On 7月1日, 午後4:30, gniibe <gni...@gmail.com> wrote:
It found that it's a bug in the test. Akr found the bug.
Here is the fix. I think that it also fix sparc issue too.
--- sigscheme-0.8.3.orig/test-c/test-gc-protect-coll.c
+++ sigscheme-0.8.3/test-c/test-gc-protect-coll.c
@@ -192,10 +192,10 @@
for (i = 0; i < N_OBJS; i++)
TST_TN_TRUE(scm_gc_protectedp(auto_objs[i]));
-#if TRY_TESTS_THAT_PASS_IN_MOST_CASES
/* unprotect again */
for (i = 0; i < N_OBJS; i++)
scm_gc_unprotect(&auto_objs[i]);
+#if TRY_TESTS_THAT_PASS_IN_MOST_CASES
for (i = 0; i < N_OBJS; i++)
TST_TN_FALSE(scm_gc_protectedp(auto_objs[i]));
#endif
@@ -227,7 +227,9 @@
TST_CASE(tst_6, "GC indirect protection via on-heap object
reference")
{
+#if TRY_TESTS_THAT_PASS_IN_MOST_CASES
ScmObj lst; /* unprotected */
+#endif
TST_TN_FALSE(scm_gc_protected_contextp());