Rollup merge of #133265 - the8472:extract-if-ranges, r=cuviper

Add a range argument to vec.extract_if

tracking issue: #43244

This adds the range argument requested in #43244 (comment)

Author: matthiaskrgr

compiler/rustc_errors/src/lib.rs

@@ -1569,18 +1569,18 @@ impl DiagCtxtInner {
                 debug!(?diagnostic);
                 debug!(?self.emitted_diagnostics);
 
-                let already_emitted_sub = |sub: &mut Subdiag| {
+                let not_yet_emitted = |sub: &mut Subdiag| {
                     debug!(?sub);
                     if sub.level != OnceNote && sub.level != OnceHelp {
-                        return false;
+                        return true;
                     }
                     let mut hasher = StableHasher::new();
                     sub.hash(&mut hasher);
                     let diagnostic_hash = hasher.finish();
                     debug!(?diagnostic_hash);
-                    !self.emitted_diagnostics.insert(diagnostic_hash)
+                    self.emitted_diagnostics.insert(diagnostic_hash)
                 };
-                diagnostic.children.extract_if(already_emitted_sub).for_each(|_| {});
+                diagnostic.children.retain_mut(not_yet_emitted);
                 if already_emitted {
                     let msg = "duplicate diagnostic emitted due to `-Z deduplicate-diagnostics=no`";
                     diagnostic.sub(Note, msg, MultiSpan::new());

compiler/rustc_lint/src/non_ascii_idents.rs

@@ -205,7 +205,7 @@ impl EarlyLintPass for NonAsciiIdents {
                     (IdentifierType::Not_NFKC, "Not_NFKC"),
                 ] {
                     let codepoints: Vec<_> =
-                        chars.extract_if(|(_, ty)| *ty == Some(id_ty)).collect();
+                        chars.extract_if(.., |(_, ty)| *ty == Some(id_ty)).collect();
                     if codepoints.is_empty() {
                         continue;
                     }
@@ -217,7 +217,7 @@ impl EarlyLintPass for NonAsciiIdents {
                 }
 
                 let remaining = chars
-                    .extract_if(|(c, _)| !GeneralSecurityProfile::identifier_allowed(*c))
+                    .extract_if(.., |(c, _)| !GeneralSecurityProfile::identifier_allowed(*c))
                     .collect::<Vec<_>>();
                 if !remaining.is_empty() {
                     cx.emit_span_lint(UNCOMMON_CODEPOINTS, sp, IdentifierUncommonCodepoints {

compiler/rustc_metadata/src/native_libs.rs

@@ -544,7 +544,7 @@ impl<'tcx> Collector<'tcx> {
             // can move them to the end of the list below.
             let mut existing = self
                 .libs
-                .extract_if(|lib| {
+                .extract_if(.., |lib| {
                     if lib.name.as_str() == passed_lib.name {
                         // FIXME: This whole logic is questionable, whether modifiers are
                         // involved or not, library reordering and kind overriding without

compiler/rustc_middle/src/ty/diagnostics.rs

@@ -309,7 +309,7 @@ pub fn suggest_constraining_type_params<'a>(
         let Some(param) = param else { return false };
 
         {
-            let mut sized_constraints = constraints.extract_if(|(_, def_id, _)| {
+            let mut sized_constraints = constraints.extract_if(.., |(_, def_id, _)| {
                 def_id.is_some_and(|def_id| tcx.is_lang_item(def_id, LangItem::Sized))
             });
             if let Some((_, def_id, _)) = sized_constraints.next() {

compiler/rustc_resolve/src/diagnostics.rs

@@ -2817,11 +2817,11 @@ fn show_candidates(
         path_strings.sort_by(|a, b| a.0.cmp(&b.0));
         path_strings.dedup_by(|a, b| a.0 == b.0);
         let core_path_strings =
-            path_strings.extract_if(|p| p.0.starts_with("core::")).collect::<Vec<_>>();
+            path_strings.extract_if(.., |p| p.0.starts_with("core::")).collect::<Vec<_>>();
         let std_path_strings =
-            path_strings.extract_if(|p| p.0.starts_with("std::")).collect::<Vec<_>>();
+            path_strings.extract_if(.., |p| p.0.starts_with("std::")).collect::<Vec<_>>();
         let foreign_crate_path_strings =
-            path_strings.extract_if(|p| !p.0.starts_with("crate::")).collect::<Vec<_>>();
+            path_strings.extract_if(.., |p| !p.0.starts_with("crate::")).collect::<Vec<_>>();
 
         // We list the `crate` local paths first.
         // Then we list the `std`/`core` paths.

compiler/rustc_resolve/src/late/diagnostics.rs

@@ -629,7 +629,7 @@ impl<'ast, 'ra: 'ast, 'tcx> LateResolutionVisitor<'_, 'ast, 'ra, 'tcx> {
         // Try to filter out intrinsics candidates, as long as we have
         // some other candidates to suggest.
         let intrinsic_candidates: Vec<_> = candidates
-            .extract_if(|sugg| {
+            .extract_if(.., |sugg| {
                 let path = path_names_to_string(&sugg.path);
                 path.starts_with("core::intrinsics::") || path.starts_with("std::intrinsics::")
             })

compiler/rustc_trait_selection/src/traits/mod.rs

@@ -447,7 +447,7 @@ pub fn normalize_param_env_or_error<'tcx>(
     // This works fairly well because trait matching does not actually care about param-env
     // TypeOutlives predicates - these are normally used by regionck.
     let outlives_predicates: Vec<_> = predicates
-        .extract_if(|predicate| {
+        .extract_if(.., |predicate| {
             matches!(predicate.kind().skip_binder(), ty::ClauseKind::TypeOutlives(..))
         })
         .collect();

library/alloc/src/collections/linked_list.rs

@@ -1939,9 +1939,7 @@ pub struct ExtractIf<
     T: 'a,
     F: 'a,
     #[unstable(feature = "allocator_api", issue = "32838")] A: Allocator = Global,
-> where
-    F: FnMut(&mut T) -> bool,
-{
+> {
     list: &'a mut LinkedList<T, A>,
     it: Option<NonNull<Node<T>>>,
     pred: F,
@@ -1979,10 +1977,7 @@ where
 }
 
 #[unstable(feature = "extract_if", reason = "recently added", issue = "43244")]
-impl<T: fmt::Debug, F> fmt::Debug for ExtractIf<'_, T, F>
-where
-    F: FnMut(&mut T) -> bool,
-{
+impl<T: fmt::Debug, F> fmt::Debug for ExtractIf<'_, T, F> {
     fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result {
         f.debug_tuple("ExtractIf").field(&self.list).finish()
     }

library/alloc/src/vec/extract_if.rs

@@ -1,3 +1,4 @@
+use core::ops::{Range, RangeBounds};
 use core::{ptr, slice};
 
 use super::Vec;
@@ -14,7 +15,7 @@ use crate::alloc::{Allocator, Global};
 /// #![feature(extract_if)]
 ///
 /// let mut v = vec![0, 1, 2];
-/// let iter: std::vec::ExtractIf<'_, _, _> = v.extract_if(|x| *x % 2 == 0);
+/// let iter: std::vec::ExtractIf<'_, _, _> = v.extract_if(.., |x| *x % 2 == 0);
 /// ```
 #[unstable(feature = "extract_if", reason = "recently added", issue = "43244")]
 #[derive(Debug)]
@@ -24,24 +25,32 @@ pub struct ExtractIf<
     T,
     F,
     #[unstable(feature = "allocator_api", issue = "32838")] A: Allocator = Global,
-> where
-    F: FnMut(&mut T) -> bool,
-{
-    pub(super) vec: &'a mut Vec<T, A>,
+> {
+    vec: &'a mut Vec<T, A>,
     /// The index of the item that will be inspected by the next call to `next`.
-    pub(super) idx: usize,
+    idx: usize,
+    /// Elements at and beyond this point will be retained. Must be equal or smaller than `old_len`.
+    end: usize,
     /// The number of items that have been drained (removed) thus far.
-    pub(super) del: usize,
+    del: usize,
     /// The original length of `vec` prior to draining.
-    pub(super) old_len: usize,
+    old_len: usize,
     /// The filter test predicate.
-    pub(super) pred: F,
+    pred: F,
 }
 
-impl<T, F, A: Allocator> ExtractIf<'_, T, F, A>
-where
-    F: FnMut(&mut T) -> bool,
-{
+impl<'a, T, F, A: Allocator> ExtractIf<'a, T, F, A> {
+    pub(super) fn new<R: RangeBounds<usize>>(vec: &'a mut Vec<T, A>, pred: F, range: R) -> Self {
+        let old_len = vec.len();
+        let Range { start, end } = slice::range(range, ..old_len);
+
+        // Guard against the vec getting leaked (leak amplification)
+        unsafe {
+            vec.set_len(0);
+        }
+        ExtractIf { vec, idx: start, del: 0, end, old_len, pred }
+    }
+
     /// Returns a reference to the underlying allocator.
     #[unstable(feature = "allocator_api", issue = "32838")]
     #[inline]
@@ -59,7 +68,7 @@ where
 
     fn next(&mut self) -> Option<T> {
         unsafe {
-            while self.idx < self.old_len {
+            while self.idx < self.end {
                 let i = self.idx;
                 let v = slice::from_raw_parts_mut(self.vec.as_mut_ptr(), self.old_len);
                 let drained = (self.pred)(&mut v[i]);
@@ -82,24 +91,15 @@ where
     }
 
     fn size_hint(&self) -> (usize, Option<usize>) {
-        (0, Some(self.old_len - self.idx))
+        (0, Some(self.end - self.idx))
     }
 }
 
 #[unstable(feature = "extract_if", reason = "recently added", issue = "43244")]
-impl<T, F, A: Allocator> Drop for ExtractIf<'_, T, F, A>
-where
-    F: FnMut(&mut T) -> bool,
-{
+impl<T, F, A: Allocator> Drop for ExtractIf<'_, T, F, A> {
     fn drop(&mut self) {
         unsafe {
             if self.idx < self.old_len && self.del > 0 {
-                // This is a pretty messed up state, and there isn't really an
-                // obviously right thing to do. We don't want to keep trying
-                // to execute `pred`, so we just backshift all the unprocessed
-                // elements and tell the vec that they still exist. The backshift
-                // is required to prevent a double-drop of the last successfully
-                // drained item prior to a panic in the predicate.
                 let ptr = self.vec.as_mut_ptr();
                 let src = ptr.add(self.idx);
                 let dst = src.sub(self.del);

library/alloc/src/vec/mod.rs

@@ -3615,12 +3615,15 @@ impl<T, A: Allocator> Vec<T, A> {
         Splice { drain: self.drain(range), replace_with: replace_with.into_iter() }
     }
 
-    /// Creates an iterator which uses a closure to determine if an element should be removed.
+    /// Creates an iterator which uses a closure to determine if element in the range should be removed.
     ///
     /// If the closure returns true, then the element is removed and yielded.
     /// If the closure returns false, the element will remain in the vector and will not be yielded
     /// by the iterator.
     ///
+    /// Only elements that fall in the provided range are considered for extraction, but any elements
+    /// after the range will still have to be moved if any element has been extracted.
+    ///
     /// If the returned `ExtractIf` is not exhausted, e.g. because it is dropped without iterating
     /// or the iteration short-circuits, then the remaining elements will be retained.
     /// Use [`retain`] with a negated predicate if you do not need the returned iterator.
@@ -3630,10 +3633,12 @@ impl<T, A: Allocator> Vec<T, A> {
     /// Using this method is equivalent to the following code:
     ///
     /// ```
+    /// # use std::cmp::min;
     /// # let some_predicate = |x: &mut i32| { *x == 2 || *x == 3 || *x == 6 };
     /// # let mut vec = vec![1, 2, 3, 4, 5, 6];
-    /// let mut i = 0;
-    /// while i < vec.len() {
+    /// # let range = 1..4;
+    /// let mut i = range.start;
+    /// while i < min(vec.len(), range.end) {
     ///     if some_predicate(&mut vec[i]) {
     ///         let val = vec.remove(i);
     ///         // your code here
@@ -3648,8 +3653,12 @@ impl<T, A: Allocator> Vec<T, A> {
     /// But `extract_if` is easier to use. `extract_if` is also more efficient,
     /// because it can backshift the elements of the array in bulk.
     ///
-    /// Note that `extract_if` also lets you mutate every element in the filter closure,
-    /// regardless of whether you choose to keep or remove it.
+    /// Note that `extract_if` also lets you mutate the elements passed to the filter closure,
+    /// regardless of whether you choose to keep or remove them.
+    ///
+    /// # Panics
+    ///
+    /// If `range` is out of bounds.
     ///
     /// # Examples
     ///
@@ -3659,25 +3668,29 @@ impl<T, A: Allocator> Vec<T, A> {
     /// #![feature(extract_if)]
     /// let mut numbers = vec![1, 2, 3, 4, 5, 6, 8, 9, 11, 13, 14, 15];
     ///
-    /// let evens = numbers.extract_if(|x| *x % 2 == 0).collect::<Vec<_>>();
+    /// let evens = numbers.extract_if(.., |x| *x % 2 == 0).collect::<Vec<_>>();
     /// let odds = numbers;
     ///
     /// assert_eq!(evens, vec![2, 4, 6, 8, 14]);
     /// assert_eq!(odds, vec![1, 3, 5, 9, 11, 13, 15]);
     /// ```
+    ///
+    /// Using the range argument to only process a part of the vector:
+    ///
+    /// ```
+    /// #![feature(extract_if)]
+    /// let mut items = vec![0, 0, 0, 0, 0, 0, 0, 1, 2, 1, 2, 1, 2];
+    /// let ones = items.extract_if(7.., |x| *x == 1).collect::<Vec<_>>();
+    /// assert_eq!(items, vec![0, 0, 0, 0, 0, 0, 0, 2, 2, 2]);
+    /// assert_eq!(ones.len(), 3);
+    /// ```
     #[unstable(feature = "extract_if", reason = "recently added", issue = "43244")]
-    pub fn extract_if<F>(&mut self, filter: F) -> ExtractIf<'_, T, F, A>
+    pub fn extract_if<F, R>(&mut self, range: R, filter: F) -> ExtractIf<'_, T, F, A>
     where
         F: FnMut(&mut T) -> bool,
+        R: RangeBounds<usize>,
     {
-        let old_len = self.len();
-
-        // Guard against us getting leaked (leak amplification)
-        unsafe {
-            self.set_len(0);
-        }
-
-        ExtractIf { vec: self, idx: 0, del: 0, old_len, pred: filter }
+        ExtractIf::new(self, filter, range)
     }
 }