Set bit error.

apache · Nov 13, 2020 · fd75933 · fd75933
1 parent 151bb2a
commit fd75933
Show file tree

Hide file tree

Showing 3 changed files with 7 additions and 66 deletions.
diff --git a/rust/arrow/src/array/builder.rs b/rust/arrow/src/array/builder.rs
@@ -334,13 +334,13 @@ impl<T: ArrowPrimitiveType> BufferBuilderTrait<T> for BufferBuilder<T> {
         self.reserve(n);
         if T::DATA_TYPE == DataType::Boolean {
             if n != 0 && v != T::default_value() {
-                unsafe {
-                    bit_util::set_bits_raw(
+                let data = unsafe {
+                    std::slice::from_raw_parts_mut(
                         self.buffer.raw_data_mut(),
-                        self.len,
-                        self.len + n,
+                        self.buffer.capacity(),
                     )
-                }
+                };
+                (self.len..self.len + n).for_each(|i| bit_util::set_bit(data, i))
             }
             self.len += n;
         } else {

diff --git a/rust/arrow/src/array/equal_json.rs b/rust/arrow/src/array/equal_json.rs
@@ -533,6 +533,8 @@ mod tests {
         "#,
         )
         .unwrap();
+        println!("{:?}", arrow_array);
+        println!("{:?}", json_array);
         assert!(arrow_array.eq(&json_array));
         assert!(json_array.eq(&arrow_array));
 

diff --git a/rust/arrow/src/util/bit_util.rs b/rust/arrow/src/util/bit_util.rs
@@ -99,36 +99,6 @@ pub unsafe fn unset_bit_raw(data: *mut u8, i: usize) {
     *data.add(i >> 3) ^= BIT_MASK[i & 7];
 }
 
-/// Sets bits in the non-inclusive range `start..end` for `data`
-///
-/// # Safety
-///
-/// Note this doesn't do any bound checking, for performance reason. The caller is
-/// responsible to guarantee that both `start` and `end` are within bounds.
-#[inline]
-pub unsafe fn set_bits_raw(data: *mut u8, start: usize, end: usize) {
-    let start_byte = (start >> 3) as isize;
-    let end_byte = (end >> 3) as isize;
-
-    let start_offset = (start & 7) as u8;
-    let end_offset = (end & 7) as u8;
-
-    // All set apart from lowest `start_offset` bits
-    let start_mask = !((1 << start_offset) - 1);
-    // All clear apart from lowest `end_offset` bits
-    let end_mask = (1 << end_offset) - 1;
-
-    if start_byte == end_byte {
-        *data.offset(start_byte) |= start_mask & end_mask;
-    } else {
-        *data.offset(start_byte) |= start_mask;
-        for i in (start_byte + 1)..end_byte {
-            *data.offset(i) = 0xFF;
-        }
-        *data.offset(end_byte) |= end_mask;
-    }
-}
-
 /// Returns the number of 1-bits in `data`
 #[inline]
 pub fn count_set_bits(data: &[u8]) -> usize {
@@ -334,37 +304,6 @@ mod tests {
         }
     }
 
-    #[test]
-    fn test_set_bits_raw() {
-        const NUM_BYTE: usize = 64;
-        const NUM_BLOCKS: usize = 12;
-        const MAX_BLOCK_SIZE: usize = 32;
-        let mut buf = vec![0; NUM_BYTE];
-
-        let mut expected = Vec::with_capacity(NUM_BYTE * 8);
-        expected.resize(NUM_BYTE * 8, false);
-
-        let mut rng = seedable_rng();
-
-        for _ in 0..NUM_BLOCKS {
-            let start = rng.gen_range(0, NUM_BYTE * 8 - MAX_BLOCK_SIZE);
-            let end = start + rng.gen_range(1, MAX_BLOCK_SIZE);
-            unsafe {
-                set_bits_raw(buf.as_mut_ptr(), start, end);
-            }
-            for i in start..end {
-                expected[i] = true;
-            }
-        }
-
-        let raw_ptr = buf.as_ptr();
-        for (i, b) in expected.iter().enumerate() {
-            unsafe {
-                assert_eq!(*b, get_bit_raw(raw_ptr, i));
-            }
-        }
-    }
-
     #[test]
     fn test_get_set_bit_roundtrip() {
         const NUM_BYTES: usize = 10;