[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: porting
From: |
Armando Di Cianno |
Subject: |
Re: porting |
Date: |
Wed, 05 Oct 2005 13:24:02 -0400 |
-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1
On 2005-10-05 12:18:10 -0400 Helge Hess <helge.hess@opengroupware.org>
wrote:
On 5. Okt 2005, at 17:33 Uhr, Armando Di Cianno wrote:
WRT GNUstep, I've had varying luck with both, possibly leaning toward
the sentiment that ffcall is more stable than libffi (again, wrt
GNUstep).
Isn't libffi included in GCC? What about using the one coming with
the
compiler, this is probably best maintained and a dependency is
removed?
Not sure how the different binary distros do it, but libffi is only
installed by gcc if gcj is configured to be on (unless you use the
different patches floating around). I'm hoping this will be fixed in
some release of gcc-4 series (making it an installable target by
itself).
Also, I'm unclear as to how much development v. maintainence libffi
receives. I know some projects like SableVM maintain their own
version of libffi, IIRC.
I've also been personally affected by a situation where libffi, wrt
GNUstep, worked on one Pentium-M (earlier rev) platform, and
absolutely failed on another Pentium-M (never rev, 400Mhz FSB).
Because of this, and misc. anecdotes I have from the past year and
half of using libffi with GNUstep, I firmly believe that ffcall is a
much better supported option, for GNUstep, at this point. Having said
that, it would be nice to have libffi working and stable, and
well-supported by GNUstep.
__Armando Di Cianno
Greets,
Helge
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.1-ecc0.1.6 (GNU/Linux)
Comment: Using the GPG bundle for GNUMail
iD8DBQFDRAwwwgiTPLI9xhcRAlh5AJ9bnEkJIeF6lLLvZ28xqNY5P+utPACaAwkK
I/as24tmtReCpgBNvs/3R/o=
=hl3V
-----END PGP SIGNATURE-----