Add functions to edit stdin path list

These would be used to create binaries that manipulate the input before
piping it back into repo2markdown.

2 files changed, 211 insertions, 0 deletions

diff --git a/src/lib.rs b/src/lib.rs
index 43c66a1..b6ed259 100644
--- a/src/lib.rs
+++ b/src/lib.rs
@@ -3,5 +3,6 @@ pub mod fenced_md_generator;
 pub mod logger;
 pub mod md_generator;
 pub mod normalizer;
+pub mod path_list_editor;
 pub mod renderer;
 mod util;
diff --git a/src/path_list_editor.rs b/src/path_list_editor.rs
new file mode 100644
index 0000000..e25b0db
--- /dev/null
+++ b/src/path_list_editor.rs
@@ -0,0 +1,210 @@
+use std::{
+    collections::HashSet,
+    ffi::{OsStr, OsString},
+    io::{Read, Write},
+    os::unix::ffi::OsStrExt,
+};
+
+pub fn append_paths<R: Read, W: Write>(
+    mut input: R,
+    mut output: W,
+    new_paths: &[OsString],
+) -> std::io::Result<()> {
+    std::io::copy(&mut input, &mut output)?;
+    for path in new_paths {
+        output.write_all(path.as_bytes())?;
+        output.write_all(b"\0")?;
+    }
+    Ok(())
+}
+
+pub fn prepend_paths<R: Read, W: Write>(
+    mut input: R,
+    mut output: W,
+    new_paths: &[OsString],
+) -> std::io::Result<()> {
+    for path in new_paths {
+        output.write_all(path.as_bytes())?;
+        output.write_all(b"\0")?;
+    }
+    std::io::copy(&mut input, &mut output)?;
+    Ok(())
+}
+
+pub fn remove_paths<R: Read, W: Write>(
+    mut input: R,
+    mut output: W,
+    unwanted_paths: &[OsString],
+) -> std::io::Result<()> {
+    let mut buf = Vec::new();
+    input.read_to_end(&mut buf)?;
+    let unwanted_paths = HashSet::<&OsString>::from_iter(unwanted_paths);
+    for segment in buf.split(|b| *b == 0) {
+        if segment.is_empty() {
+            continue;
+        }
+        let os_string_segment = OsStr::from_bytes(segment).to_os_string();
+        if !unwanted_paths.contains(&os_string_segment) {
+            output.write_all(segment)?;
+            output.write_all(b"\0")?;
+        }
+    }
+    Ok(())
+}
+
+#[cfg(test)]
+mod tests {
+    use std::ffi::OsString;
+
+    use super::{append_paths, prepend_paths, remove_paths};
+
+    #[test]
+    fn append_paths_works_with_empty_stdin_and_args() {
+        let input = b"";
+        let mut output = Vec::new();
+
+        append_paths(&input[..], &mut output, &[]).unwrap();
+
+        assert_eq!(output, b"");
+    }
+
+    #[test]
+    fn append_paths_works_with_empty_args() {
+        let input = b"README.md\0";
+        let mut output = Vec::new();
+
+        append_paths(&input[..], &mut output, &[]).unwrap();
+
+        assert_eq!(output, b"README.md\0");
+    }
+
+    #[test]
+    fn append_paths_works_with_empty_stdin() {
+        let input = b"";
+        let mut output = Vec::new();
+
+        append_paths(&input[..], &mut output, &[OsString::from("README.md")]).unwrap();
+
+        assert_eq!(output, b"README.md\0");
+    }
+
+    #[test]
+    fn append_paths_works_with_unique_stdin_and_args() {
+        let input = b"a.rs\0b.rs\0";
+        let mut output = Vec::new();
+
+        append_paths(&input[..], &mut output, &[OsString::from("c.rs")]).unwrap();
+
+        assert_eq!(output, b"a.rs\0b.rs\0c.rs\0");
+    }
+
+    #[test]
+    fn append_paths_keeps_duplicate_entries() {
+        let input = b"a.rs\0b.rs\0";
+        let mut output = Vec::new();
+
+        append_paths(&input[..], &mut output, &[OsString::from("a.rs")]).unwrap();
+
+        assert_eq!(output, b"a.rs\0b.rs\0a.rs\0");
+    }
+
+    #[test]
+    fn prepend_paths_works_with_empty_stdin_and_args() {
+        let input = b"";
+        let mut output = Vec::new();
+
+        prepend_paths(&input[..], &mut output, &[]).unwrap();
+
+        assert_eq!(output, b"");
+    }
+
+    #[test]
+    fn prepend_paths_works_with_empty_args() {
+        let input = b"README.md\0";
+        let mut output = Vec::new();
+
+        prepend_paths(&input[..], &mut output, &[]).unwrap();
+
+        assert_eq!(output, b"README.md\0");
+    }
+
+    #[test]
+    fn prepend_paths_works_with_empty_stdin() {
+        let input = b"";
+        let mut output = Vec::new();
+
+        prepend_paths(&input[..], &mut output, &[OsString::from("README.md")]).unwrap();
+
+        assert_eq!(output, b"README.md\0");
+    }
+
+    #[test]
+    fn prepend_paths_works_with_unique_stdin_and_args() {
+        let input = b"a.rs\0b.rs\0";
+        let mut output = Vec::new();
+
+        prepend_paths(&input[..], &mut output, &[OsString::from("c.rs")]).unwrap();
+
+        assert_eq!(output, b"c.rs\0a.rs\0b.rs\0");
+    }
+
+    #[test]
+    fn prepend_paths_keeps_duplicate_entries() {
+        let input = b"a.rs\0b.rs\0";
+        let mut output = Vec::new();
+
+        prepend_paths(&input[..], &mut output, &[OsString::from("a.rs")]).unwrap();
+
+        assert_eq!(output, b"a.rs\0a.rs\0b.rs\0");
+    }
+
+    #[test]
+    fn remove_paths_works_with_empty_stdin_and_args() {
+        let input = b"";
+        let mut output = Vec::new();
+
+        remove_paths(&input[..], &mut output, &[]).unwrap();
+
+        assert_eq!(output, b"");
+    }
+
+    #[test]
+    fn remove_paths_works_with_empty_args() {
+        let input = b"README.md\0";
+        let mut output = Vec::new();
+
+        remove_paths(&input[..], &mut output, &[]).unwrap();
+
+        assert_eq!(output, b"README.md\0");
+    }
+
+    #[test]
+    fn remove_paths_works_with_empty_stdin() {
+        let input = b"";
+        let mut output = Vec::new();
+
+        remove_paths(&input[..], &mut output, &[OsString::from("README.md")]).unwrap();
+
+        assert_eq!(output, b"");
+    }
+
+    #[test]
+    fn remove_paths_filters_matching_entries() {
+        let input = b"a.rs\0b.rs\0c.rs\0";
+        let mut output = Vec::new();
+
+        remove_paths(&input[..], &mut output, &[OsString::from("b.rs")]).unwrap();
+
+        assert_eq!(output, b"a.rs\0c.rs\0");
+    }
+
+    #[test]
+    fn remove_paths_removes_all_matching_entries() {
+        let input = b"a.rs\0b.rs\0b.rs\0c.rs\0";
+        let mut output = Vec::new();
+
+        remove_paths(&input[..], &mut output, &[OsString::from("b.rs")]).unwrap();
+
+        assert_eq!(output, b"a.rs\0c.rs\0");
+    }
+}