GCC bug workaround

[John Darrington]

I have noticed a really wierd problem, which I think is
a bug in gcc version 4.4.5

When compiled with -O2 tests [185, 189] (the ODS reader tests) fail.

Investigation shows that the stack is corrupt on return from
the function casereader_count_cases__ and the program counter
returns to the wrong address and crashes.

It seems to be a rather wierd bug.

The following patch seems to workaround the symptoms.



diff --git a/src/data/casereader.c b/src/data/casereader.c
index c8050c6..3698201 100644
--- a/src/data/casereader.c
+++ b/src/data/casereader.c
@@ -241,6 +241,9 @@ casereader_count_cases__ (const struct casereader *reader,
 {
   struct casereader *clone;
   casenumber n_cases;
+#if (__GNUC__ == 4 ) && (__GNUC_MINOR__ == 4)
+  volatile int x = 1; x++;
+#endif
 
   clone = casereader_clone (reader);
   n_cases = casereader_advance (clone, max_cases);


Maybe someone can check this doesn't happen with later compilers.

-- 
PGP Public key ID: 1024D/2DE827B3 
fingerprint = 8797 A26D 0854 2EAB 0285  A290 8A67 719C 2DE8 27B3
See http://keys.gnupg.net or any PGP keyserver for public key.

signature.asc
Description: Digital signature