One good turn deserves another. I just committed this patch:

  Index: dso.c
  ===================================================================
  RCS file: /home/cvs/apr/dso/unix/dso.c,v
  retrieving revision 1.45
  retrieving revision 1.46
  diff -u -r1.45 -r1.46
  --- dso.c     2001/11/20 08:18:38     1.45
  +++ dso.c     2001/11/20 16:31:25     1.46
  @@ -117,7 +117,7 @@

#elif defined(DSO_USE_DYLD)
NSObjectFileImage image;
- NSModule os_handle;
+ NSModule os_handle = NULL;
char* err_msg = NULL;
if (NSCreateObjectFileImageFromFile(path, &image) != NSObjectFileImageSuccess) {
err_msg = "cannot create object file image";



So DSO should be considerably happier on Darwin now. :)

-Fitz


On Tuesday, November 20, 2001, at 01:46 AM, Sander Temme wrote:

We were not returning the pointer to the module structure that we just
pulled out of the DSO file. No wonder the module structure we tried to use
in mod_so.c:load_module was total garbage! The following patch fixes that
problem.

Index: dso.c
===================================================================
RCS file: /home/cvspublic/apr/dso/unix/dso.c,v
retrieving revision 1.44
diff -u -r1.44 dso.c
--- dso.c       2001/11/13 21:15:33     1.44
+++ dso.c       2001/11/20 07:21:02
@@ -209,6 +209,7 @@
         handle->errormsg = "cannot resolve symbol";
        return APR_EINIT;
     }
+    *ressym = retval;
     return APR_SUCCESS;
 #elif defined(DSO_USE_DLFCN)

Enjoy,

Sander



Reply via email to