Use bundled ack instead of grep

Our ack is slightly modified to use `'\0'` characters instead of `':'` in the --nogroup listing. This makes it easier to handle both `':'` characters and newlines in filenames, but we aren't really doing that yet.

spec/app/project-spec.coffee

@@ -139,6 +139,5 @@ describe "Project", ->
       it "works on evil filenames", ->
         project.setPath(require.resolve('fixtures/evil-files'))
         waitsForPromise ->
-          project.scan /(a)+/, ({path, match, range}) ->
-            matches.push({path, match, range})
-
+          project.scan /evil/, ({path, match, range}) ->
+            #console.log path

spec/fixtures/evil-files/file with spaces.txt

@@ -0,0 +1 @@
+I am evil because there are spaces in my name

spec/fixtures/evil-files/goddam linefeeds

@@ -1 +0,0 @@
-you know how we do it

spec/fixtures/evil-files/goddam newlines

@@ -0,0 +1 @@
+I am evil because there's a newline in my name

spec/fixtures/evil-files/quote".js

@@ -1 +0,0 @@
-// I am evil because there's a " in my filename. Why you do that!?

spec/fixtures/evil-files/quote".txt

@@ -0,0 +1 @@
+I am evil because there's a " in my filename. Why you do that!?

spec/fixtures/evil-files/utfă.md

@@ -1,3 +1 @@
-# Hello Word
-
-This is markdown.
+I am evil because there's a UTF-8 character in my name

src/app/project.coffee

@@ -138,20 +138,14 @@ class Project
 
   scan: (regex, iterator) ->
     regex = new RegExp(regex.source, 'g')
-    command = [
-      "find \"#{@getPath()}\" -type f -print0" # find all paths in the project's working directory
-      "grep --text --perl-regexp --invert-match --regexp=\"#{@ignorePathRegex()}\"" # accept only non-ignored paths, separated by \0 (find doesn't support pcre)
-      "perl -0pi -e 's/\\n\\z\//'" # delete grep's trailing newline because it screws up xargs
-      "xargs -0 grep --null --perl-regexp --with-filename --line-number --recursive --regexp=\"#{regex.source}\"" # run grep on each filtered file
-    ].join(" | ")
-
-    ChildProcess.exec command, bufferLines: true, stdout: (data) ->
+    command = "#{require.resolve('ack')} --all-types --match \"#{regex.source}\" \"#{@getPath()}\""
+    ChildProcess.exec command , bufferLines: true, stdout: (data) ->
       for grepLine in data.split('\n') when grepLine.length
-        nullCharIndex = grepLine.indexOf('\0')
-        colonIndex = grepLine.indexOf(':')
-        path = grepLine.substring(0, nullCharIndex)
-        row = parseInt(grepLine.substring(nullCharIndex + 1, colonIndex)) - 1
-        line = grepLine.substring(colonIndex + 1)
+        pathEndIndex = grepLine.indexOf('\0')
+        lineNumberEndIndex = grepLine.indexOf('\0', pathEndIndex + 1)
+        path = grepLine.substring(0, pathEndIndex)
+        row = parseInt(grepLine.substring(pathEndIndex + 1, lineNumberEndIndex)) - 1
+        line = grepLine.substring(lineNumberEndIndex + 1)
         while match = regex.exec(line)
           range = new Range([row, match.index], [row, match.index + match[0].length])
           iterator({path, match, range})

src/stdlib/require.coffee

@@ -92,6 +92,7 @@ resolve = (file) ->
   return file
 
 __expand = (path) ->
+  return path if __isFile path
   for ext, handler of exts
     if __exists "#{path}.#{ext}"
       return "#{path}.#{ext}"
@@ -104,6 +105,9 @@ __expand = (path) ->
 __exists = (path) ->
   $native.exists path
 
+__isFile = (path) ->
+  $native.isFile path
+
 __coffeeCache = (filePath) ->
   {CoffeeScript} = require 'coffee-script'
   tmpPath = "/tmp/atom-compiled-scripts"