Initial work

spiraldb · Aug 15, 2024 · 47b2283 · 47b2283
1 parent 89c9e0f
commit 47b2283
Show file tree

Hide file tree

Showing 3 changed files with 92 additions and 60 deletions.
diff --git a/fuzz/Cargo.toml b/fuzz/Cargo.toml
@@ -2,19 +2,23 @@
 name = "vortex-fuzz"
 version = "0.0.0"
 publish = false
-edition = "2021"
+edition = { workspace = true }
 license = { workspace = true }
 
 [package.metadata]
 cargo-fuzz = true
 
 [dependencies]
-libfuzzer-sys = "0.4"
+libfuzzer-sys = { version = "0.4" }
 vortex-array = { workspace = true, features = ["arbitrary"] }
 vortex-dtype = { workspace = true }
 vortex-sampling-compressor = { workspace = true }
 vortex-scalar = { workspace = true }
 
+[lib]
+name = "vortex_fuzz"
+path = "src/lib.rs"
+
 
 [[bin]]
 name = "fuzz_target_1"

diff --git a/fuzz/fuzz_targets/fuzz_target_1.rs b/fuzz/fuzz_targets/fuzz_target_1.rs
@@ -2,81 +2,39 @@
 
 use std::collections::HashSet;
 
-use libfuzzer_sys::arbitrary::{Arbitrary, Unstructured};
 use libfuzzer_sys::{fuzz_target, Corpus};
 use vortex::compute::slice;
 use vortex::compute::unary::scalar_at;
 use vortex::encoding::EncodingId;
 use vortex::Array;
-use vortex_sampling_compressor::compressors::alp::ALPCompressor;
-use vortex_sampling_compressor::compressors::bitpacked::BitPackedCompressor;
-use vortex_sampling_compressor::compressors::dict::DictCompressor;
-use vortex_sampling_compressor::compressors::r#for::FoRCompressor;
-use vortex_sampling_compressor::compressors::roaring_bool::RoaringBoolCompressor;
-use vortex_sampling_compressor::compressors::roaring_int::RoaringIntCompressor;
-use vortex_sampling_compressor::compressors::runend::DEFAULT_RUN_END_COMPRESSOR;
-use vortex_sampling_compressor::compressors::sparse::SparseCompressor;
-use vortex_sampling_compressor::compressors::zigzag::ZigZagCompressor;
+use vortex_fuzz::FuzzArrayAction;
 use vortex_sampling_compressor::compressors::CompressorRef;
 use vortex_sampling_compressor::SamplingCompressor;
 use vortex_scalar::{PValue, Scalar, ScalarValue};
 
-fuzz_target!(|data: &[u8]| -> Corpus {
-    let mut u = Unstructured::new(data);
-
-    let array = Array::arbitrary(&mut u).unwrap();
+fuzz_target!(|fuzz_action: FuzzArrayAction| -> Corpus {
+    let FuzzArrayAction { array, action } = fuzz_action;
 
     // TODO(adamg): We actually might want to test empty things, but I'm punting this issue for now
     if array.is_empty() {
         return Corpus::Reject;
     };
-    match u.int_in_range(0..=9).unwrap() {
-        0 => {
-            let start = u.choose_index(array.len()).unwrap();
-            let stop = u.choose_index(array.len() - start).unwrap() + start;
-            let slice = slice(&array, start, stop).unwrap();
-            assert_slice(&array, &slice, start);
-        }
-        1 => match fuzz_compress(&array, &ALPCompressor) {
-            Some(compressed_array) => assert_array_eq(&array, &compressed_array),
-            None => return Corpus::Reject,
-        },
-        2 => match fuzz_compress(&array, &BitPackedCompressor) {
-            Some(compressed_array) => assert_array_eq(&array, &compressed_array),
-            None => return Corpus::Reject,
-        },
-        3 => match fuzz_compress(&array, &DictCompressor) {
-            Some(compressed_array) => assert_array_eq(&array, &compressed_array),
-            None => return Corpus::Reject,
-        },
-        4 => match fuzz_compress(&array, &FoRCompressor) {
-            Some(compressed_array) => assert_array_eq(&array, &compressed_array),
-            None => return Corpus::Reject,
-        },
-        5 => match fuzz_compress(&array, &RoaringBoolCompressor) {
-            Some(compressed_array) => assert_array_eq(&array, &compressed_array),
-            None => return Corpus::Reject,
-        },
-        6 => match fuzz_compress(&array, &RoaringIntCompressor) {
-            Some(compressed_array) => assert_array_eq(&array, &compressed_array),
-            None => return Corpus::Reject,
-        },
-        7 => match fuzz_compress(&array, &DEFAULT_RUN_END_COMPRESSOR) {
-            Some(compressed_array) => assert_array_eq(&array, &compressed_array),
-            None => return Corpus::Reject,
-        },
-        8 => match fuzz_compress(&array, &SparseCompressor) {
-            Some(compressed_array) => assert_array_eq(&array, &compressed_array),
-            None => return Corpus::Reject,
-        },
-        9 => match fuzz_compress(&array, &ZigZagCompressor) {
-            Some(compressed_array) => assert_array_eq(&array, &compressed_array),
+
+    match action {
+        vortex_fuzz::Action::Compress(c) => match fuzz_compress(&array, c.as_ref()) {
+            Some(compressed_array) => {
+                assert_array_eq(&array, &compressed_array);
+                Corpus::Keep
+            }
             None => return Corpus::Reject,
         },
-        _ => unreachable!(),
+        vortex_fuzz::Action::Slice(range) => {
+            let slice = slice(&array, range.start, range.end).unwrap();
+            assert_slice(&array, &slice, range.start);
+            Corpus::Keep
+        }
+        vortex_fuzz::Action::NoOp => Corpus::Reject,
     }
-
-    Corpus::Keep
 });
 
 fn fuzz_compress(array: &Array, compressor_ref: CompressorRef<'_>) -> Option<Array> {

diff --git a/fuzz/src/lib.rs b/fuzz/src/lib.rs
@@ -0,0 +1,70 @@
+use std::ops::Range;
+
+use libfuzzer_sys::arbitrary::{Arbitrary, Result, Unstructured};
+use vortex::Array;
+use vortex_sampling_compressor::compressors::alp::ALPCompressor;
+use vortex_sampling_compressor::compressors::bitpacked::BitPackedCompressor;
+use vortex_sampling_compressor::compressors::dict::DictCompressor;
+use vortex_sampling_compressor::compressors::r#for::FoRCompressor;
+use vortex_sampling_compressor::compressors::roaring_bool::RoaringBoolCompressor;
+use vortex_sampling_compressor::compressors::roaring_int::RoaringIntCompressor;
+use vortex_sampling_compressor::compressors::runend::DEFAULT_RUN_END_COMPRESSOR;
+use vortex_sampling_compressor::compressors::sparse::SparseCompressor;
+use vortex_sampling_compressor::compressors::zigzag::ZigZagCompressor;
+use vortex_sampling_compressor::compressors::EncodingCompressor;
+
+pub struct FuzzArrayAction {
+    pub array: Array,
+    pub action: Action,
+}
+
+impl std::fmt::Debug for FuzzArrayAction {
+    fn fmt(&self, f: &mut std::fmt::Formatter<'_>) -> std::fmt::Result {
+        f.debug_struct("FuzzArrayAction")
+            .field("action", &self.action)
+            .field("array", &self.array)
+            .finish()
+    }
+}
+
+#[derive()]
+pub enum Action {
+    NoOp,
+    Compress(Box<dyn EncodingCompressor>),
+    Slice(Range<usize>),
+}
+
+impl std::fmt::Debug for Action {
+    fn fmt(&self, f: &mut std::fmt::Formatter<'_>) -> std::fmt::Result {
+        match self {
+            Self::NoOp => write!(f, "NoOp"),
+            Self::Slice(arg0) => f.debug_tuple("Slice").field(arg0).finish(),
+            Self::Compress(c) => write!(f, "Compress({})", c.id()),
+        }
+    }
+}
+
+impl<'a> Arbitrary<'a> for FuzzArrayAction {
+    fn arbitrary(u: &mut Unstructured<'a>) -> Result<Self> {
+        let array = Array::arbitrary(u)?;
+        let action = match u.int_in_range(0..=9)? {
+            0 => {
+                let start = u.choose_index(array.len())?;
+                let stop = u.choose_index(array.len() - start).unwrap() + start;
+                Action::Slice(start..stop)
+            }
+            1 => Action::Compress(Box::new(ALPCompressor) as _),
+            2 => Action::Compress(Box::new(BitPackedCompressor) as _),
+            3 => Action::Compress(Box::new(DictCompressor) as _),
+            4 => Action::Compress(Box::new(FoRCompressor) as _),
+            5 => Action::Compress(Box::new(RoaringBoolCompressor) as _),
+            6 => Action::Compress(Box::new(RoaringIntCompressor) as _),
+            7 => Action::Compress(Box::new(DEFAULT_RUN_END_COMPRESSOR) as _),
+            8 => Action::Compress(Box::new(SparseCompressor) as _),
+            9 => Action::Compress(Box::new(ZigZagCompressor) as _),
+            _ => Action::NoOp,
+        };
+
+        Ok(Self { array, action })
+    }
+}