instead of creating temp dso objects on the stack and moving them to
the heap if dlopen/dlsym are used, use static objects to begin with,
and just donate them to malloc if we no longer need them.

these changes also make it so clock_gettime(CLOCK_REALTIME, &ts) works
even on pre-2.6 kernels, emulated via the gettimeofday syscall. there
is no cost for the fallback check, as it falls under the error case
that already must be checked for storing the error code in errno, but
which would normally be hidden inside __syscall_ret.

we cannot report failure after forking, so the idea is to ensure prior
to fork that fd 0,1,2 exist. this will prevent dup2 from possibly
hitting a resource limit and failing in the child process. fcntl
rather than dup2 is used prior to forking to avoid race conditions.

-1, not 0, indicates failure

some program was undefining AF_NETLINK and thereby breaking AF_ROUTE...

this resolves an issue reported by Vasiliy Kulikov

fread was calling f->read without checking that the file was in
reading mode. this could:
1. crash, if f->read was a null pointer
2. cause unwanted blocking on a terminal already at eof
3. allow reading on a write-only file

1. my interpretation of subject sequence definition was wrong. adjust
parser to conform to the standard.

2. some code for handling tail overflow case was missing (forgot to
finish writing it).

3. typo (= instead of ==) caused ERANGE to wrongly behave like EINVAL

stopping without letting the parser see a stop character prevented
getting a result. so treat all high chars as the null character and
pass them into the parser.

also eliminated ugly tmp var using compound literals.

this fixes a number of bugs in integer parsing due to lazy haphazard
wrapping, as well as some misinterpretations of the standard. the new
parser is able to work character-at-a-time or on whole strings, making
it easy to support the wide functions without unbounded space for
conversion. it will also be possible to update scanf to use the new
parser.

also support (and restrict to subsets) older chinese sets, and
explicitly refuse to convert to cjk (since there's no code for it yet)

these are useless legacy functions but some old software contains
cruft that expects them to exist...

this broke most uses of iconv in real-world programs, especially
glib's iconv wrappers.

STREAMS are utterly useless as far as I can tell, but some software
was apparently broken by the presence of stropts.h but lack of macros
it's supposed to define...

this should not be necessary - the invalid bit patterns cannot be
created except through type punning. however, some broken gnu software
is passing them to printf and triggering dangerous stack-smashing, so
let's catch them anyway...

hopefully this resolves the rest of the issues with hideously
nonportable hacks in programs that use gnulib.

this is a really ugly and backwards function, but its presence will
prevent lots of broken gnulib software from trying to define its own
version of fpurge and thereby failing to build or worse.

these interfaces are required to be thread-safe even though they are
not state-free. the random number sequence is shared across all
threads.

per POSIX: The mprotect() function shall change the access protections
to be that specified by prot for those whole pages containing any part
of the address space of the process starting at address addr and
continuing for len bytes.

on the other hand, linux mprotect fails with EINVAL if the base
address and/or length is not page-aligned, so we have to align them
before making the syscall.