On 23 Jul 2012, at 07:54, Leif Halvard Silli wrote: > John W Kennedy, Wed, 18 Jul 2012 14:48:15 -0400: >> On Jul 18, 2012, at 4:21 AM, Leif Halvard Silli wrote: >>> On my OS X 10.7 computer, then TextEdit does sniff UTF-8 (without the >>> BOM). >> >> It does indeed have a sniffing feature, though it also appears to use >> the com.apple.TextEncoding extended attribute, when available (and >> which it, itself, will create, when saving). > > Ah, I see I was wrong about TextEdit on OS X 10.5, then: For UTF-8 > files that it *itself* saves, then it does — despite the lack of a BOM > - open them without defaulting to a legacy encoding. > > (Wonder if the the so called 'resource fork' - the meta data companion > that OS X since version 10.6 abandoned - plays into this?)
It is still there, called "extended attributes": http://my.safaribooksonline.com/book/-/9780132778831/chapter-4dot-data-management/ch04lev1sec5 > But if saved by some app (that probably doesn't add the extended > TextEncoding attribute that you mention), then TextEdit for OS X 10.5 > fails to open it as UTF-8. On 10.7, one can choose in Preferences -> Open and Save -> Opening files: Automatic. Hopefully it opens all files. If set to UTF-8, it probably refutes files will illegal character sequences. Hans