[Haskell-cafe] Current situation regarding global IORefs

[Robert Dockins]

On Apr 21, 2006, at 10:34 AM, Brian Hulley wrote:
> Robert Dockins wrote:
>> On Apr 21, 2006, at 9:56 AM, Brian Hulley wrote:
>>
>>> Hi -
>>> I've run into the global mutable state problem described in http://
>>> www.haskell.org/hawiki/GlobalMutableState
>>> Since the page was last edited in March last year, I'm wondering if
>>> there have been any developments or further thoughts on how to
>>> safely create top level IORefs since they are absolutely essential
>>> for the library I'm writing.
>>>
>>> For my library, which implements a GUI, I have a Manager module
>>> which keeps track of which control currently has the keyboard focus
>>> etc, and I don't want to have to pass round the state of the
>>> manager to every control since this would be monstrously
>>> inconvenient and a total waste of space/time, so at the moment I'm
>>> reduced to:
>>>
>>>     module Manager where
>>>     keyboard :: IORef (Maybe Control)
>>>     {-# NOINLINE keyboard #-}
>>>     keyboard = unsafePerformIO $ newIORef Nothing
>>>
>>> The problem is that I don't know if this is guaranteed to be
>>> completely safe for all Haskell compilers or even for all future
>>> versions of ghc (?)
>>
>> RE: the technique itself, you should also compile the module with -
>> fno-cse.
>
> Thanks
>
>>
>> RE: the design, Isn't that bit of state local to a dialog/window/
>> control group or something?  I understand that top level state is a
>> problem in general that needs some sort of solution, but I'm not sure
>> it's the right hammer here....
>
> There is only one GUI for the application and only one control in  
> it can have the keyboard focus so it seems natural to use global  
> state here

I'd suggest you consider not making those assumptions... they are the  
kinds of assumptions that can make later code reuse and maintenance  
more difficult than it should be.  (Obviously, if code reuse/ 
maintenance is a low priority then it doesn't matter).

> , but I suppose I could also look into using a state monad. The  
> advantage (perhaps also disadvantage ;-) ) of global state is that  
> it allows me to easily convert all my old C++ singleton classes to  
> Haskell modules...

<ramble type="somewhat coherent">
Ahhh... the singleton pattern.  There is a debate among OO theorists  
about whether the singleton pattern is actually a good idea.  I tend  
to side with those who say that it is Just Wrong.  The reality is  
that "singletons" are only unique within some scope (OS process, VM,  
sandbox, whatever).  "Global" state is similar; it is always bounded  
by _something_.  I think its always better to make the boundaries  
explicit and aligned with the problem domain rather than implicit,  
because the implicit boundaries sometimes/often don't do what you  
want.  As soon as you have an even slightly unusual execution  
environment, your assumptions can be violated (eg, within Java  
application containers *shudder*).  I have to imagine using, eg, HS  
plugins with modules containing top-level state could cause all sorts  
of havoc.
</ramble>

>> As far as I know, the only recent developments in this area are a
>> rumor from the Simons that they are working on some sort of thread-
>> local state which (under some sets of design decisions) can fill the
>> needs of top level state.  If you press them, they might be willing
>> to give some details about this.
>
> I was kind of hoping that there would just be a safe, simple way to  
> create a top level monomorphic IORef without having to use a pragma  
> etc.

I don't think that exists currently.

> Thanks, Brian.


Rob Dockins

Speak softly and drive a Sherman tank.
Laugh hard; it's a long way to the bank.
           -- TMBG