On Tue, 2011-10-04 at 16:48 -0700, Jason Dagit wrote:

On Tue, Oct 4, 2011 at 4:26 PM, Duncan Coutts wrote:

...

On Tue, 2011-10-04 at 12:56 -0700, Jason Dagit wrote:

...

On Tue, Oct 4, 2011 at 11:36 AM, Duncan Coutts wrote:

...

On Tue, 2011-10-04 at 10:48 -0700, Jason Dagit wrote:

...

On Tue, Oct 4, 2011 at 7:54 AM, Duncan Coutts wrote:

...

On Mon, 2011-10-03 at 22:15 -0700, Jason Dagit wrote: > I have a library that uses stdcall on windows and ccall on other > platforms. I found this bug report: > http://hackage.haskell.org/trac/c2hs/ticket/17 > > Does c2hs handle this case correctly yet? The bug has not been close > although a patch was submitted. I'm not sure how to interpret that. > I'd prefer to stick to things that are cabal installable, so if you don't mind making a new release that would be my preferred solution.

By the way, do let us know if the current version actually works now for your use case.

Will do. I told the author of the OpenCL bindings that I would try his bindings (hopefully) by this weekend. I know he uses c2hs but I don't know if he uses the fun macro. Judah also has a binding (hsOpenCL) that does use the fun macro. So even if the bindings I plan on trying out do not make use of this feature, I can try Judah's bindings and report back.

Great thanks. BTW, it should not matter whether you uses 'fun' or 'call' hooks. What is important is that the C header file declare the calling convention using GNU C attribute syntax. Duncan