patch: gst-2.95h on Mac OS X 10.3.9

Hi all,

I'm running the now fairly old 10.3.9 version of Mac OS X, and I found
that unfortunately gst-2.95h didn't compile out-of-the-box for me.

The two changes that I had to make were:

* change the header file includes in socketx.h to the older way of
   getting the prototypes for select(2)

* change part of a .section declaration in darwin_closure.S to remove
   mention of the live_support section attribute, which isn't available
   on OS X 10.3.9.

After these changes, gst compiles and runs well for me. Thanks for a
great piece of software! I'm looking forward to getting into it.

The patch is both attached to this message, and available at
http://www.lshift.net/~tonyg/smalltalk-2.95h-for-osx-10.3.9.patch

Regards,
   Tony

diff -ru smalltalk-2.95h/lib-src/socketx.h smalltalk-2.95h-modified/lib-src/socketx.h
--- smalltalk-2.95h/lib-src/socketx.h 2007-11-05 08:22:48.000000000 +0000
+++ smalltalk-2.95h-modified/lib-src/socketx.h 2008-01-01 18:36:39.000000000 +0000
@@ -103,8 +103,10 @@
 
 #else /* !__MSVCRT__ */
 
+#include <sys/types.h>
+#include <sys/time.h>
+#include <unistd.h>
 #include <errno.h>
-#include <sys/select.h>
 #include <sys/socket.h>
 #include <netdb.h>
 #include <netinet/in.h>
diff -ru smalltalk-2.95h/libffi/src/powerpc/darwin_closure.S smalltalk-2.95h-modified/libffi/src/powerpc/darwin_closure.S
--- smalltalk-2.95h/libffi/src/powerpc/darwin_closure.S 2006-03-17 08:00:52.000000000 +0000
+++ smalltalk-2.95h-modified/libffi/src/powerpc/darwin_closure.S 2007-12-30 05:43:47.000000000 +0000
@@ -246,7 +246,7 @@
 /* END(ffi_closure_ASM)  */
 
 .data
-.section __TEXT,__eh_frame,coalesced,no_toc+strip_static_syms+live_support
+.section __TEXT,__eh_frame,coalesced,no_toc+strip_static_syms
 EH_frame1:
  .set L$set$0,LECIE1-LSCIE1
  .long L$set$0 ; Length of Common Information Entry

_______________________________________________
help-smalltalk mailing list
[hidden email]
http://lists.gnu.org/mailman/listinfo/help-smalltalk

> * change the header file includes in socketx.h to the older way of
>   getting the prototypes for select(2)
>
> * change part of a .section declaration in darwin_closure.S to remove
>   mention of the live_support section attribute, which isn't available
>   on OS X 10.3.9.

Thanks!  The second part is, I think, wrong, but since gst doesn't use
closures I don't care much.

Paolo


_______________________________________________
help-smalltalk mailing list
[hidden email]
http://lists.gnu.org/mailman/listinfo/help-smalltalk

Paolo Bonzini wrote:
I stand corrected.  darwin_closure.S does not use
".subsections_via_symbols" so it is okay to remove the live+support
attribute.  I would have applied the patch anyway.

Paolo


_______________________________________________
help-smalltalk mailing list
[hidden email]
http://lists.gnu.org/mailman/listinfo/help-smalltalk

Hi Paolo,

Paolo Bonzini wrote:
> Thanks!  The second part is, I think, wrong, but since gst doesn't use
> closures I don't care much.

Yes, it seemed a bit worrying to me, as well. I did a bit of searching
to see if anyone else had run into it, and the Python ctypes library
did, around March 2006.

 From http://osdir.com/ml/python.ctypes/2006-03/msg00055.html:

"OS X 10.3 doesn't know the live_support flag, and since its only an
optimisation, it can be omitted."

The OS X assembler manual page suggests it's something a little more
important than a simple optimisation, but it's not a very clear section
of the manual.

You may also be interested in
http://mail.python.org/pipermail/python-bugs-list/2006-May/033334.html.

Regards,
   Tony



_______________________________________________
help-smalltalk mailing list
[hidden email]
http://lists.gnu.org/mailman/listinfo/help-smalltalk