[GHC] #13693: RTS cannot be reinitialized reliably after hs_exit()

13 May 2017

      #13693: RTS cannot be reinitialized reliably after hs_exit()
-------------------------------------+-------------------------------------
           Reporter:  dfeuer         |             Owner:  (none)
               Type:  bug            |            Status:  new
           Priority:  normal         |         Milestone:  8.4.1
          Component:  Runtime        |           Version:  6.10.1
  System                             |
           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:
-------------------------------------+-------------------------------------
 #2863 added the FFI spec violation to the docs, but as far as I can tell,
 there is and was no ticket to actually fix the problems, which seem to be:

 1. If the runtime is deinitialized by an `hs_exit()` matching the
 ''first'' `hs_init()`, then it cannot be reinitialized reliably using
 `hs_init()`.

 2. This condition is not detected, and leads to unpredictable behavior.

 I'll add a link to ''this'' ticket to the user's guide; even if the ticket
 is ultimately closed as `WONTFIX`, it will be good to have a persistent
 reference.

--
Ticket URL: http://ghc.haskell.org/trac/ghc/ticket/13693
GHC http://www.haskell.org/ghc/
The Glasgow Haskell Compiler