[Top][All Lists]

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

gnustep-base-1.9.1 on ppc-darwin

From: Carl Eugen Hoyos
Subject: gnustep-base-1.9.1 on ppc-darwin
Date: Mon, 24 May 2004 01:11:47 +0200 (CEST)

I tested gnustep-base-1.9.1 on powerpc-darwin 7.3, OSX 10.3.3

Compiling apple-gnu-gnu with Apples latest gcc (build 1640) is very
difficult because of issues like missing definitions of layout and
Category. I wrote a report in savannah. After "solving" all these problems
(I deleted the offending parts), the resulting library didn't work.

As far as I remember, this has never been different for apple-gnu-gnu.

Then I tried with GNU compilers and runtime: This is more difficult now,
because you have to patch configure to get a gnu-objc-runtime and
libffi for 3.4.0 on darwin=-(

Good news is that the library compiles and works both with the releases
gcc-3.3.3 and gcc-3.4.0 using library-combo gnu-gnu-gnu and libffi. I used
libffi-3.4.0 also for gcc-3.3.3, although it doesn't install correctly: I
had to move ffitarget.h into prefix/include after installation!

Tests seem to work: Only the test nsinvocation fails for
"NS_MESSAGE" and "NS_INVOCATION" (The same symptom as on win32).

Installation of dlcompat isn't necessary anymore, because something
similar is now part of darwin-7.3;-)

Two of the warnings while compiling were very interesting:

 Compiling file NSRunLoop.m ...
In file included from NSRunLoop.m:51:
/usr/include/poll.h:73:2: warning: #warning "poll() functionality for Mac
OS X is implemented via an emulation layer on top of select(), not in the
kernel directly.  It is recommended that programs running under OS X 10.3
prefer select() over poll().  Configure scripts should look for the
_POLL_EMUL_H_ define (instead of _POLL_H_ or _SYS_POLL_H_) and avoid
implementations where poll is not implemented in the kernel."

 Compiling file objc-load.m ...
In file included from dynamic-load.h:21,
                 from objc-load.m:54:
/usr/include/dlfcn.h:35:2: warning: #warning "You are using dlopen(), a
legacy API. Please use the Mach-O dylib loading APIs if at all possible"

I tested ffcall-1.9 as well: One test for callback ("make check") fails
for all compilers: Where should this be reported?

If installed anyway, the test nsconnection_client fails for large structs
- again as on win32;-)

Hope that helps someone, Carl Eugen Hoyos, ce AT ag DOT or DOT at

reply via email to

[Prev in Thread] Current Thread [Next in Thread]