[GHC] #9628: Add Annotations to the AST to simplify source to source conversions

[GHC]

#9628: Add Annotations to the AST to simplify source to source conversions
-------------------------------------+-------------------------------------
              Reporter:  alanz       |            Owner:  alanz
                  Type:  feature     |           Status:  new
  request                            |        Milestone:
              Priority:  normal      |          Version:  7.9
             Component:  Compiler    |         Keywords:
            Resolution:              |     Architecture:  Unknown/Multiple
      Operating System:              |       Difficulty:  Unknown
  Unknown/Multiple                   |       Blocked By:
       Type of failure:              |  Related Tickets:
  None/Unknown                       |
             Test Case:              |
              Blocking:              |
Differential Revisions:  D297        |
-------------------------------------+-------------------------------------

Comment (by simonpj):

 Thank you. Now I understand that all this is focused solely on the source-
 location of keywords.  (Previously I thought that there might be other
 client-specific information involved.)

 Could we then just take a moment to review the pros and cons of the
 proposed approach (the `ApiAnns` approach) compared with the earlier idea
 of just adding `SrcSpan` info to the `HsSyn` syntax tree itself.  That is,
 like "Richard Eisenberg's response" and "SPJ response to concern re extra
 noise in AST" in the wiki page.  I'll call that the "in-tree approach".
 Note that using record update syntax, a pass like the renamer would never
 need to explicitly match or mention the keyword location fields.

 Advantages of the `ApiAnns` approach
  * Very non-invasive
  * Programs that don't need the information don't touch it; this seems
 like the strongest advantage

 Advantages of the in-tree approach
  * The things you want are right there; no extra mappings to carry around,
 no lookups, no `AnnColon2` stuff.
  * Seems simple and uniform with the existing source locations.  Eg. if a
 keyword 'do' could so be `mdo`, you might have a `Located Bool` to
 indicate.  But if it's just `do` then we'd get `Located ()`, and then we
 drop it altogether.  The in-tree story amounts to not dropping it.

 This is really a matter for our customers.  Alan (Mr HaRE) is one.  Do we
 have other known customers for this feature?  What do they say?  Or don't
 they case?

--
Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/9628#comment:32>
GHC <http://www.haskell.org/ghc/>
The Glasgow Haskell Compiler