Initial commit: noxis JVM bytecode obfuscator

Workspace of 9 crates (noxis, noxis-core, noxis-classfile, noxis-analysis,
noxis-passes, noxis-runtime, noxis-jar, noxis-verify, noxis-structure-check).
AGPL-3.0-only license, Rust 2024 edition, unsafe_code denied.

crates/noxis-structure-check/Cargo.toml

@@ -0,0 +1,13 @@
+[package]
+name = "noxis-structure-check"
+description = "Автогейт структурных правил проекта при компиляции"
+version.workspace = true
+edition.workspace = true
+rust-version.workspace = true
+license.workspace = true
+
+[lib]
+path = "src/lib.rs"
+
+[lints]
+workspace = true

crates/noxis-structure-check/build.rs

@@ -0,0 +1,9 @@
+//! Bootstrap gate: this crate cannot depend on itself, so it includes the
+//! shared rules source directly instead of going through the crate API.
+
+#[path = "src/rules.rs"]
+mod rules;
+
+fn main() {
+    rules::check();
+}

crates/noxis-structure-check/src/lib.rs

@@ -0,0 +1,79 @@
+//! Compile-time gate for the project structure rules.
+//!
+//! Every crate calls [`check`] from its `build.rs`, so a file that grows past
+//! 250 lines or a directory that fills up with a fifth file fails the build.
+
+mod rules;
+
+pub use rules::{MAX_DIR_FILES, MAX_FILE_LINES, check, scan};
+
+#[cfg(test)]
+mod tests {
+    #![allow(clippy::expect_used, clippy::unwrap_used)]
+    use std::fs;
+    use std::path::{Path, PathBuf};
+
+    use super::scan;
+
+    struct TempDir(PathBuf);
+
+    impl TempDir {
+        fn new(tag: &str) -> Self {
+            let path =
+                std::env::temp_dir().join(format!("noxis-structure-{tag}-{}", std::process::id()));
+            let _ = fs::remove_dir_all(&path);
+            fs::create_dir_all(&path).expect("create temp dir");
+            Self(path)
+        }
+    }
+
+    impl Drop for TempDir {
+        fn drop(&mut self) {
+            let _ = fs::remove_dir_all(&self.0);
+        }
+    }
+
+    fn violations(dir: &Path) -> Vec<String> {
+        let mut out = Vec::new();
+        scan(dir, &mut out);
+        out
+    }
+
+    #[test]
+    fn rejects_file_over_line_limit() {
+        let tmp = TempDir::new("long");
+        fs::write(tmp.0.join("lib.rs"), "fn main() {}\n".repeat(300)).expect("write");
+        let found = violations(&tmp.0);
+        assert_eq!(found.len(), 1, "{found:?}");
+        assert!(found[0].contains("lines >"), "{found:?}");
+    }
+
+    #[test]
+    fn rejects_directory_over_file_limit() {
+        let tmp = TempDir::new("many");
+        for index in 0..super::MAX_DIR_FILES + 1 {
+            fs::write(tmp.0.join(format!("m{index}.rs")), "").expect("write");
+        }
+        let found = violations(&tmp.0);
+        assert_eq!(found.len(), 1, "{found:?}");
+        assert!(found[0].contains("files >"), "{found:?}");
+    }
+
+    #[test]
+    fn accepts_compact_tree() {
+        let tmp = TempDir::new("ok");
+        fs::create_dir(tmp.0.join("sub")).expect("mkdir");
+        fs::write(tmp.0.join("lib.rs"), "pub fn f() {}\n").expect("write");
+        fs::write(tmp.0.join("sub").join("a.rs"), "pub fn a() {}\n").expect("write");
+        let found = violations(&tmp.0);
+        assert!(found.is_empty(), "{found:?}");
+    }
+
+    #[test]
+    fn ignores_non_limited_extensions() {
+        let tmp = TempDir::new("ext");
+        fs::write(tmp.0.join("notes.txt"), "line\n".repeat(300)).expect("write");
+        let found = violations(&tmp.0);
+        assert!(found.is_empty(), "{found:?}");
+    }
+}

crates/noxis-structure-check/src/rules.rs

@@ -0,0 +1,165 @@
+//! Project structure rules, invoked from every crate's `build.rs`.
+//!
+//! Rules:
+//! * every `.rs` and `.md` file is at most [`MAX_FILE_LINES`] lines long;
+//! * every directory inside a crate `src/` or the workspace `docs/` holds at
+//!   most [`MAX_DIR_FILES`] files — subdirectories are unlimited.
+//!
+//! Deliberately excluded: `Cargo.lock`, `.gitignore`, the legacy `info/` tree,
+//! and crate roots (they always own `Cargo.toml` plus `build.rs`).
+
+use std::ffi::OsStr;
+use std::fs;
+use std::path::{Path, PathBuf};
+
+/// Maximum number of lines in a source (`rs`) or documentation (`md`) file.
+pub const MAX_FILE_LINES: usize = 250;
+
+/// Maximum number of files per directory; subdirectories are not counted.
+pub const MAX_DIR_FILES: usize = 4;
+
+/// Extensions subject to [`MAX_FILE_LINES`].
+const LINE_LIMITED_EXTS: [&str; 2] = ["rs", "md"];
+
+/// Directory names the scan never descends into.
+const SKIPPED_DIRS: [&str; 2] = ["target", ".git"];
+
+/// Scans the calling crate's `src/`/`tests/`/`benches/` plus the workspace
+/// `docs/` tree and panics with a report on the first violation.
+///
+/// # Panics
+///
+/// Panics whenever a rule is violated — that is the point of the gate.
+pub fn check() {
+    let manifest = manifest_dir();
+    let mut violations = Vec::new();
+
+    // `src/` обязателен; `tests/` и `benches/` проверяются, если созданы —
+    // иначе правило «250 строк» обходится выносом кода в тесты.
+    let mut roots = vec![manifest.join("src"), manifest.join("tests")];
+    roots.push(manifest.join("benches"));
+
+    if let Some(root) = workspace_root(&manifest) {
+        roots.push(root.join("docs"));
+    }
+
+    for root in &roots {
+        watch(root);
+        scan(root, &mut violations);
+    }
+
+    if violations.is_empty() {
+        return;
+    }
+
+    let mut report = String::from("project structure rules violated:\n");
+    for violation in &violations {
+        report.push_str("  - ");
+        report.push_str(violation);
+        report.push('\n');
+    }
+    report.push_str("split the file, or move code into a subdirectory (rules: docs/README.md)");
+    panic!("{report}");
+}
+
+/// Re-runs the build script whenever a checked tree changes.
+///
+/// Without this cargo only invalidates `build.rs` on crate-local changes, so
+/// editing `docs/` would never be re-validated.
+fn watch(root: &Path) {
+    if root.is_dir() {
+        println!("cargo:rerun-if-changed={}", root.display());
+    }
+}
+
+/// Collects every rule violation found under `root` into `violations`.
+pub fn scan(root: &Path, violations: &mut Vec<String>) {
+    if !root.is_dir() {
+        return;
+    }
+    let entries = match fs::read_dir(root) {
+        Ok(entries) => entries,
+        Err(err) => {
+            violations.push(format!("{}: unreadable: {err}", root.display()));
+            return;
+        }
+    };
+
+    let mut files = 0usize;
+    let mut subdirs = Vec::new();
+
+    for entry in entries.flatten() {
+        let path = entry.path();
+        if path.is_dir() {
+            let name = entry.file_name();
+            if SKIPPED_DIRS
+                .iter()
+                .any(|s| name.as_os_str() == OsStr::new(s))
+            {
+                continue;
+            }
+            subdirs.push(path);
+            continue;
+        }
+        files += 1;
+        check_line_limit(&path, violations);
+    }
+
+    if files > MAX_DIR_FILES {
+        violations.push(format!(
+            "{}: {files} files > {MAX_DIR_FILES} (subdirectories are allowed)",
+            root.display()
+        ));
+    }
+
+    for subdir in subdirs {
+        scan(&subdir, violations);
+    }
+}
+
+fn check_line_limit(path: &Path, violations: &mut Vec<String>) {
+    let ext = path.extension().and_then(OsStr::to_str).unwrap_or_default();
+    if !LINE_LIMITED_EXTS.contains(&ext) {
+        return;
+    }
+    let lines = count_lines(path);
+    if lines > MAX_FILE_LINES {
+        violations.push(format!(
+            "{}: {lines} lines > {MAX_FILE_LINES}",
+            path.display()
+        ));
+    }
+}
+
+/// Counts lines without requiring the file to be valid UTF-8.
+fn count_lines(path: &Path) -> usize {
+    let Ok(bytes) = fs::read(path) else {
+        return 0;
+    };
+    if bytes.is_empty() {
+        return 0;
+    }
+    let newlines = bytes.iter().filter(|&&b| b == b'\n').count();
+    newlines + usize::from(bytes.last() != Some(&b'\n'))
+}
+
+fn manifest_dir() -> PathBuf {
+    match std::env::var_os("CARGO_MANIFEST_DIR") {
+        Some(value) => PathBuf::from(value),
+        None => panic!("CARGO_MANIFEST_DIR is unset; run this check from a build script"),
+    }
+}
+
+/// Walks up from `start` until a manifest declaring `[workspace]` is found.
+fn workspace_root(start: &Path) -> Option<PathBuf> {
+    let mut current = Some(start);
+    while let Some(dir) = current {
+        if let Ok(text) = fs::read_to_string(dir.join("Cargo.toml"))
+            && text.lines().any(|line| line.trim() == "[workspace]")
+        {
+            return Some(dir.to_path_buf());
+        }
+        current = dir.parent();
+    }
+    None
+}