arla-0.32 glitches (with patches)
David Sanderson
david at transarc.ibm.com
Fri Mar 17 00:10:34 CET 2000
There is a problem with the generated *_ExecuteRequest RPC code where if
an unknown opcode comes in, _result contains random garbage as the error
number.
This should make _result deterministic in this case:
----------------------------------------------------------------------
*** ydr/output.c0 Wed Mar 1 18:29:46 2000
--- ydr/output.c Thu Mar 16 17:44:48 2000
***************
*** 1941,1946 ****
--- 1941,1947 ----
listiter (func_list, gencase, &c);
fprintf (c_file, "default:\n"
+ "_result = RXGEN_OPCODE;\n"
"fprintf (stderr, \"Ignoring %%d\\n\", opcode);\n"
"}\n"
"return _result;\n"
----------------------------------------------------------------------
and this should (hopefully) let the RXGEN_OPCODE macro be visible when
compiling. It was at least enough to get the objects in the rxdef/
directory to compile.
----------------------------------------------------------------------
*** rx/rx.h0 Fri Mar 3 03:55:15 2000
--- rx/rx.h Thu Mar 16 13:41:41 2000
***************
*** 31,36 ****
--- 31,37 ----
#include "../rx/rx_event.h"
#include "../rx/rx_queue.h"
#include "../rx/rx_packet.h"
+ #include "../rx/rxgencon.h"
#else /* KERNEL */
#ifdef HAVE_CONFIG_H
#include <config.h>
***************
*** 43,48 ****
--- 44,50 ----
#include "rx_clock.h"
#include "rx_event.h"
#include "rx_pkt.h"
+ #include "rxgencon.h"
#endif /* KERNEL */
----------------------------------------------------------------------
I also encountered a build problem with arla-0.32 where 'Bool' was
already defined as a macro, leading to a typedef error. This fixed it
for me. I saw this problem on Solaris 2.6 but did not take time to
track down where the Bool macro was getting defined.
----------------------------------------------------------------------
*** util/bool.h0 Sun Nov 9 18:48:45 1997
--- util/bool.h Thu Mar 16 12:25:59 2000
***************
*** 53,59 ****
--- 53,61 ----
#else /* FALSE */
+ #ifndef Bool
typedef int Bool;
+ #endif
#endif
----------------------------------------------------------------------
David Sanderson
IBM Transarc Lab
david at transarc.ibm.com
More information about the Arla-drinkers
mailing list