[GHC] #12855: ghc-8.0.2_rc1 inlines string literals too aggressively, breaks tests of http-client-0.4.30

[GHC]

#12855: ghc-8.0.2_rc1 inlines string literals too aggressively, breaks tests of
http-client-0.4.30
-------------------------------------+-------------------------------------
           Reporter:  slyfox         |             Owner:
               Type:  bug            |            Status:  new
           Priority:  normal         |         Milestone:  8.0.2
          Component:  Compiler       |           Version:  8.0.1
           Keywords:                 |  Operating System:  Unknown/Multiple
       Architecture:                 |   Type of failure:  None/Unknown
  Unknown/Multiple                   |
          Test Case:                 |        Blocked By:
           Blocking:                 |   Related Tickets:
Differential Rev(s):                 |         Wiki Page:
-------------------------------------+-------------------------------------
 GHC version 8.0.1.20161117

 {{{#!hs
 -- $ cat BodySpec.hs
 -- $ ghc --make -O1 -package={base,bytestring} BodySpec.hs -o a && ./a
 {- # OPTIONS_GHC -Wall -Werror #-}
 module Main (main) where

 import qualified Data.ByteString as S
 import qualified Data.ByteString.Char8 as S8

 main :: IO ()
 main = (S8.concat (map S.singleton (S.unpack (S8.pack "<foo>"))) ==
 S8.empty) `seq` return ()
 }}}

 {{{
 $ ghc --make -O1 -package={base,bytestring} BodySpec.hs -o a && ./a
 Segmentation fault (core dumped)
 }}}

 The cause is similar to #12757: low-level optimisations inline string
 literal (or it's wrapping value? i don't know) in multiple places. That
 breaks pinter-chasing in bytestring package.

 {{{
 $ strings a | fgrep '<foo>'
 <foo>
 <foo>
 <foo>
 <foo>
 }}}

--
Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/12855>
GHC <http://www.haskell.org/ghc/>
The Glasgow Haskell Compiler