Merge branch 'main' into safer-appendvalue-bytes-view

Author: alamb

arrow-buffer/src/util/bit_chunk_iterator.rs

@@ -221,7 +221,10 @@ pub struct BitChunks<'a> {
 impl<'a> BitChunks<'a> {
     /// Create a new [`BitChunks`] from a byte array, and an offset and length in bits
     pub fn new(buffer: &'a [u8], offset: usize, len: usize) -> Self {
-        assert!(ceil(offset + len, 8) <= buffer.len() * 8);
+        assert!(
+            ceil(offset + len, 8) <= buffer.len(),
+            "offset + len out of bounds"
+        );
 
         let byte_offset = offset / 8;
         let bit_offset = offset % 8;
@@ -476,6 +479,57 @@ mod tests {
         assert_eq!(0x7F, bitchunks.remainder_bits());
     }
 
+    #[test]
+    #[should_panic(expected = "offset + len out of bounds")]
+    fn test_out_of_bound_should_panic_length_is_more_than_buffer_length() {
+        const ALLOC_SIZE: usize = 4 * 1024;
+        let input = vec![0xFF_u8; ALLOC_SIZE];
+
+        let buffer: Buffer = Buffer::from_vec(input);
+
+        // We are reading more than exists in the buffer
+        buffer.bit_chunks(0, (ALLOC_SIZE + 1) * 8);
+    }
+
+    #[test]
+    #[should_panic(expected = "offset + len out of bounds")]
+    fn test_out_of_bound_should_panic_length_is_more_than_buffer_length_but_not_when_not_using_ceil()
+     {
+        const ALLOC_SIZE: usize = 4 * 1024;
+        let input = vec![0xFF_u8; ALLOC_SIZE];
+
+        let buffer: Buffer = Buffer::from_vec(input);
+
+        // We are reading more than exists in the buffer
+        buffer.bit_chunks(0, (ALLOC_SIZE * 8) + 1);
+    }
+
+    #[test]
+    #[should_panic(expected = "offset + len out of bounds")]
+    fn test_out_of_bound_should_panic_when_offset_is_not_zero_and_length_is_the_entire_buffer_length()
+     {
+        const ALLOC_SIZE: usize = 4 * 1024;
+        let input = vec![0xFF_u8; ALLOC_SIZE];
+
+        let buffer: Buffer = Buffer::from_vec(input);
+
+        // We are reading more than exists in the buffer
+        buffer.bit_chunks(8, ALLOC_SIZE * 8);
+    }
+
+    #[test]
+    #[should_panic(expected = "offset + len out of bounds")]
+    fn test_out_of_bound_should_panic_when_offset_is_not_zero_and_length_is_the_entire_buffer_length_with_ceil()
+     {
+        const ALLOC_SIZE: usize = 4 * 1024;
+        let input = vec![0xFF_u8; ALLOC_SIZE];
+
+        let buffer: Buffer = Buffer::from_vec(input);
+
+        // We are reading more than exists in the buffer
+        buffer.bit_chunks(1, ALLOC_SIZE * 8);
+    }
+
     #[test]
     #[allow(clippy::assertions_on_constants)]
     fn test_unaligned_bit_chunk_iterator() {

parquet-variant-compute/src/variant_array.rs

@@ -208,7 +208,7 @@ impl ExtensionType for VariantType {
 /// assert_eq!(variant_array.value(0), Variant::from("such wow"));
 /// ```
 ///
-#[derive(Clone, Debug)]
+#[derive(Debug, Clone, PartialEq)]
 pub struct VariantArray {
     /// Reference to the underlying StructArray
     inner: StructArray,
@@ -741,7 +741,7 @@ impl From<ShreddedVariantFieldArray> for StructArray {
 /// (partial shredding).
 ///
 /// [Parquet Variant Shredding Spec]: https://github.com/apache/parquet-format/blob/master/VariantShredding.md#value-shredding
-#[derive(Clone, Debug)]
+#[derive(Debug, Clone, PartialEq)]
 pub struct ShreddingState {
     value: Option<BinaryViewArray>,
     typed_value: Option<ArrayRef>,
@@ -1467,4 +1467,27 @@ mod test {
             Variant::ShortString(ShortString::try_new("norm").unwrap())
         );
     }
+
+    #[test]
+    fn test_variant_equality() {
+        let v_iter = [None, Some(Variant::BooleanFalse), Some(Variant::Null), None];
+        let v = VariantArray::from_iter(v_iter.clone());
+
+        {
+            let v_copy = v.clone();
+            assert_eq!(v, v_copy);
+        }
+
+        {
+            let v_iter_reversed = v_iter.iter().cloned().rev();
+            let v_reversed = VariantArray::from_iter(v_iter_reversed);
+
+            assert_ne!(v, v_reversed);
+        }
+
+        {
+            let v_sliced = v.slice(0, 1);
+            assert_ne!(v, v_sliced);
+        }
+    }
 }