Hmbown/CodeWhale · error

Pinned source is not bounded text

Error message

Pinned source is not bounded text

What it means

After reading the pinned blob with `git cat-file blob`, `context_blob` enforces two bounds: the content must be at most `MAX_CONTEXT_FILE_BYTES` and must contain no NUL byte. This guarantees the source context handed to review passes is bounded plain text, not a huge or binary payload.

Solutions

  1. Exclude large/binary files from review context (respect .gitattributes or a size filter) before calling `source_context`
  2. Add the offending file to .gitignore or remove it from the repository
  3. Increase `MAX_CONTEXT_FILE_BYTES` only if the pipeline can genuinely afford larger contexts

Example fix

// before
for path in changed_files { ctx.push(source_context(ws, commit, path)?); }
// after
for path in changed_files {
    if is_probably_binary(ws, commit, &path) { continue; }
    if blob_size(ws, commit, &path)? > MAX_CONTEXT_FILE_BYTES { continue; }
    ctx.push(source_context(ws, commit, &path)?);
}
Defensive patterns

Strategy: validation

Validate before calling

let size = git(["cat-file", "-s", blob_id])?; if size > MAX_CONTEXT_FILE_BYTES { skip(path); } let head = read_first_bytes(ws, blob_id, 8192)?; if head.contains(&0) { skip_binary(path); }

Try / catch

match source_context(ws, commit, path) { Ok(src) => Some(src), Err(e) if e.to_string().contains("bounded text") => { warn!("skipping oversized/binary {path}"); None }, Err(e) => return Err(e) }

Prevention

When it happens

Trigger: The pinned blob exceeds `MAX_CONTEXT_FILE_BYTES`, or the blob is binary (contains a NUL byte) — e.g. a minified bundle, image, or compiled artifact was committed and included in the review context.

Common situations: Reviewing PRs that commit build artifacts, vendor trees, or generated minified JS; large data fixtures; accidentally committed binaries that pass the regular-file check.

Understand the failure class

Background: "File too large" / "file size exceeds limit" errors: why libraries cap file sizes and how to fix them — this error's family across 46 libraries.

Related errors

AI-assisted analysis of Hmbown/CodeWhale@73e0f67d83 (2026-09-22). Data as JSON: /api/errors/9868e696676d3145. Report an issue: GitHub.

Appendix: source

Thrown at crates/tui/src/tools/review_pr.rs:448

        .context("No pinned source blob")?;
    let fields = header.split_whitespace().collect::<Vec<_>>();
    anyhow::ensure!(
        returned_path == path
            && fields.len() == 4
            && matches!(fields[0], "100644" | "100755")
            && fields[1] == "blob"
            && commit_id(fields[2])
            && fields[3]
                .parse::<usize>()
                .is_ok_and(|size| size <= MAX_CONTEXT_FILE_BYTES),
        "Pinned source is missing, non-regular or exceeds the context limit"
    );
    let source = run_command(
        workspace,
        Program::Git,
        &["cat-file".into(), "blob".into(), fields[2].into()],
    )?;
    anyhow::ensure!(
        source.len() <= MAX_CONTEXT_FILE_BYTES && !source.contains('\0'),
        "Pinned source is not bounded text"
    );
    Ok(source)
}

fn read_bounded(reader: impl Read, limit: usize) -> std::io::Result<Vec<u8>> {
    let mut bytes = Vec::new();
    reader.take(limit as u64 + 1).read_to_end(&mut bytes)?;
    Ok(bytes)
}

fn run_command(workspace: &Path, program: Program, args: &[String]) -> Result<String> {
    let mut command = match program {
        Program::Gh => Gh::command().context("PR review requires GitHub CLI on PATH")?,
        Program::Git => Git::review_command(workspace)?,
    };
    command

View on GitHub (pinned to 73e0f67d83)