Hello,
I'm trying to build Chicken using cmake under GNU/Linux. I can't find
a way to enable/disable the compilation with/without libffi.
What I get is (from chicken-2.522):
$ ccmake .
Page 1 of 1
BUILD_TESTINGON
CMAKE_B
Mario Domenech Goulart wrote:
Hello,
I'm trying to build Chicken using cmake under GNU/Linux. I can't find
a way to enable/disable the compilation with/without libffi.
There is no such option in the CMake build. If ffi.h exists on the
system, -lffi is added to the linker flags. Is there
Hi Brandon,
On Tue, 20 Feb 2007 11:58:58 -0800 "Brandon J. Van Every" <[EMAIL PROTECTED]>
wrote:
> Mario Domenech Goulart wrote:
> >
> > I'm trying to build Chicken using cmake under GNU/Linux. I can't find
> > a way to enable/disable the compilation with/without libffi.
> >
>
> There is no su
Mario Domenech Goulart scripsit:
> Borked systems (mine, for example). :-) I don't know why, but I could
> not compile Chicken with libffi on my machine. I may be wrong, but
> I'm pretty sure it's a local problem, since I could build it fine on
> another system. Anyway, using configure's --witho
Mario Domenech Goulart wrote:
Hi Brandon,
On Tue, 20 Feb 2007 11:58:58 -0800 "Brandon J. Van Every" <[EMAIL PROTECTED]>
wrote:
Mario Domenech Goulart wrote:
I'm trying to build Chicken using cmake under GNU/Linux. I can't find
a way to enable/disable the compilation with/without lib
On 2/20/07, Brandon J. Van Every <[EMAIL PROTECTED]> wrote:
Borked systems (mine, for example). :-) I don't know why, but I could
not compile Chicken with libffi on my machine. I may be wrong, but
I'm pretty sure it's a local problem, since I could build it fine on
another system. Anyway, using
felix winkelmann wrote:
The problem is that there exist many versions of libffi, and some are
simply
broken or badly installed. Being able to disable should be possible in
these
cases. I will look into CMakeLists.txt and make it optional.
Are *old* libffi's the problem? If so, we could tes
Brandon J. Van Every scripsit:
> Or is it some horrible set of parallel ill-defined borkings that no
> one in their mind would want to deal with?
Pretty much. As Felix said, it isn't always libffi itself but the
way it is installed.
> In which case, I might ask why use libffi at all?
Because i
John Cowan wrote:
Brandon J. Van Every scripsit:
Or, why use a system libffi. If it's that bad, it would be safer to
put a "known good" version in Chicken itself, as we do with PCRE.
Because it has to be synchronized with gcc; it has intimate knowledge
of gcc's calling conventions on
On 2/21/07, Brandon J. Van Every <[EMAIL PROTECTED]> wrote:
Is there a simple way to test that the libffi actually works when compiled,
linked, and run? Because then we could implement a TRY_RUN. The simple
test would have to be a valid representation of what's required to run,
though. Other
John, could you try the current darcs head? I have
added a user option (WITHOUT_LIBFFI).
cheers,
felix
___
Chicken-users mailing list
Chicken-users@nongnu.org
http://lists.nongnu.org/mailman/listinfo/chicken-users
On Fri, 23 Feb 2007 08:30:21 +0100 "felix winkelmann" <[EMAIL PROTECTED]> wrote:
> John, could you try the current darcs head? I have
> added a user option (WITHOUT_LIBFFI).
Tried it on GNU/Linux with cmake 2.4-patch 5. It seems to be ok.
Thanks.
Best wishes,
Mario
_
felix winkelmann scripsit:
> John, could you try the current darcs head? I have
> added a user option (WITHOUT_LIBFFI).
I built and tested successfully both with and without, so my libffi
seems to be working again, probably a result of the new drop of gcc 3.x.x.
--
There is / One art
On 23 Feb 2007 11:26:36 -0300, Mario Domenech Goulart
<[EMAIL PROTECTED]> wrote:
On Fri, 23 Feb 2007 08:30:21 +0100 "felix winkelmann" <[EMAIL PROTECTED]> wrote:
> John, could you try the current darcs head? I have
> added a user option (WITHOUT_LIBFFI).
Tried it on GNU/Linux with cmake 2.4-pat
14 matches
Mail list logo