xff", true, false)); |
| 546 |
assert_eq!("He/llo", parse_str_nilterm(b"He/llo\x00World\xff", true, true)); |
| 547 |
assert_eq!("He/llo%00World%ff", parse_str_nilterm(b"He/llo\x00World\xff", false, true)); |
| 548 |
assert_eq!("_Hello%1fWorld%80", parse_filename(b"/Hello\x1fWorld\x80")); |
| 549 |
} |
| 550 |
let mut path = PathBuf::new(); |
| 551 |
parse_pathname(b"", &mut path); |
| 552 |
assert!(path.is_relative()); |
| 553 |
assert_eq!("", path.to_str().unwrap()); |
| 554 |
parse_pathname(b"/", &mut path); |
| 555 |
assert!(path.is_relative()); |
| 556 |
assert_eq!("", path.to_str().unwrap()); |
| 557 |
parse_pathname(br"\", &mut path); |
| 558 |
assert!(path.is_relative()); |
| 559 |
assert_eq!("", path.to_str().unwrap()); |
| 560 |
parse_pathname(br".", &mut path); |
| 561 |
assert!(path.is_relative()); |
| 562 |
assert_eq!("", path.to_str().unwrap()); |
| 563 |
parse_pathname(br"..", &mut path); |
| 564 |
assert!(path.is_relative()); |
| 565 |
assert_eq!("", path.to_str().unwrap()); |
| 566 |
parse_pathname(br"./..", &mut path); |
| 567 |
assert!(path.is_relative()); |
| 568 |
assert_eq!("", path.to_str().unwrap()); |
| 569 |
parse_pathname(br".\..", &mut path); |
| 570 |
assert!(path.is_relative()); |
| 571 |
assert_eq!("", path.to_str().unwrap()); |
| 572 |
parse_pathname(br"/..\./", &mut path); |
| 573 |
assert!(path.is_relative()); |
| 574 |
assert_eq!("", path.to_str().unwrap()); |
| 575 |
parse_pathname(br"\../.\", &mut path); |
| 576 |
assert!(path.is_relative()); |
| 577 |
assert_eq!("", path.to_str().unwrap()); |
| 578 |
parse_pathname(br"foo/bar\baz", &mut path); |
| 579 |
assert!(path.is_relative()); |
| 580 |
let expect = format!("foo{}bar{}baz", MAIN_SEPARATOR, MAIN_SEPARATOR); |
| 581 |
assert_eq!(expect, path.to_str().unwrap()); |
| 582 |
path.clear(); |
| 583 |
parse_pathname(br"\foo/bar\baz/", &mut path); |
| 584 |
assert!(path.is_relative()); |
| 585 |
let expect = format!("foo{}bar{}baz", MAIN_SEPARATOR, MAIN_SEPARATOR); |
| 586 |
assert_eq!(expect, path.to_str().unwrap()); |
| 587 |
path.clear(); |
| 588 |
parse_pathname(br"/foo\bar/baz\", &mut path); |
| 589 |
assert!(path.is_relative()); |
| 590 |
let expect = format!("foo{}bar{}baz", MAIN_SEPARATOR, MAIN_SEPARATOR); |
| 591 |
assert_eq!(expect, path.to_str().unwrap()); |
| 592 |
path.clear(); |
| 593 |
parse_pathname(b"foo\xffbar\xffbaz", &mut path); |
| 594 |
assert!(path.is_relative()); |
| 595 |
let expect = format!("foo{}bar{}baz", MAIN_SEPARATOR, MAIN_SEPARATOR); |
| 596 |
assert_eq!(expect, path.to_str().unwrap()); |
| 597 |
path.clear(); |
| 598 |
parse_pathname(b"\xfffoo\xffb\x91ar\xffbaz\xff", &mut path); |
| 599 |
assert!(path.is_relative()); |
| 600 |
let expect = format!("foo{}b%91ar{}baz", MAIN_SEPARATOR, MAIN_SEPARATOR); |
| 601 |
assert_eq!(expect, path.to_str().unwrap()); |
| 602 |
path.clear(); |
| 603 |
} |
| 604 |
|
| 605 |
#[test] |
| 606 |
fn path_parser_to_str_works() { |
| 607 |
let mut path = String::new(); |
| 608 |
parse_pathname_to_str(b"", &mut path); |
| 609 |
assert!(!path.starts_with('/')); |
| 610 |
assert_eq!("", &path); |
| 611 |
parse_pathname_to_str(b"/", &mut path); |
| 612 |
assert!(!path.starts_with('/')); |
| 613 |
assert_eq!("", &path); |
| 614 |
parse_pathname_to_str(br"\", &mut path); |
| 615 |
assert!(!path.starts_with('/')); |
| 616 |
assert_eq!("", &path); |
| 617 |
parse_pathname_to_str(br".", &mut path); |
| 618 |
assert!(!path.starts_with('/')); |
| 619 |
assert_eq!("", &path); |
| 620 |
parse_pathname_to_str(br"..", &mut path); |
| 621 |
assert!(!path.starts_with('/')); |
| 622 |
assert_eq!("", &path); |
| 623 |
parse_pathname_to_str(br"./..", &mut path); |
| 624 |
assert!(!path.starts_with('/')); |
| 625 |
assert_eq!("", &path); |
| 626 |
parse_pathname_to_str(br".\..", &mut path); |
| 627 |
assert!(!path.starts_with('/')); |
| 628 |
assert_eq!("", &path); |
| 629 |
parse_pathname_to_str(br"/..\./", &mut path); |
| 630 |
assert!(!path.starts_with('/')); |
| 631 |
assert_eq!("", &path); |
| 632 |
parse_pathname_to_str(br"\../.\", &mut path); |
| 633 |
assert!(!path.starts_with('/')); |
| 634 |
assert_eq!("", &path); |
| 635 |
parse_pathname_to_str(br"foo/bar\baz", &mut path); |
| 636 |
assert!(!path.starts_with('/')); |
| 637 |
let expect = "foo/bar/baz"; |
| 638 |
assert_eq!(expect, &path); |
| 639 |
path.clear(); |
| 640 |
parse_pathname_to_str(br"\foo/bar\baz/", &mut path); |
| 641 |
assert!(!path.starts_with('/')); |
| 642 |
let expect = "foo/bar/baz"; |
| 643 |
assert_eq!(expect, &path); |
| 644 |
path.clear(); |
| 645 |
parse_pathname_to_str(br"/foo\bar/baz\", &mut path); |
| 646 |
assert!(!path.starts_with('/')); |
| 647 |
let expect = "foo/bar/baz"; |
| 648 |
assert_eq!(expect, &path); |
| 649 |
path.clear(); |
| 650 |
parse_pathname_to_str(b"foo\xffbar\xffbaz", &mut path); |
| 651 |
assert!(!path.starts_with('/')); |
| 652 |
let expect = "foo/bar/baz"; |
| 653 |
assert_eq!(expect, &path); |
| 654 |
path.clear(); |
| 655 |
parse_pathname_to_str(b"\xfffoo\xffb\x91ar\xffbaz\xff", &mut path); |
| 656 |
assert!(!path.starts_with('/')); |
| 657 |
let expect = "foo/b%91ar/baz"; |
| 658 |
assert_eq!(expect, &path); |
| 659 |
path.clear(); |
| 660 |
} |
| 661 |
} |