UTF-8 in source files?

["Howland Craig D (Craig)" <howland at LGSInnovations dot com>]

Jeff:
     What character sets are OK to be in source files?  (How does CVS
handle anything non-ASCII?  Can UTF-8 be used, only 7-bit ASCII, does
CVS police it for you if there are limitations?)
     I'd assume that plain ASCII only is the answer.  If that is
true, whether CVS helps to notify about accidental submissions is the
next one that is needed.
     I was editing on RHEL5, which supports UTF-8, and pasted in some
text that has some things like a +- sign, an infinity sign, etc.
I didn't think about it until later when I looked at the file in
Cygwin, when it became apparent since it doesn't have the multi-byte
stuff, yet.  But even it properly shows the +- symbol.  (XP's WordPad
shows it almost the same as vim under Cygwin.)  I noticed due to the
infinity and a couple other things being obviously messed up.  This
made me think of the bigger question.  I would not normally plan to
put in anything other than standard ASCII, but the question arises
because it can fairly easily happen.
     I have attached a sample that has the UTF-8, both a source file
and a patch file made from it.
				Craig

Attachment: sample.c
Description: sample.c

Attachment: sample.patch
Description: sample.patch