db_mid2l_insert(): Move assert to mdb.c.
mdb_cursor_set(): Previous assert cannot fail now.
mdb_cursor_put(): Check mc/key and return EINVAL.
mdb_cursor_dbi(): No error return, so just segfault if cursor==NULL.

This should likely be reverted when all callers handle these errors.

mdb_cursor_put(): Use free sub-page space added by a
previous put(existing DUPSORT item), if any.
Tweak NUMKEYS code to avoid a new signedness warning.

It was inaccurate and misplaced. Verify it in mdb.c instead.

I.e. d69d2ce2, it left out some changes.

It and the MDB_msg_func can now return >= 0 for success.
Always return any MDB_msg_func() error result.

Use a sub-DB for DUPSORT item #1/#2 per key if needed: Not a sub-
page too big for a node, nor an overflow page (which not all DUPSORT
code checks for). Move "insert" code, to avoid non-loop goto upwards.

(This is the commit which needs the change to xdata.mv_size in
commit 9d6e4a91 "page sizes".)

Handle keys stored by a liblmdb with a bigger MDB_MAXKEYSIZE.
mdb_get/mdb_del(absent key bigger than our MDB_MAXKEYSIZE)
now return MDB_NOTFOUND instead of MDB_BAD_VALSIZE.

Change me_nodemax to not count the mp_ptrs[] entry. That's mostly
how it was used. Compare node sizes ">" me_nodemax instead of ">=".
The ">=" was a workaround for confusing sizes with and without the
mp_ptrs[] entry, but broke for nodes with size (old me_nodemax-1).

Explicitly make me_nodemax even. An odd value could break the
comparisons. It was even anyway because MDB_MINKEYS == 2.

Do not give uneven sizes to the sub-page and the node it lives in.

Don't set dkey.mv_size if mdb won't clear it before next iteration.

Drop unused computed/assigned values.
mdb_page_alloc(): Make "mop_len>0 implies mop!=NULL" provable.

Support MDB_NOSUBDIR. Catch more errors.  Write messages
to stderr, not stdout which the data too may be piped to.

On near match, return the matched data.