From c6643b50ea7290daf0a9f0e33d1f27c6f52796ce Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?Le=C3=B3n=20Orell=20Valerian=20Liehr?= Date: Tue, 4 Jul 2023 19:40:48 +0200 Subject: [PATCH 1/5] Revert the lexing of c_str_literals --- compiler/rustc_lexer/src/lib.rs | 7 ---- .../rfcs/rfc-3348-c-string-literals/basic.rs | 4 +- .../rfc-3348-c-string-literals/basic.stderr | 25 ++++++++++++ .../rfcs/rfc-3348-c-string-literals/gate.rs | 6 ++- .../rfc-3348-c-string-literals/gate.stderr | 35 ++++++++++------ .../rfc-3348-c-string-literals/no-nuls.rs | Bin 565 -> 760 bytes .../rfc-3348-c-string-literals/no-nuls.stderr | Bin 674 -> 4477 bytes .../rfc-3348-c-string-literals/non-ascii.rs | 4 +- .../non-ascii.stderr | 38 ++++++++++++++++++ 9 files changed, 96 insertions(+), 23 deletions(-) create mode 100644 tests/ui/rfcs/rfc-3348-c-string-literals/basic.stderr create mode 100644 tests/ui/rfcs/rfc-3348-c-string-literals/non-ascii.stderr diff --git a/compiler/rustc_lexer/src/lib.rs b/compiler/rustc_lexer/src/lib.rs index d511d2b1280..29335a8c0f4 100644 --- a/compiler/rustc_lexer/src/lib.rs +++ b/compiler/rustc_lexer/src/lib.rs @@ -367,13 +367,6 @@ pub fn advance_token(&mut self) -> Token { Some(|terminated| Byte { terminated }), ), - // c-string literal, raw c-string literal or identifier. - 'c' => self.c_or_byte_string( - |terminated| CStr { terminated }, - |n_hashes| RawCStr { n_hashes }, - None, - ), - // Identifier (this should be checked after other variant that can // start as identifier). c if is_id_start(c) => self.ident_or_unknown_prefix(), diff --git a/tests/ui/rfcs/rfc-3348-c-string-literals/basic.rs b/tests/ui/rfcs/rfc-3348-c-string-literals/basic.rs index e4b07ab8108..3fc5fd481ea 100644 --- a/tests/ui/rfcs/rfc-3348-c-string-literals/basic.rs +++ b/tests/ui/rfcs/rfc-3348-c-string-literals/basic.rs @@ -1,4 +1,6 @@ -// run-pass +// FIXME(c_str_literals): This should be `run-pass` +// known-bug: #113333 +// edition: 2021 #![feature(c_str_literals)] diff --git a/tests/ui/rfcs/rfc-3348-c-string-literals/basic.stderr b/tests/ui/rfcs/rfc-3348-c-string-literals/basic.stderr new file mode 100644 index 00000000000..571c319d8c5 --- /dev/null +++ b/tests/ui/rfcs/rfc-3348-c-string-literals/basic.stderr @@ -0,0 +1,25 @@ +error: prefix `c` is unknown + --> $DIR/basic.rs:8:27 + | +LL | assert_eq!(b"test\0", c"test".to_bytes_with_nul()); + | ^ unknown prefix + | + = note: prefixed identifiers and literals are reserved since Rust 2021 +help: consider inserting whitespace here + | +LL | assert_eq!(b"test\0", c "test".to_bytes_with_nul()); + | + + +error: no rules expected the token `"test"` + --> $DIR/basic.rs:8:28 + | +LL | assert_eq!(b"test\0", c"test".to_bytes_with_nul()); + | -^^^^^ + | | + | no rules expected this token in macro call + | help: missing comma here + | + = note: while trying to match sequence start + +error: aborting due to 2 previous errors + diff --git a/tests/ui/rfcs/rfc-3348-c-string-literals/gate.rs b/tests/ui/rfcs/rfc-3348-c-string-literals/gate.rs index b27da26ed23..ddd6d9a25da 100644 --- a/tests/ui/rfcs/rfc-3348-c-string-literals/gate.rs +++ b/tests/ui/rfcs/rfc-3348-c-string-literals/gate.rs @@ -1,4 +1,6 @@ // gate-test-c_str_literals +// known-bug: #113333 +// edition: 2021 macro_rules! m { ($t:tt) => {} @@ -6,8 +8,8 @@ macro_rules! m { fn main() { c"foo"; - //~^ ERROR: `c".."` literals are experimental + // FIXME(c_str_literals): This should be ``c".."` literals are experimental` m!(c"test"); - //~^ ERROR: `c".."` literals are experimental + // FIXME(c_str_literals): This should be ``c".."` literals are experimental` } diff --git a/tests/ui/rfcs/rfc-3348-c-string-literals/gate.stderr b/tests/ui/rfcs/rfc-3348-c-string-literals/gate.stderr index bc0c537aada..8de36ca4a6e 100644 --- a/tests/ui/rfcs/rfc-3348-c-string-literals/gate.stderr +++ b/tests/ui/rfcs/rfc-3348-c-string-literals/gate.stderr @@ -1,21 +1,32 @@ -error[E0658]: `c".."` literals are experimental - --> $DIR/gate.rs:8:5 +error: prefix `c` is unknown + --> $DIR/gate.rs:10:5 | LL | c"foo"; - | ^^^^^^ + | ^ unknown prefix | - = note: see issue #105723 for more information - = help: add `#![feature(c_str_literals)]` to the crate attributes to enable + = note: prefixed identifiers and literals are reserved since Rust 2021 +help: consider inserting whitespace here + | +LL | c "foo"; + | + -error[E0658]: `c".."` literals are experimental - --> $DIR/gate.rs:11:8 +error: prefix `c` is unknown + --> $DIR/gate.rs:13:8 | LL | m!(c"test"); - | ^^^^^^^ + | ^ unknown prefix | - = note: see issue #105723 for more information - = help: add `#![feature(c_str_literals)]` to the crate attributes to enable + = note: prefixed identifiers and literals are reserved since Rust 2021 +help: consider inserting whitespace here + | +LL | m!(c "test"); + | + -error: aborting due to 2 previous errors +error: expected one of `!`, `.`, `::`, `;`, `?`, `{`, `}`, or an operator, found `"foo"` + --> $DIR/gate.rs:10:6 + | +LL | c"foo"; + | ^^^^^ expected one of 8 possible tokens + +error: aborting due to 3 previous errors -For more information about this error, try `rustc --explain E0658`. diff --git a/tests/ui/rfcs/rfc-3348-c-string-literals/no-nuls.rs b/tests/ui/rfcs/rfc-3348-c-string-literals/no-nuls.rs index 7bc6097f124aabfded4a9e9791611cfe4fbf9f78..96945f125da71a9097f6f26c75a328f097448568 100644 GIT binary patch literal 760 zcmcJNO-sZu5Qg`hUoma>kP6avy>(F#*#$vQdJuM{o6V+eFij;NsFeP9W4gLhJlZ*h zcamo&GjEc>i)`QI`QBDJL|Jy>0woEs12?>tIV{pehENo2N=A%T4AVjzRS0gFB0|T{ z2$d2V!X=KOL*Rmj+caGIqa&?vpYB)VS2;Y^T!XG#D-N({u#;8@pfyp1IM$&HE6Vz|75 zu3sW|Kibf;!t3k1hm|vd^MGO6{N({>hu?dnKfSxeX+CuLli#$j9$@dIcremQ8dzg_ P7LFCq9*JKyoYbRF@$lQ` delta 167 zcmeytx|L<3h80&`oPujmkbjVs!enbkX+SIX)z%kSR(@bGAg5XJ85yzvbed2f%Rn4Ih5|y!^t&kXMB|mMR^m zC*}$sgNqA1njIh?T!n-4`ec1PSwG?VxvSyk(pPYyY5h8yfz60(HfJOirhJBiX_C{z z97uvBMe@#!QjF(>%Svg)VH^buD#{eO;L6-D!MGWx8Nb2(!rZUQsTsW>Nv=#dI`KIO z=-&ZEp^rfI5Qa}Zp|GK~Es4ZK;-DW88Zz6?f8?3})iUf$HaJRs6yzR`5S?vekb?%` zNmo5VqA3(U2zxMyJq#SX=_-O(vc=@TEk6GtlV~#^nqOIWKZQotw#oT$jeCafw@y%Qo3Q zPSCPEM;jBMN(!DFx@K1cWyzSFfW04Mhd1Y+KEI4HjfNE;5i#rWCql1B(8Y@nzCYRR fyP*I|{vI!G(MUg&)j`5rY|)T>DjOU+cR0qrx?#zt diff --git a/tests/ui/rfcs/rfc-3348-c-string-literals/non-ascii.rs b/tests/ui/rfcs/rfc-3348-c-string-literals/non-ascii.rs index 82e8e2090d7..066505c23df 100644 --- a/tests/ui/rfcs/rfc-3348-c-string-literals/non-ascii.rs +++ b/tests/ui/rfcs/rfc-3348-c-string-literals/non-ascii.rs @@ -1,4 +1,6 @@ -// run-pass +// FIXME(c_str_literals): This should be `run-pass` +// known-bug: #113333 +// edition: 2021 #![feature(c_str_literals)] diff --git a/tests/ui/rfcs/rfc-3348-c-string-literals/non-ascii.stderr b/tests/ui/rfcs/rfc-3348-c-string-literals/non-ascii.stderr new file mode 100644 index 00000000000..47361fb61d2 --- /dev/null +++ b/tests/ui/rfcs/rfc-3348-c-string-literals/non-ascii.stderr @@ -0,0 +1,38 @@ +error: prefix `c` is unknown + --> $DIR/non-ascii.rs:9:9 + | +LL | c"\xEF\x80🦀\u{1F980}".to_bytes_with_nul(), + | ^ unknown prefix + | + = note: prefixed identifiers and literals are reserved since Rust 2021 +help: consider inserting whitespace here + | +LL | c "\xEF\x80🦀\u{1F980}".to_bytes_with_nul(), + | + + +error: out of range hex escape + --> $DIR/non-ascii.rs:9:11 + | +LL | c"\xEF\x80🦀\u{1F980}".to_bytes_with_nul(), + | ^^^^ must be a character in the range [\x00-\x7f] + +error: out of range hex escape + --> $DIR/non-ascii.rs:9:15 + | +LL | c"\xEF\x80🦀\u{1F980}".to_bytes_with_nul(), + | ^^^^ must be a character in the range [\x00-\x7f] + +error: no rules expected the token `"\xEF\x80🦀\u{1F980}"` + --> $DIR/non-ascii.rs:9:10 + | +LL | c"\xEF\x80🦀\u{1F980}".to_bytes_with_nul(), + | -^^^^^^^^^^^^^^^^^^^^ + | | + | no rules expected this token in macro call + | help: missing comma here + | +note: while trying to match `,` + --> $SRC_DIR/core/src/macros/mod.rs:LL:COL + +error: aborting due to 4 previous errors + From 5b25f9d8bdb4a5ffdaed4fb6ebdc521c6eba1b0f Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?Le=C3=B3n=20Orell=20Valerian=20Liehr?= Date: Tue, 4 Jul 2023 19:45:55 +0200 Subject: [PATCH 2/5] Revert "fix ptr cast" This reverts commit 2f459f7f140307b5abbb7ea81440ed1843b490e7. --- library/std/src/sys/unix/args.rs | 14 ++++++-------- 1 file changed, 6 insertions(+), 8 deletions(-) diff --git a/library/std/src/sys/unix/args.rs b/library/std/src/sys/unix/args.rs index f8fa81e6ef1..7b6f8810cba 100644 --- a/library/std/src/sys/unix/args.rs +++ b/library/std/src/sys/unix/args.rs @@ -242,15 +242,13 @@ pub fn args() -> Args { let mut res = Vec::new(); unsafe { - let process_info_sel = - sel_registerName(c"processInfo".as_ptr() as *const libc::c_uchar); - let arguments_sel = sel_registerName(c"arguments".as_ptr() as *const libc::c_uchar); - let utf8_sel = sel_registerName(c"UTF8String".as_ptr() as *const libc::c_uchar); - let count_sel = sel_registerName(c"count".as_ptr() as *const libc::c_uchar); - let object_at_sel = - sel_registerName(c"objectAtIndex:".as_ptr() as *const libc::c_uchar); + let process_info_sel = sel_registerName(c"processInfo".as_ptr()); + let arguments_sel = sel_registerName(c"arguments".as_ptr()); + let utf8_sel = sel_registerName(c"UTF8String".as_ptr()); + let count_sel = sel_registerName(c"count".as_ptr()); + let object_at_sel = sel_registerName(c"objectAtIndex:".as_ptr()); - let klass = objc_getClass(c"NSProcessInfo".as_ptr() as *const libc::c_uchar); + let klass = objc_getClass(c"NSProcessInfo".as_ptr()); let info = objc_msgSend(klass, process_info_sel); let args = objc_msgSend(info, arguments_sel); From 9dbe67fc8c722bc8df4a6a792677a93a073773f4 Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?Le=C3=B3n=20Orell=20Valerian=20Liehr?= Date: Tue, 4 Jul 2023 19:45:58 +0200 Subject: [PATCH 3/5] Revert "use c literals in library" This reverts commit f212ba6d6d60963c8101bb24fc3e53fca80c046f. --- library/std/src/lib.rs | 1 - library/std/src/sys/unix/args.rs | 12 ++++++------ library/std/src/sys/unix/fs.rs | 2 +- library/std/src/sys/unix/mod.rs | 7 ++++--- library/std/src/sys/unix/process/process_common.rs | 9 +++++---- library/std/src/sys/unix/thread.rs | 3 ++- library/std/src/sys/windows/c.rs | 4 ++-- library/std/src/sys/windows/compat.rs | 6 +++--- library/std/src/sys/windows/mod.rs | 4 ++-- src/tools/tidy/src/deps.rs | 1 + 10 files changed, 26 insertions(+), 23 deletions(-) diff --git a/library/std/src/lib.rs b/library/std/src/lib.rs index da08c018d0e..72b9ad3480b 100644 --- a/library/std/src/lib.rs +++ b/library/std/src/lib.rs @@ -241,7 +241,6 @@ #![feature(allocator_internals)] #![feature(allow_internal_unsafe)] #![feature(allow_internal_unstable)] -#![feature(c_str_literals)] #![feature(c_unwind)] #![feature(cfg_target_thread_local)] #![feature(concat_idents)] diff --git a/library/std/src/sys/unix/args.rs b/library/std/src/sys/unix/args.rs index 7b6f8810cba..eafd6821f54 100644 --- a/library/std/src/sys/unix/args.rs +++ b/library/std/src/sys/unix/args.rs @@ -242,13 +242,13 @@ pub fn args() -> Args { let mut res = Vec::new(); unsafe { - let process_info_sel = sel_registerName(c"processInfo".as_ptr()); - let arguments_sel = sel_registerName(c"arguments".as_ptr()); - let utf8_sel = sel_registerName(c"UTF8String".as_ptr()); - let count_sel = sel_registerName(c"count".as_ptr()); - let object_at_sel = sel_registerName(c"objectAtIndex:".as_ptr()); + let process_info_sel = sel_registerName("processInfo\0".as_ptr()); + let arguments_sel = sel_registerName("arguments\0".as_ptr()); + let utf8_sel = sel_registerName("UTF8String\0".as_ptr()); + let count_sel = sel_registerName("count\0".as_ptr()); + let object_at_sel = sel_registerName("objectAtIndex:\0".as_ptr()); - let klass = objc_getClass(c"NSProcessInfo".as_ptr()); + let klass = objc_getClass("NSProcessInfo\0".as_ptr()); let info = objc_msgSend(klass, process_info_sel); let args = objc_msgSend(info, arguments_sel); diff --git a/library/std/src/sys/unix/fs.rs b/library/std/src/sys/unix/fs.rs index 9cf5cfcc8d5..fbc7f04ce9a 100644 --- a/library/std/src/sys/unix/fs.rs +++ b/library/std/src/sys/unix/fs.rs @@ -1097,7 +1097,7 @@ pub fn file_attr(&self) -> io::Result { cfg_has_statx! { if let Some(ret) = unsafe { try_statx( fd, - c"".as_ptr() as *const c_char, + b"\0" as *const _ as *const c_char, libc::AT_EMPTY_PATH | libc::AT_STATX_SYNC_AS_STAT, libc::STATX_ALL, ) } { diff --git a/library/std/src/sys/unix/mod.rs b/library/std/src/sys/unix/mod.rs index 1b72e21a832..326f1481e19 100644 --- a/library/std/src/sys/unix/mod.rs +++ b/library/std/src/sys/unix/mod.rs @@ -1,5 +1,6 @@ #![allow(missing_docs, nonstandard_style)] +use crate::ffi::CStr; use crate::io::ErrorKind; pub use self::rand::hashmap_random_keys; @@ -74,7 +75,7 @@ pub unsafe fn init(argc: isize, argv: *const *const u8, sigpipe: u8) { // thread-id for the main thread and so renaming the main thread will rename the // process and we only want to enable this on platforms we've tested. if cfg!(target_os = "macos") { - thread::Thread::set_name(&c"main"); + thread::Thread::set_name(&CStr::from_bytes_with_nul_unchecked(b"main\0")); } unsafe fn sanitize_standard_fds() { @@ -121,7 +122,7 @@ unsafe fn sanitize_standard_fds() { if pfd.revents & libc::POLLNVAL == 0 { continue; } - if open64(c"/dev/null".as_ptr().cast(), libc::O_RDWR, 0) == -1 { + if open64("/dev/null\0".as_ptr().cast(), libc::O_RDWR, 0) == -1 { // If the stream is closed but we failed to reopen it, abort the // process. Otherwise we wouldn't preserve the safety of // operations on the corresponding Rust object Stdin, Stdout, or @@ -151,7 +152,7 @@ unsafe fn sanitize_standard_fds() { use libc::open64; for fd in 0..3 { if libc::fcntl(fd, libc::F_GETFD) == -1 && errno() == libc::EBADF { - if open64(c"/dev/null".as_ptr().cast(), libc::O_RDWR, 0) == -1 { + if open64("/dev/null\0".as_ptr().cast(), libc::O_RDWR, 0) == -1 { // If the stream is closed but we failed to reopen it, abort the // process. Otherwise we wouldn't preserve the safety of // operations on the corresponding Rust object Stdin, Stdout, or diff --git a/library/std/src/sys/unix/process/process_common.rs b/library/std/src/sys/unix/process/process_common.rs index 5f316b12b62..640648e8707 100644 --- a/library/std/src/sys/unix/process/process_common.rs +++ b/library/std/src/sys/unix/process/process_common.rs @@ -24,11 +24,11 @@ if #[cfg(target_os = "fuchsia")] { // fuchsia doesn't have /dev/null } else if #[cfg(target_os = "redox")] { - const DEV_NULL: &CStr = c"null:"; + const DEV_NULL: &str = "null:\0"; } else if #[cfg(target_os = "vxworks")] { - const DEV_NULL: &CStr = c"/null"; + const DEV_NULL: &str = "/null\0"; } else { - const DEV_NULL: &CStr = c"/dev/null"; + const DEV_NULL: &str = "/dev/null\0"; } } @@ -474,7 +474,8 @@ pub fn to_child_stdio(&self, readable: bool) -> io::Result<(ChildStdio, Option= Win10 1607 // https://docs.microsoft.com/en-us/windows/win32/api/processthreadsapi/nf-processthreadsapi-setthreaddescription @@ -354,7 +354,7 @@ pub fn WaitOnAddress( } compat_fn_with_fallback! { - pub static NTDLL: &CStr = c"ntdll"; + pub static NTDLL: &CStr = ansi_str!("ntdll"); pub fn NtCreateKeyedEvent( KeyedEventHandle: LPHANDLE, diff --git a/library/std/src/sys/windows/compat.rs b/library/std/src/sys/windows/compat.rs index 649cc4bfdbf..4fe95d41116 100644 --- a/library/std/src/sys/windows/compat.rs +++ b/library/std/src/sys/windows/compat.rs @@ -228,9 +228,9 @@ pub fn option() -> Option { /// Load all needed functions from "api-ms-win-core-synch-l1-2-0". pub(super) fn load_synch_functions() { fn try_load() -> Option<()> { - const MODULE_NAME: &CStr = c"api-ms-win-core-synch-l1-2-0"; - const WAIT_ON_ADDRESS: &CStr = c"WaitOnAddress"; - const WAKE_BY_ADDRESS_SINGLE: &CStr = c"WakeByAddressSingle"; + const MODULE_NAME: &CStr = ansi_str!("api-ms-win-core-synch-l1-2-0"); + const WAIT_ON_ADDRESS: &CStr = ansi_str!("WaitOnAddress"); + const WAKE_BY_ADDRESS_SINGLE: &CStr = ansi_str!("WakeByAddressSingle"); // Try loading the library and all the required functions. // If any step fails, then they all fail. diff --git a/library/std/src/sys/windows/mod.rs b/library/std/src/sys/windows/mod.rs index b11c8962203..bcc172b0fae 100644 --- a/library/std/src/sys/windows/mod.rs +++ b/library/std/src/sys/windows/mod.rs @@ -1,6 +1,6 @@ #![allow(missing_docs, nonstandard_style)] -use crate::ffi::{OsStr, OsString}; +use crate::ffi::{CStr, OsStr, OsString}; use crate::io::ErrorKind; use crate::mem::MaybeUninit; use crate::os::windows::ffi::{OsStrExt, OsStringExt}; @@ -51,7 +51,7 @@ pub unsafe fn init(_argc: isize, _argv: *const *const u8, _sigpipe: u8) { // Normally, `thread::spawn` will call `Thread::set_name` but since this thread already // exists, we have to call it ourselves. - thread::Thread::set_name(&c"main"); + thread::Thread::set_name(&CStr::from_bytes_with_nul_unchecked(b"main\0")); } // SAFETY: must be called only once during runtime cleanup. diff --git a/src/tools/tidy/src/deps.rs b/src/tools/tidy/src/deps.rs index 1f6c0b75a09..ecc84c1618c 100644 --- a/src/tools/tidy/src/deps.rs +++ b/src/tools/tidy/src/deps.rs @@ -131,6 +131,7 @@ "crossbeam-epoch", "crossbeam-utils", "crypto-common", + "cstr", "datafrog", "derive_more", "digest", From 3788b7ab32a6398e7563557f7b32f741f43704a3 Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?Le=C3=B3n=20Orell=20Valerian=20Liehr?= Date: Tue, 4 Jul 2023 19:53:44 +0200 Subject: [PATCH 4/5] Revert "use new c literals instead of cstr! macro" This reverts commit a17561ffc90c900cb7d0e96b00c6381244764ef7. --- Cargo.lock | 11 +++++ compiler/rustc_codegen_llvm/Cargo.toml | 1 + compiler/rustc_codegen_llvm/src/allocator.rs | 4 +- compiler/rustc_codegen_llvm/src/back/lto.rs | 2 +- compiler/rustc_codegen_llvm/src/back/write.rs | 16 +++---- compiler/rustc_codegen_llvm/src/base.rs | 6 ++- compiler/rustc_codegen_llvm/src/builder.rs | 16 +++++-- compiler/rustc_codegen_llvm/src/consts.rs | 7 +-- compiler/rustc_codegen_llvm/src/context.rs | 44 +++++++++---------- .../rustc_codegen_llvm/src/debuginfo/gdb.rs | 3 +- .../src/debuginfo/metadata.rs | 9 ++-- .../rustc_codegen_llvm/src/debuginfo/mod.rs | 7 +-- compiler/rustc_codegen_llvm/src/lib.rs | 1 - 13 files changed, 75 insertions(+), 52 deletions(-) diff --git a/Cargo.lock b/Cargo.lock index cdf6a4cc2de..54d74ab6a75 100644 --- a/Cargo.lock +++ b/Cargo.lock @@ -766,6 +766,16 @@ dependencies = [ "typenum", ] +[[package]] +name = "cstr" +version = "0.2.8" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "c11a39d776a3b35896711da8a04dc1835169dcd36f710878187637314e47941b" +dependencies = [ + "proc-macro2", + "quote", +] + [[package]] name = "ctrlc" version = "3.4.0" @@ -3016,6 +3026,7 @@ name = "rustc_codegen_llvm" version = "0.0.0" dependencies = [ "bitflags 1.3.2", + "cstr", "libc", "measureme", "object", diff --git a/compiler/rustc_codegen_llvm/Cargo.toml b/compiler/rustc_codegen_llvm/Cargo.toml index 39ff3a0ba2d..ad51f2d0958 100644 --- a/compiler/rustc_codegen_llvm/Cargo.toml +++ b/compiler/rustc_codegen_llvm/Cargo.toml @@ -8,6 +8,7 @@ test = false [dependencies] bitflags = "1.0" +cstr = "0.2" libc = "0.2" measureme = "10.0.0" object = { version = "0.31.1", default-features = false, features = [ diff --git a/compiler/rustc_codegen_llvm/src/allocator.rs b/compiler/rustc_codegen_llvm/src/allocator.rs index ad0636894b7..a57508815d6 100644 --- a/compiler/rustc_codegen_llvm/src/allocator.rs +++ b/compiler/rustc_codegen_llvm/src/allocator.rs @@ -77,7 +77,7 @@ pub(crate) unsafe fn codegen( llvm::LLVMRustGetOrInsertFunction(llmod, callee.as_ptr().cast(), callee.len(), ty); llvm::LLVMRustSetVisibility(callee, llvm::Visibility::Hidden); - let llbb = llvm::LLVMAppendBasicBlockInContext(llcx, llfn, c"entry".as_ptr().cast()); + let llbb = llvm::LLVMAppendBasicBlockInContext(llcx, llfn, "entry\0".as_ptr().cast()); let llbuilder = llvm::LLVMCreateBuilderInContext(llcx); llvm::LLVMPositionBuilderAtEnd(llbuilder, llbb); @@ -129,7 +129,7 @@ pub(crate) unsafe fn codegen( attributes::apply_to_llfn(callee, llvm::AttributePlace::Function, &[no_return]); llvm::LLVMRustSetVisibility(callee, llvm::Visibility::Hidden); - let llbb = llvm::LLVMAppendBasicBlockInContext(llcx, llfn, c"entry".as_ptr().cast()); + let llbb = llvm::LLVMAppendBasicBlockInContext(llcx, llfn, "entry\0".as_ptr().cast()); let llbuilder = llvm::LLVMCreateBuilderInContext(llcx); llvm::LLVMPositionBuilderAtEnd(llbuilder, llbb); diff --git a/compiler/rustc_codegen_llvm/src/back/lto.rs b/compiler/rustc_codegen_llvm/src/back/lto.rs index 94885b40cc1..d7dd98d7938 100644 --- a/compiler/rustc_codegen_llvm/src/back/lto.rs +++ b/compiler/rustc_codegen_llvm/src/back/lto.rs @@ -601,7 +601,7 @@ pub(crate) fn run_pass_manager( llvm::LLVMRustAddModuleFlag( module.module_llvm.llmod(), llvm::LLVMModFlagBehavior::Error, - c"LTOPostLink".as_ptr().cast(), + "LTOPostLink\0".as_ptr().cast(), 1, ); } diff --git a/compiler/rustc_codegen_llvm/src/back/write.rs b/compiler/rustc_codegen_llvm/src/back/write.rs index 998e3b300da..0f5e975445f 100644 --- a/compiler/rustc_codegen_llvm/src/back/write.rs +++ b/compiler/rustc_codegen_llvm/src/back/write.rs @@ -931,16 +931,16 @@ unsafe fn embed_bitcode( let llglobal = llvm::LLVMAddGlobal( llmod, common::val_ty(llconst), - c"rustc.embedded.module".as_ptr().cast(), + "rustc.embedded.module\0".as_ptr().cast(), ); llvm::LLVMSetInitializer(llglobal, llconst); let section = if is_apple { - c"__LLVM,__bitcode" + "__LLVM,__bitcode\0" } else if is_aix { - c".ipa" + ".ipa\0" } else { - c".llvmbc" + ".llvmbc\0" }; llvm::LLVMSetSection(llglobal, section.as_ptr().cast()); llvm::LLVMRustSetLinkage(llglobal, llvm::Linkage::PrivateLinkage); @@ -950,15 +950,15 @@ unsafe fn embed_bitcode( let llglobal = llvm::LLVMAddGlobal( llmod, common::val_ty(llconst), - c"rustc.embedded.cmdline".as_ptr().cast(), + "rustc.embedded.cmdline\0".as_ptr().cast(), ); llvm::LLVMSetInitializer(llglobal, llconst); let section = if is_apple { - c"__LLVM,__cmdline" + "__LLVM,__cmdline\0" } else if is_aix { - c".info" + ".info\0" } else { - c".llvmcmd" + ".llvmcmd\0" }; llvm::LLVMSetSection(llglobal, section.as_ptr().cast()); llvm::LLVMRustSetLinkage(llglobal, llvm::Linkage::PrivateLinkage); diff --git a/compiler/rustc_codegen_llvm/src/base.rs b/compiler/rustc_codegen_llvm/src/base.rs index 2f7eb08ad3d..5b2bbdb4bde 100644 --- a/compiler/rustc_codegen_llvm/src/base.rs +++ b/compiler/rustc_codegen_llvm/src/base.rs @@ -19,6 +19,8 @@ use crate::llvm; use crate::value::Value; +use cstr::cstr; + use rustc_codegen_ssa::base::maybe_create_entry_wrapper; use rustc_codegen_ssa::mono_item::MonoItemExt; use rustc_codegen_ssa::traits::*; @@ -108,11 +110,11 @@ fn module_codegen(tcx: TyCtxt<'_>, cgu_name: Symbol) -> ModuleCodegen BackendTypes for Builder<'_, 'll, 'tcx> { type Value = as BackendTypes>::Value; @@ -1002,13 +1007,14 @@ fn resume(&mut self, exn0: &'ll Value, exn1: &'ll Value) { } fn cleanup_pad(&mut self, parent: Option<&'ll Value>, args: &[&'ll Value]) -> Funclet<'ll> { + let name = cstr!("cleanuppad"); let ret = unsafe { llvm::LLVMBuildCleanupPad( self.llbuilder, parent, args.as_ptr(), args.len() as c_uint, - c"cleanuppad".as_ptr(), + name.as_ptr(), ) }; Funclet::new(ret.expect("LLVM does not have support for cleanuppad")) @@ -1022,13 +1028,14 @@ fn cleanup_ret(&mut self, funclet: &Funclet<'ll>, unwind: Option<&'ll BasicBlock } fn catch_pad(&mut self, parent: &'ll Value, args: &[&'ll Value]) -> Funclet<'ll> { + let name = cstr!("catchpad"); let ret = unsafe { llvm::LLVMBuildCatchPad( self.llbuilder, parent, args.as_ptr(), args.len() as c_uint, - c"catchpad".as_ptr(), + name.as_ptr(), ) }; Funclet::new(ret.expect("LLVM does not have support for catchpad")) @@ -1040,13 +1047,14 @@ fn catch_switch( unwind: Option<&'ll BasicBlock>, handlers: &[&'ll BasicBlock], ) -> &'ll Value { + let name = cstr!("catchswitch"); let ret = unsafe { llvm::LLVMBuildCatchSwitch( self.llbuilder, parent, unwind, handlers.len() as c_uint, - c"catchswitch".as_ptr(), + name.as_ptr(), ) }; let ret = ret.expect("LLVM does not have support for catchswitch"); diff --git a/compiler/rustc_codegen_llvm/src/consts.rs b/compiler/rustc_codegen_llvm/src/consts.rs index 2087754c66b..df52f50f86f 100644 --- a/compiler/rustc_codegen_llvm/src/consts.rs +++ b/compiler/rustc_codegen_llvm/src/consts.rs @@ -8,6 +8,7 @@ use crate::type_::Type; use crate::type_of::LayoutLlvmExt; use crate::value::Value; +use cstr::cstr; use rustc_codegen_ssa::traits::*; use rustc_hir::def_id::DefId; use rustc_middle::middle::codegen_fn_attrs::{CodegenFnAttrFlags, CodegenFnAttrs}; @@ -481,9 +482,9 @@ fn codegen_static(&self, def_id: DefId, is_mutable: bool) { .all(|&byte| byte == 0); let sect_name = if all_bytes_are_zero { - c"__DATA,__thread_bss" + cstr!("__DATA,__thread_bss") } else { - c"__DATA,__thread_data" + cstr!("__DATA,__thread_data") }; llvm::LLVMSetSection(g, sect_name.as_ptr()); } @@ -512,7 +513,7 @@ fn codegen_static(&self, def_id: DefId, is_mutable: bool) { let val = llvm::LLVMMetadataAsValue(self.llcx, meta); llvm::LLVMAddNamedMetadataOperand( self.llmod, - c"wasm.custom_sections".as_ptr().cast(), + "wasm.custom_sections\0".as_ptr().cast(), val, ); } diff --git a/compiler/rustc_codegen_llvm/src/context.rs b/compiler/rustc_codegen_llvm/src/context.rs index 287a22bc9a6..e1e0a442845 100644 --- a/compiler/rustc_codegen_llvm/src/context.rs +++ b/compiler/rustc_codegen_llvm/src/context.rs @@ -8,6 +8,7 @@ use crate::type_::Type; use crate::value::Value; +use cstr::cstr; use rustc_codegen_ssa::base::{wants_msvc_seh, wants_wasm_eh}; use rustc_codegen_ssa::traits::*; use rustc_data_structures::base_n; @@ -223,42 +224,36 @@ pub unsafe fn create_module<'ll>( // If skipping the PLT is enabled, we need to add some module metadata // to ensure intrinsic calls don't use it. if !sess.needs_plt() { - llvm::LLVMRustAddModuleFlag( - llmod, - llvm::LLVMModFlagBehavior::Warning, - c"RtLibUseGOT".as_ptr().cast(), - 1, - ); + let avoid_plt = "RtLibUseGOT\0".as_ptr().cast(); + llvm::LLVMRustAddModuleFlag(llmod, llvm::LLVMModFlagBehavior::Warning, avoid_plt, 1); } // Enable canonical jump tables if CFI is enabled. (See https://reviews.llvm.org/D65629.) if sess.is_sanitizer_cfi_canonical_jump_tables_enabled() && sess.is_sanitizer_cfi_enabled() { + let canonical_jump_tables = "CFI Canonical Jump Tables\0".as_ptr().cast(); llvm::LLVMRustAddModuleFlag( llmod, llvm::LLVMModFlagBehavior::Override, - c"CFI Canonical Jump Tables".as_ptr().cast(), + canonical_jump_tables, 1, ); } // Enable LTO unit splitting if specified or if CFI is enabled. (See https://reviews.llvm.org/D53891.) if sess.is_split_lto_unit_enabled() || sess.is_sanitizer_cfi_enabled() { + let enable_split_lto_unit = "EnableSplitLTOUnit\0".as_ptr().cast(); llvm::LLVMRustAddModuleFlag( llmod, llvm::LLVMModFlagBehavior::Override, - c"EnableSplitLTOUnit".as_ptr().cast(), + enable_split_lto_unit, 1, ); } // Add "kcfi" module flag if KCFI is enabled. (See https://reviews.llvm.org/D119296.) if sess.is_sanitizer_kcfi_enabled() { - llvm::LLVMRustAddModuleFlag( - llmod, - llvm::LLVMModFlagBehavior::Override, - c"kcfi".as_ptr().cast(), - 1, - ); + let kcfi = "kcfi\0".as_ptr().cast(); + llvm::LLVMRustAddModuleFlag(llmod, llvm::LLVMModFlagBehavior::Override, kcfi, 1); } // Control Flow Guard is currently only supported by the MSVC linker on Windows. @@ -270,7 +265,7 @@ pub unsafe fn create_module<'ll>( llvm::LLVMRustAddModuleFlag( llmod, llvm::LLVMModFlagBehavior::Warning, - c"cfguard".as_ptr() as *const _, + "cfguard\0".as_ptr() as *const _, 1, ) } @@ -279,7 +274,7 @@ pub unsafe fn create_module<'ll>( llvm::LLVMRustAddModuleFlag( llmod, llvm::LLVMModFlagBehavior::Warning, - c"cfguard".as_ptr() as *const _, + "cfguard\0".as_ptr() as *const _, 2, ) } @@ -297,26 +292,26 @@ pub unsafe fn create_module<'ll>( llvm::LLVMRustAddModuleFlag( llmod, behavior, - c"branch-target-enforcement".as_ptr().cast(), + "branch-target-enforcement\0".as_ptr().cast(), bti.into(), ); llvm::LLVMRustAddModuleFlag( llmod, behavior, - c"sign-return-address".as_ptr().cast(), + "sign-return-address\0".as_ptr().cast(), pac_ret.is_some().into(), ); let pac_opts = pac_ret.unwrap_or(PacRet { leaf: false, key: PAuthKey::A }); llvm::LLVMRustAddModuleFlag( llmod, behavior, - c"sign-return-address-all".as_ptr().cast(), + "sign-return-address-all\0".as_ptr().cast(), pac_opts.leaf.into(), ); llvm::LLVMRustAddModuleFlag( llmod, behavior, - c"sign-return-address-with-bkey".as_ptr().cast(), + "sign-return-address-with-bkey\0".as_ptr().cast(), u32::from(pac_opts.key == PAuthKey::B), ); } else { @@ -332,7 +327,7 @@ pub unsafe fn create_module<'ll>( llvm::LLVMRustAddModuleFlag( llmod, llvm::LLVMModFlagBehavior::Override, - c"cf-protection-branch".as_ptr().cast(), + "cf-protection-branch\0".as_ptr().cast(), 1, ) } @@ -340,7 +335,7 @@ pub unsafe fn create_module<'ll>( llvm::LLVMRustAddModuleFlag( llmod, llvm::LLVMModFlagBehavior::Override, - c"cf-protection-return".as_ptr().cast(), + "cf-protection-return\0".as_ptr().cast(), 1, ) } @@ -349,7 +344,7 @@ pub unsafe fn create_module<'ll>( llvm::LLVMRustAddModuleFlag( llmod, llvm::LLVMModFlagBehavior::Error, - c"Virtual Function Elim".as_ptr().cast(), + "Virtual Function Elim\0".as_ptr().cast(), 1, ); } @@ -481,13 +476,14 @@ pub fn coverage_context(&self) -> Option<&coverageinfo::CrateCoverageContext<'ll } pub(crate) fn create_used_variable_impl(&self, name: &'static CStr, values: &[&'ll Value]) { + let section = cstr!("llvm.metadata"); let array = self.const_array(self.type_ptr_to(self.type_i8()), values); unsafe { let g = llvm::LLVMAddGlobal(self.llmod, self.val_ty(array), name.as_ptr()); llvm::LLVMSetInitializer(g, array); llvm::LLVMRustSetLinkage(g, llvm::Linkage::AppendingLinkage); - llvm::LLVMSetSection(g, c"llvm.metadata".as_ptr()); + llvm::LLVMSetSection(g, section.as_ptr()); } } } diff --git a/compiler/rustc_codegen_llvm/src/debuginfo/gdb.rs b/compiler/rustc_codegen_llvm/src/debuginfo/gdb.rs index 8be54b7eb71..37f30917609 100644 --- a/compiler/rustc_codegen_llvm/src/debuginfo/gdb.rs +++ b/compiler/rustc_codegen_llvm/src/debuginfo/gdb.rs @@ -38,6 +38,7 @@ pub fn get_or_insert_gdb_debug_scripts_section_global<'ll>(cx: &CodegenCx<'ll, ' unsafe { llvm::LLVMGetNamedGlobal(cx.llmod, c_section_var_name.as_ptr().cast()) }; section_var.unwrap_or_else(|| { + let section_name = b".debug_gdb_scripts\0"; let mut section_contents = Vec::new(); // Add the pretty printers for the standard library first. @@ -70,7 +71,7 @@ pub fn get_or_insert_gdb_debug_scripts_section_global<'ll>(cx: &CodegenCx<'ll, ' let section_var = cx .define_global(section_var_name, llvm_type) .unwrap_or_else(|| bug!("symbol `{}` is already defined", section_var_name)); - llvm::LLVMSetSection(section_var, c".debug_gdb_scripts".as_ptr().cast()); + llvm::LLVMSetSection(section_var, section_name.as_ptr().cast()); llvm::LLVMSetInitializer(section_var, cx.const_bytes(section_contents)); llvm::LLVMSetGlobalConstant(section_var, llvm::True); llvm::LLVMSetUnnamedAddress(section_var, llvm::UnnamedAddr::Global); diff --git a/compiler/rustc_codegen_llvm/src/debuginfo/metadata.rs b/compiler/rustc_codegen_llvm/src/debuginfo/metadata.rs index 4b88ab8a97a..ad07d637106 100644 --- a/compiler/rustc_codegen_llvm/src/debuginfo/metadata.rs +++ b/compiler/rustc_codegen_llvm/src/debuginfo/metadata.rs @@ -20,6 +20,7 @@ }; use crate::value::Value; +use cstr::cstr; use rustc_codegen_ssa::debuginfo::type_names::cpp_like_debuginfo; use rustc_codegen_ssa::debuginfo::type_names::VTableNameKind; use rustc_codegen_ssa::traits::*; @@ -811,6 +812,7 @@ pub fn build_compile_unit_di_node<'ll, 'tcx>( let name_in_debuginfo = name_in_debuginfo.to_string_lossy(); let work_dir = tcx.sess.opts.working_dir.to_string_lossy(FileNameDisplayPreference::Remapped); + let flags = "\0"; let output_filenames = tcx.output_filenames(()); let split_name = if tcx.sess.target_can_use_split_dwarf() { output_filenames @@ -847,7 +849,7 @@ pub fn build_compile_unit_di_node<'ll, 'tcx>( producer.as_ptr().cast(), producer.len(), tcx.sess.opts.optimize != config::OptLevel::No, - c"".as_ptr().cast(), + flags.as_ptr().cast(), 0, // NB: this doesn't actually have any perceptible effect, it seems. LLVM will instead // put the path supplied to `MCSplitDwarfFile` into the debug info of the final @@ -876,7 +878,8 @@ pub fn build_compile_unit_di_node<'ll, 'tcx>( ); let val = llvm::LLVMMetadataAsValue(debug_context.llcontext, gcov_metadata); - llvm::LLVMAddNamedMetadataOperand(debug_context.llmod, c"llvm.gcov".as_ptr(), val); + let llvm_gcov_ident = cstr!("llvm.gcov"); + llvm::LLVMAddNamedMetadataOperand(debug_context.llmod, llvm_gcov_ident.as_ptr(), val); } // Insert `llvm.ident` metadata on the wasm targets since that will @@ -889,7 +892,7 @@ pub fn build_compile_unit_di_node<'ll, 'tcx>( ); llvm::LLVMAddNamedMetadataOperand( debug_context.llmod, - c"llvm.ident".as_ptr(), + cstr!("llvm.ident").as_ptr(), llvm::LLVMMDNodeInContext(debug_context.llcontext, &name_metadata, 1), ); } diff --git a/compiler/rustc_codegen_llvm/src/debuginfo/mod.rs b/compiler/rustc_codegen_llvm/src/debuginfo/mod.rs index c2f16cad3fc..0858f6b5f8e 100644 --- a/compiler/rustc_codegen_llvm/src/debuginfo/mod.rs +++ b/compiler/rustc_codegen_llvm/src/debuginfo/mod.rs @@ -113,7 +113,7 @@ pub fn finalize(&self, sess: &Session) { llvm::LLVMRustAddModuleFlag( self.llmod, llvm::LLVMModFlagBehavior::Warning, - c"Dwarf Version".as_ptr().cast(), + "Dwarf Version\0".as_ptr().cast(), dwarf_version, ); } else { @@ -121,16 +121,17 @@ pub fn finalize(&self, sess: &Session) { llvm::LLVMRustAddModuleFlag( self.llmod, llvm::LLVMModFlagBehavior::Warning, - c"CodeView".as_ptr().cast(), + "CodeView\0".as_ptr().cast(), 1, ) } // Prevent bitcode readers from deleting the debug info. + let ptr = "Debug Info Version\0".as_ptr(); llvm::LLVMRustAddModuleFlag( self.llmod, llvm::LLVMModFlagBehavior::Warning, - c"Debug Info Version".as_ptr().cast(), + ptr.cast(), llvm::LLVMRustDebugMetadataVersion(), ); } diff --git a/compiler/rustc_codegen_llvm/src/lib.rs b/compiler/rustc_codegen_llvm/src/lib.rs index 24968e00cc8..24ba28bbc82 100644 --- a/compiler/rustc_codegen_llvm/src/lib.rs +++ b/compiler/rustc_codegen_llvm/src/lib.rs @@ -11,7 +11,6 @@ #![feature(let_chains)] #![feature(never_type)] #![feature(impl_trait_in_assoc_type)] -#![feature(c_str_literals)] #![recursion_limit = "256"] #![allow(rustc::potential_query_instability)] #![deny(rustc::untranslatable_diagnostic)] From 22fd6a6abf1f97b488278cd142cc7fe721beb92a Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?Le=C3=B3n=20Orell=20Valerian=20Liehr?= Date: Tue, 4 Jul 2023 20:37:17 +0200 Subject: [PATCH 5/5] Add regression test --- .../edition-2015-2018-lexing.rs | 24 +++++++++++++++++++ 1 file changed, 24 insertions(+) create mode 100644 tests/ui/rfcs/rfc-3348-c-string-literals/edition-2015-2018-lexing.rs diff --git a/tests/ui/rfcs/rfc-3348-c-string-literals/edition-2015-2018-lexing.rs b/tests/ui/rfcs/rfc-3348-c-string-literals/edition-2015-2018-lexing.rs new file mode 100644 index 00000000000..2a4cd600426 --- /dev/null +++ b/tests/ui/rfcs/rfc-3348-c-string-literals/edition-2015-2018-lexing.rs @@ -0,0 +1,24 @@ +// Regression test for issue #113235. + +// check-pass +// revisions: edition2015 edition2018 +//[edition2015] edition: 2015 +//[edition2018] edition: 2018 + +// Make sure that in pre-2021 editions we continue to parse the snippet +// `c"hello"` as an identifier followed by a (normal) string literal and +// allow the code below to compile. +// Prefixes including `c` as used by C string literals are only reserved +// in edition 2021 and onward. +// +// Consider checking out rust-2021/reserved-prefixes-migration.rs as well. + +macro_rules! parse { + (c $e:expr) => { + $e + }; +} + +fn main() { + let _: &'static str = parse!(c"hello"); +}