Rollup merge of rust-lang#140143 - thaliaarchi:move-env-pal, r=joboet

Move `sys::pal::os::Env` into `sys::env`

Although `Env` (as `Vars`), `Args`, path functions, and OS constants are publicly exposed via `std::env`, their implementations are each self-contained. Keep them separate in `std::sys` and make a new module, `sys::env`, for `Env`.

Also fix `unsafe_op_in_unsafe_fn` for Unix and update the `!DynSend` and `!DynSync` impls which had grown out of sync with the platforms (see rust-lang#48005 for discussion on that).

r? joboet

Tracked in rust-lang#117276.

Author: matthiaskrgr

compiler/rustc_data_structures/src/marker.rs

@@ -39,8 +39,15 @@ impls_dyn_send_neg!(
     [std::io::StderrLock<'_>]
 );
 
-#[cfg(any(unix, target_os = "hermit", target_os = "wasi", target_os = "solid_asp3"))]
-// Consistent with `std`, `os_imp::Env` is `!Sync` in these platforms
+#[cfg(any(
+    unix,
+    target_os = "hermit",
+    all(target_vendor = "fortanix", target_env = "sgx"),
+    target_os = "solid_asp3",
+    target_os = "wasi",
+    target_os = "xous"
+))]
+// Consistent with `std`, `env_imp::Env` is `!Sync` in these platforms
 impl !DynSend for std::env::VarsOs {}
 
 macro_rules! already_send {
@@ -106,8 +113,15 @@ impls_dyn_sync_neg!(
     [std::sync::mpsc::Sender<T> where T]
 );
 
-#[cfg(any(unix, target_os = "hermit", target_os = "wasi", target_os = "solid_asp3"))]
-// Consistent with `std`, `os_imp::Env` is `!Sync` in these platforms
+#[cfg(any(
+    unix,
+    target_os = "hermit",
+    all(target_vendor = "fortanix", target_env = "sgx"),
+    target_os = "solid_asp3",
+    target_os = "wasi",
+    target_os = "xous"
+))]
+// Consistent with `std`, `env_imp::Env` is `!Sync` in these platforms
 impl !DynSync for std::env::VarsOs {}
 
 macro_rules! already_sync {

library/std/src/env.rs

@@ -13,7 +13,7 @@
 use crate::error::Error;
 use crate::ffi::{OsStr, OsString};
 use crate::path::{Path, PathBuf};
-use crate::sys::os as os_imp;
+use crate::sys::{env as env_imp, os as os_imp};
 use crate::{fmt, io, sys};
 
 /// Returns the current working directory as a [`PathBuf`].
@@ -96,7 +96,7 @@ pub struct Vars {
 /// [`env::vars_os()`]: vars_os
 #[stable(feature = "env", since = "1.0.0")]
 pub struct VarsOs {
-    inner: os_imp::Env,
+    inner: env_imp::Env,
 }
 
 /// Returns an iterator of (variable, value) pairs of strings, for all the
@@ -150,7 +150,7 @@ pub fn vars() -> Vars {
 #[must_use]
 #[stable(feature = "env", since = "1.0.0")]
 pub fn vars_os() -> VarsOs {
-    VarsOs { inner: os_imp::env() }
+    VarsOs { inner: env_imp::env() }
 }
 
 #[stable(feature = "env", since = "1.0.0")]
@@ -259,7 +259,7 @@ pub fn var_os<K: AsRef<OsStr>>(key: K) -> Option<OsString> {
 }
 
 fn _var_os(key: &OsStr) -> Option<OsString> {
-    os_imp::getenv(key)
+    env_imp::getenv(key)
 }
 
 /// The error type for operations interacting with environment variables.
@@ -363,7 +363,7 @@ impl Error for VarError {
 #[stable(feature = "env", since = "1.0.0")]
 pub unsafe fn set_var<K: AsRef<OsStr>, V: AsRef<OsStr>>(key: K, value: V) {
     let (key, value) = (key.as_ref(), value.as_ref());
-    unsafe { os_imp::setenv(key, value) }.unwrap_or_else(|e| {
+    unsafe { env_imp::setenv(key, value) }.unwrap_or_else(|e| {
         panic!("failed to set environment variable `{key:?}` to `{value:?}`: {e}")
     })
 }
@@ -434,7 +434,7 @@ pub unsafe fn set_var<K: AsRef<OsStr>, V: AsRef<OsStr>>(key: K, value: V) {
 #[stable(feature = "env", since = "1.0.0")]
 pub unsafe fn remove_var<K: AsRef<OsStr>>(key: K) {
     let key = key.as_ref();
-    unsafe { os_imp::unsetenv(key) }
+    unsafe { env_imp::unsetenv(key) }
         .unwrap_or_else(|e| panic!("failed to remove environment variable `{key:?}`: {e}"))
 }
 

library/std/src/sys/args/mod.rs

@@ -2,6 +2,16 @@
 
 #![forbid(unsafe_op_in_unsafe_fn)]
 
+#[cfg(any(
+    all(target_family = "unix", not(any(target_os = "espidf", target_os = "vita"))),
+    target_family = "windows",
+    target_os = "hermit",
+    target_os = "uefi",
+    target_os = "wasi",
+    target_os = "xous",
+))]
+mod common;
+
 cfg_if::cfg_if! {
     if #[cfg(any(
         all(target_family = "unix", not(any(target_os = "espidf", target_os = "vita"))),

library/std/src/sys/args/uefi.rs

@@ -1,14 +1,11 @@
 use r_efi::protocols::loaded_image;
 
+pub use super::common::Args;
 use crate::env::current_exe;
 use crate::ffi::OsString;
 use crate::iter::Iterator;
 use crate::sys::pal::helpers;
 
-#[path = "common.rs"]
-mod common;
-pub use common::Args;
-
 pub fn args() -> Args {
     let lazy_current_exe = || Vec::from([current_exe().map(Into::into).unwrap_or_default()]);
 

library/std/src/sys/args/unix.rs

@@ -5,16 +5,13 @@
 
 #![allow(dead_code)] // runtime init functions not used during testing
 
+pub use super::common::Args;
 use crate::ffi::CStr;
 #[cfg(target_os = "hermit")]
 use crate::os::hermit::ffi::OsStringExt;
 #[cfg(not(target_os = "hermit"))]
 use crate::os::unix::ffi::OsStringExt;
 
-#[path = "common.rs"]
-mod common;
-pub use common::Args;
-
 /// One-time global initialization.
 pub unsafe fn init(argc: isize, argv: *const *const u8) {
     unsafe { imp::init(argc, argv) }

library/std/src/sys/args/wasi.rs

@@ -1,12 +1,9 @@
 #![forbid(unsafe_op_in_unsafe_fn)]
 
+pub use super::common::Args;
 use crate::ffi::{CStr, OsStr, OsString};
 use crate::os::wasi::ffi::OsStrExt;
 
-#[path = "common.rs"]
-mod common;
-pub use common::Args;
-
 /// Returns the command line arguments
 pub fn args() -> Args {
     Args::new(maybe_args().unwrap_or(Vec::new()))

library/std/src/sys/args/windows.rs

@@ -6,6 +6,7 @@
 #[cfg(test)]
 mod tests;
 
+pub use super::common::Args;
 use crate::ffi::{OsStr, OsString};
 use crate::num::NonZero;
 use crate::os::windows::prelude::*;
@@ -18,10 +19,6 @@ use crate::sys_common::AsInner;
 use crate::sys_common::wstr::WStrUnits;
 use crate::{io, iter, ptr};
 
-#[path = "common.rs"]
-mod common;
-pub use common::Args;
-
 pub fn args() -> Args {
     // SAFETY: `GetCommandLineW` returns a pointer to a null terminated UTF-16
     // string so it's safe for `WStrUnits` to use.

library/std/src/sys/args/xous.rs

@@ -1,10 +1,7 @@
+pub use super::common::Args;
 use crate::sys::pal::os::get_application_parameters;
 use crate::sys::pal::os::params::ArgumentList;
 
-#[path = "common.rs"]
-mod common;
-pub use common::Args;
-
 pub fn args() -> Args {
     let Some(params) = get_application_parameters() else {
         return Args::new(vec![]);

library/std/src/sys/env/common.rs

@@ -0,0 +1,48 @@
+use crate::ffi::OsString;
+use crate::{fmt, vec};
+
+pub struct Env {
+    iter: vec::IntoIter<(OsString, OsString)>,
+}
+
+// FIXME(https://github.com/rust-lang/rust/issues/114583): Remove this when <OsStr as Debug>::fmt matches <str as Debug>::fmt.
+pub struct EnvStrDebug<'a> {
+    slice: &'a [(OsString, OsString)],
+}
+
+impl fmt::Debug for EnvStrDebug<'_> {
+    fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result {
+        f.debug_list()
+            .entries(self.slice.iter().map(|(a, b)| (a.to_str().unwrap(), b.to_str().unwrap())))
+            .finish()
+    }
+}
+
+impl Env {
+    pub(super) fn new(env: Vec<(OsString, OsString)>) -> Self {
+        Env { iter: env.into_iter() }
+    }
+
+    pub fn str_debug(&self) -> impl fmt::Debug + '_ {
+        EnvStrDebug { slice: self.iter.as_slice() }
+    }
+}
+
+impl fmt::Debug for Env {
+    fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result {
+        f.debug_list().entries(self.iter.as_slice()).finish()
+    }
+}
+
+impl !Send for Env {}
+impl !Sync for Env {}
+
+impl Iterator for Env {
+    type Item = (OsString, OsString);
+    fn next(&mut self) -> Option<(OsString, OsString)> {
+        self.iter.next()
+    }
+    fn size_hint(&self) -> (usize, Option<usize>) {
+        self.iter.size_hint()
+    }
+}

library/std/src/sys/env/hermit.rs

@@ -0,0 +1,72 @@
+use core::slice::memchr;
+
+pub use super::common::Env;
+use crate::collections::HashMap;
+use crate::ffi::{CStr, OsStr, OsString, c_char};
+use crate::io;
+use crate::os::hermit::ffi::OsStringExt;
+use crate::sync::Mutex;
+
+static ENV: Mutex<Option<HashMap<OsString, OsString>>> = Mutex::new(None);
+
+pub fn init(env: *const *const c_char) {
+    let mut guard = ENV.lock().unwrap();
+    let map = guard.insert(HashMap::new());
+
+    if env.is_null() {
+        return;
+    }
+
+    unsafe {
+        let mut environ = env;
+        while !(*environ).is_null() {
+            if let Some((key, value)) = parse(CStr::from_ptr(*environ).to_bytes()) {
+                map.insert(key, value);
+            }
+            environ = environ.add(1);
+        }
+    }
+
+    fn parse(input: &[u8]) -> Option<(OsString, OsString)> {
+        // Strategy (copied from glibc): Variable name and value are separated
+        // by an ASCII equals sign '='. Since a variable name must not be
+        // empty, allow variable names starting with an equals sign. Skip all
+        // malformed lines.
+        if input.is_empty() {
+            return None;
+        }
+        let pos = memchr::memchr(b'=', &input[1..]).map(|p| p + 1);
+        pos.map(|p| {
+            (
+                OsStringExt::from_vec(input[..p].to_vec()),
+                OsStringExt::from_vec(input[p + 1..].to_vec()),
+            )
+        })
+    }
+}
+
+/// Returns a vector of (variable, value) byte-vector pairs for all the
+/// environment variables of the current process.
+pub fn env() -> Env {
+    let guard = ENV.lock().unwrap();
+    let env = guard.as_ref().unwrap();
+
+    let result = env.iter().map(|(key, value)| (key.clone(), value.clone())).collect();
+
+    Env::new(result)
+}
+
+pub fn getenv(k: &OsStr) -> Option<OsString> {
+    ENV.lock().unwrap().as_ref().unwrap().get(k).cloned()
+}
+
+pub unsafe fn setenv(k: &OsStr, v: &OsStr) -> io::Result<()> {
+    let (k, v) = (k.to_owned(), v.to_owned());
+    ENV.lock().unwrap().as_mut().unwrap().insert(k, v);
+    Ok(())
+}
+
+pub unsafe fn unsetenv(k: &OsStr) -> io::Result<()> {
+    ENV.lock().unwrap().as_mut().unwrap().remove(k);
+    Ok(())
+}