Matthew Maurer | 813ee27 | 2021-08-13 14:57:18 -0700 | [diff] [blame] | 1 | use core::char; |
| 2 | use core::fmt; |
| 3 | |
| 4 | /// Representation of a demangled symbol name. |
| 5 | pub struct Demangle<'a> { |
| 6 | inner: &'a str, |
| 7 | /// The number of ::-separated elements in the original name. |
| 8 | elements: usize, |
| 9 | } |
| 10 | |
| 11 | /// De-mangles a Rust symbol into a more readable version |
| 12 | /// |
| 13 | /// All Rust symbols by default are mangled as they contain characters that |
| 14 | /// cannot be represented in all object files. The mangling mechanism is similar |
| 15 | /// to C++'s, but Rust has a few specifics to handle items like lifetimes in |
| 16 | /// symbols. |
| 17 | /// |
| 18 | /// This function will take a **mangled** symbol and return a value. When printed, |
| 19 | /// the de-mangled version will be written. If the symbol does not look like |
| 20 | /// a mangled symbol, the original value will be written instead. |
| 21 | /// |
| 22 | /// # Examples |
| 23 | /// |
| 24 | /// ``` |
| 25 | /// use rustc_demangle::demangle; |
| 26 | /// |
| 27 | /// assert_eq!(demangle("_ZN4testE").to_string(), "test"); |
| 28 | /// assert_eq!(demangle("_ZN3foo3barE").to_string(), "foo::bar"); |
| 29 | /// assert_eq!(demangle("foo").to_string(), "foo"); |
| 30 | /// ``` |
| 31 | |
| 32 | // All Rust symbols are in theory lists of "::"-separated identifiers. Some |
| 33 | // assemblers, however, can't handle these characters in symbol names. To get |
| 34 | // around this, we use C++-style mangling. The mangling method is: |
| 35 | // |
| 36 | // 1. Prefix the symbol with "_ZN" |
| 37 | // 2. For each element of the path, emit the length plus the element |
| 38 | // 3. End the path with "E" |
| 39 | // |
| 40 | // For example, "_ZN4testE" => "test" and "_ZN3foo3barE" => "foo::bar". |
| 41 | // |
| 42 | // We're the ones printing our backtraces, so we can't rely on anything else to |
| 43 | // demangle our symbols. It's *much* nicer to look at demangled symbols, so |
| 44 | // this function is implemented to give us nice pretty output. |
| 45 | // |
| 46 | // Note that this demangler isn't quite as fancy as it could be. We have lots |
| 47 | // of other information in our symbols like hashes, version, type information, |
| 48 | // etc. Additionally, this doesn't handle glue symbols at all. |
| 49 | pub fn demangle(s: &str) -> Result<(Demangle, &str), ()> { |
| 50 | // First validate the symbol. If it doesn't look like anything we're |
| 51 | // expecting, we just print it literally. Note that we must handle non-Rust |
| 52 | // symbols because we could have any function in the backtrace. |
| 53 | let inner = if s.starts_with("_ZN") { |
| 54 | &s[3..] |
| 55 | } else if s.starts_with("ZN") { |
| 56 | // On Windows, dbghelp strips leading underscores, so we accept "ZN...E" |
| 57 | // form too. |
| 58 | &s[2..] |
| 59 | } else if s.starts_with("__ZN") { |
| 60 | // On OSX, symbols are prefixed with an extra _ |
| 61 | &s[4..] |
| 62 | } else { |
| 63 | return Err(()); |
| 64 | }; |
| 65 | |
| 66 | // only work with ascii text |
| 67 | if inner.bytes().any(|c| c & 0x80 != 0) { |
| 68 | return Err(()); |
| 69 | } |
| 70 | |
| 71 | let mut elements = 0; |
| 72 | let mut chars = inner.chars(); |
| 73 | let mut c = chars.next().ok_or(())?; |
| 74 | while c != 'E' { |
| 75 | // Decode an identifier element's length. |
| 76 | if !c.is_digit(10) { |
| 77 | return Err(()); |
| 78 | } |
| 79 | let mut len = 0usize; |
| 80 | while let Some(d) = c.to_digit(10) { |
| 81 | len = len |
| 82 | .checked_mul(10) |
| 83 | .and_then(|len| len.checked_add(d as usize)) |
| 84 | .ok_or(())?; |
| 85 | c = chars.next().ok_or(())?; |
| 86 | } |
| 87 | |
| 88 | // `c` already contains the first character of this identifier, skip it and |
| 89 | // all the other characters of this identifier, to reach the next element. |
| 90 | for _ in 0..len { |
| 91 | c = chars.next().ok_or(())?; |
| 92 | } |
| 93 | |
| 94 | elements += 1; |
| 95 | } |
| 96 | |
| 97 | Ok((Demangle { inner, elements }, chars.as_str())) |
| 98 | } |
| 99 | |
| 100 | // Rust hashes are hex digits with an `h` prepended. |
| 101 | fn is_rust_hash(s: &str) -> bool { |
| 102 | s.starts_with('h') && s[1..].chars().all(|c| c.is_digit(16)) |
| 103 | } |
| 104 | |
| 105 | impl<'a> fmt::Display for Demangle<'a> { |
| 106 | fn fmt(&self, f: &mut fmt::Formatter) -> fmt::Result { |
| 107 | // Alright, let's do this. |
| 108 | let mut inner = self.inner; |
| 109 | for element in 0..self.elements { |
| 110 | let mut rest = inner; |
| 111 | while rest.chars().next().unwrap().is_digit(10) { |
| 112 | rest = &rest[1..]; |
| 113 | } |
| 114 | let i: usize = inner[..(inner.len() - rest.len())].parse().unwrap(); |
| 115 | inner = &rest[i..]; |
| 116 | rest = &rest[..i]; |
| 117 | // Skip printing the hash if alternate formatting |
| 118 | // was requested. |
| 119 | if f.alternate() && element + 1 == self.elements && is_rust_hash(&rest) { |
| 120 | break; |
| 121 | } |
| 122 | if element != 0 { |
| 123 | f.write_str("::")?; |
| 124 | } |
| 125 | if rest.starts_with("_$") { |
| 126 | rest = &rest[1..]; |
| 127 | } |
| 128 | loop { |
| 129 | if rest.starts_with('.') { |
| 130 | if let Some('.') = rest[1..].chars().next() { |
| 131 | f.write_str("::")?; |
| 132 | rest = &rest[2..]; |
| 133 | } else { |
| 134 | f.write_str(".")?; |
| 135 | rest = &rest[1..]; |
| 136 | } |
| 137 | } else if rest.starts_with('$') { |
| 138 | let (escape, after_escape) = if let Some(end) = rest[1..].find('$') { |
| 139 | (&rest[1..=end], &rest[end + 2..]) |
| 140 | } else { |
| 141 | break; |
| 142 | }; |
| 143 | |
| 144 | // see src/librustc_codegen_utils/symbol_names/legacy.rs for these mappings |
| 145 | let unescaped = match escape { |
| 146 | "SP" => "@", |
| 147 | "BP" => "*", |
| 148 | "RF" => "&", |
| 149 | "LT" => "<", |
| 150 | "GT" => ">", |
| 151 | "LP" => "(", |
| 152 | "RP" => ")", |
| 153 | "C" => ",", |
| 154 | |
| 155 | _ => { |
| 156 | if escape.starts_with('u') { |
| 157 | let digits = &escape[1..]; |
| 158 | let all_lower_hex = digits.chars().all(|c| match c { |
| 159 | '0'..='9' | 'a'..='f' => true, |
| 160 | _ => false, |
| 161 | }); |
| 162 | let c = u32::from_str_radix(digits, 16) |
| 163 | .ok() |
| 164 | .and_then(char::from_u32); |
| 165 | if let (true, Some(c)) = (all_lower_hex, c) { |
| 166 | // FIXME(eddyb) do we need to filter out control codepoints? |
| 167 | if !c.is_control() { |
| 168 | c.fmt(f)?; |
| 169 | rest = after_escape; |
| 170 | continue; |
| 171 | } |
| 172 | } |
| 173 | } |
| 174 | break; |
| 175 | } |
| 176 | }; |
| 177 | f.write_str(unescaped)?; |
| 178 | rest = after_escape; |
| 179 | } else if let Some(i) = rest.find(|c| c == '$' || c == '.') { |
| 180 | f.write_str(&rest[..i])?; |
| 181 | rest = &rest[i..]; |
| 182 | } else { |
| 183 | break; |
| 184 | } |
| 185 | } |
| 186 | f.write_str(rest)?; |
| 187 | } |
| 188 | |
| 189 | Ok(()) |
| 190 | } |
| 191 | } |
| 192 | |
| 193 | #[cfg(test)] |
| 194 | mod tests { |
| 195 | use std::prelude::v1::*; |
| 196 | |
| 197 | macro_rules! t { |
| 198 | ($a:expr, $b:expr) => { |
| 199 | assert!(ok($a, $b)) |
| 200 | }; |
| 201 | } |
| 202 | |
| 203 | macro_rules! t_err { |
| 204 | ($a:expr) => { |
| 205 | assert!(ok_err($a)) |
| 206 | }; |
| 207 | } |
| 208 | |
| 209 | macro_rules! t_nohash { |
| 210 | ($a:expr, $b:expr) => {{ |
| 211 | assert_eq!(format!("{:#}", ::demangle($a)), $b); |
| 212 | }}; |
| 213 | } |
| 214 | |
| 215 | fn ok(sym: &str, expected: &str) -> bool { |
| 216 | match ::try_demangle(sym) { |
| 217 | Ok(s) => { |
| 218 | if s.to_string() == expected { |
| 219 | true |
| 220 | } else { |
| 221 | println!("\n{}\n!=\n{}\n", s, expected); |
| 222 | false |
| 223 | } |
| 224 | } |
| 225 | Err(_) => { |
| 226 | println!("error demangling"); |
| 227 | false |
| 228 | } |
| 229 | } |
| 230 | } |
| 231 | |
| 232 | fn ok_err(sym: &str) -> bool { |
| 233 | match ::try_demangle(sym) { |
| 234 | Ok(_) => { |
| 235 | println!("succeeded in demangling"); |
| 236 | false |
| 237 | } |
| 238 | Err(_) => ::demangle(sym).to_string() == sym, |
| 239 | } |
| 240 | } |
| 241 | |
| 242 | #[test] |
| 243 | fn demangle() { |
| 244 | t_err!("test"); |
| 245 | t!("_ZN4testE", "test"); |
| 246 | t_err!("_ZN4test"); |
| 247 | t!("_ZN4test1a2bcE", "test::a::bc"); |
| 248 | } |
| 249 | |
| 250 | #[test] |
| 251 | fn demangle_dollars() { |
| 252 | t!("_ZN4$RP$E", ")"); |
| 253 | t!("_ZN8$RF$testE", "&test"); |
| 254 | t!("_ZN8$BP$test4foobE", "*test::foob"); |
| 255 | t!("_ZN9$u20$test4foobE", " test::foob"); |
| 256 | t!("_ZN35Bar$LT$$u5b$u32$u3b$$u20$4$u5d$$GT$E", "Bar<[u32; 4]>"); |
| 257 | } |
| 258 | |
| 259 | #[test] |
| 260 | fn demangle_many_dollars() { |
| 261 | t!("_ZN13test$u20$test4foobE", "test test::foob"); |
| 262 | t!("_ZN12test$BP$test4foobE", "test*test::foob"); |
| 263 | } |
| 264 | |
| 265 | #[test] |
| 266 | fn demangle_osx() { |
| 267 | t!( |
| 268 | "__ZN5alloc9allocator6Layout9for_value17h02a996811f781011E", |
| 269 | "alloc::allocator::Layout::for_value::h02a996811f781011" |
| 270 | ); |
| 271 | t!("__ZN38_$LT$core..option..Option$LT$T$GT$$GT$6unwrap18_MSG_FILE_LINE_COL17haf7cb8d5824ee659E", "<core::option::Option<T>>::unwrap::_MSG_FILE_LINE_COL::haf7cb8d5824ee659"); |
| 272 | t!("__ZN4core5slice89_$LT$impl$u20$core..iter..traits..IntoIterator$u20$for$u20$$RF$$u27$a$u20$$u5b$T$u5d$$GT$9into_iter17h450e234d27262170E", "core::slice::<impl core::iter::traits::IntoIterator for &'a [T]>::into_iter::h450e234d27262170"); |
| 273 | } |
| 274 | |
| 275 | #[test] |
| 276 | fn demangle_windows() { |
| 277 | t!("ZN4testE", "test"); |
| 278 | t!("ZN13test$u20$test4foobE", "test test::foob"); |
| 279 | t!("ZN12test$RF$test4foobE", "test&test::foob"); |
| 280 | } |
| 281 | |
| 282 | #[test] |
| 283 | fn demangle_elements_beginning_with_underscore() { |
| 284 | t!("_ZN13_$LT$test$GT$E", "<test>"); |
| 285 | t!("_ZN28_$u7b$$u7b$closure$u7d$$u7d$E", "{{closure}}"); |
| 286 | t!("_ZN15__STATIC_FMTSTRE", "__STATIC_FMTSTR"); |
| 287 | } |
| 288 | |
| 289 | #[test] |
| 290 | fn demangle_trait_impls() { |
| 291 | t!( |
| 292 | "_ZN71_$LT$Test$u20$$u2b$$u20$$u27$static$u20$as$u20$foo..Bar$LT$Test$GT$$GT$3barE", |
| 293 | "<Test + 'static as foo::Bar<Test>>::bar" |
| 294 | ); |
| 295 | } |
| 296 | |
| 297 | #[test] |
| 298 | fn demangle_without_hash() { |
| 299 | let s = "_ZN3foo17h05af221e174051e9E"; |
| 300 | t!(s, "foo::h05af221e174051e9"); |
| 301 | t_nohash!(s, "foo"); |
| 302 | } |
| 303 | |
| 304 | #[test] |
| 305 | fn demangle_without_hash_edgecases() { |
| 306 | // One element, no hash. |
| 307 | t_nohash!("_ZN3fooE", "foo"); |
| 308 | // Two elements, no hash. |
| 309 | t_nohash!("_ZN3foo3barE", "foo::bar"); |
| 310 | // Longer-than-normal hash. |
| 311 | t_nohash!("_ZN3foo20h05af221e174051e9abcE", "foo"); |
| 312 | // Shorter-than-normal hash. |
| 313 | t_nohash!("_ZN3foo5h05afE", "foo"); |
| 314 | // Valid hash, but not at the end. |
| 315 | t_nohash!("_ZN17h05af221e174051e93fooE", "h05af221e174051e9::foo"); |
| 316 | // Not a valid hash, missing the 'h'. |
| 317 | t_nohash!("_ZN3foo16ffaf221e174051e9E", "foo::ffaf221e174051e9"); |
| 318 | // Not a valid hash, has a non-hex-digit. |
| 319 | t_nohash!("_ZN3foo17hg5af221e174051e9E", "foo::hg5af221e174051e9"); |
| 320 | } |
| 321 | |
| 322 | #[test] |
| 323 | fn demangle_thinlto() { |
| 324 | // One element, no hash. |
| 325 | t!("_ZN3fooE.llvm.9D1C9369", "foo"); |
| 326 | t!("_ZN3fooE.llvm.9D1C9369@@16", "foo"); |
| 327 | t_nohash!( |
| 328 | "_ZN9backtrace3foo17hbb467fcdaea5d79bE.llvm.A5310EB9", |
| 329 | "backtrace::foo" |
| 330 | ); |
| 331 | } |
| 332 | |
| 333 | #[test] |
| 334 | fn demangle_llvm_ir_branch_labels() { |
| 335 | t!("_ZN4core5slice77_$LT$impl$u20$core..ops..index..IndexMut$LT$I$GT$$u20$for$u20$$u5b$T$u5d$$GT$9index_mut17haf9727c2edfbc47bE.exit.i.i", "core::slice::<impl core::ops::index::IndexMut<I> for [T]>::index_mut::haf9727c2edfbc47b.exit.i.i"); |
| 336 | t_nohash!("_ZN4core5slice77_$LT$impl$u20$core..ops..index..IndexMut$LT$I$GT$$u20$for$u20$$u5b$T$u5d$$GT$9index_mut17haf9727c2edfbc47bE.exit.i.i", "core::slice::<impl core::ops::index::IndexMut<I> for [T]>::index_mut.exit.i.i"); |
| 337 | } |
| 338 | |
| 339 | #[test] |
| 340 | fn demangle_ignores_suffix_that_doesnt_look_like_a_symbol() { |
| 341 | t_err!("_ZN3fooE.llvm moocow"); |
| 342 | } |
| 343 | |
| 344 | #[test] |
| 345 | fn dont_panic() { |
| 346 | ::demangle("_ZN2222222222222222222222EE").to_string(); |
| 347 | ::demangle("_ZN5*70527e27.ll34csaғE").to_string(); |
| 348 | ::demangle("_ZN5*70527a54.ll34_$b.1E").to_string(); |
| 349 | ::demangle( |
| 350 | "\ |
| 351 | _ZN5~saäb4e\n\ |
| 352 | 2734cOsbE\n\ |
| 353 | 5usage20h)3\0\0\0\0\0\0\07e2734cOsbE\ |
| 354 | ", |
| 355 | ) |
| 356 | .to_string(); |
| 357 | } |
| 358 | |
| 359 | #[test] |
| 360 | fn invalid_no_chop() { |
| 361 | t_err!("_ZNfooE"); |
| 362 | } |
| 363 | |
| 364 | #[test] |
| 365 | fn handle_assoc_types() { |
| 366 | t!("_ZN151_$LT$alloc..boxed..Box$LT$alloc..boxed..FnBox$LT$A$C$$u20$Output$u3d$R$GT$$u20$$u2b$$u20$$u27$a$GT$$u20$as$u20$core..ops..function..FnOnce$LT$A$GT$$GT$9call_once17h69e8f44b3723e1caE", "<alloc::boxed::Box<alloc::boxed::FnBox<A, Output=R> + 'a> as core::ops::function::FnOnce<A>>::call_once::h69e8f44b3723e1ca"); |
| 367 | } |
| 368 | |
| 369 | #[test] |
| 370 | fn handle_bang() { |
| 371 | t!( |
| 372 | "_ZN88_$LT$core..result..Result$LT$$u21$$C$$u20$E$GT$$u20$as$u20$std..process..Termination$GT$6report17hfc41d0da4a40b3e8E", |
| 373 | "<core::result::Result<!, E> as std::process::Termination>::report::hfc41d0da4a40b3e8" |
| 374 | ); |
| 375 | } |
| 376 | |
| 377 | #[test] |
| 378 | fn demangle_utf8_idents() { |
| 379 | t_nohash!( |
| 380 | "_ZN11utf8_idents157_$u10e1$$u10d0$$u10ed$$u10db$$u10d4$$u10da$$u10d0$$u10d3$_$u10d2$$u10d4$$u10db$$u10e0$$u10d8$$u10d4$$u10da$$u10d8$_$u10e1$$u10d0$$u10d3$$u10d8$$u10da$$u10d8$17h21634fd5714000aaE", |
| 381 | "utf8_idents::საჭმელად_გემრიელი_სადილი" |
| 382 | ); |
| 383 | } |
| 384 | |
| 385 | #[test] |
| 386 | fn demangle_issue_60925() { |
| 387 | t_nohash!( |
| 388 | "_ZN11issue_609253foo37Foo$LT$issue_60925..llv$u6d$..Foo$GT$3foo17h059a991a004536adE", |
| 389 | "issue_60925::foo::Foo<issue_60925::llvm::Foo>::foo" |
| 390 | ); |
| 391 | } |
| 392 | } |