This is the mail archive of the gcc-patches@gcc.gnu.org mailing list for the GCC project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

3.4 PATCH: Fix many Tru64 UNIX V4.0F libffi testsuite failures


Currently, there are many libffi testsuite failures on alpha-dec-osf4.0f
(unlike alpha-dec-osf5.1b):

FAIL: libffi.call/closure_fn0.c (test for excess errors)
Excess errors:
In file included from /vol/gnu/src/gcc/gcc/libffi/testsuite/libffi.call/closure_fn0.c:10:
/vol/gnu/src/gcc/gcc/libffi/testsuite/libffi.call/ffitest.h: In function `allocate_mmap':
/vol/gnu/src/gcc/gcc/libffi/testsuite/libffi.call/ffitest.h:70: warning: comparison between pointer and integer

The following trivial patch fixes this.  With this change, testsuite
results look good again and are mostly the same for Tru64 UNIX V4.0F and
V5.1B:

--- 4.0f-gcc-local/mail-report.log	Wed Jan  7 21:50:30 2004
+++ 5.1b-gcc-local/mail-report.log	Wed Jan  7 08:25:42 2004
@@ -1,21 +1,20 @@
 cat <<'EOF' |
 LAST_UPDATED: Mon Dec 22 20:19:27 UTC 2003
 
-Native configuration is alpha-dec-osf4.0f
+Native configuration is alpha-dec-osf5.1b
 
 		=== libffi tests ===
 
 
 Running target unix
-FAIL: libffi.call/cls_align_sint64.c output pattern test, is 12 ld 4951 127 ld 1: 9320 ld 13
-FAIL: libffi.call/cls_align_uint64.c output pattern test, is 12 ld 4951 127 ld 1: 9320 ld 13
-FAIL: libffi.call/cls_ulonglong.c output pattern test, is lu: lu
-FAIL: libffi.special/unwindtest.cc (test for excess errors)
-WARNING: libffi.special/unwindtest.cc compilation failed to produce executable
+FAIL: libffi.call/cls_align_longdouble.c output pattern test, is 12 0 0 224 0 4: 236 0 4
+FAIL: libffi.call/float.c execution test
+FAIL: libffi.call/float2.c execution test
+FAIL: libffi.special/unwindtest.cc execution test
 
 		=== libffi Summary ===
 
-# of expected passes		206
+# of expected passes		207
 # of unexpected failures	4
 # of unsupported tests		2

The float.c/float2.c test failures are already known (PR target/11398), and
the first three tests seem to fail on Tru64 UNIX V4.0F since printf doesn't
support %lld/%llu (it wouldn't be different from %ld/%lu anyway on that
platform).

Ok for mainline?

	Rainer

-----------------------------------------------------------------------------
Rainer Orth, Faculty of Technology, Bielefeld University


Wed Jan  7 20:36:43 2004  Rainer Orth  <ro@TechFak.Uni-Bielefeld.DE>

	* testsuite/libffi.call/ffitest.h (allocate_mmap): Cast MAP_FAILED
	to void *.

Index: libffi/testsuite/libffi.call/ffitest.h
===================================================================
RCS file: /cvs/gcc/gcc/libffi/testsuite/libffi.call/ffitest.h,v
retrieving revision 1.3
diff -u -p -r1.3 ffitest.h
--- libffi/testsuite/libffi.call/ffitest.h	26 Nov 2003 13:46:10 -0000	1.3
+++ libffi/testsuite/libffi.call/ffitest.h	7 Jan 2004 20:46:11 -0000
@@ -67,7 +67,7 @@ allocate_mmap (size_t size)
 	       MAP_PRIVATE, dev_zero_fd, 0);
 #endif
 
-  if (page == MAP_FAILED)
+  if (page == (void *) MAP_FAILED)
     {
       perror ("virtual memory exhausted");
       exit (1);


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]