This is the mail archive of the
libffi-discuss@sourceware.org
mailing list for the libffi project.
Re: libffi on i386 Plan9/APE
- From: Andrew Haley <aph at redhat dot com>
- To: Jens Staal <staal1978 at gmail dot com>
- Cc: libffi-discuss at sourceware dot org
- Date: Mon, 03 Dec 2012 09:27:17 +0000
- Subject: Re: libffi on i386 Plan9/APE
- References: <1431145.hgiI9FWO1y@krypton>
On 12/03/2012 05:58 AM, Jens Staal wrote:
> I managed to sort-of build libffi on Plan9 under the ANSI-POSIX ENVIRONMENT.
> On this OS only a static library can be built.
>
> The configure choked on a couple of things assuming GNU coreutils-behavior
> (the command "cut" does not exist, and I needed to sed away a couple of
> options from ls and mv), so I ended up manually adjusting files that failed to
> be configured (so there is lots of room for improvements on my port).
>
> Two files under x86 (unix.S and sysv.S) are GAS ASM, which is different from
> Plan9 ASM so they could not be compiled into objects. Are there C variants of
> those files available?
No, and it's not possible to do so.
> Are those two object files essential for libffi function?
Yes. They are the core of libffi. They do the arg passing from
memory to registers and back.
> What are the relevant tests to do of libffi.a ?
There's a full DejaGNU testsuite provided, run with "make check".
Andrew.