The WT_DHANDLE_BTREE() macro is used to determine whether a dhandle is backed by a btree. Currently it returns true for the dhandle of the local file of a tiered table. This seems incorrect. The local file by itself is not a btree; a tiered table's btree is spread across the local file and some number of other objects.

For this ticket we should:

1. Determine whether there is ever a need to treat the dhandle of a local file of a tiered table as a btree handle.
2. If not, change the WT_DHANDLE_BTREE() macro to return false on the local file of a tiered table. This might be done by adding a test for !WT_SUFFIX_MATCH(dhandle->name, ".wtobj") to the macro.

A simple way to address #1 would be to make the change and run a lot of tiered storage tests and see if anything breaks.

This was prompted by WT-10823, where we needed to include the check for a wtobj suffix to stop __wt_conn_btree_apply_all() from trying (and failing) to operate on tiered local files.