The encoding 646 once again

Fri Mar 15 07:34:00 GMT 2002

On Fri, Mar 15, 2002 at 08:48:07AM -0700, Tom Tromey wrote:
> >>>>> "Martin" == Martin Kahlert <martin.kahlert@infineon.com> writes:
>> Martin> Once upon a time i complaint about Sun Solaris' inability to
> Martin> provide a working iconv implementation which caused my
> Martin> application to try to find the 646 encoding about some million
> Martin> times - it was not that fast.
>> Martin> Now i have to give gcj --encoding=UTF-8 in order to work on
> Martin> Solaris otherwise it says hello.java:0: unknown encoding:
> Martin> `646'.
>> Does using --encoding=ASCII work for you? My understanding is that
> 646 is ASCII. If `--encoding=ASCII' doesn't work then we'd need to
> add more code. If it does work then we just have to add alias
> capability. I think that is probably reasonable.

ACII does not work.
By using truss i found out, that gcj searches for
open("/usr/lib/iconv/UCS-2%UCS-2.so", O_RDONLY) Err#2 ENOENT
open("/usr/lib/iconv/UCS-2%UTF2.so", O_RDONLY) Err#2 ENOENT
So the only valid values on this system are these:
KOI8-R, UCS-4, UTF-7 and UTF-8. Nothing else.
Why can't we alias 646 simply to UTF-8?
Thanks,
Martin.
-- 
The early bird catches the worm. If you want something else for 
breakfast, get up later.