From 76511ddca04d9ec24194327593ec4f627c7eaba2 Mon Sep 17 00:00:00 2001
From: Blaine Bublitz <blaine.bublitz@gmail.com>
Date: Sun, 27 Sep 2015 14:22:30 -0700
Subject: [PATCH] add option to strip BOM, default to true - closes #83

---
 README.md                         |  4 ++++
 lib/src/getContents/bufferFile.js |  8 +++++++-
 lib/src/getContents/streamFile.js |  8 ++++++--
 lib/src/index.js                  |  1 +
 test/src.js                       | 22 +++++++++++++++++++++-
 5 files changed, 39 insertions(+), 4 deletions(-)

diff --git a/README.md b/README.md
index dd1412aa..510589dd 100644
--- a/README.md
+++ b/README.md
@@ -50,6 +50,10 @@ fs.src(['*.js', '!b*.js'])
     - `false` will disable writing the file to disk via `.dest()`.
 
   - since - `Date` or `number` if you only want files that have been modified since the time specified.
+
+  - stripBOM - `true` or `false` if you want the BOM to be stripped on UTF-8 encoded files.
+    - Default value is `true`.
+
   - passthrough - `true` or `false` if you want a duplex stream which passes items through and emits globbed files.
     - Default is `false`.
 
diff --git a/lib/src/getContents/bufferFile.js b/lib/src/getContents/bufferFile.js
index ebd225e0..3c0cfa06 100644
--- a/lib/src/getContents/bufferFile.js
+++ b/lib/src/getContents/bufferFile.js
@@ -8,7 +8,13 @@ function bufferFile(file, opt, cb) {
     if (err) {
       return cb(err);
     }
-    file.contents = stripBom(data);
+
+    if (opt.stripBOM){
+      file.contents = stripBom(data);
+    } else {
+      file.contents = data;
+    }
+
     cb(null, file);
   });
 }
diff --git a/lib/src/getContents/streamFile.js b/lib/src/getContents/streamFile.js
index 23504d8c..4ad5984f 100644
--- a/lib/src/getContents/streamFile.js
+++ b/lib/src/getContents/streamFile.js
@@ -4,8 +4,12 @@ var fs = require('graceful-fs');
 var stripBom = require('strip-bom-stream');
 
 function streamFile(file, opt, cb) {
-  file.contents = fs.createReadStream(file.path)
-    .pipe(stripBom());
+  file.contents = fs.createReadStream(file.path);
+
+  if (opt.stripBOM) {
+    file.contents = file.contents.pipe(stripBom());
+  }
+
   cb(null, file);
 }
 
diff --git a/lib/src/index.js b/lib/src/index.js
index 136b8ced..a1341b91 100644
--- a/lib/src/index.js
+++ b/lib/src/index.js
@@ -21,6 +21,7 @@ function src(glob, opt) {
   var options = assign({
     read: true,
     buffer: true,
+    stripBOM: true,
     sourcemaps: false,
     passthrough: false,
     followSymlinks: true
diff --git a/test/src.js b/test/src.js
index 2dc3c606..f9f7d781 100644
--- a/test/src.js
+++ b/test/src.js
@@ -121,7 +121,7 @@ describe('source stream', function() {
     stream.write(expectedFile);
   });
 
-  it('should strip BOM from UTF-8-encoded files', function(done) {
+  it('should strip BOM from UTF-8-encoded files by default', function(done) {
     var expectedPath = path.join(__dirname, './fixtures/bom-utf8.txt');
     var expectedContent = fs.readFileSync(expectedPath)
       // U+FEFF takes up 3 bytes in UTF-8: http://mothereff.in/utf-8#%EF%BB%BF
@@ -143,6 +143,26 @@ describe('source stream', function() {
     stream.pipe(bufferStream);
   });
 
+  it('should not strip BOM from UTF-8-encoded files if option is false', function(done) {
+    var expectedPath = path.join(__dirname, './fixtures/bom-utf8.txt');
+    var expectedContent = fs.readFileSync(expectedPath);
+
+    var onEnd = function(){
+      buffered.length.should.equal(1);
+      should.exist(buffered[0].stat);
+      buffered[0].path.should.equal(expectedPath);
+      buffered[0].isBuffer().should.equal(true);
+      bufEqual(buffered[0].contents, expectedContent).should.equal(true);
+      done();
+    };
+
+    var stream = vfs.src('./fixtures/bom-utf8.txt', {cwd: __dirname, stripBOM: false});
+
+    var buffered = [];
+    bufferStream = through.obj(dataWrap(buffered.push.bind(buffered)), onEnd);
+    stream.pipe(bufferStream);
+  });
+
   it('should not strip anything that looks like a UTF-8-encoded BOM from UTF-16-BE-encoded files', function(done) {
     // Note: this goes for any non-UTF-8 encoding, but testing for UTF-16-BE
     // and UTF-16-LE is enough to demonstrate this is done properly.