280789 - add a note about 'UTF-8 BOM' (especially for Windows users)

Status: RESOLVED FIXED

(Developer Documentation Graveyard :: General, defect)

Description

[Jungshik Shin]

This is a spin-off of bug 280656. 
Some Windows editors (including but not limited to Notepad and Wordpad) add
'UTF-8 BOM' (three bytes representing U+FEFF : EF BB BF) whenever a file is
saved in UTF-8. This can interfere with lxr and other mozilla.org tools. It's
also possible that some tools (e.g. compilers) breat on coming across UTF-8 BOMs. 

Therefore, we ned to make it clear that UTF-8 BOM needs to be removed before
landing files. Somewhere under /hacking (e.g. 
http://www.mozilla.org/hacking/mozilla-style-guide.html ) would be a good place
for that. 

In addition, we may refer to a simple perl script for the removal of UTF-8
BOM(s). Somehow the following (perl -pi nukebom.pl file ) doesn't work on
Windows (http://www.nntp.perl.org/group/perl.unicode/2307 ):


-----------------------------
# nukebom.pl 
$INC{ "bytes.pm" }++ if $] < 5.006;
use bytes;

s/^\xEF\xBB\xBF//s;
--------------------

Comment 1

[OstGote!]

BOM "Byte Order Mark"

Comment 2

[Kevin Boyd]

Any suggestions on how to accomplish this for my users, who have no shell access (nor any shell skills), and only have access to OpenOffice for editing their signatures?

(When Thunderbird imports the sig.txt file, it's dropping the UTF-8 BOM right in the signature, which is being VERY annoying because it forces a dialog box every single time they start/send an email.)

Comment 3

[Jungshik Shin]

Kevin, this bug has nothing to do with end users of thunderbird/firefox. It's about files landed in mozilla.org's cvs server which happen to have UTF-8 BOM at the beginning. 

I'm not sure what your problem is, but why don't you file a new bug and add me to Cc?