Rollup merge of #121194 - beetrees:rustc-raw-args, r=petrochenkov

Refactor pre-getopts command line argument handling

Rebased version of #111658. I've also fixed the Windows CI failure (although I don't have access to Windows to test it myself).

compiler/rustc_driver_impl/src/args.rs

@@ -1,9 +1,7 @@
-use std::error;
-use std::fmt;
-use std::fs;
-use std::io;
+use std::{env, error, fmt, fs, io};
 
 use rustc_session::EarlyDiagCtxt;
+use rustc_span::ErrorGuaranteed;
 
 /// Expands argfiles in command line arguments.
 #[derive(Default)]
@@ -86,7 +84,7 @@ impl Expander {
     fn read_file(path: &str) -> Result<String, Error> {
         fs::read_to_string(path).map_err(|e| {
             if e.kind() == io::ErrorKind::InvalidData {
-                Error::Utf8Error(Some(path.to_string()))
+                Error::Utf8Error(path.to_string())
             } else {
                 Error::IOError(path.to_string(), e)
             }
@@ -94,23 +92,53 @@ impl Expander {
     }
 }
 
+/// Replaces any `@file` arguments with the contents of `file`, with each line of `file` as a
+/// separate argument.
+///
 /// **Note:** This function doesn't interpret argument 0 in any special way.
 /// If this function is intended to be used with command line arguments,
 /// `argv[0]` must be removed prior to calling it manually.
 #[allow(rustc::untranslatable_diagnostic)] // FIXME: make this translatable
-pub fn arg_expand_all(early_dcx: &EarlyDiagCtxt, at_args: &[String]) -> Vec<String> {
+pub fn arg_expand_all(
+    early_dcx: &EarlyDiagCtxt,
+    at_args: &[String],
+) -> Result<Vec<String>, ErrorGuaranteed> {
     let mut expander = Expander::default();
+    let mut result = Ok(());
     for arg in at_args {
         if let Err(err) = expander.arg(arg) {
-            early_dcx.early_fatal(format!("Failed to load argument file: {err}"));
+            result = Err(early_dcx.early_err(format!("failed to load argument file: {err}")));
         }
     }
-    expander.finish()
+    result.map(|()| expander.finish())
+}
+
+/// Gets the raw unprocessed command-line arguments as Unicode strings, without doing any further
+/// processing (e.g., without `@file` expansion).
+///
+/// This function is identical to [`env::args()`] except that it emits an error when it encounters
+/// non-Unicode arguments instead of panicking.
+pub fn raw_args(early_dcx: &EarlyDiagCtxt) -> Result<Vec<String>, ErrorGuaranteed> {
+    let mut res = Ok(Vec::new());
+    for (i, arg) in env::args_os().enumerate() {
+        match arg.into_string() {
+            Ok(arg) => {
+                if let Ok(args) = &mut res {
+                    args.push(arg);
+                }
+            }
+            Err(arg) => {
+                res =
+                    Err(early_dcx.early_err(format!("argument {i} is not valid Unicode: {arg:?}")))
+            }
+        }
+    }
+    res
 }
 
 #[derive(Debug)]
-pub enum Error {
-    Utf8Error(Option<String>),
+enum Error {
+    Utf8Error(String),
     IOError(String, io::Error),
     ShellParseError(String),
 }
@@ -118,10 +146,9 @@ pub enum Error {
 impl fmt::Display for Error {
     fn fmt(&self, fmt: &mut fmt::Formatter<'_>) -> fmt::Result {
         match self {
-            Error::Utf8Error(None) => write!(fmt, "Utf8 error"),
-            Error::Utf8Error(Some(path)) => write!(fmt, "Utf8 error in {path}"),
-            Error::IOError(path, err) => write!(fmt, "IO Error: {path}: {err}"),
-            Error::ShellParseError(path) => write!(fmt, "Invalid shell-style arguments in {path}"),
+            Error::Utf8Error(path) => write!(fmt, "UTF-8 error in {path}"),
+            Error::IOError(path, err) => write!(fmt, "IO error: {path}: {err}"),
+            Error::ShellParseError(path) => write!(fmt, "invalid shell-style arguments in {path}"),
         }
     }
 }

compiler/rustc_driver_impl/src/lib.rs

@@ -292,7 +292,7 @@ fn run_compiler(
     // the compiler with @empty_file as argv[0] and no more arguments.
     let at_args = at_args.get(1..).unwrap_or_default();
 
-    let args = args::arg_expand_all(&default_early_dcx, at_args);
+    let args = args::arg_expand_all(&default_early_dcx, at_args)?;
 
     let Some(matches) = handle_options(&default_early_dcx, &args) else { return Ok(()) };
 
@@ -1515,15 +1515,7 @@ pub fn main() -> ! {
     let mut callbacks = TimePassesCallbacks::default();
     let using_internal_features = install_ice_hook(DEFAULT_BUG_REPORT_URL, |_| ());
     let exit_code = catch_with_exit_code(|| {
-        let args = env::args_os()
-            .enumerate()
-            .map(|(i, arg)| {
-                arg.into_string().unwrap_or_else(|arg| {
-                    early_dcx.early_fatal(format!("argument {i} is not valid Unicode: {arg:?}"))
-                })
-            })
-            .collect::<Vec<_>>();
-        RunCompiler::new(&args, &mut callbacks)
+        RunCompiler::new(&args::raw_args(&early_dcx)?, &mut callbacks)
             .set_using_internal_features(using_internal_features)
             .run()
     });