Use OsStr in tracked environment variables

Author: madsmtm

compiler/rustc_interface/messages.ftl

@@ -4,9 +4,6 @@ interface_cant_emit_mir =
 interface_emoji_identifier =
     identifiers cannot contain emoji: `{$ident}`
 
-interface_env_var_not_unicode =
-    cannot read environment variable "{$key}" with value "{$var}", since it contains non-unicode data
-
 interface_error_writing_dependencies =
     error writing dependencies to `{$path}`: {$error}
 

compiler/rustc_interface/src/errors.rs

@@ -1,4 +1,3 @@
-use std::ffi::OsString;
 use std::io;
 use std::path::Path;
 
@@ -22,13 +21,6 @@ pub struct EmojiIdentifier {
     pub ident: Symbol,
 }
 
-#[derive(Diagnostic)]
-#[diag(interface_env_var_not_unicode)]
-pub struct EnvVarNotUnicode {
-    pub key: Symbol,
-    pub var: OsString,
-}
-
 #[derive(Diagnostic)]
 #[diag(interface_mixed_bin_crate)]
 pub struct MixedBinCrate;

compiler/rustc_interface/src/passes.rs

@@ -1,6 +1,5 @@
 use std::any::Any;
-use std::env::VarError;
-use std::ffi::OsString;
+use std::ffi::{OsStr, OsString};
 use std::io::{self, BufWriter, Write};
 use std::path::{Path, PathBuf};
 use std::sync::{Arc, LazyLock};
@@ -322,22 +321,29 @@ fn early_lint_checks(tcx: TyCtxt<'_>, (): ()) {
     )
 }
 
-fn env_var(tcx: TyCtxt<'_>, key: Symbol) -> Option<Symbol> {
-    let var = match env::var(key.as_str()) {
-        Ok(var) => Some(Symbol::intern(&var)),
-        Err(VarError::NotPresent) => None,
-        Err(VarError::NotUnicode(var)) => {
-            tcx.dcx().emit_err(errors::EnvVarNotUnicode { key, var });
-            None
-        }
-    };
+fn env_var_os<'tcx>(tcx: TyCtxt<'tcx>, key: &'tcx OsStr) -> Option<&'tcx OsStr> {
+    let value = env::var_os(key);
+
+    let value_tcx = value.as_deref().map(|value| {
+        let encoded_bytes = tcx.arena.alloc_slice(value.as_encoded_bytes());
+        debug_assert_eq!(value.as_encoded_bytes(), encoded_bytes);
+        // SAFETY: The bytes came from `as_encoded_bytes`, and we assume that
+        // `alloc_slice` is implemented correctly, and passes the same bytes
+        // back (debug asserted above).
+        unsafe { OsStr::from_encoded_bytes_unchecked(encoded_bytes) }
+    });
+
     // Also add the variable to Cargo's dependency tracking
     //
     // NOTE: This only works for passes run before `write_dep_info`. See that
     // for extension points for configuring environment variables to be
     // properly change-tracked.
-    tcx.sess.psess.env_depinfo.borrow_mut().insert((key, var));
-    var
+    tcx.sess.psess.env_depinfo.borrow_mut().insert((
+        Symbol::intern(&key.to_string_lossy()),
+        value.and_then(|value| value.to_str()).map(|value| Symbol::intern(&value)),
+    ));
+
+    value_tcx
 }
 
 // Returns all the paths that correspond to generated files.
@@ -707,7 +713,7 @@ pub static DEFAULT_QUERY_PROVIDERS: LazyLock<Providers> = LazyLock::new(|| {
         |tcx, _| tcx.arena.alloc_from_iter(tcx.resolutions(()).stripped_cfg_items.steal());
     providers.resolutions = |tcx, ()| tcx.resolver_for_lowering_raw(()).1;
     providers.early_lint_checks = early_lint_checks;
-    providers.env_var = env_var;
+    providers.env_var_os = env_var_os;
     proc_macro_decls::provide(providers);
     rustc_const_eval::provide(providers);
     rustc_middle::hir::provide(providers);

compiler/rustc_middle/src/query/mod.rs

@@ -6,6 +6,7 @@
 
 #![allow(unused_parens)]
 
+use std::ffi::OsStr;
 use std::mem;
 use std::ops::Deref;
 use std::path::PathBuf;
@@ -130,9 +131,7 @@ rustc_queries! {
     /// NOTE: This currently does not work with dependency info in the
     /// analysis, codegen and linking passes, place extra code at the top of
     /// `rustc_interface::passes::write_dep_info` to make that work.
-    ///
-    /// Will emit an error and return `None` if the variable is not UTF-8.
-    query env_var(key: Symbol) -> Option<Symbol> {
+    query env_var_os(key: &'tcx OsStr) -> Option<&'tcx OsStr> {
         // Environment variables are global state
         eval_always
         desc { "get the value of an environment variable" }

compiler/rustc_middle/src/util/mod.rs

@@ -3,9 +3,14 @@ pub mod call_kind;
 pub mod common;
 pub mod find_self_call;
 
+use std::env::VarError;
+use std::ffi::OsStr;
+
 pub use call_kind::{CallDesugaringKind, CallKind, call_kind};
 pub use find_self_call::find_self_call;
 
+use crate::ty::TyCtxt;
+
 #[derive(Default, Copy, Clone)]
 pub struct Providers {
     pub queries: rustc_middle::query::Providers,
@@ -29,3 +34,13 @@ impl std::ops::Deref for Providers {
         &self.queries
     }
 }
+
+impl<'tcx> TyCtxt<'tcx> {
+    pub fn env_var(self, key: &'tcx OsStr) -> Result<&'tcx str, VarError> {
+        if let Some(value) = self.env_var_os(key.as_ref()) {
+            value.to_str().ok_or_else(|| VarError::NotUnicode(value.to_os_string()))
+        } else {
+            Err(VarError::NotPresent)
+        }
+    }
+}