> If you have some code that triggers the bug, I'd be interested > so we can put it in the regression suite.
Hi Simon, I've whittled it down to the following program. It triggers the seg fault. The Foo data type is not necessary, but it does make the program more self contained. You ought to be able to replace Foo (..) with values from the Prelude. -- begin Haskell code -- module Main where import StablePtr (newStablePtr, freeStablePtr) data Foo = A | B | C | D main :: IO () main = do aSPtr <- newStablePtr A bSPtr <- newStablePtr B cSPtr <- newStablePtr C cSPtr' <- newStablePtr C freeStablePtr aSPtr freeStablePtr bSPtr freeStablePtr cSPtr freeStablePtr cSPtr' aSPtr <- newStablePtr A bSPtr <- newStablePtr B cSPtr <- newStablePtr C dSPtr <- newStablePtr D print "Hello World" -- end Haskell code -- Cheers, Bernie. _______________________________________________ Glasgow-haskell-bugs mailing list [EMAIL PROTECTED] http://www.haskell.org/mailman/listinfo/glasgow-haskell-bugs