Scopes: Inconsistent use of "relative" decoding

[jridgewell]

I've only worked on the original scopes so far, but my understanding of the spec appears to be different than @hbenl and @szuend's implementations.

Fields:

1. Line
   • Says relative: Note: this is the point in the original code where the scope starts. line is relative to the line of the preceding "start/end original scope" item.
   • hbenl agrees
   • szuend agrees
2. Column
   • Absolute for now: Scopes: Make Original Source's column relative #75
3. Kind
   • Absolute, everyone agrees
4. Field Flags
   • Absolute, everyone agrees
5. Name
   • Says relative: Note: This offset is relative to the offset of the last scope name or absolute if this is the first name
   • hbenl treats it as absolute
   • szuend treats it as absolute 1, 2
6. Vars
   • Says relative, but meaning is unclear: DATA relative offset into names field for the original symbol name defined in this scope
   • I interpret this to mean Scope 2's vars are relative to Scope 1's (if there were any). Not just that Scope 2's 2nd var is relative to Scope 2's 1st var.
   • hbenl treats them as absolute
   • szuend treats them as absolute 1, 2

[hbenl]

I've used absolute indices into the names array so far because of #70.