[RFC] Improve Dwarf 5 .debug_names Type Lookup/Parsing Speed

I think I am ready to open all the PRs for this work and get feedback.

With the previously discussed idea of using a DW_FORM_flag_present, the debug_names size increase is reduced to 19% (down from 30%). Total object size is now increased by 0.62% (down from 0.92%).

The complete PR is here

* 6f62cfd1b509 [lldb][DWARFIndex] Use IDX_parent to implement GetFullyQualifiedType query
* 9ca21052af4b [lldb][DWARFUnit] Implement PeekDIEName query
* 58b6d9b4038c [llvm][DebugNames] Implement Entry::GetParentEntry query
* b99113215416 [AsmPrinter][DebugNames] Implement DW_IDX_parent entries
* fa08a9dbc97d [lldb][[DWARFDeclContext] Add function to extract qualified names as vector
* 683cfd5217e5 [lldb][DWARFIndex][nfc] Factor out fully qualified name query
* 8ec587dc1414 [AccelTable][nfc] Add helper function to cast Data

Sadly I am pretty lost in Github-land when it comes to stacked patches, so the idea is to open PRs as dependencies are merged.
The bottom two commits already have PRs.

The most important commits are:

* b99113215416 [AsmPrinter][DebugNames] Implement DW_IDX_parent entries
* 6f62cfd1b509 [lldb][DWARFIndex] Use IDX_parent to implement GetFullyQualifiedType query

The description of that first commit is fairly detailed.