[bug-gnu-libiconv] Please restore "UTF8" as alias for UTF-8 charset

[Stuart Caie]

Hello libiconv team,

please consider restoring "UTF8" as an alias for "UTF-8", which you made 
HPUX-only with the release of libtool 1.13.

It is not an HPUX-only alias. Removing it from vanilla libiconv reduced 
compatibility with other iconv implementations which support the "UTF8" 
name:

- glibc: 
https://sourceware.org/git/?p=glibc.git;a=blob;f=iconv/gconv_builtin.h;h=b96258129a3f192a2b8a30bd3d9c05a9ffd62901;hb=HEAD#l46
- newlib: supports "utf_8" and "utf8" but NOT "utf-8": 
https://sourceware.org/git/gitweb.cgi?p=newlib-cygwin.git;a=blob;f=newlib/libc/iconv/encoding.aliases;hb=HEAD#l45
- uclibc: 
https://cgit.uclibc-ng.org/cgi/cgit/uclibc-ng.git/tree/libiconv/iconv.c#n73
- musl: https://git.musl-libc.org/cgit/musl/tree/src/locale/iconv.c#n39
- dietlibc only supports "UTF-8"
- Bionic: 
https://android.googlesource.com/platform/bionic/+/master/libc/bionic/iconv.cpp#65
- FreeBSD 10+: 
https://web.mit.edu/freebsd/head/share/i18n/esdb/UTF/UTF.alias
- NetBSD: 
http://cvsweb.netbsd.org/bsdweb.cgi/src/share/i18n/esdb/UTF/UTF.alias
- OpenBSD does not have an iconv() function and uses vanilla libiconv
- Darwin patches UTF8 back into libiconv: 
https://opensource.apple.com/source/libiconv/libiconv-51.200.6/libiconv/lib/encodings.def.auto.html
- Cygwin patches UTF8 back into libiconv: 
https://github.com/cygwinports/libiconv/blob/master/1.14-aliases.patch
  ("for better compatibility with glibc" 
https://cygwin.com/ml/cygwin/2015-02/msg00482.html)

I want to write software that can translate data to the UTF-8 charset. 
For that, I need a de-jure reliable name for that charset, as the 
parameters to iconv_open() are implementation-defined.

I picked "UTF8" after reviewing a lot of implementations, including 
pre-1.13 libiconv. I don't think I'm the only person on the planet to 
choose that, there are likely programs out there that chose "UTF8" as 
well, and would break if they were linked to libiconv >= 1.13

What I now need to is loop through possible names, including "UTF-8", 
"UTF8" and "UTF_8", until one of them works. But I'd prefer that every 
iconv_open() implementation support both "UTF-8" and "UTF8" and I could 
rely on either.

Regards
Stuart