[web-devel] [conduit 0.2.x] generalizing sourceIO?

[Michael Snoyman]

On Wed, Feb 8, 2012 at 9:04 AM, Michael Snoyman <michael at snoyman.com> wrote:
> On Tue, Feb 7, 2012 at 11:18 PM, Erik de Castro Lopo
> <mle+hs at mega-nerd.com> wrote:
>> Erik de Castro Lopo wrote:
>>
>>> Would it not be possible to achieve what you want with sourceState:
>>>
>>>     http://hackage.haskell.org/packages/archive/conduit/0.2.0/doc/html/Data-Conduit.html#v:sourceState
>>
>> No it wouldn't, but maybe that calls for a function sourceStateIO that
>> has all the features of sourceIO and sourceState.
>
> That might make sense. There are definitely cases where it would be
> useful. I don't want to change the type of sourceIO to the suggested
> type, since:
>
> 1. It complicates things for the common case
> 2. Returning a tuple means that when we return IOClosed, we still need
> to return a state, even though it doesn't make sense.
>
> This should be a fairly minor change, I think I'll throw it in soon.
>

How's this?

https://github.com/snoyberg/conduit/commit/7ab2a34d5baa40b7f98cec194305161527ecf370

Michael