sysdep/hasutmp{,x}addrv6.h-yes.c update test

detects if 'struct utmp{,x}' has '.ut_addr_v6' in compilation phase

More utmp stuff! Fill utmp(x)'s ut_addr_v6 field when it exists

- Add sysdeps to test ut_addr_v6 in utmp and utmpx
- We already have the remote IP as a string, we just add a
decoding step in logsys_utmp(x) to fill the addr field.
- We assume arpa/inet.h and AF_INET6 exist on the system.
If the build fails on (ancient) systems, sysdeps will need to
be added to test for them.

Signed-off-by: Laurent Bercot <ska@appnovation.com>

Small build system and utmp fixes

closes #65

Traditional utmpx mechanisms, at least on Linux, like to
have the ut_host and ut_user fields voided out to indicate
a logout.

 Also, some systems do not define _PATH_WTMPX, but define
WTMPX_FILE instead. Add support for WTMPX_FILE.

 This patch makes it possible to build tinyssh against utmps,
providing secure utmp functionality with the musl libc.

Signed-off-by: Laurent Bercot <ska@appnovation.com>

 Linker command line order is important: -la -lb means that
liba can depend on libb, but not the other way around. The
way make-tinyssh.sh constructs the libs variable from the
LIBS environment variable should not reverse the order.
Also, local libs like libtinysshcrypto.a should be listed
first, so they correctly resolve local symbols (this is
important when linking against skalibs, which exports symbols
like uint32_unpack_big() with a different signature).

Signed-off-by: Laurent Bercot <ska@appnovation.com>

    Merge pull request #64 from petrkutalek/feature/update_systemd_unit
    Instruct Systemd to accept status code 111 as a successful exit

fixes build on big-endian platforms

use PACKET_LIMIT/2 as maximum packet size
workaround for miscalculated packet_length

tinysshnoneauthd

add -e param., allow to execute custom program instead of shell

but don't close the connection

fixes #39

README.md update

honor "max packet size" requested from the client