[GHC] #14254: The Binary instance for TypeRep smells a bit expensive

GHC ghc-devs at haskell.org
Mon Oct 9 04:09:36 UTC 2017


#14254: The Binary instance for TypeRep smells a bit expensive
-------------------------------------+-------------------------------------
        Reporter:  dfeuer            |                Owner:  (none)
            Type:  bug               |               Status:  patch
        Priority:  normal            |            Milestone:  8.4.1
       Component:  Compiler          |              Version:  8.2.1
      Resolution:                    |             Keywords:  Typeable
Operating System:  Unknown/Multiple  |         Architecture:
 Type of failure:  Compile-time      |  Unknown/Multiple
  performance bug                    |            Test Case:
      Blocked By:                    |             Blocking:
 Related Tickets:                    |  Differential Rev(s):  Phab:D3998,
       Wiki Page:                    |  Phab:D4078, Phab:D4082
-------------------------------------+-------------------------------------
Changes (by dfeuer):

 * differential:  Phab:D3998 => Phab:D3998, Phab:D4078, Phab:D4082


Comment:

 I believe the simplest thing right now that will actually make
 deserialization linear is Phab:D4082. If at some point GHC allows unlifted
 types in kinds (which Richard thinks is probably not hard), then it will
 get even simpler and even cheaper.

 Phab:D4078 implements Richard's idea of starting off by digging down to
 the constructor, and as a side effect reduces the number of tags needed to
 express deeply nested applications. But I very much doubt the
 complexity/benefit ratio supports that approach, unless we change the
 structure of `TypeRep` to match.

-- 
Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/14254#comment:9>
GHC <http://www.haskell.org/ghc/>
The Glasgow Haskell Compiler


More information about the ghc-tickets mailing list