for release

CONTRIBUTING.md

@@ -12,7 +12,9 @@ If you'd like binaries for any of the platforms listed
 ## Support for other languages
 
 If you'd like me to add support for other languages, just open an issue and tell
-me all file extensions associated with its build artifacts.
+me all file extensions associated with its build artifacts, plus common names
+for build directories, and names/extensions of any configuration files typically
+used.
 
 Alternately, if you'd like to fork it and open a PR yourself, the relevant regex is
 [here](https://github.com/vmchale/file-sniffer/blob/master/src/walk_parallel/single_threaded.rs#L73).

README.md

@@ -4,7 +4,7 @@
 
 If you do a significant amount of programming, you'll probably end up with
 build artifacts scattered about. `sn` is a tool to help you find those
-artifacts. 
+artifacts.
 
 `sn` is also a replacement for `du`. It has far nicer
 output, saner commands and defaults, and it even runs faster on big directories
@@ -82,14 +82,14 @@ To turn off colorized output:
 export CLICOLOR=0
 ```
 
-### Comparison
+### Comparison (or, 100 Things I Hate About du)
 
 #### Reasons to use `du`
 
   * Reads disk usage, not file sizes
   * Optionally dereferences symlinks
-  * Slightly faster on small directories 
-  * Well-supported
+  * Slightly faster on small directories
+  * Stable and well-supported
 
 #### Reasons to use `sn`
 
@@ -105,6 +105,29 @@ export CLICOLOR=0
   * Extensible in Rust
   * Benefits from upstream improvements in Rust ecosystem
 
+#### Benchmark results
+
+| Directory | Tool | Command | Time |
+| --------- || ---- | ------- | ---- |
+| Source | sn | `sn p` | 60.74 ms |
+| Source | sn | `sn a` | 99.92 ms |
+| Source | du | `du -hacd2` | 88.28 ms |
+| Build | sn | `sn p`| 185.2 ms |
+| Build | sn | `sn a` | 271.9 ms |
+| Build | du | `du -hacd2` | 195.5 ms |
+| Project | sn | `sn p` | 36.68 ms |
+| Project | sn | `sn a` | 42.90 ms |
+| Project | du | `du -hacd2` | 35.53 ms |
+
+These commands are all essentially equivalent in function, except that `sn p`
+may use more threads than `sn a` or `du`.
+
+Results were obtained using Gabriel Gonzalez's [bench](https://github.com/Gabriel439/bench)
+tool. "Source" was my programming directory alone, comprising data, source code,
+and version control; around 600MB total. "Project" was a single polyglot project,
+plus artifacts; around 1GB total. "Build" was my programming directory, with
+current projects built; around 4GB total.
+
 #### Screenshots (alacritty + solarized dark)
 
 ##### The Tin Summer
@@ -119,7 +142,8 @@ export CLICOLOR=0
 
 Currently, `sn` looks for files that either have an extension associated with
 build artifacts, or executable files that are ignored by version control. It also looks for "build
-directories", like `.stack-work`, `elm-stuff`, etc. and it considers *all* their
+directories", like `.stack-work`, `elm-stuff`, etc. and if it finds a
+configuration file like `tweet-hs.cabal`, it considers *all* their
 contents to be build artifacts.
 
 #### Languages Supported
@@ -136,4 +160,5 @@ The *intent* is to support basically anything, so feel free to open a PR or star
   - [x] Vimscript
   - [x] Idris
   - [x] FORTRAN
+  - [ ] Ruby
   - [ ] C

TODO.md

@@ -37,9 +37,15 @@
   - [x] fix bugs w/ excludes & overzealous use of .gitignores
   - [x] multiple included paths
 - [x] let it run on a single file
+- [x] don't call `is_project_dir()` three times
 - [x] improve ergonomics (and speed) by guessing language of project
   directory
 
+# Bugs
+
+- [ ] when using `--all` with `ar`, it should not recurse arbitrarily far.
+- [ ] when `--all` is used with `sort`, it *should* recurse arbitrarily far.
+
 # UI/Ergonomics
 
 - [ ] silent flag to ignore warnings?
@@ -65,6 +71,12 @@
   - [ ] elm
   - [ ] python
 
+# Code maintenance
+
+- [ ] make `read_all()` take a struct.
+- [ ] strip out machinery for `with_gitignore` and `artifact_regex`
+
+
 # Performance
 
 - [ ] parity with du without threading
@@ -97,4 +109,4 @@
   - [ ] make an error type & use that to organize things
 - [x] change french/german binary name
   - [ ] upsteam PR to clap-rs?
-- [ ] fix build.rs
+- [x] fix build.rs

build.rs

@@ -11,7 +11,12 @@ use clap::App;
 fn main() {
 
     // load configuration
+    #[cfg(feature = "english")]
     let yaml = load_yaml!("src/cli/options-en.yml");
+    #[cfg(feature = "francais")]
+    let yaml = load_yaml!("src/cli/options-fr.yml");
+    #[cfg(feature = "deutsch")]
+    let yaml = load_yaml!("src/cli/options-de.yml");
     let mut app = App::from_yaml(yaml).version(crate_version!());
 
     // generate bash completions if desired

src/cli/options-en.yml

@@ -191,10 +191,6 @@ subcommands:
                 short: o
                 long: sort
                 help: Sort results by size
-            - gitignore:
-                short: g
-                long: no-gitignore
-                help: Don't bother using .gitignore or darcs boring file information for faster traversal
             - depth: 
                 short: d
                 long: depth

src/test.rs

@@ -20,7 +20,7 @@ fn bench_cli_options(b: &mut Bencher) {
     b.iter(|| {
         App::from_yaml(yaml)
             .version(crate_version!())
-            .get_matches_from(vec!["sn", "ar", "-g", "."])
+            .get_matches_from(vec!["sn", "ar", "."])
     })
 }
 

src/utils.rs

@@ -9,10 +9,20 @@ use self::num_cpus::get;
 
 /// Gather the information from `.gitignore`, `.ignore`, and darcs `boring` files in a given
 /// directory, and assemble a `RegexSet` from it.
-pub fn mk_ignores(in_paths: &PathBuf, maybe_gitignore: &Option<RegexSet>) -> Option<RegexSet> {
+pub fn mk_ignores(in_paths: &PathBuf, maybe_ignore: &Option<RegexSet>) -> Option<RegexSet> {
 
-    if let Some(ref gitignore) = *maybe_gitignore {
-        Some(gitignore.to_owned())
+    if let Some(ref ignore) = *maybe_ignore {
+        Some(ignore.to_owned())
+    } else if let (ignore_path, Ok(mut file)) =
+        {
+            let mut ignore_path = in_paths.clone();
+            ignore_path.push(".ignore");
+            (ignore_path.clone(), File::open(ignore_path.clone()))
+        } {
+        let mut contents = String::new();
+        file.read_to_string(&mut contents)
+            .expect("File read failed."); // ok because we check that the file exists
+        Some(file_contents_to_regex(&contents, &ignore_path))
     } else if let (gitignore_path, Ok(mut file)) =
         {
             let mut gitignore_path = in_paths.clone();
@@ -33,16 +43,6 @@ pub fn mk_ignores(in_paths: &PathBuf, maybe_gitignore: &Option<RegexSet>) -> Opt
         file.read_to_string(&mut contents)
             .expect("File read failed."); // ok because we check that the file exists
         Some(darcs_contents_to_regex(&contents, &darcs_path))
-    } else if let (ignore_path, Ok(mut file)) =
-        {
-            let mut ignore_path = in_paths.clone();
-            ignore_path.push(".ignore");
-            (ignore_path.clone(), File::open(ignore_path.clone()))
-        } {
-        let mut contents = String::new();
-        file.read_to_string(&mut contents)
-            .expect("File read failed."); // ok because we check that the file exists
-        Some(file_contents_to_regex(&contents, &ignore_path))
     } else {
         None
     }

src/walk_parallel/single_threaded.rs

@@ -19,9 +19,19 @@ fn glob_exists(s: &str) -> bool {
     glob(s).unwrap().filter_map(Result::ok).count() != 0 // ok because panic on IO Errors is good?
 }
 
-/// Helper function to identify project directories. The heuristic is as follows: 
+/// Helper function to identify project directories. The heuristic is as follows:
 ///
-/// 1. For `.stack-work`, look for a 
+/// 1. For `.stack-work`, look for a `.cabal` file or a `package.yaml` file in the parent
+///    directory.
+/// 2. For `target`, look for a `Cargo.toml` file in the parent directory.
+/// 3. For `elm-stuff`, look for `elm-package.json` in the parent directory.
+/// 4. For `build`, `dist`, look for a `.cabal`, `setup.py` or `cabal.project` file.
+/// 5. For `dist-newstyle`, look for a `.cabal` or `cabal.project` file.
+/// 6. For `nimcache`, look for a `.nim` file in the parent directory.
+/// 6. Otherwise, if `setup.py` is in the parent directory and it ends with `.egg-info`, return
+///    true.
+/// 7. In all other cases, return false, but still proceed into the directory to search files by
+///    extension.
 pub fn is_project_dir(p: &str, name: &str) -> bool {
     // for project directories
     lazy_static! {
@@ -41,6 +51,10 @@ pub fn is_project_dir(p: &str, name: &str) -> bool {
                 parent_string.push_str("/../*.cabal");
                 parent_path.exists() || hpack.exists() || glob_exists(&parent_string)
             }
+            "nimcache" => {
+                parent_string.push_str("/../*.nim");
+                glob_exists(&parent_string)
+            }
             "target" => {
                 parent_path.push("../Cargo.toml");
                 parent_path.exists()
@@ -250,7 +264,6 @@ pub fn read_size(
                 if path_type.is_file() {
                     // if this fails, it's probably because `path` is a broken symlink
                     if let Ok(metadata) = val.metadata() {
-                        // faster on Windows
                         if !artifacts_only ||
                             {
                                 is_artifact(
@@ -272,7 +285,6 @@ pub fn read_size(
                     let dir_size = if artifacts_only &&
                         is_project_dir(path_string, val.file_name().to_str().unwrap())
                     {
-                        // REGEX_PROJECT_DIR.is_match(path_string) {
                         read_size(
                             &path,
                             depth + 1,
@@ -414,9 +426,21 @@ pub fn read_all(
                 // otherwise, go deeper
                 else if path_type.is_dir() {
                     if let Some(d) = max_depth {
-                        if depth + 1 >= d ||
-                            (artifacts_only &&
-                                 is_project_dir(path_string, val.file_name().to_str().unwrap()))
+                        if depth + 1 >= d && (!with_gitignore || !artifacts_only) {
+                            let dir_size = {
+                                read_size(
+                                    &path,
+                                    depth + 1,
+                                    artifact_regex,
+                                    excludes,
+                                    &gitignore,
+                                    with_gitignore,
+                                    artifacts_only,
+                                )
+                            };
+                            tree.push(path_string.to_string(), dir_size, None, depth + 1, true);
+                        } else if artifacts_only &&
+                                   is_project_dir(path_string, val.file_name().to_str().unwrap())
                         {
                             let dir_size = {
                                 read_size(
@@ -425,16 +449,8 @@ pub fn read_all(
                                     artifact_regex,
                                     excludes,
                                     &gitignore,
-                                    with_gitignore &&
-                                        !is_project_dir(
-                                            path_string,
-                                            val.file_name().to_str().unwrap(),
-                                        ),
-                                    artifacts_only &&
-                                        !is_project_dir(
-                                            path_string,
-                                            val.file_name().to_str().unwrap(),
-                                        ), // FIXME only compute this once
+                                    false,
+                                    false,
                                 )
                             };
                             tree.push(path_string.to_string(), dir_size, None, depth + 1, true);