[GHC] #10334: __ctzdi2/si2/__clzdi2/si2 unknown symbols in ghc-prim on non-shared libs platform

[GHC]

#10334: __ctzdi2/si2/__clzdi2/si2 unknown symbols in ghc-prim on non-shared libs
platform
-------------------------------------+-------------------------------------
        Reporter:  kgardas           |                   Owner:
            Type:  bug               |                  Status:  new
        Priority:  normal            |               Milestone:
       Component:  Core Libraries    |                 Version:  7.11
      Resolution:                    |                Keywords:
Operating System:  Unknown/Multiple  |            Architecture:
 Type of failure:  None/Unknown      |  Unknown/Multiple
      Blocked By:                    |               Test Case:
 Related Tickets:                    |                Blocking:
                                     |  Differential Revisions:
-------------------------------------+-------------------------------------
Changes (by ekmett):

 * owner:  ekmett =>


Comment:

 I'd encourage a fix, but I'm going to relinquish control of the ticket as
 it is more about how the linker/RTS works than the structure of the core
 libraries per se.

 If we need a good `ctz` and `clz`, I have implementations in various
 libraries we could adapt back into C/C-- and contribute from various
 projects.

 e.g.
 https://github.com/ekmett/bits/blob/ded1403737ce26e819c1d1c1facb41a9e776eb46/cbits/debruijn.c#L2
 gives usable De Bruijn multiplication tables and

 https://github.com/ekmett/bits/blob/ded1403737ce26e819c1d1c1facb41a9e776eb46/src/Data/Bits/Extras.hs#L56

 provides a working portable `ctz`. Right below it is a decent `clz` that
 uses our existing `popCount` primitive. You may need to adjust the answers
 by 1 to comply with the `__builtin_clz` and `__builtin_ctz` primitives, I
 forget.

--
Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/10334#comment:1>
GHC <http://www.haskell.org/ghc/>
The Glasgow Haskell Compiler