Auto merge of rust-lang#124716 - matthiaskrgr:rollup-ni58ie1, r=matthiaskrgr

Rollup of 4 pull requests

Successful merges:

 - rust-lang#122441 (Improve several `Read` implementations)
 - rust-lang#124584 (Various improvements to entrypoint code)
 - rust-lang#124699 (Use `unchecked_sub` in `split_at`)
 - rust-lang#124715 (interpret, miri: uniform treatments of intrinsics/functions with and without return block)

r? `@ghost`
`@rustbot` modify labels: rollup

Author: bors

compiler/rustc_ast/src/entry.rs

@@ -4,11 +4,35 @@ use rustc_span::Symbol;
 
 #[derive(Debug)]
 pub enum EntryPointType {
+    /// This function is not an entrypoint.
     None,
+    /// This is a function called `main` at the root level.
+    /// ```
+    /// fn main() {}
+    /// ```
     MainNamed,
+    /// This is a function with the `#[rustc_main]` attribute.
+    /// Used by the testing harness to create the test entrypoint.
+    /// ```ignore (clashes with test entrypoint)
+    /// #[rustc_main]
+    /// fn main() {}
+    /// ```
     RustcMainAttr,
+    /// This is a function with the `#[start]` attribute.
+    /// ```ignore (clashes with test entrypoint)
+    /// #[start]
+    /// fn main() {}
+    /// ```
     Start,
-    OtherMain, // Not an entry point, but some other function named main
+    /// This function is **not** an entrypoint but simply named `main` (not at the root).
+    /// This is only used for diagnostics.
+    /// ```
+    /// #[allow(dead_code)]
+    /// mod meow {
+    ///     fn main() {}
+    /// }
+    /// ```
+    OtherMain,
 }
 
 pub fn entry_point_type(

compiler/rustc_builtin_macros/src/test_harness.rs

@@ -266,7 +266,7 @@ fn generate_test_harness(
 ///
 /// By default this expands to
 ///
-/// ```ignore UNSOLVED (I think I still need guidance for this one. Is it correct? Do we try to make it run? How do we nicely fill it out?)
+/// ```ignore (messes with test internals)
 /// #[rustc_main]
 /// pub fn main() {
 ///     extern crate test;

compiler/rustc_const_eval/src/const_eval/machine.rs

@@ -467,19 +467,6 @@ impl<'mir, 'tcx> interpret::Machine<'mir, 'tcx> for CompileTimeInterpreter<'mir,
         let intrinsic_name = ecx.tcx.item_name(instance.def_id());
 
         // CTFE-specific intrinsics.
-        let Some(ret) = target else {
-            // Handle diverging intrinsics. We can't handle any of them (that are not already
-            // handled above), but check if there is a fallback body.
-            if ecx.tcx.intrinsic(instance.def_id()).unwrap().must_be_overridden {
-                throw_unsup_format!(
-                    "intrinsic `{intrinsic_name}` is not supported at compile-time"
-                );
-            }
-            return Ok(Some(ty::Instance {
-                def: ty::InstanceDef::Item(instance.def_id()),
-                args: instance.args,
-            }));
-        };
         match intrinsic_name {
             sym::ptr_guaranteed_cmp => {
                 let a = ecx.read_scalar(&args[0])?;
@@ -559,7 +546,8 @@ impl<'mir, 'tcx> interpret::Machine<'mir, 'tcx> for CompileTimeInterpreter<'mir,
             }
         }
 
-        ecx.go_to_block(ret);
+        // Intrinsic is done, jump to next block.
+        ecx.return_to_block(target)?;
         Ok(None)
     }
 

compiler/rustc_const_eval/src/interpret/intrinsics.rs

@@ -113,10 +113,6 @@ impl<'mir, 'tcx: 'mir, M: Machine<'mir, 'tcx>> InterpCx<'mir, 'tcx, M> {
     ) -> InterpResult<'tcx, bool> {
         let instance_args = instance.args;
         let intrinsic_name = self.tcx.item_name(instance.def_id());
-        let Some(ret) = ret else {
-            // We don't support any intrinsic without return place.
-            return Ok(false);
-        };
 
         match intrinsic_name {
             sym::caller_location => {
@@ -376,7 +372,7 @@ impl<'mir, 'tcx: 'mir, M: Machine<'mir, 'tcx>> InterpCx<'mir, 'tcx, M> {
                     };
 
                     M::panic_nounwind(self, &msg)?;
-                    // Skip the `go_to_block` at the end.
+                    // Skip the `return_to_block` at the end (we panicked, we do not return).
                     return Ok(true);
                 }
             }
@@ -437,11 +433,12 @@ impl<'mir, 'tcx: 'mir, M: Machine<'mir, 'tcx>> InterpCx<'mir, 'tcx, M> {
                 self.write_scalar(Scalar::from_target_usize(align.bytes(), self), dest)?;
             }
 
+            // Unsupported intrinsic: skip the return_to_block below.
             _ => return Ok(false),
         }
 
         trace!("{:?}", self.dump_place(&dest.clone().into()));
-        self.go_to_block(ret);
+        self.return_to_block(ret)?;
         Ok(true)
     }
 

compiler/rustc_passes/messages.ftl

@@ -49,12 +49,6 @@ passes_attr_crate_level =
 passes_attr_only_in_functions =
     `{$attr}` attribute can only be used on functions
 
-passes_attr_only_on_main =
-    `{$attr}` attribute can only be used on `fn main()`
-
-passes_attr_only_on_root_main =
-    `{$attr}` attribute can only be used on root `fn main()`
-
 passes_both_ffi_const_and_pure =
     `#[ffi_const]` function cannot be `#[ffi_pure]`
 

compiler/rustc_passes/src/entry.rs

@@ -18,10 +18,10 @@ use crate::errors::{
 struct EntryContext<'tcx> {
     tcx: TyCtxt<'tcx>,
 
-    /// The function that has attribute named `main`.
-    attr_main_fn: Option<(LocalDefId, Span)>,
+    /// The function has the `#[rustc_main]` attribute.
+    rustc_main_fn: Option<(LocalDefId, Span)>,
 
-    /// The function that has the attribute 'start' on it.
+    /// The function that has the attribute `#[start]` on it.
     start_fn: Option<(LocalDefId, Span)>,
 
     /// The functions that one might think are `main` but aren't, e.g.
@@ -42,10 +42,10 @@ fn entry_fn(tcx: TyCtxt<'_>, (): ()) -> Option<(DefId, EntryFnType)> {
     }
 
     let mut ctxt =
-        EntryContext { tcx, attr_main_fn: None, start_fn: None, non_main_fns: Vec::new() };
+        EntryContext { tcx, rustc_main_fn: None, start_fn: None, non_main_fns: Vec::new() };
 
     for id in tcx.hir().items() {
-        find_item(id, &mut ctxt);
+        check_and_search_item(id, &mut ctxt);
     }
 
     configure_main(tcx, &ctxt)
@@ -56,7 +56,16 @@ fn attr_span_by_symbol(ctxt: &EntryContext<'_>, id: ItemId, sym: Symbol) -> Opti
     attr::find_by_name(attrs, sym).map(|attr| attr.span)
 }
 
-fn find_item(id: ItemId, ctxt: &mut EntryContext<'_>) {
+fn check_and_search_item(id: ItemId, ctxt: &mut EntryContext<'_>) {
+    if !matches!(ctxt.tcx.def_kind(id.owner_id), DefKind::Fn) {
+        for attr in [sym::start, sym::rustc_main] {
+            if let Some(span) = attr_span_by_symbol(ctxt, id, attr) {
+                ctxt.tcx.dcx().emit_err(AttrOnlyInFunctions { span, attr });
+            }
+        }
+        return;
+    }
+
     let at_root = ctxt.tcx.opt_local_parent(id.owner_id.def_id) == Some(CRATE_DEF_ID);
 
     let attrs = ctxt.tcx.hir().attrs(id.hir_id());
@@ -65,26 +74,20 @@ fn find_item(id: ItemId, ctxt: &mut EntryContext<'_>) {
         at_root,
         ctxt.tcx.opt_item_name(id.owner_id.to_def_id()),
     );
+
     match entry_point_type {
-        EntryPointType::None => (),
-        _ if !matches!(ctxt.tcx.def_kind(id.owner_id), DefKind::Fn) => {
-            for attr in [sym::start, sym::rustc_main] {
-                if let Some(span) = attr_span_by_symbol(ctxt, id, attr) {
-                    ctxt.tcx.dcx().emit_err(AttrOnlyInFunctions { span, attr });
-                }
-            }
-        }
-        EntryPointType::MainNamed => (),
+        EntryPointType::None => {}
+        EntryPointType::MainNamed => {}
         EntryPointType::OtherMain => {
             ctxt.non_main_fns.push(ctxt.tcx.def_span(id.owner_id));
         }
         EntryPointType::RustcMainAttr => {
-            if ctxt.attr_main_fn.is_none() {
-                ctxt.attr_main_fn = Some((id.owner_id.def_id, ctxt.tcx.def_span(id.owner_id)));
+            if ctxt.rustc_main_fn.is_none() {
+                ctxt.rustc_main_fn = Some((id.owner_id.def_id, ctxt.tcx.def_span(id.owner_id)));
             } else {
                 ctxt.tcx.dcx().emit_err(MultipleRustcMain {
                     span: ctxt.tcx.def_span(id.owner_id.to_def_id()),
-                    first: ctxt.attr_main_fn.unwrap().1,
+                    first: ctxt.rustc_main_fn.unwrap().1,
                     additional: ctxt.tcx.def_span(id.owner_id.to_def_id()),
                 });
             }
@@ -107,10 +110,11 @@ fn find_item(id: ItemId, ctxt: &mut EntryContext<'_>) {
 fn configure_main(tcx: TyCtxt<'_>, visitor: &EntryContext<'_>) -> Option<(DefId, EntryFnType)> {
     if let Some((def_id, _)) = visitor.start_fn {
         Some((def_id.to_def_id(), EntryFnType::Start))
-    } else if let Some((local_def_id, _)) = visitor.attr_main_fn {
+    } else if let Some((local_def_id, _)) = visitor.rustc_main_fn {
         let def_id = local_def_id.to_def_id();
         Some((def_id, EntryFnType::Main { sigpipe: sigpipe(tcx) }))
     } else {
+        // The actual resolution of main happens in the resolver, this here
         if let Some(main_def) = tcx.resolutions(()).main_def
             && let Some(def_id) = main_def.opt_fn_def_id()
         {

compiler/rustc_passes/src/errors.rs

@@ -1206,22 +1206,6 @@ pub struct NakedFunctionsMustUseNoreturn {
     pub last_span: Span,
 }
 
-#[derive(Diagnostic)]
-#[diag(passes_attr_only_on_main)]
-pub struct AttrOnlyOnMain {
-    #[primary_span]
-    pub span: Span,
-    pub attr: Symbol,
-}
-
-#[derive(Diagnostic)]
-#[diag(passes_attr_only_on_root_main)]
-pub struct AttrOnlyOnRootMain {
-    #[primary_span]
-    pub span: Span,
-    pub attr: Symbol,
-}
-
 #[derive(Diagnostic)]
 #[diag(passes_attr_only_in_functions)]
 pub struct AttrOnlyInFunctions {

library/core/src/slice/mod.rs

@@ -9,7 +9,7 @@
 use crate::cmp::Ordering::{self, Equal, Greater, Less};
 use crate::fmt;
 use crate::hint;
-use crate::intrinsics::exact_div;
+use crate::intrinsics::{exact_div, unchecked_sub};
 use crate::mem::{self, SizedTypeProperties};
 use crate::num::NonZero;
 use crate::ops::{Bound, OneSidedRange, Range, RangeBounds};
@@ -1983,7 +1983,7 @@ impl<T> [T] {
         );
 
         // SAFETY: Caller has to check that `0 <= mid <= self.len()`
-        unsafe { (from_raw_parts(ptr, mid), from_raw_parts(ptr.add(mid), len - mid)) }
+        unsafe { (from_raw_parts(ptr, mid), from_raw_parts(ptr.add(mid), unchecked_sub(len, mid))) }
     }
 
     /// Divides one mutable slice into two at an index, without doing bounds checking.
@@ -2035,7 +2035,12 @@ impl<T> [T] {
         //
         // `[ptr; mid]` and `[mid; len]` are not overlapping, so returning a mutable reference
         // is fine.
-        unsafe { (from_raw_parts_mut(ptr, mid), from_raw_parts_mut(ptr.add(mid), len - mid)) }
+        unsafe {
+            (
+                from_raw_parts_mut(ptr, mid),
+                from_raw_parts_mut(ptr.add(mid), unchecked_sub(len, mid)),
+            )
+        }
     }
 
     /// Divides one slice into two at an index, returning `None` if the slice is

library/std/src/io/cursor.rs

@@ -364,6 +364,27 @@ where
         self.pos += n as u64;
         Ok(())
     }
+
+    fn read_to_end(&mut self, buf: &mut Vec<u8>) -> io::Result<usize> {
+        let content = self.remaining_slice();
+        let len = content.len();
+        buf.try_reserve(len)?;
+        buf.extend_from_slice(content);
+        self.pos += len as u64;
+
+        Ok(len)
+    }
+
+    fn read_to_string(&mut self, buf: &mut String) -> io::Result<usize> {
+        let content =
+            crate::str::from_utf8(self.remaining_slice()).map_err(|_| io::Error::INVALID_UTF8)?;
+        let len = content.len();
+        buf.try_reserve(len)?;
+        buf.push_str(content);
+        self.pos += len as u64;
+
+        Ok(len)
+    }
 }
 
 #[stable(feature = "rust1", since = "1.0.0")]

library/std/src/io/impls.rs

@@ -329,8 +329,9 @@ impl Read for &[u8] {
     #[inline]
     fn read_to_string(&mut self, buf: &mut String) -> io::Result<usize> {
         let content = str::from_utf8(self).map_err(|_| io::Error::INVALID_UTF8)?;
-        buf.push_str(content);
         let len = self.len();
+        buf.try_reserve(len)?;
+        buf.push_str(content);
         *self = &self[len..];
         Ok(len)
     }
@@ -473,14 +474,8 @@ impl<A: Allocator> Read for VecDeque<u8, A> {
 
     #[inline]
     fn read_to_string(&mut self, buf: &mut String) -> io::Result<usize> {
-        // We have to use a single contiguous slice because the `VecDequeue` might be split in the
-        // middle of an UTF-8 character.
-        let len = self.len();
-        let content = self.make_contiguous();
-        let string = str::from_utf8(content).map_err(|_| io::Error::INVALID_UTF8)?;
-        buf.push_str(string);
-        self.clear();
-        Ok(len)
+        // SAFETY: We only append to the buffer
+        unsafe { io::append_to_string(buf, |buf| self.read_to_end(buf)) }
     }
 }
 

library/std/src/io/mod.rs

@@ -384,7 +384,10 @@ where
 {
     let mut g = Guard { len: buf.len(), buf: buf.as_mut_vec() };
     let ret = f(g.buf);
-    if str::from_utf8(&g.buf[g.len..]).is_err() {
+
+    // SAFETY: the caller promises to only append data to `buf`
+    let appended = g.buf.get_unchecked(g.len..);
+    if str::from_utf8(appended).is_err() {
         ret.and_then(|_| Err(Error::INVALID_UTF8))
     } else {
         g.len = g.buf.len();

library/std/src/process.rs

@@ -486,6 +486,10 @@ impl Read for ChildStderr {
     fn is_read_vectored(&self) -> bool {
         self.inner.is_read_vectored()
     }
+
+    fn read_to_end(&mut self, buf: &mut Vec<u8>) -> io::Result<usize> {
+        self.inner.read_to_end(buf)
+    }
 }
 
 impl AsInner<AnonPipe> for ChildStderr {

src/tools/miri/src/lib.rs

@@ -95,6 +95,7 @@ pub use rustc_const_eval::interpret::*;
 #[doc(no_inline)]
 pub use rustc_const_eval::interpret::{self, AllocMap, PlaceTy, Provenance as _};
 
+pub use crate::shims::EmulateItemResult;
 pub use crate::shims::env::{EnvVars, EvalContextExt as _};
 pub use crate::shims::foreign_items::{DynSym, EvalContextExt as _};
 pub use crate::shims::intrinsics::EvalContextExt as _;

src/tools/miri/src/shims/alloc.rs

@@ -4,7 +4,6 @@ use rustc_ast::expand::allocator::AllocatorKind;
 use rustc_target::abi::{Align, Size};
 
 use crate::*;
-use shims::foreign_items::EmulateForeignItemResult;
 
 /// Check some basic requirements for this allocation request:
 /// non-zero size, power-of-two alignment.
@@ -55,12 +54,12 @@ pub trait EvalContextExt<'mir, 'tcx: 'mir>: crate::MiriInterpCxExt<'mir, 'tcx> {
     fn emulate_allocator(
         &mut self,
         default: impl FnOnce(&mut MiriInterpCx<'mir, 'tcx>) -> InterpResult<'tcx>,
-    ) -> InterpResult<'tcx, EmulateForeignItemResult> {
+    ) -> InterpResult<'tcx, EmulateItemResult> {
         let this = self.eval_context_mut();
 
         let Some(allocator_kind) = this.tcx.allocator_kind(()) else {
             // in real code, this symbol does not exist without an allocator
-            return Ok(EmulateForeignItemResult::NotSupported);
+            return Ok(EmulateItemResult::NotSupported);
         };
 
         match allocator_kind {
@@ -70,11 +69,11 @@ pub trait EvalContextExt<'mir, 'tcx: 'mir>: crate::MiriInterpCxExt<'mir, 'tcx> {
                 // and not execute any Miri shim. Somewhat unintuitively doing so is done
                 // by returning `NotSupported`, which triggers the `lookup_exported_symbol`
                 // fallback case in `emulate_foreign_item`.
-                return Ok(EmulateForeignItemResult::NotSupported);
+                return Ok(EmulateItemResult::NotSupported);
             }
             AllocatorKind::Default => {
                 default(this)?;
-                Ok(EmulateForeignItemResult::NeedsJumping)
+                Ok(EmulateItemResult::NeedsJumping)
             }
         }
     }