BOM in files

[codeseven]

BOM in files

[johnpapa]

I'm not sure where that setting is or why its causing a problem. Can you elaborate?

[kostia]

@johnpapa

Same problem. Unusable in production with Express.js + UglifyJS2, since UglifyJS2 as of the current released version (https://npmjs.org/package/uglify-js) can't handle BOM files (see mishoo/UglifyJS#402).

BOM or Byte order mark is the way many Windows system are saving UTF8 files (https://en.wikipedia.org/wiki/UTF-8#Byte_order_mark). Toaster files however do have BOM-UTF8:

~/develop/kati/bower_components/toastr master grep -rl $'\xEF\xBB\xBF' .
./toastr.css
./toastr.js
./toastr.min.css
./toastr.min.js

This is in general a bad practice to use exotic or OS-specific encoding. I would appreciate if you could re-save the files in plain UTF8. The simplest way to accomplish that:

awk '{if(NR==1)sub(/^\xef\xbb\xbf/,"");print}' toastr.js > tmp.js
mv tmp.js toaster.js

# The diff would look like that:
diff --git a/bower_components/toastr/toastr.js b/bower_components/toastr/toastr.js
index 67dac2e..b491e26 100644
--- a/bower_components/toastr/toastr.js
+++ b/bower_components/toastr/toastr.js
@@ -1,4 +1,4 @@
-<U+FEFF>/*
+/*
  * Toastr
  * Version 2.0.1
  * Copyright 2012 John Papa and Hans Fjällemark.

If you like I could create a PR for this.

[Crempa]

oh sorry I don't replied, I totally forgot to this issue
@kostia is right (thanks for detailed description), I solved it with pre-compillation filter in my assets management system which removes all BOMs from files, but I think this is problem anyway