Change .gdb_index de-duplication implementation

While investigating PR symtab/29179, I found that one Ada test failed
because, although a certain symbol was present in the index, with the
new DWARF reader it pointed to a different CU than was chosen by
earlier versions of gdb.

This patch changes how symbol de-duplication is done, deferring the
process until the entire symbol table has been constructed.  This way,
it's possible to always choose the lower-numbered CU among duplicates,
which is how gdb (implicitly) previously worked.

Bug: https://sourceware.org/bugzilla/show_bug.cgi?id=29179