From 5db81020006d2920fc9c62ffc0f4322f90bffa04 Mon Sep 17 00:00:00 2001
From: Mads Marquart <mads@marquart.dk>
Date: Sun, 24 Nov 2024 02:36:21 +0100
Subject: [PATCH] Use OsStr in tracked environment variables

---
 compiler/rustc_interface/messages.ftl  |  3 ---
 compiler/rustc_interface/src/errors.rs |  8 -------
 compiler/rustc_interface/src/passes.rs | 32 +++++++++++++++-----------
 compiler/rustc_middle/src/query/mod.rs |  5 ++--
 4 files changed, 21 insertions(+), 27 deletions(-)

diff --git a/compiler/rustc_interface/messages.ftl b/compiler/rustc_interface/messages.ftl
index 7c98f07586350..47dfbc1d7fbf1 100644
--- a/compiler/rustc_interface/messages.ftl
+++ b/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}
 
diff --git a/compiler/rustc_interface/src/errors.rs b/compiler/rustc_interface/src/errors.rs
index e7ef7d2f375fa..939980a932fdb 100644
--- a/compiler/rustc_interface/src/errors.rs
+++ b/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;
diff --git a/compiler/rustc_interface/src/passes.rs b/compiler/rustc_interface/src/passes.rs
index d9d82d3f0146e..66431458959d9 100644
--- a/compiler/rustc_interface/src/passes.rs
+++ b/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 std::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<'tcx>(tcx: TyCtxt<'tcx>, key: &'tcx OsStr) -> Option<&'tcx OsStr> {
+    let value = std::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.map(|value| Symbol::intern(&value.to_string_lossy())),
+    ));
+
+    value_tcx
 }
 
 // Returns all the paths that correspond to generated files.
diff --git a/compiler/rustc_middle/src/query/mod.rs b/compiler/rustc_middle/src/query/mod.rs
index e31849426e08c..174be2b80e047 100644
--- a/compiler/rustc_middle/src/query/mod.rs
+++ b/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(key: &'tcx OsStr) -> Option<&'tcx OsStr> {
         // Environment variables are global state
         eval_always
         desc { "get the value of an environment variable" }