On Fri, Jul 23, 2010 at 9:39 AM, OLOPS Daniel (MM)
<[email protected]> wrote:
> - Compile my on glib (that's the kind of voodoo I would prefer to avoid!): My
> compilation attempt was unsuccessful, so I installed an older version
> (2.12.4) from IBM which doesn't face the "-qthreaded" problem, but even so
> Amanda complaint about invalid magic numbers in Perl modules.
This is the kind of voodoo you'll have to do, too -- unless you can
get your hands on the IBM compiler. I'm not surprised that another
version of glib from IBM was compiled with the IBM compiler.
As for the random_data redefinition, looking in config.log I see this program:
| #include <stdlib.h>
| #if HAVE_RANDOM_H
| # include <random.h>
| #endif
|
|
| int
| main ()
| {
| if (sizeof (struct random_data))
| return 0;
| ;
| return 0;
| }
fail with
conftest.c: In function 'main':
conftest.c:149: error: invalid application of 'sizeof' to incomplete
type 'struct random_data'
which suggests that struct random_data is *not* defined in stdlib.h..
yet apparently it is? twice?
Can you look at stdlib.h around line 57 and see if there's anything in
particular that might cause this discrepancy?
The related code is in gnulib, so it might help to upgrade the gnulib
included with the Amanda source. Check out the gnulib source
(http://www.gnu.org/software/gnulib/), then, from the root of the
Amanda tree:
GNULIB_TOOL=/path/to/gnulib/tool ./gnulib/regenerate/regenerate
Note that gnulib, glib, and glibc are completely unrelated to one
another, despite the letters they share in common.
Dustin
--
Open Source Storage Engineer
http://www.zmanda.com
