In `h_sequence__ma` the same bug occurs, and it manifests as a crash
in the same way, so I've added a test for it.  In `h_permutation__ma`
it evidently exists in the same form, but I haven't figured out how to
reproduce it; in that case I added a fix to the implementation, but no
test.

I committed the fix for bug #19 without a test because I didn't know
how our test worked yet; here's a test.

A somewhat more desirable way to do this would be to commit the
test *first*, marked as "incomplete" with `g_test_incomplete()` (an
expected failure).  However, `g_test_incomplete()` does not handle
segfaults!  There's no way to mark a segfaulting test as an "expected
segfault".  So if you want to verify that this test reveals the bug,
you'll need to `git show thiscommit | patch -p1` or something, in a
tree that doesn't have the fix applied.  Or you can comment out the
fix, I guess.

NEWS:
* Switching endianness mid-byte no longer potentially re-reads bytes.
* bit_offset now consistently refers to the number of bits already
  read.
* HParsedTokens now have a bit_length field; this is a size_t.  This
  may be removed for memory reasons.

The bit writer has not yet been updated to match; the result of
switching bit writer endianness in the middle of a byte remains
undefined.