Fix fuzzing setup after Scalar ordering changes and fix io fuzzer (#2098)

Author: robert3005

.github/workflows/fuzz.yml

@@ -30,7 +30,7 @@ jobs:
       - name: Archive crash artifacts
         uses: actions/upload-artifact@v4
         with:
-          name: fuzzing-crash-artifacts
+          name: io-fuzzing-crash-artifacts
           path: fuzz/artifacts
       - name: Persist corpus
         shell: bash
@@ -42,8 +42,8 @@ jobs:
           AWS_SECRET_ACCESS_KEY: ${{ secrets.R2_FUZZ_SECRET_ACCESS_KEY }}
           AWS_REGION: 'us-east-1'
 
-  fuzz:
-    name: 'fuzz'
+  ops_fuzz:
+    name: 'Array Operations Fuzz'
     runs-on: ubuntu-latest
     steps:
       - uses: actions/checkout@v4
@@ -66,7 +66,7 @@ jobs:
       - name: Archive crash artifacts
         uses: actions/upload-artifact@v4
         with:
-          name: fuzzing-crash-artifacts
+          name: operations-fuzzing-crash-artifacts
           path: fuzz/artifacts
       - name: Persist corpus
         shell: bash

fuzz/fuzz_targets/file_io.rs

@@ -9,6 +9,7 @@ use libfuzzer_sys::{fuzz_target, Corpus};
 use vortex_array::array::ChunkedArray;
 use vortex_array::compute::{compare, Operator};
 use vortex_array::{ArrayDType, ArrayData, IntoArrayData, IntoArrayVariant, IntoCanonical};
+use vortex_dtype::Nullability;
 use vortex_file::{Scan, VortexOpenOptions, VortexWriteOptions};
 use vortex_sampling_compressor::ALL_ENCODINGS_CONTEXT;
 
@@ -17,6 +18,10 @@ fuzz_target!(|array_data: ArrayData| -> Corpus {
         return Corpus::Reject;
     }
 
+    if array_data.dtype().nullability() != Nullability::NonNullable {
+        return Corpus::Reject;
+    }
+
     let runtime = tokio::runtime::Builder::new_current_thread()
         .enable_all()
         .build()

fuzz/src/sort.rs

@@ -27,7 +27,7 @@ pub fn sort_canonical_array(array: &ArrayData) -> ArrayData {
                 )
                 .map(|(b, v)| v.then_some(b))
                 .collect::<Vec<_>>();
-            sort_opt_slice(&mut opt_values);
+            opt_values.sort();
             BoolArray::from_iter(opt_values).into_array()
         }
         DType::Primitive(p, _) => {
@@ -57,7 +57,7 @@ pub fn sort_canonical_array(array: &ArrayData) -> ArrayData {
             let mut opt_values = utf8
                 .with_iterator(|iter| iter.map(|v| v.map(|u| u.to_vec())).collect::<Vec<_>>())
                 .unwrap();
-            sort_opt_slice(&mut opt_values);
+            opt_values.sort();
             VarBinViewArray::from_iter(opt_values, array.dtype().clone()).into_array()
         }
         DType::Struct(..) => {
@@ -88,17 +88,7 @@ fn sort_primitive_slice<T: NativePType>(s: &mut [Option<T>]) {
     s.sort_by(|a, b| match (a, b) {
         (Some(v), Some(w)) => v.total_compare(*w),
         (None, None) => Ordering::Equal,
-        (None, Some(_)) => Ordering::Greater,
-        (Some(_), None) => Ordering::Less,
-    });
-}
-
-/// Reverse sorting of Option<T> such that None is last (Greatest)
-fn sort_opt_slice<T: Ord>(s: &mut [Option<T>]) {
-    s.sort_by(|a, b| match (a, b) {
-        (Some(v), Some(w)) => v.cmp(w),
-        (None, None) => Ordering::Equal,
-        (None, Some(_)) => Ordering::Greater,
-        (Some(_), None) => Ordering::Less,
+        (None, Some(_)) => Ordering::Less,
+        (Some(_), None) => Ordering::Greater,
     });
 }