This is the mail archive of the
gcc-patches@gcc.gnu.org
mailing list for the GCC project.
Re: Merge cpplib and front end hashtables, part 1
On Tue, May 22, 2001 at 01:43:58AM +0100, Jason Merrill wrote:
> >>>>> "Zack" == Zack Weinberg <zackw@Stanford.EDU> writes:
>
> > Character sets that are not proper supersets of ASCII are trickier.
> > For UCS2-encoded Unicode, we can look for byte order marks.
>
> I don't think we need to care about character sets that are not proper
> supersets of ASCII (in which I include ASCII with escapes, like SJIS).
> Text files should be stored in a texty encoding.
I don't have a problem with requiring the use of an external converter
for sufficiently exotic file formats, but I think that we should make
an effort to accept all formats which are actually used in the real
world for source code. It's kind of like always knowing what to do
with DOS line terminators.
In particular, I think we should be able to cope with UCS2, UCS4, and
EBCDIC, as long as the system provides an iconv library that can also
cope. We should provide a fallback iconv in libiberty, but it would
be okay if that only knew from UTF8 and ISO8859.n (maybe only
ISO8859.1 even).
--
zw One can't, of course, perform Kaspar Hauser experiments on the
offspring of one's conspecifics.
-- Jerry Fodor, _The Mind Doesn't Work That Way_