On Wed, 2009-03-25 at 08:54 +0000, Simon Peyton-Jones wrote:
> [Narrowing to cvs-ghc]
> 
> So Axel says "we are doomed" and Manuel says "it's easy".
> 
> I don't want GHC 6.10.2 to "fundamentally break" GtkHs.
> 
> Let's resolve this.

We all discussed this today during the #ghc meeting. We think it's
actually not so bad as it initially looked.

Axel is going to test a quick fix and if all is ok then he'll organise a
point release with the Gtk2Hs release manager. The quick fix will be to
use Foreign.Concurrent.mkForeignPtr. The fix that seems plausible for
the next major release of gtk2hs would use ordinary newForeignPtr again
but would use a different way of setting (non-signal) callbacks so that
the finalisers that free said callbacks are in C and not in Haskell (ie
using hs_free_fun_ptr).

If that's not an accurate summary then hopefully Axel will correct me.

Duncan

_______________________________________________
Cvs-ghc mailing list
[email protected]
http://www.haskell.org/mailman/listinfo/cvs-ghc

Reply via email to