[GHC] #7229: Detecting if a process was killed by a signal is impossible

[GHC]

#7229: Detecting if a process was killed by a signal is impossible
--------------------------------------+------------------------------------
        Reporter:  benmachine         |            Owner:
            Type:  bug                |           Status:  new
        Priority:  high               |        Milestone:  7.8.1
       Component:  libraries/process  |          Version:
      Resolution:                     |         Keywords:
Operating System:  Unknown/Multiple   |     Architecture:  Unknown/Multiple
 Type of failure:  None/Unknown       |       Difficulty:  Unknown
       Test Case:                     |       Blocked By:
        Blocking:                     |  Related Tickets:
--------------------------------------+------------------------------------

Comment (by hvr):

 So, the implementation of the last proposal ammmounts to two simple
 modifications:

 a patch to `process`:

 {{{#!diff
 diff --git a/cbits/runProcess.c b/cbits/runProcess.c
 index 3462cfc..3a6f2f3 100644
 --- a/cbits/runProcess.c
 +++ b/cbits/runProcess.c
 @@ -22,12 +22,12 @@
     UNIX versions
 -------------------------------------------------------------------------
 */

 -//
 -// If a process terminates with a signal, the exit status we return to
 -// via the System.Process API follows the Unix shell convention of
 -// (128 + signal).
 -//
 -#define TERMSIG_STATUS(r) ((r) | 0x80)
 +// If a process was terminated by a signal, the exit status we return
 +// via the System.Process API is (signum << 8); this avoids collision
 +// with normal process termination status codes, as according to
 +// http://pubs.opengroup.org/onlinepubs/9699919799/functions/wait.html
 +// WEXITSTATUS(s) returns an 8-bit value. See also #7229.
 +#define TERMSIG_STATUS(r) (r << 8)

  static long max_fd = 0;
 }}}

 ...and a patch for `base` (provided by duncan):

 {{{#!diff
 diff --git a/GHC/TopHandler.lhs b/GHC/TopHandler.lhs
 index 9e4bc07..b0c5d0d 100644
 --- a/GHC/TopHandler.lhs
 +++ b/GHC/TopHandler.lhs
 @@ -180,7 +180,25 @@ flushStdHandles = do
  -- we have to use unsafeCoerce# to get the 'IO a' result type, since the
  -- compiler doesn't let us declare that as the result type of a foreign
 export.
  safeExit :: Int -> IO a
 +#ifdef mingw32_HOST_OS
  safeExit r = unsafeCoerce# (shutdownHaskellAndExit $ fromIntegral r)
 +#else
 +-- On Unix we use an encoding for the ExitCode:
 +--   high 7 bits signal, low 8 bits normal exit code.
 +-- People can also use ExitCode that do not correspond to Unix exit()
 +-- codes and we just use a replacement.
 +safeExit r
 +  | sig /= 0 && code == 0 && other == 0
 +  = unsafeCoerce# (shutdownHaskellAndSignal $ fromIntegral sig)
 +  | sig == 0              && other == 0
 +  = unsafeCoerce# (shutdownHaskellAndExit $ fromIntegral code)
 +  | otherwise
 +  = unsafeCoerce# (shutdownHaskellAndExit 0xff)
 +  where
 +    sig   = (r .&. 0x7f00) `shiftR` 8
 +    code  =  r .&. 0x00ff
 +    other =  r .&. (complement 0x7fff)
 +#endif
 }}}


 If there are no objections left, I'll apply the changes above, `validate`
 (tweaking the testsuite if necessary) and push the changes.

--
Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/7229#comment:29>
GHC <http://www.haskell.org/ghc/>
The Glasgow Haskell Compiler