Update CONTRIBUTING.md

Parsing was failing for malformed wav files
where the final chunk was not padded to a
word boundary.

CONTRIBUTING.md

@@ -0,0 +1,64 @@
+# Contributing
+
+Contributions to `bwavfile` are welcome!
+
+This project is currently in a maintenance phase, new features are entertained if we
+believe there is some demand for them but for the most part our current priorities 
+are:
+
+* Addressing bugs.
+* Keeping the codebase modern as the Rust language evolves.
+
+This being said, there are some features that we've been wanting for some time and 
+new contributors are welcome to take a swing at these:
+
+* Reading `levl` metadata for the generation of waveform overviews.
+* Reading `smpl` metadata for reading sampler data, note assignments, loops etc.
+
+## Adding New Features
+
+If you are adding a large amount of new functionality, please weigh the amount of 
+work you're doing against the potential benefit, the burden on the maintainers to
+review the work, and the technical debt incurred.
+
+In general,
+
+* new features should address new technologies; do not expend large amounts of
+  effort to implement features that are primarily of historical interest.
+* new features should address the needs of professional users in a music or media
+  production environment.
+
+Features that implement new reading functionality must, when submitted, include 
+test WAV files created empirically by third-party software. 
+
+## Regarding use of Agents
+
+`bwavfile` is an open-source project that is offered free for no commerical gain, and 
+is developed and maintained for educational and creative reasons.
+
+If you use an agent or LLM to produce code for it you are missing out on the benefits 
+of contributing to an open-source project, particularly community, collaboration with 
+other developers and designers, and being able to learn and experiment without the 
+burden of deadlines or worrying about business cases or profits.
+
+This project is supposed to be fun, do not let machines have fun for you.
+
+We can't prevent you from using LLMs to contribute to this project but we ask you 
+abide by the following eitiquette when doing so:
+
+* All communication with the maintainers must be written by a human in their own
+  voice. Never use an LLM to craft thread comments, discussion posts, issues, emails
+  or other correspondence with other developers or the maintainers. 
+* PRs must be submitted by a person. Do not allow an agent to submit its own PRs to
+  this project.
+* Especially if you are a new contributor to this project, please submit only one PR
+  at a time and please restrict the subject matter of the PR to a specific unit,
+  module or tool. All submissions have to be reviewed and understood by the
+  maintainers before they can be merged.
+
+Obviously we can't verify if you follow all of these rules but certain telltale 
+traits of LLM-predicted text or code will raise a flag: lack of brevity in 
+descriptions or code comments, large amounts of text describing your process or 
+steps that add little to understanding the changes you've made, use of an 
+obsequious tone or being excessively accomodating, immediately doing requests 
+without further discussion or clarifications.

Cargo.lock

@@ -1,6 +1,6 @@
 # This file is automatically @generated by Cargo.
 # It is not intended for manual editing.
-version = 3
+version = 4
 
 [[package]]
 name = "ansi_term"
@@ -30,7 +30,7 @@ checksum = "bef38d45163c2f1dde094a7dfd33ccf595c92905c8f8f4fdc18d06fb1037718a"
 
 [[package]]
 name = "bwavfile"
-version = "2.0.0"
+version = "2.0.2"
 dependencies = [
  "byteorder",
  "clap",

Cargo.toml

@@ -1,6 +1,6 @@
 [package]
 name = "bwavfile"
-version = "2.0.1"
+version = "2.0.2"
 authors = ["Jamie Hardt <jamiehardt@me.com>", "Ian Hobson <ian.r.hobson@gmail.com>"]
 edition = "2018"
 license = "MIT"

src/chunks.rs

@@ -37,7 +37,7 @@ where
         self.write_u16::<LittleEndian>(format.block_alignment)?;
         self.write_u16::<LittleEndian>(format.bits_per_sample)?;
         if let Some(ext) = format.extended_format {
-            let cb_size = 24u16;
+            let cb_size = 22u16;
             self.write_u16::<LittleEndian>(cb_size)?;
             self.write_u16::<LittleEndian>(ext.valid_bits_per_sample)?;
             self.write_u32::<LittleEndian>(ext.channel_mask)?;

src/fmt.rs

@@ -1,3 +1,5 @@
+#![allow(dead_code)]
+
 use crate::common_format::{CommonFormat, WAVE_UUID_BFORMAT_PCM, WAVE_UUID_PCM};
 use crate::Sample;
 
@@ -8,7 +10,6 @@ use byteorder::LittleEndian;
 use byteorder::ReadBytesExt;
 
 // Need more test cases for ADMAudioID
-#[allow(dead_code)]
 
 /// ADM Audio ID record.
 ///
@@ -404,7 +405,7 @@ where
         format: WaveFmt,
         into: &mut [i32],
     ) -> Result<usize, std::io::Error> {
-        assert!(into.len() % format.channel_count as usize == 0);
+        assert!(into.len().is_multiple_of(format.channel_count as usize));
 
         for frame in into {
             *frame = match (format.valid_bits_per_sample(), format.bits_per_sample) {
@@ -424,7 +425,7 @@ where
         format: WaveFmt,
         into: &mut [f32],
     ) -> Result<usize, std::io::Error> {
-        assert!(into.len() % format.channel_count as usize == 0);
+        assert!(into.len().is_multiple_of(format.channel_count as usize));
         todo!()
     }
 }

src/parser.rs

@@ -18,6 +18,7 @@ use super::fourcc::{BW64_SIG, DATA_SIG, DS64_SIG, RF64_SIG, RIFF_SIG, WAVE_SIG};
 const RF64_SIZE_MARKER: u32 = 0xFF_FF_FF_FF;
 
 #[derive(Debug)]
+#[allow(dead_code)]
 pub enum Event {
     StartParse,
     ReadHeader {
@@ -258,7 +259,7 @@ impl<R: Read + Seek> Parser<R> {
 
             state = State::ReadyForChunk {
                 at: at + 8 + this_displacement,
-                remaining: remaining - 8 - this_displacement,
+                remaining: remaining.saturating_sub(8 + this_displacement),
             }
         }
 
@@ -292,3 +293,35 @@ impl<R: Read + Seek> Parser<R> {
         }
     }
 }
+
+#[cfg(test)]
+mod tests {
+    use super::*;
+    use std::io::Cursor;
+
+    #[test]
+    fn parser_handles_missing_trailing_pad_on_final_odd_chunk() {
+        let mut buf = Vec::new();
+        buf.extend_from_slice(b"RIFF");
+        buf.extend_from_slice(&39u32.to_le_bytes());
+        buf.extend_from_slice(b"WAVE");
+        buf.extend_from_slice(b"fmt ");
+        buf.extend_from_slice(&16u32.to_le_bytes());
+        buf.extend_from_slice(&[0u8; 16]);
+        buf.extend_from_slice(b"data");
+        buf.extend_from_slice(&3u32.to_le_bytes());
+        buf.extend_from_slice(&[0xAA, 0xBB, 0xCC]);
+        assert_eq!(buf.len(), 47);
+
+        let chunks = Parser::make(Cursor::new(&buf))
+            .unwrap()
+            .into_chunk_list()
+            .expect("parser should accept a missing trailing pad on the final odd-length chunk");
+
+        assert_eq!(chunks.len(), 2);
+        assert_eq!(chunks[0].signature, FourCC::from(*b"fmt "));
+        assert_eq!(chunks[0].length, 16);
+        assert_eq!(chunks[1].signature, DATA_SIG);
+        assert_eq!(chunks[1].length, 3);
+    }
+}

src/wavereader.rs

@@ -110,7 +110,7 @@ impl<R: Read + Seek> AudioFrameReader<R> {
         let common_format = self.format.common_format();
         let bits_per_sample = self.format.bits_per_sample;
 
-        if buffer.len() % channel_count != 0 {
+        if !buffer.len().is_multiple_of(channel_count) {
             return Err(Error::InvalidBufferSize {
                 buffer_size: buffer.len(),
                 channel_count: self.format.channel_count,
@@ -245,7 +245,7 @@ impl<R: Read + Seek> WaveReader<R> {
     /// will return an `Err(errors::Error)` immediately if there is a structural
     /// inconsistency that makes the stream unreadable or if it's missing
     /// essential components that make interpreting the audio data impossible.
-
+    ///
     /// ```rust
     /// use std::fs::File;
     /// use std::io::{Error,ErrorKind};

src/wavewriter.rs

@@ -51,7 +51,7 @@ where
         let format = &self.inner.inner.format;
         let channel_count = format.channel_count as usize;
 
-        if buffer.len() % channel_count != 0 {
+        if !buffer.len().is_multiple_of(channel_count) {
             return Err(Error::InvalidBufferSize {
                 buffer_size: buffer.len(),
                 channel_count: format.channel_count,
@@ -339,7 +339,7 @@ where
         assert!(data.len() < u32::MAX as usize);
         self.inner.write_u32::<LittleEndian>(data.len() as u32)?;
         self.inner.write_all(data)?;
-        if data.len() % 2 == 0 {
+        if data.len().is_multiple_of(2) {
             self.increment_form_length(8 + data.len() as u64)?;
         } else {
             self.inner.write_u8(0)?;