fix: Panic on trailing omitted FORMAT records (#417)

* Handle trailing omitted FORMAT records * Remove newline * fix: Wrong if statement * fmt --------- Co-authored-by: Fenner Macrae <[email protected]> Co-authored-by: Johannes Koester <[email protected]>
rust-bio · Mar 27, 2024 · 9f575ee · 9f575ee
1 parent b6aeba4
commit 9f575ee
Show file tree

Hide file tree

Showing 3 changed files with 30 additions and 0 deletions.
diff --git a/src/bcf/mod.rs b/src/bcf/mod.rs
@@ -1548,6 +1548,24 @@ mod tests {
         );
     }
 
+    #[test]
+    fn test_trailing_omitted_format_fields() {
+        let mut reader = Reader::from_path("test/test_trailing_omitted_format.vcf").unwrap();
+        let first_record = reader
+            .records()
+            .next()
+            .unwrap()
+            .expect("Fail to read record");
+
+        let expected: Vec<&[u8]> = Vec::new();
+        assert_eq!(*first_record.format(b"STR").string().unwrap(), expected,);
+        assert_eq!(
+            *first_record.format(b"INT").integer().unwrap(),
+            vec![&[i32::missing()]],
+        );
+        assert!(first_record.format(b"FLT").float().unwrap()[0][0].is_nan(),);
+    }
+
     // #[test]
     // fn test_buffer_lifetime() {
     //     let mut reader = Reader::from_path("test/obs-cornercase.vcf").unwrap();

diff --git a/src/bcf/record.rs b/src/bcf/record.rs
@@ -1483,6 +1483,9 @@ impl<'a, 'b, B: BorrowMut<Buffer> + Borrow<Buffer> + 'b> Format<'a, B> {
     /// memory.
     pub fn string(mut self) -> Result<BufferBacked<'b, Vec<&'b [u8]>, B>> {
         self.data(htslib::BCF_HT_STR).map(|ret| {
+            if ret == 0 {
+                return BufferBacked::new(Vec::new(), self.buffer);
+            }
             BufferBacked::new(
                 unsafe {
                     slice::from_raw_parts(self.buffer.borrow_mut().inner as *const u8, ret as usize)

diff --git a/test/test_trailing_omitted_format.vcf b/test/test_trailing_omitted_format.vcf
@@ -0,0 +1,9 @@
+##fileformat=VCFv4.3
+##contig=<ID=chr1,length=10000>
+##INFO=<ID=FOO,Number=1,Type=Integer,Description="Some field">
+##FORMAT=<ID=GT,Number=1,Type=String,Description="Genotype">
+##FORMAT=<ID=STR,Number=1,Type=String,Description="Some string field">
+##FORMAT=<ID=INT,Number=1,Type=Integer,Description="Some integer field">
+##FORMAT=<ID=FLT,Number=1,Type=Float,Description="Some float field">
+#CHROM	POS	ID	REF	ALT	QUAL	FILTER	INFO	FORMAT	SAMPLE1
+chr1	1234	.	t	a	.	.	FOO=1	GT:STR:FLT:INT	.