Rollup merge of rust-lang#139710 - thaliaarchi:move-args-pal, r=joboet

Move `args` into `std::sys`

Move platform definitions of `args` into `std::sys`, as part of rust-lang#117276.

cc ``@joboet``

Author: ChrisDenton

Diff for: library/std/src/sys/args/common.rs

@@ -0,0 +1,43 @@
+use crate::ffi::OsString;
+use crate::{fmt, vec};
+
+pub struct Args {
+    iter: vec::IntoIter<OsString>,
+}
+
+impl !Send for Args {}
+impl !Sync for Args {}
+
+impl Args {
+    pub(super) fn new(args: Vec<OsString>) -> Self {
+        Args { iter: args.into_iter() }
+    }
+}
+
+impl fmt::Debug for Args {
+    fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result {
+        self.iter.as_slice().fmt(f)
+    }
+}
+
+impl Iterator for Args {
+    type Item = OsString;
+    fn next(&mut self) -> Option<OsString> {
+        self.iter.next()
+    }
+    fn size_hint(&self) -> (usize, Option<usize>) {
+        self.iter.size_hint()
+    }
+}
+
+impl ExactSizeIterator for Args {
+    fn len(&self) -> usize {
+        self.iter.len()
+    }
+}
+
+impl DoubleEndedIterator for Args {
+    fn next_back(&mut self) -> Option<OsString> {
+        self.iter.next_back()
+    }
+}

Diff for: library/std/src/sys/pal/hermit/args.rs renamed to library/std/src/sys/args/hermit.rs

@@ -1,8 +1,12 @@
 use crate::ffi::{CStr, OsString, c_char};
 use crate::os::hermit::ffi::OsStringExt;
+use crate::ptr;
 use crate::sync::atomic::Ordering::{Acquire, Relaxed, Release};
 use crate::sync::atomic::{AtomicIsize, AtomicPtr};
-use crate::{fmt, ptr, vec};
+
+#[path = "common.rs"]
+mod common;
+pub use common::Args;
 
 static ARGC: AtomicIsize = AtomicIsize::new(0);
 static ARGV: AtomicPtr<*const u8> = AtomicPtr::new(ptr::null_mut());
@@ -27,40 +31,5 @@ pub fn args() -> Args {
         })
         .collect();
 
-    Args { iter: args.into_iter() }
-}
-
-pub struct Args {
-    iter: vec::IntoIter<OsString>,
-}
-
-impl fmt::Debug for Args {
-    fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result {
-        self.iter.as_slice().fmt(f)
-    }
-}
-
-impl !Send for Args {}
-impl !Sync for Args {}
-
-impl Iterator for Args {
-    type Item = OsString;
-    fn next(&mut self) -> Option<OsString> {
-        self.iter.next()
-    }
-    fn size_hint(&self) -> (usize, Option<usize>) {
-        self.iter.size_hint()
-    }
-}
-
-impl ExactSizeIterator for Args {
-    fn len(&self) -> usize {
-        self.iter.len()
-    }
-}
-
-impl DoubleEndedIterator for Args {
-    fn next_back(&mut self) -> Option<OsString> {
-        self.iter.next_back()
-    }
+    Args::new(args)
 }

Diff for: library/std/src/sys/args/mod.rs

@@ -0,0 +1,34 @@
+//! Platform-dependent command line arguments abstraction.
+
+#![forbid(unsafe_op_in_unsafe_fn)]
+
+cfg_if::cfg_if! {
+    if #[cfg(all(target_family = "unix", not(any(target_os = "espidf", target_os = "vita"))))] {
+        mod unix;
+        pub use unix::*;
+    } else if #[cfg(target_family = "windows")] {
+        mod windows;
+        pub use windows::*;
+    } else if #[cfg(target_os = "hermit")] {
+        mod hermit;
+        pub use hermit::*;
+    } else if #[cfg(all(target_vendor = "fortanix", target_env = "sgx"))] {
+        mod sgx;
+        pub use sgx::*;
+    } else if #[cfg(target_os = "uefi")] {
+        mod uefi;
+        pub use uefi::*;
+    } else if #[cfg(target_os = "wasi")] {
+        mod wasi;
+        pub use wasi::*;
+    } else if #[cfg(target_os = "xous")] {
+        mod xous;
+        pub use xous::*;
+    } else if #[cfg(target_os = "zkvm")] {
+        mod zkvm;
+        pub use zkvm::*;
+    } else {
+        mod unsupported;
+        pub use unsupported::*;
+    }
+}

Diff for: library/std/src/sys/pal/sgx/args.rs renamed to library/std/src/sys/args/sgx.rs

@@ -1,8 +1,10 @@
-use super::abi::usercalls::alloc;
-use super::abi::usercalls::raw::ByteBuffer;
+#![allow(fuzzy_provenance_casts)] // FIXME: this module systematically confuses pointers and integers
+
 use crate::ffi::OsString;
 use crate::sync::atomic::{AtomicUsize, Ordering};
 use crate::sys::os_str::Buf;
+use crate::sys::pal::abi::usercalls::alloc;
+use crate::sys::pal::abi::usercalls::raw::ByteBuffer;
 use crate::sys_common::FromInner;
 use crate::{fmt, slice};
 

Diff for: library/std/src/sys/pal/uefi/args.rs renamed to library/std/src/sys/args/uefi.rs

@@ -1,14 +1,13 @@
 use r_efi::protocols::loaded_image;
 
-use super::helpers;
 use crate::env::current_exe;
 use crate::ffi::OsString;
 use crate::iter::Iterator;
-use crate::{fmt, vec};
+use crate::sys::pal::helpers;
 
-pub struct Args {
-    parsed_args_list: vec::IntoIter<OsString>,
-}
+#[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()]);
@@ -22,51 +21,17 @@ pub fn args() -> Args {
     let lp_size = unsafe { (*protocol.as_ptr()).load_options_size } as usize;
     // Break if we are sure that it cannot be UTF-16
     if lp_size < size_of::<u16>() || lp_size % size_of::<u16>() != 0 {
-        return Args { parsed_args_list: lazy_current_exe().into_iter() };
+        return Args::new(lazy_current_exe());
     }
     let lp_size = lp_size / size_of::<u16>();
 
     let lp_cmd_line = unsafe { (*protocol.as_ptr()).load_options as *const u16 };
     if !lp_cmd_line.is_aligned() {
-        return Args { parsed_args_list: lazy_current_exe().into_iter() };
+        return Args::new(lazy_current_exe());
     }
     let lp_cmd_line = unsafe { crate::slice::from_raw_parts(lp_cmd_line, lp_size) };
 
-    Args {
-        parsed_args_list: parse_lp_cmd_line(lp_cmd_line)
-            .unwrap_or_else(lazy_current_exe)
-            .into_iter(),
-    }
-}
-
-impl fmt::Debug for Args {
-    fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result {
-        self.parsed_args_list.as_slice().fmt(f)
-    }
-}
-
-impl Iterator for Args {
-    type Item = OsString;
-
-    fn next(&mut self) -> Option<OsString> {
-        self.parsed_args_list.next()
-    }
-
-    fn size_hint(&self) -> (usize, Option<usize>) {
-        self.parsed_args_list.size_hint()
-    }
-}
-
-impl ExactSizeIterator for Args {
-    fn len(&self) -> usize {
-        self.parsed_args_list.len()
-    }
-}
-
-impl DoubleEndedIterator for Args {
-    fn next_back(&mut self) -> Option<OsString> {
-        self.parsed_args_list.next_back()
-    }
+    Args::new(parse_lp_cmd_line(lp_cmd_line).unwrap_or_else(lazy_current_exe))
 }
 
 /// Implements the UEFI command-line argument parsing algorithm.

Diff for: library/std/src/sys/pal/unix/args.rs renamed to library/std/src/sys/args/unix.rs

@@ -5,13 +5,16 @@
 
 #![allow(dead_code)] // runtime init functions not used during testing
 
-use crate::ffi::{CStr, OsString};
+use crate::ffi::CStr;
 use crate::os::unix::ffi::OsStringExt;
-use crate::{fmt, vec};
+
+#[path = "common.rs"]
+mod common;
+pub use common::Args;
 
 /// One-time global initialization.
 pub unsafe fn init(argc: isize, argv: *const *const u8) {
-    imp::init(argc, argv)
+    unsafe { imp::init(argc, argv) }
 }
 
 /// Returns the command line arguments
@@ -55,42 +58,7 @@ pub fn args() -> Args {
         vec.push(OsStringExt::from_vec(cstr.to_bytes().to_vec()));
     }
 
-    Args { iter: vec.into_iter() }
-}
-
-pub struct Args {
-    iter: vec::IntoIter<OsString>,
-}
-
-impl !Send for Args {}
-impl !Sync for Args {}
-
-impl fmt::Debug for Args {
-    fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result {
-        self.iter.as_slice().fmt(f)
-    }
-}
-
-impl Iterator for Args {
-    type Item = OsString;
-    fn next(&mut self) -> Option<OsString> {
-        self.iter.next()
-    }
-    fn size_hint(&self) -> (usize, Option<usize>) {
-        self.iter.size_hint()
-    }
-}
-
-impl ExactSizeIterator for Args {
-    fn len(&self) -> usize {
-        self.iter.len()
-    }
-}
-
-impl DoubleEndedIterator for Args {
-    fn next_back(&mut self) -> Option<OsString> {
-        self.iter.next_back()
-    }
+    Args::new(vec)
 }
 
 #[cfg(any(
@@ -141,7 +109,7 @@ mod imp {
     pub unsafe fn init(argc: isize, argv: *const *const u8) {
         // on GNU/Linux if we are main then we will init argv and argc twice, it "duplicates work"
         // BUT edge-cases are real: only using .init_array can break most emulators, dlopen, etc.
-        really_init(argc, argv);
+        unsafe { really_init(argc, argv) };
     }
 
     /// glibc passes argc, argv, and envp to functions in .init_array, as a non-standard extension.
@@ -159,9 +127,7 @@ mod imp {
             argv: *const *const u8,
             _envp: *const *const u8,
         ) {
-            unsafe {
-                really_init(argc as isize, argv);
-            }
+            unsafe { really_init(argc as isize, argv) };
         }
         init_wrapper
     };
@@ -228,16 +194,3 @@ mod imp {
         (argc as isize, argv.cast())
     }
 }
-
-#[cfg(any(target_os = "espidf", target_os = "vita"))]
-mod imp {
-    use crate::ffi::c_char;
-    use crate::ptr;
-
-    #[inline(always)]
-    pub unsafe fn init(_argc: isize, _argv: *const *const u8) {}
-
-    pub fn argc_argv() -> (isize, *const *const c_char) {
-        (0, ptr::null())
-    }
-}

Diff for: library/std/src/sys/pal/unsupported/args.rs renamed to library/std/src/sys/args/unsupported.rs

@@ -0,0 +1,29 @@
+#![forbid(unsafe_op_in_unsafe_fn)]
+
+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()))
+}
+
+fn maybe_args() -> Option<Vec<OsString>> {
+    unsafe {
+        let (argc, buf_size) = wasi::args_sizes_get().ok()?;
+        let mut argv = Vec::with_capacity(argc);
+        let mut buf = Vec::with_capacity(buf_size);
+        wasi::args_get(argv.as_mut_ptr(), buf.as_mut_ptr()).ok()?;
+        argv.set_len(argc);
+        let mut ret = Vec::with_capacity(argc);
+        for ptr in argv {
+            let s = CStr::from_ptr(ptr.cast());
+            ret.push(OsStr::from_bytes(s.to_bytes()).to_owned());
+        }
+        Some(ret)
+    }
+}