Typos in GCC output [was: Warning: Missing charsets ...]

[stephen at xemacs.org]

Nix writes:

 > It is conditionalized off the locale. GCC makes the (reasonable)
 > assumption

You can call it "reasonable assumption" if you like, I call it
"tyranny of the majority."

With all due respect to Joseph Myers, this was an ill-advised change.
I mean, if this is such a great idea, in UTF-8 locales shouldn't you
change the parser to allow the C notation for strings to use balanced
quotes, and in fr_FR locales, guillemots?  That's not just
typographically correct, it would make things a lot easier for Emacs
font-lock, you know!

 > that if you're in a UTF-8 locale, your tools, display devices, and
 > so on can handle UTF-8.

It's NOT an issue of "being able to handle" UTF-8---XEmacs handles
UTF-8 just fine, and if it knows it's coming it will display it
prettily, too.  The problem is that GCC is making the assumption that
there's nothing in the pipeline able to handle *more* than just UTF-8,
that might be expecting something else for any of a number of reasons.

It may very well be that imposing this pain on XEmacs and other
multilingual apps is the right thing to do; backward compatibility is
the only reason not to do it, and backward compatibility cannot be a
justification for permanent stasis.  But defending it with "reasonable
assumption" ... I expect better of y'all.