[GHC] #1409: Allow recursively dependent modules transparently (without .hs-boot or anything)

[GHC]

#1409: Allow recursively dependent modules transparently (without .hs-boot or
anything)
-------------------------------------+-------------------------------------
        Reporter:  Isaac Dupree      |                Owner:  (none)
            Type:  feature request   |               Status:  new
        Priority:  normal            |            Milestone:  ⊥
       Component:  Compiler          |              Version:  6.10.2
      Resolution:                    |             Keywords:  backpack
Operating System:  Unknown/Multiple  |         Architecture:
                                     |  Unknown/Multiple
 Type of failure:  None/Unknown      |            Test Case:
      Blocked By:                    |             Blocking:
 Related Tickets:  #9256             |  Differential Rev(s):
       Wiki Page:                    |
-------------------------------------+-------------------------------------

Comment (by atravers):

 So there's still some interest after all.

 I had thought about providing some details about my prototype
 implementation (e.g. how it avoids the "all-in-one" approach) but having
 read [a draft of] the Melton-Tempero paper (and various others over the
 break) I now pause to consider a question:

 ''Are recursive imports taking recursion a step too far?''

 To use a automotive analogy - oil works well for certain parts of the
 typical car (e.g. engine and gearbox) but that doesn't mean oil would work
 just as well for other parts of the car (e.g. brakes and clutches).

 With 2020 on the horizon, and drawing some inspiration from a few of those
 other papers (and, well, Douglas Adams), I now suggest the following:

 1. Support for recursive imports be **withdrawn** from the new Haskell
 2020 standard;

 2. A new proposal for recursive imports be added to the Haskell-proposals
 page of the Haskell wiki site, where a concise description can be made
 along with a list of subsequent problems, for the intrepid to find
 solutions to.

 3. If those solutions are viable, then recursive imports would be
 reinstated in a post-2020 Haskell standard.

 The intent is that the proposal-page be ''the'' central point of reference
 for everything related to implementing recursive imports instead of being
 scattered far and wide as it is now - a veritable clearing-house for old
 and new ideas and suggestions.

-- 
Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/1409#comment:80>
GHC <http://www.haskell.org/ghc/>
The Glasgow Haskell Compiler