When reusing the main write txn

Only free when cleaning up operation controls, not response controls.
(This only makes a difference on search requests, which may have
multiple responses with the same set of controls.)

Free writer txn on env_close

Make sure ntags isn't incremented if we're skippnig the tag

This commit was accidentally omitted from previous push

Only in back-mdb; back-bdb needs work but it's deprecated;
not worth the effort. In particular txn, retry after deadlocks
makes the whole thing too messy.

Change retry to num times 60.  Testing shows that on a known dataset, this has the same growth behavior as 2.4.39, while num times 20 resulted in significant growth.

Leave it in main env struct, it doesn't need to be rolled back
with nested txns.

Don't walk the readers table unless we really need to.

Can't allow special chars in RDN, not all OSs/filesystems support it.

bypass cleanup handlers for suppressed search result msgs

Only restore op->o_req_dn if no one else changed it on us.

Check for duplicate scopes in search_aliases, not later.

This reverts commit cac00c6a.

Aliases pointing within the original search scope could cause
the same entry to be returned multiple times. Check for subtree
scope and duplicate scopes.