Merge pull request #36 from epage/bin

 feat: Ignore binary files

Cargo.toml

@@ -27,12 +27,12 @@ ignore = "0.4"
 phf = { version = "0.7", features = ["unicase"] }
 regex = "1.0"
 lazy_static = "1.2.0"
-grep-searcher = "0.1"
 serde = "1.0"
 serde_derive = "1.0"
 serde_json = "1.0"
 itertools = "0.8"
 unicase = "1.1"
+bstr = "0.2"
 
 [dev-dependencies]
 assert_fs = "0.10"

benches/file.rs

@@ -18,6 +18,7 @@ fn process_empty(b: &mut test::Bencher) {
             sample_path.path(),
             &corrections,
             true,
+            false,
             typos::report::print_silent,
         )
     });
@@ -37,6 +38,7 @@ fn process_no_tokens(b: &mut test::Bencher) {
             sample_path.path(),
             &corrections,
             true,
+            false,
             typos::report::print_silent,
         )
     });
@@ -56,6 +58,7 @@ fn process_single_token(b: &mut test::Bencher) {
             sample_path.path(),
             &corrections,
             true,
+            false,
             typos::report::print_silent,
         )
     });
@@ -75,6 +78,7 @@ fn process_sherlock(b: &mut test::Bencher) {
             sample_path.path(),
             &corrections,
             true,
+            false,
             typos::report::print_silent,
         )
     });
@@ -94,6 +98,7 @@ fn process_code(b: &mut test::Bencher) {
             sample_path.path(),
             &corrections,
             true,
+            false,
             typos::report::print_silent,
         )
     });
@@ -113,6 +118,7 @@ fn process_corpus(b: &mut test::Bencher) {
             sample_path.path(),
             &corrections,
             true,
+            false,
             typos::report::print_silent,
         )
     });

src/lib.rs

@@ -12,15 +12,22 @@ pub use crate::dict::*;
 use std::fs::File;
 use std::io::Read;
 
+use bstr::ByteSlice;
+
 pub fn process_file(
     path: &std::path::Path,
     dictionary: &Dictionary,
     ignore_hex: bool,
+    binary: bool,
     report: report::Report,
 ) -> Result<(), failure::Error> {
     let mut buffer = Vec::new();
     File::open(path)?.read_to_end(&mut buffer)?;
-    for (line_idx, line) in grep_searcher::LineIter::new(b'\n', &buffer).enumerate() {
+    if !binary && buffer.find_byte(b'\0').is_some() {
+        return Ok(());
+    }
+
+    for (line_idx, line) in buffer.lines().enumerate() {
         let line_num = line_idx + 1;
         for ident in tokens::Identifier::parse(line) {
             if !ignore_hex && is_hex(ident.token()) {

src/main.rs

@@ -55,6 +55,12 @@ struct Options {
     /// The approximate number of threads to use.
     threads: usize,
 
+    #[structopt(long, raw(overrides_with = r#""no-binary""#))]
+    /// Search binary files.
+    binary: bool,
+    #[structopt(long, raw(overrides_with = r#""binary""#), raw(hidden = "true"))]
+    no_binary: bool,
+
     #[structopt(long, raw(overrides_with = r#""no-hidden""#))]
     /// Search hidden files and directories.
     hidden: bool,
@@ -118,6 +124,15 @@ impl Options {
         }
     }
 
+    pub fn binary(&self) -> Option<bool> {
+        match (self.binary, self.no_binary) {
+            (true, false) => Some(true),
+            (false, true) => Some(false),
+            (false, false) => None,
+            (_, _) => unreachable!("StructOpt should make this impossible"),
+        }
+    }
+
     pub fn ignore_hidden(&self) -> Option<bool> {
         match (self.hidden, self.no_hidden) {
             (true, false) => Some(false),
@@ -183,6 +198,7 @@ fn run() -> Result<(), failure::Error> {
 
     let dictionary = typos::Dictionary::new();
     let ignore_hex = options.ignore_hex().unwrap_or(true);
+    let binary = options.binary().unwrap_or(false);
 
     let first_path = &options
         .path
@@ -207,6 +223,7 @@ fn run() -> Result<(), failure::Error> {
                 entry.path(),
                 &dictionary,
                 ignore_hex,
+                binary,
                 options.format.report(),
             )?;
         }