Thu, 17 Feb 2000 12:04:52 +0100, Sven Panne <[EMAIL PROTECTED]>
pisze:
> Not exactly. The real problem (pointed out to me by Manuel Chakravarty
> some time ago) is exemplified by the following code snippet:
[...]
I had exactly that thing in mind.
Thu, 17 Feb 2000 14:21:12 +0100, Sven Panne <[EMAIL PROTECTED]>
pisze:
> BTW, the foreign object interface will change in future versions
> of GHC:
OK, I transformed my code to use it, works great. (It happened that I
don't need to worry about ForeignObj converted to Addr being finalized
afterwards.)
I use C2HS libraries now (but the c2hs program needs so *MUCH*
memory and time to parse Perl headers that I can't use it here;
fortunately there is no need to marsh C structures). I hope that
GHC's FFI libraries and C2HS will be somewhat unified?
I find c2hs much more convenient than H/Direct when interfacing to
a C library. There is too much glue specific to each function or to
the library being interfaced that I don't believe in a powerful enough
metalanguage that will describe the interface in a language-independent
and convenient way. There should be real Haskell code and bits of C
code to glue everything together, with powerful FFI libraries and
simple direct gateways to the C world.
BTW, greyarea.is.tsukuba.ac.jp is inaccessible. Is it known where
it will be back, or is there any newer version than c2hs-0.7.3?
I will use gtk+hs in one project from March to May.
--
__("< Marcin Kowalczyk * [EMAIL PROTECTED] http://qrczak.ids.net.pl/
\__/ GCS/M d- s+:-- a22 C+++$ UL++>++++$ P+++ L++>++++$ E-
^^ W++ N+++ o? K? w(---) O? M- V? PS-- PE++ Y? PGP+ t
QRCZAK 5? X- R tv-- b+>++ DI D- G+ e>++++ h! r--%>++ y-
