Exceptions vs. Errors (Re: Readline read_history and write_history addition)

[Yitzchak Gale]

I wrote:
>> unfortunately, bracket is currently not available for MonadIO,
>> nor is there any way to emulate it AFIK.
>> (This is a "maybe" for HaskellPrime:
>> http://hackage.haskell.org/trac/haskell-prime/ticket/110)

Judah Jacobson wrote:
> Following is what I've been using to solve that problem.  I can add it
> to that HaskellPrime ticket if people think it's useful.

Very nice! I think you should not only add it to the ticket
(if that is still relevant, I am not sure what is going on with
Haskell' right now), but you should also submit a patch for the
mtl library. We need a better name than IO1, though.

> Note that in bracket1, the "after" action must run in IO.  In
> practice, that hasn't been a problem for me.  In fact, since the
> "after" clause might run in response to an asynchronous exception, I
> don't see how it could be sequenced with an arbitrary monad, anyway.

Agreed.

The real fix would be to provide a lower-level primitive
for block and unblock. The concurrency paper "Asynchronous
exceptions in Haskell" states (sec. 4.2) that the reason for the type

block :: IO a -> IO a

rather than the more obvious

block :: IO ()

is clumsiness and fragility. That may be so. But that type
is too high-level for a primitive. In GHC, the implementation
of block is:

block (IO io) = IO $ blockAsyncExceptions# io

It should really be:

block x = do
  IO blockAsyncExceptions#
  ret <- x
  IO unblockAsyncExceptions#
  return ret

in which case we could then supply implementations for
other monads as well.

Regards,
Yitz