Use the same structure we use in the history.

Each function in the history stores a cleanup function, rather than storing
extra structures.  In particular, we save writes and file offsets using
this.

Append a "+" if you want to start forking as normal.

Useful for showing exactly where we inserted failures, and how long each
one took.

Correctly mark which calls we failed.

Catch fcntl, particularly fcntl locks.

If 'localhost' exists in /etc/hosts with a v4 and a v6 address, the hack
which joins two addrinfo entries together is not necessary.  Detect to see
if 'localhost' returned a v6 address, and if so, just return the single
linked list result from getaddrinfo()

(Reworked by Rusty Russell)
Signed-off-by: Sam Vilain <sam@vilain.net>

move daemon-with-notify to daemon_with_notify as dashes aren't allowed in CCAN module names (and ccanlint segfaults on them)

The type checking function does not use its parameter:

    ccan/opt/opt.h: In function ‘_check_is_entry’:
        ccan/opt/opt.h:328:53: warning: unused parameter ‘e’

Annotate the function using UNUSED from compiler.h.  This commit also
adds compiler.h as a dependency for opt.

OPT_ENDTABLE does not initalise all the elements in stuct opt_table.
When compliling with -Wextra -Wmissing-field-initializers is enabled,
which produces the following warning:

        iviewiir.c:299:9: error: missing field 'cb' initializer
              [-Wmissing-field-initializers]
                OPT_ENDTABLE
                ^
        In file included from iviewiir.c:12:
        ./ccan/opt/opt.h:82:38: note: instantiated from:
        #define OPT_ENDTABLE { NULL, OPT_END }
                                             ^
By changing the definition of OPT_ENDTABLE to initalise all 7 elements
of struct opt_table, the warning is silenced.

It's cool, but making it standalone is slightly simpler and means license is
pure MIT.

Ben Hutchings points out that struct sockaddr isn't large enough to hold
a struct sockaddr_in6.  Our example works because we only access the family
field, but anyone extending it might be surprised; use a union instead.

This can happen if the documentation doesn't have a properly formatted summary line,
and it causes the examples_relevant test to segfault.

Brad Hards fixed a two places where score->error was not \n-terminated,
and then I found a few more, so make it automatic.

This confuses the output, and makes us sometimes see spurious
failure.  At least we will still catch valgrind errors in
children of failtest, thanks to --error-exitcode=101.

Reordering header makes it more readable, and we don't rely on
getopt_long any more.