Re: catting random bytes causes character craziness
"Stephen J. Turnbull" <stephen(a)xemacs.org> writes:
>>>>> "Jamie" == Jamie Zawinski
<jwz(a)jwz.org> writes:
Jamie> This is clearly a bug: no matter what binary characters got
Jamie> dumped into my shell buffer, they should damned well not be
Jamie> being displayed as Hebrew, even if I *did* have those fonts
Jamie> installed.
Welcome to the wonderful world of Mule.
There's no such thing as a binary character. All octets are
binary;
the character interpretation is up to the user.
Except, last time I checked, XEmacs/Mule completely ignores the user's
hints as to how he'd prefer to handle the "binary octets". I bet
Jamie isn't running in a Hebrew locale. Or in an "ISO 2022 locale"
(if there's such a thing), or anything like it.
There would be much less complaints about Mule if it accepted the
user's locale settings out of the box. Despite the general brokenness
of the entire locale model, it is a better default to accept it than
to completely ignore it.
Eg, suppose you were catting a text file in Tel Aviv.
I prefer to imagine I lived in a world where ISO 2022 was removed from
XEmacs years ago. :-)