0.2.1

Fixes #99.

CHANGELOG.md

@@ -1,3 +1,26 @@
+0.2.1
+=====
+Feature enhancements:
+
+* Added or improved file type filtering for Clojure and SystemVerilog.
+* [FEATURE #89](https://github.com/BurntSushi/ripgrep/issues/89):
+  Add a --null flag that outputs a NUL byte after every file path.
+
+Bug fixes:
+
+* [BUG #98](https://github.com/BurntSushi/ripgrep/issues/98):
+  Fix a bug in single threaded mode when if opening a file failed, ripgrep
+  quit instead of continuing the search.
+* [BUG #99](https://github.com/BurntSushi/ripgrep/issues/99):
+  Fix another bug in single threaded mode where empty lines were being printed
+  by mistake.
+* [BUG #105](https://github.com/BurntSushi/ripgrep/issues/105):
+  Fix an off-by-one error with --column.
+* [BUG #106](https://github.com/BurntSushi/ripgrep/issues/106):
+  Fix a bug where a whitespace only line in a gitignore file caused ripgrep
+  to panic (i.e., crash).
+
+
 0.2.0
 =====
 Feature enhancements:

Cargo.lock

@@ -1,6 +1,6 @@
 [root]
 name = "ripgrep"
-version = "0.2.0"
+version = "0.2.1"
 dependencies = [
  "deque 0.3.1 (registry+https://github.com/rust-lang/crates.io-index)",
  "docopt 0.6.85 (registry+https://github.com/rust-lang/crates.io-index)",

Cargo.toml

@@ -1,6 +1,6 @@
 [package]
 name = "ripgrep"
-version = "0.2.0"  #:version
+version = "0.2.1"  #:version
 authors = ["Andrew Gallant <jamslam@gmail.com>"]
 description = """
 Line oriented search tool using Rust's regex library. Combines the raw

doc/rg.1

@@ -239,6 +239,14 @@ Note that .ignore files will continue to be respected.
 .RS
 .RE
 .TP
+.B \-\-null
+Whenever a file name is printed, follow it with a NUL byte.
+This includes printing filenames before matches, and when printing a
+list of matching files such as with \-\-count, \-\-files\-with\-matches
+and \-\-files.
+.RS
+.RE
+.TP
 .B \-p, \-\-pretty
 Alias for \-\-color=always \-\-heading \-n.
 .RS

doc/rg.1.md

@@ -155,6 +155,12 @@ the raw speed of grep.
 : Don't respect version control ignore files (e.g., .gitignore).
   Note that .ignore files will continue to be respected.
 
+--null
+: Whenever a file name is printed, follow it with a NUL byte.
+  This includes printing filenames before matches, and when printing
+  a list of matching files such as with --count, --files-with-matches
+  and --files.
+
 -p, --pretty
 : Alias for --color=always --heading -n.
 

pkg/brew/ripgrep.rb

@@ -1,15 +1,15 @@
 require  'formula'
 class Ripgrep < Formula
-  version '0.1.17'
+  version '0.2.0'
   desc "Search tool like grep and The Silver Searcher."
   homepage "https://github.com/BurntSushi/ripgrep"
 
   if Hardware::CPU.is_64_bit?
     url "https://github.com/BurntSushi/ripgrep/releases/download/#{version}/ripgrep-#{version}-x86_64-apple-darwin.tar.gz"
-    sha256 "cb7b551a08849cef6ef8f17229224f094299692981976a3c5873c93f68c8fa1a"
+    sha256 "f55ef5dac04178bcae0d6c5ba2d09690d326e8c7c3f28e561025b04e1ab81d80"
   else
     url "https://github.com/BurntSushi/ripgrep/releases/download/#{version}/ripgrep-#{version}-i686-apple-darwin.tar.gz"
-    sha256 "0e936874b9f3fd661c5566e7f8fe18343baa5e9371e57d8d71000e9234fc376b"
+    sha256 "d901d55ccb48c19067f563d42652dfd8642bf50d28a40c0e2a4d3e866857a93b"
   end
 
   def install

src/args.rs

@@ -155,6 +155,12 @@ Less common options:
         Don't respect version control ignore files (e.g., .gitignore).
         Note that .ignore files will continue to be respected.
 
+    --null
+        Whenever a file name is printed, follow it with a NUL byte.
+        This includes printing filenames before matches, and when printing
+        a list of matching files such as with --count, --files-with-matches
+        and --files.
+
     -p, --pretty
         Alias for --color=always --heading -n.
 
@@ -224,6 +230,7 @@ pub struct RawArgs {
     flag_no_line_number: bool,
     flag_no_mmap: bool,
     flag_no_filename: bool,
+    flag_null: bool,
     flag_pretty: bool,
     flag_quiet: bool,
     flag_regexp: Vec<String>,
@@ -269,6 +276,7 @@ pub struct Args {
     no_ignore: bool,
     no_ignore_parent: bool,
     no_ignore_vcs: bool,
+    null: bool,
     quiet: bool,
     replace: Option<Vec<u8>>,
     text: bool,
@@ -399,6 +407,7 @@ impl RawArgs {
             no_ignore_vcs:
                 // --no-ignore implies --no-ignore-vcs
                 self.flag_no_ignore_vcs || no_ignore,
+            null: self.flag_null,
             quiet: self.flag_quiet,
             replace: self.flag_replace.clone().map(|s| s.into_bytes()),
             text: text,
@@ -553,6 +562,7 @@ impl Args {
             .heading(self.heading)
             .line_per_match(self.line_per_match)
             .quiet(self.quiet)
+            .null(self.null)
             .with_filename(self.with_filename);
         if let Some(ref rep) = self.replace {
             p = p.replace(rep.clone());
@@ -673,15 +683,20 @@ impl Args {
     pub fn walker(&self, path: &Path) -> Result<walk::Iter> {
         let wd = WalkDir::new(path).follow_links(self.follow);
         let mut ig = Ignore::new();
-        ig.ignore_hidden(!self.hidden);
-        ig.no_ignore(self.no_ignore);
-        ig.no_ignore_vcs(self.no_ignore_vcs);
-        ig.add_types(self.types.clone());
-        if !self.no_ignore_parent {
-            try!(ig.push_parents(path));
-        }
-        if let Some(ref overrides) = self.glob_overrides {
-            ig.add_override(overrides.clone());
+        // Only register ignore rules if this is a directory. If it's a file,
+        // then it was explicitly given by the end user, so we always search
+        // it.
+        if path.is_dir() {
+            ig.ignore_hidden(!self.hidden);
+            ig.no_ignore(self.no_ignore);
+            ig.no_ignore_vcs(self.no_ignore_vcs);
+            ig.add_types(self.types.clone());
+            if !self.no_ignore_parent {
+                try!(ig.push_parents(path));
+            }
+            if let Some(ref overrides) = self.glob_overrides {
+                ig.add_override(overrides.clone());
+            }
         }
         Ok(walk::Iter::new(ig, wd))
     }

src/gitignore.rs

@@ -283,12 +283,15 @@ impl GitignoreBuilder {
         from: P,
         mut line: &str,
     ) -> Result<(), Error> {
-        if line.is_empty() || line.starts_with("#") {
+        if line.starts_with("#") {
             return Ok(());
         }
         if !line.ends_with("\\ ") {
             line = line.trim_right();
         }
+        if line.is_empty() {
+            return Ok(());
+        }
         let mut pat = Pattern {
             from: from.as_ref().to_path_buf(),
             original: line.to_string(),
@@ -426,4 +429,10 @@ mod tests {
     not_ignored!(ignot11, ROOT, "#foo", "#foo");
     not_ignored!(ignot12, ROOT, "\n\n\n", "foo");
     not_ignored!(ignot13, ROOT, "foo/**", "foo", true);
+
+    // See: https://github.com/BurntSushi/ripgrep/issues/106
+    #[test]
+    fn regression_106() {
+        Gitignore::from_str("/", " ").unwrap();
+    }
 }

src/main.rs

@@ -22,8 +22,8 @@ extern crate winapi;
 
 use std::error::Error;
 use std::fs::File;
-use std::io::{self, Write};
-use std::path::{Path, PathBuf};
+use std::io;
+use std::path::Path;
 use std::process;
 use std::result;
 use std::sync::{Arc, Mutex};
@@ -89,16 +89,15 @@ fn run(args: Args) -> Result<u64> {
     let args = Arc::new(args);
     let paths = args.paths();
     let threads = cmp::max(1, args.threads() - 1);
+    let isone =
+        paths.len() == 1 && (paths[0] == Path::new("-") || paths[0].is_file());
     if args.files() {
         return run_files(args.clone());
     }
     if args.type_list() {
         return run_types(args.clone());
     }
-    if paths.len() == 1 && (paths[0] == Path::new("-") || paths[0].is_file()) {
-        return run_one(args.clone(), &paths[0]);
-    }
-    if threads == 1 {
+    if threads == 1 || isone {
         return run_one_thread(args.clone());
     }
 
@@ -158,32 +157,35 @@ fn run_one_thread(args: Arc<Args>) -> Result<u64> {
         match_count: 0,
     };
     let paths = args.paths();
-    let filesep = args.file_separator();
     let mut term = args.stdout();
 
     let mut paths_searched: u64 = 0;
     for p in paths {
         if p == Path::new("-") {
-            if worker.match_count > 0 {
-                if let Some(ref sep) = filesep {
-                    let _ = term.write_all(sep);
-                    let _ = term.write_all(b"\n");
-                }
-            }
             paths_searched += 1;
             let mut printer = args.printer(&mut term);
+            if worker.match_count > 0 {
+                if let Some(sep) = args.file_separator() {
+                    printer = printer.file_separator(sep);
+                }
+            }
             worker.do_work(&mut printer, WorkReady::Stdin);
         } else {
             for ent in try!(args.walker(p)) {
-                if worker.match_count > 0 {
-                    if let Some(ref sep) = filesep {
-                        let _ = term.write_all(sep);
-                        let _ = term.write_all(b"\n");
-                    }
-                }
                 paths_searched += 1;
                 let mut printer = args.printer(&mut term);
-                let file = try!(File::open(ent.path()));
+                if worker.match_count > 0 {
+                    if let Some(sep) = args.file_separator() {
+                        printer = printer.file_separator(sep);
+                    }
+                }
+                let file = match File::open(ent.path()) {
+                    Ok(file) => file,
+                    Err(err) => {
+                        eprintln!("{}: {}", ent.path().display(), err);
+                        continue;
+                    }
+                };
                 worker.do_work(&mut printer, WorkReady::DirFile(ent, file));
             }
         }
@@ -196,25 +198,6 @@ fn run_one_thread(args: Arc<Args>) -> Result<u64> {
     Ok(worker.match_count)
 }
 
-fn run_one(args: Arc<Args>, path: &Path) -> Result<u64> {
-    let mut worker = Worker {
-        args: args.clone(),
-        inpbuf: args.input_buffer(),
-        grep: args.grep(),
-        match_count: 0,
-    };
-    let term = args.stdout();
-    let mut printer = args.printer(term);
-    let work =
-        if path == Path::new("-") {
-            WorkReady::Stdin
-        } else {
-            WorkReady::PathFile(path.to_path_buf(), try!(File::open(path)))
-        };
-    worker.do_work(&mut printer, work);
-    Ok(worker.match_count)
-}
-
 fn run_files(args: Arc<Args>) -> Result<u64> {
     let term = args.stdout();
     let mut printer = args.printer(term);
@@ -253,7 +236,6 @@ enum Work {
 enum WorkReady {
     Stdin,
     DirFile(DirEntry, File),
-    PathFile(PathBuf, File),
 }
 
 struct MultiWorker {
@@ -328,17 +310,6 @@ impl Worker {
                     self.search(printer, path, file)
                 }
             }
-            WorkReady::PathFile(path, file) => {
-                let mut path = &*path;
-                if let Some(p) = strip_prefix("./", path) {
-                    path = p;
-                }
-                if self.args.mmap() {
-                    self.search_mmap(printer, path, &file)
-                } else {
-                    self.search(printer, path, file)
-                }
-            }
         };
         match result {
             Ok(count) => {

src/out.rs

@@ -48,8 +48,6 @@ impl Out {
 
     /// If set, the separator is printed between matches from different files.
     /// By default, no separator is printed.
-    ///
-    /// If sep is empty, then no file separator is printed.
     pub fn file_separator(mut self, sep: Vec<u8>) -> Out {
         self.file_separator = Some(sep);
         self

src/printer.rs

@@ -25,6 +25,8 @@ pub struct Printer<W> {
     /// printed via the match directly, but occasionally we need to insert them
     /// ourselves (for example, to print a context separator).
     eol: u8,
+    /// A file separator to show before any matches are printed.
+    file_separator: Option<Vec<u8>>,
     /// Whether to show file name as a heading or not.
     ///
     /// N.B. If with_filename is false, then this setting has no effect.
@@ -33,6 +35,9 @@ pub struct Printer<W> {
     line_per_match: bool,
     /// Whether to suppress all output.
     quiet: bool,
+    /// Whether to print NUL bytes after a file path instead of new lines
+    /// or `:`.
+    null: bool,
     /// A string to use as a replacement of each match in a matching line.
     replace: Option<Vec<u8>>,
     /// Whether to prefix each match with the corresponding file name.
@@ -48,9 +53,11 @@ impl<W: Terminal + Send> Printer<W> {
             column: false,
             context_separator: "--".to_string().into_bytes(),
             eol: b'\n',
+            file_separator: None,
             heading: false,
             line_per_match: false,
             quiet: false,
+            null: false,
             replace: None,
             with_filename: false,
         }
@@ -75,6 +82,13 @@ impl<W: Terminal + Send> Printer<W> {
         self
     }
 
+    /// If set, the separator is printed before any matches. By default, no
+    /// separator is printed.
+    pub fn file_separator(mut self, sep: Vec<u8>) -> Printer<W> {
+        self.file_separator = Some(sep);
+        self
+    }
+
     /// Whether to show file name as a heading or not.
     ///
     /// N.B. If with_filename is false, then this setting has no effect.
@@ -89,6 +103,13 @@ impl<W: Terminal + Send> Printer<W> {
         self
     }
 
+    /// Whether to cause NUL bytes to follow file paths instead of other
+    /// visual separators (like `:`, `-` and `\n`).
+    pub fn null(mut self, yes: bool) -> Printer<W> {
+        self.null = yes;
+        self
+    }
+
     /// When set, all output is suppressed.
     pub fn quiet(mut self, yes: bool) -> Printer<W> {
         self.quiet = yes;
@@ -146,14 +167,22 @@ impl<W: Terminal + Send> Printer<W> {
     pub fn path<P: AsRef<Path>>(&mut self, path: P) {
         let path = strip_prefix("./", path.as_ref()).unwrap_or(path.as_ref());
         self.write_path(path);
-        self.write_eol();
+        if self.null {
+            self.write(b"\x00");
+        } else {
+            self.write_eol();
+        }
     }
 
     /// Prints the given path and a count of the number of matches found.
     pub fn path_count<P: AsRef<Path>>(&mut self, path: P, count: u64) {
         if self.with_filename {
             self.write_path(path);
-            self.write(b":");
+            if self.null {
+                self.write(b"\x00");
+            } else {
+                self.write(b":");
+            }
         }
         self.write(count.to_string().as_bytes());
         self.write_eol();
@@ -186,7 +215,7 @@ impl<W: Terminal + Send> Printer<W> {
             let column =
                 if self.column {
                     Some(re.find(&buf[start..end])
-                           .map(|(s, _)| s + 1).unwrap_or(0) as u64)
+                           .map(|(s, _)| s).unwrap_or(0) as u64)
                 } else {
                     None
                 };
@@ -211,10 +240,15 @@ impl<W: Terminal + Send> Printer<W> {
         column: Option<u64>,
     ) {
         if self.heading && self.with_filename && !self.has_printed {
+            self.write_file_sep();
             self.write_heading(path.as_ref());
         } else if !self.heading && self.with_filename {
             self.write_path(path.as_ref());
-            self.write(b":");
+            if self.null {
+                self.write(b"\x00");
+            } else {
+                self.write(b":");
+            }
         }
         if let Some(line_number) = line_number {
             self.line_number(line_number, b':');
@@ -261,10 +295,15 @@ impl<W: Terminal + Send> Printer<W> {
         line_number: Option<u64>,
     ) {
         if self.heading && self.with_filename && !self.has_printed {
+            self.write_file_sep();
             self.write_heading(path.as_ref());
         } else if !self.heading && self.with_filename {
             self.write_path(path.as_ref());
-            self.write(b"-");
+            if self.null {
+                self.write(b"\x00");
+            } else {
+                self.write(b"-");
+            }
         }
         if let Some(line_number) = line_number {
             self.line_number(line_number, b'-');
@@ -281,7 +320,11 @@ impl<W: Terminal + Send> Printer<W> {
             let _ = self.wtr.attr(Attr::Bold);
         }
         self.write_path(path.as_ref());
-        self.write_eol();
+        if self.null {
+            self.write(b"\x00");
+        } else {
+            self.write_eol();
+        }
         if self.wtr.supports_color() {
             let _ = self.wtr.reset();
         }
@@ -324,4 +367,15 @@ impl<W: Terminal + Send> Printer<W> {
         let eol = self.eol;
         self.write(&[eol]);
     }
+
+    fn write_file_sep(&mut self) {
+        if self.quiet {
+            return;
+        }
+        if let Some(ref sep) = self.file_separator {
+            self.has_printed = true;
+            let _ = self.wtr.write_all(sep);
+            let _ = self.wtr.write_all(b"\n");
+        }
+    }
 }

src/types.rs

@@ -18,7 +18,7 @@ const TYPE_EXTENSIONS: &'static [(&'static str, &'static [&'static str])] = &[
     ("awk", &["*.awk"]),
     ("c", &["*.c", "*.h", "*.H"]),
     ("cbor", &["*.cbor"]),
-    ("clojure", &["*.clj", "*.cljs"]),
+    ("clojure", &["*.clj", "*.cljc", "*.cljs", "*.cljx"]),
     ("cmake", &["CMakeLists.txt"]),
     ("coffeescript", &["*.coffee"]),
     ("cpp", &[
@@ -70,6 +70,7 @@ const TYPE_EXTENSIONS: &'static [(&'static str, &'static [&'static str])] = &[
     ("scala", &["*.scala"]),
     ("sh", &["*.bash", "*.csh", "*.ksh", "*.sh", "*.tcsh"]),
     ("sql", &["*.sql"]),
+    ("sv", &["*.v", "*.vg", "*.sv", "*.svh", "*.h"]),
     ("swift", &["*.swift"]),
     ("tex", &["*.tex", "*.cls", "*.sty"]),
     ("ts", &["*.ts", "*.tsx"]),

tests/tests.rs

@@ -54,6 +54,13 @@ fn path(unix: &str) -> String {
     }
 }
 
+fn sort_lines(lines: &str) -> String {
+    let mut lines: Vec<String> =
+        lines.trim().lines().map(|s| s.to_owned()).collect();
+    lines.sort();
+    format!("{}\n", lines.join("\n"))
+}
+
 sherlock!(single_file, |wd: WorkDir, mut cmd| {
     let lines: String = wd.stdout(&mut cmd);
     let expected = "\
@@ -86,8 +93,8 @@ sherlock!(columns, |wd: WorkDir, mut cmd: Command| {
     cmd.arg("--column");
     let lines: String = wd.stdout(&mut cmd);
     let expected = "\
-58:For the Doctor Watsons of this world, as opposed to the Sherlock
-50:be, to a very large extent, the result of luck. Sherlock Holmes
+57:For the Doctor Watsons of this world, as opposed to the Sherlock
+49:be, to a very large extent, the result of luck. Sherlock Holmes
 ";
     assert_eq!(lines, expected);
 });
@@ -716,6 +723,36 @@ clean!(regression_93, r"(\d{1,3}\.){3}\d{1,3}", ".",
     assert_eq!(lines, "foo:192.168.1.1\n");
 });
 
+// See: https://github.com/BurntSushi/ripgrep/issues/99
+clean!(regression_99, "test", ".",
+|wd: WorkDir, mut cmd: Command| {
+    wd.create("foo1", "test");
+    wd.create("foo2", "zzz");
+    wd.create("bar", "test");
+    cmd.arg("-j1").arg("--heading");
+
+    let lines: String = wd.stdout(&mut cmd);
+    assert_eq!(sort_lines(&lines), sort_lines("bar\ntest\n\nfoo1\ntest\n"));
+});
+
+// See: https://github.com/BurntSushi/ripgrep/issues/105
+clean!(regression_105_part1, "test", ".", |wd: WorkDir, mut cmd: Command| {
+    wd.create("foo", "zztest");
+    cmd.arg("--vimgrep");
+
+    let lines: String = wd.stdout(&mut cmd);
+    assert_eq!(lines, "foo:1:3:zztest\n");
+});
+
+// See: https://github.com/BurntSushi/ripgrep/issues/105
+clean!(regression_105_part2, "test", ".", |wd: WorkDir, mut cmd: Command| {
+    wd.create("foo", "zztest");
+    cmd.arg("--column");
+
+    let lines: String = wd.stdout(&mut cmd);
+    assert_eq!(lines, "foo:3:zztest\n");
+});
+
 // See: https://github.com/BurntSushi/ripgrep/issues/20
 sherlock!(feature_20, "Sherlock", ".", |wd: WorkDir, mut cmd: Command| {
     cmd.arg("--no-filename");
@@ -752,6 +789,48 @@ sherlock:be, to a very large extent, the result of luck. Sherlock Holmes
     assert_eq!(lines, expected);
 });
 
+// See: https://github.com/BurntSushi/ripgrep/issues/89
+sherlock!(feature_89_files_with_matches, "Sherlock", ".",
+|wd: WorkDir, mut cmd: Command| {
+    cmd.arg("--null").arg("--files-with-matches");
+
+    let lines: String = wd.stdout(&mut cmd);
+    assert_eq!(lines, "sherlock\x00");
+});
+
+// See: https://github.com/BurntSushi/ripgrep/issues/89
+sherlock!(feature_89_count, "Sherlock", ".",
+|wd: WorkDir, mut cmd: Command| {
+    cmd.arg("--null").arg("--count");
+
+    let lines: String = wd.stdout(&mut cmd);
+    assert_eq!(lines, "sherlock\x002\n");
+});
+
+// See: https://github.com/BurntSushi/ripgrep/issues/89
+sherlock!(feature_89_files, "NADA", ".",
+|wd: WorkDir, mut cmd: Command| {
+    cmd.arg("--null").arg("--files");
+
+    let lines: String = wd.stdout(&mut cmd);
+    assert_eq!(lines, "sherlock\x00");
+});
+
+// See: https://github.com/BurntSushi/ripgrep/issues/89
+sherlock!(feature_89_match, "Sherlock", ".",
+|wd: WorkDir, mut cmd: Command| {
+    cmd.arg("--null").arg("-C1");
+
+    let lines: String = wd.stdout(&mut cmd);
+    let expected = "\
+sherlock\x00For the Doctor Watsons of this world, as opposed to the Sherlock
+sherlock\x00Holmeses, success in the province of detective work must always
+sherlock\x00be, to a very large extent, the result of luck. Sherlock Holmes
+sherlock\x00can extract a clew from a wisp of straw or a flake of cigar ash;
+";
+    assert_eq!(lines, expected);
+});
+
 #[test]
 fn binary_nosearch() {
     let wd = WorkDir::new("binary_nosearch");