[GHC] #13423: Exception to I/O hack in demand analyzer too broad
#13423: Exception to I/O hack in demand analyzer too broad
-------------------------------------+-------------------------------------
Reporter: dfeuer | Owner: (none)
Type: bug | Status: new
Priority: normal | Milestone: 8.4.1
Component: Compiler | Version: 8.1
Keywords: | Operating System: Unknown/Multiple
Architecture: | Type of failure: Other
Unknown/Multiple |
Test Case: | Blocked By:
Blocking: | Related Tickets:
Differential Rev(s): | Wiki Page:
-------------------------------------+-------------------------------------
The I/O hack in the demand analyzer triggers only for expressions that are
not direct applications of primops (see `io_hack_reqd` in
`stranal/DmdAnal.hs`). This seems rather arbitrary, and can produce
surprising results. Consider
{{{#!hs
fish :: Int -> IORef Int -> IORef Int -> IO ()
fish n ref next = do
writeIORef ref n
writeIORef next n
x <- getLine
case readMaybe x of
Just True -> fish (n + 1) ref next
_ -> pure ()
}}}
If `next` is bottom, I would expect this to write `n` to `ref` and then
throw an exception. But we actually get something else. First, the demand
signature for `fish` ends up being
`Str=
#13423: Exception to I/O hack in demand analyzer too broad -------------------------------------+------------------------------------- Reporter: dfeuer | Owner: (none) Type: bug | Status: new Priority: normal | Milestone: 8.4.1 Component: Compiler | Version: 8.1 Resolution: | Keywords: Operating System: Unknown/Multiple | Architecture: | Unknown/Multiple Type of failure: Other | Test Case: Blocked By: | Blocking: Related Tickets: | Differential Rev(s): Wiki Page: | -------------------------------------+------------------------------------- Comment (by rwbarton): Is this really wrong in this case though? In a single-threaded context there is no observable difference between writing an `IORef` and then diverging, and diverging immediately; and in a multi-threaded context the ordering of `writeIORef` with respect to anything else is not guaranteed. -- Ticket URL: http://ghc.haskell.org/trac/ghc/ticket/13423#comment:1 GHC http://www.haskell.org/ghc/ The Glasgow Haskell Compiler
Is this really wrong in this case though? In a single-threaded context
#13423: Exception to I/O hack in demand analyzer too broad -------------------------------------+------------------------------------- Reporter: dfeuer | Owner: (none) Type: bug | Status: new Priority: normal | Milestone: 8.4.1 Component: Compiler | Version: 8.1 Resolution: | Keywords: Operating System: Unknown/Multiple | Architecture: | Unknown/Multiple Type of failure: Other | Test Case: Blocked By: | Blocking: Related Tickets: | Differential Rev(s): Wiki Page: | -------------------------------------+------------------------------------- Comment (by dfeuer): Replying to [comment:1 rwbarton]: there is no observable difference between writing an `IORef` and then diverging, and diverging immediately; and in a multi-threaded context the ordering of `writeIORef` with respect to anything else is not guaranteed. That's not true. Someone could write {{{#!hs fish n ref next `catch` \(e :: E) -> readIORef ref >>= ... }}} -- Ticket URL: http://ghc.haskell.org/trac/ghc/ticket/13423#comment:2 GHC http://www.haskell.org/ghc/ The Glasgow Haskell Compiler
#13423: Exception to I/O hack in demand analyzer too broad -------------------------------------+------------------------------------- Reporter: dfeuer | Owner: (none) Type: bug | Status: new Priority: normal | Milestone: 8.4.1 Component: Compiler | Version: 8.1 Resolution: | Keywords: Operating System: Unknown/Multiple | Architecture: | Unknown/Multiple Type of failure: Other | Test Case: Blocked By: | Blocking: Related Tickets: | Differential Rev(s): Wiki Page: | -------------------------------------+------------------------------------- Comment (by dfeuer): Oh, if you mean diverging in the stronger sense of falling into a hole without throwing an exception, then that's true, but it's not the situation I was talking about. -- Ticket URL: http://ghc.haskell.org/trac/ghc/ticket/13423#comment:3 GHC http://www.haskell.org/ghc/ The Glasgow Haskell Compiler
#13423: Exception to I/O hack in demand analyzer too broad -------------------------------------+------------------------------------- Reporter: dfeuer | Owner: (none) Type: bug | Status: new Priority: normal | Milestone: 8.4.1 Component: Compiler | Version: 8.1 Resolution: | Keywords: Operating System: Unknown/Multiple | Architecture: | Unknown/Multiple Type of failure: Other | Test Case: Blocked By: | Blocking: Related Tickets: | Differential Rev(s): Wiki Page: | -------------------------------------+------------------------------------- Comment (by dfeuer): In any case, you can just swap out `writeIORef` for any other genuinely side-effecting primop and get the same sort of result, so that's not really too relevant. -- Ticket URL: http://ghc.haskell.org/trac/ghc/ticket/13423#comment:4 GHC http://www.haskell.org/ghc/ The Glasgow Haskell Compiler
participants (1)
-
GHC