Detect ZIP archives as binary content

CHANGELOG.md

@@ -19,6 +19,7 @@
 - Map `BUILD` case sensitively to Python (Starlark) for Bazel, see #3576 (@vorburger)
 
 ## Bugfixes
+- Treat ZIP archives as binary content based on their magic header, see #0000 (@officialasishkumar)
 - Fix i686 `.deb` package using incorrect architecture name (`i686` instead of `i386`), preventing installation on Debian. Closes #3611, see #3650 (@Sim-hu)
 - Fix inconsistent `.deb` MUSL package names (aarch64-musl used `arm64` instead of `musl-linux-arm64`, and `musleabihf` target missed `bat-musl` prefix). Closes #3482, see #3642 (@mvanhorn)
 - Fix incorrect text width computation when using `--binary=as-text` with non-printable characters in caret notation, see #3640 and #3631 (@eyupcanakman)

src/input.rs

@@ -261,11 +261,7 @@ impl<'a> InputReader<'a> {
         let mut first_line = vec![];
         reader.read_until(b'\n', &mut first_line).ok();
 
-        let content_type = if first_line.is_empty() {
-            None
-        } else {
-            Some(content_inspector::inspect(&first_line[..]))
-        };
+        let content_type = inspect_content_type(&first_line);
 
         if content_type == Some(ContentType::UTF_16LE) {
             read_utf16_line(&mut reader, &mut first_line, 0x00, 0x0A).ok();
@@ -319,6 +315,25 @@ impl<'a> InputReader<'a> {
     }
 }
 
+fn inspect_content_type(first_line: &[u8]) -> Option<ContentType> {
+    if first_line.is_empty() {
+        return None;
+    }
+
+    let content_type = content_inspector::inspect(first_line);
+    if content_type == ContentType::UTF_8 && has_zip_signature(first_line) {
+        Some(ContentType::BINARY)
+    } else {
+        Some(content_type)
+    }
+}
+
+fn has_zip_signature(bytes: &[u8]) -> bool {
+    [b"PK\x03\x04", b"PK\x05\x06", b"PK\x07\x08"]
+        .into_iter()
+        .any(|signature| bytes.starts_with(signature))
+}
+
 fn read_utf16_line<R: BufRead>(
     reader: &mut R,
     buf: &mut Vec<u8>,
@@ -374,6 +389,22 @@ fn basic() {
     assert!(buffer.is_empty());
 }
 
+#[test]
+fn zip_magic_headers_are_treated_as_binary() {
+    for content in [b"PK\x03\x04hello", b"PK\x05\x06hello", b"PK\x07\x08hello"] {
+        let reader = InputReader::new(&content[..]);
+        assert_eq!(Some(ContentType::BINARY), reader.content_type);
+    }
+}
+
+#[test]
+fn non_zip_pk_prefix_is_not_treated_as_binary() {
+    assert_eq!(
+        Some(ContentType::UTF_8),
+        inspect_content_type(b"PK\x03\x03hello")
+    );
+}
+
 #[test]
 fn utf16le() {
     let content = b"\xFF\xFE\x73\x00\x0A\x00\x64\x00";

tests/integration_tests.rs

@@ -2093,6 +2093,24 @@ fn header_binary() {
         .stderr("");
 }
 
+#[test]
+fn header_zip_file_is_binary() {
+    let tmp_dir = tempdir().expect("can create temporary directory");
+    let tmp_path = tmp_dir.path().join("test.zip");
+    std::fs::write(&tmp_path, b"PK\x03\x04hello").expect("can write temporary file");
+
+    bat()
+        .arg(&tmp_path)
+        .arg("--decorations=always")
+        .arg("--style=header")
+        .arg("-r=0:0")
+        .arg("--file-name=test.zip")
+        .assert()
+        .success()
+        .stdout("File: test.zip   <BINARY>\n")
+        .stderr("");
+}
+
 #[test]
 fn header_full_binary() {
     bat()