[Mono-bugs] [Bug 78035][Maj] Changed - make check fails due to errors in delegate1 and delegate2 tests

bugzilla-daemon at bugzilla.ximian.com bugzilla-daemon at bugzilla.ximian.com
Tue Apr 11 21:07:53 EDT 2006


Please do not reply to this email- if you want to comment on the bug, go to the
URL shown below and enter your comments there.

Changed by matt at use.net.

http://bugzilla.ximian.com/show_bug.cgi?id=78035

--- shadow/78035	2006-04-10 06:56:45.000000000 -0400
+++ shadow/78035.tmp.5046	2006-04-11 21:07:52.000000000 -0400
@@ -152,6 +152,47 @@
 ------- Additional Comments From lupus at ximian.com  2006-04-10 06:56 -------
 mono_bitset_sub() has been fixed to not access beyond the allocated
 buffer. The mono_unload_interface_ids() code is correct, since the
 destination bitset must be always bigger or equal to the image bitset.
 It would be useful if the reporter ran the tests in gdb and report
 the output of a failing test (I can't reproduce the error either).
+
+------- Additional Comments From matt at use.net  2006-04-11 21:07 -------
+I verified that the mono_bitset_sub() issue is no longer reported, but
+I still see this issue:
+
+==6045== Source and destination overlap in strncpy(0x62C4874,
+0x62C4876, 6)
+==6045==    at 0x4A1A3F3: strncpy (mc_replace_strmem.c:116)
+==6045==    by 0x49019C: canonicalize_path (image.c:115)
+==6045==    by 0x490F1F: mono_image_open_full (image.c:924)
+==6045==    by 0x48CBFE: mono_assembly_open_full (assembly.c:1191)
+==6045==    by 0x418985: mono_main (driver.c:933)
+==6045==    by 0x5518CFF: (below main) (in /lib/libc-2.3.5.so)
+
+Looking at the code at image.c:115, it seems like this is a true
+positive in some cases. You may not be seeing this message because I
+am using a development version of valgrind.
+
+The test output is:
+Test.F from delegate: 2
+Test.F from delegate: 3
+Test.F from delegate: 4
+Test.F from delegate: 1
+Async Callback STATE2
+Async Callback STATE3
+Async Callback STATE4
+Async Callback STATE1
+Result = 1
+cant execute EndInvoke twice ... OK
+completed1: True
+completed2: True
+completed3: True
+completed4: True
+
+-
+
+Not sure what exactly you'd like me to do in gdb. The overlapping
+strncpy reported appears to be causing the memory corruption that
+causes this to fail. Whether the corruption causes the test(s) to fail
+will depend highly on the memory layout as determined by gcc version,
+optimization level, kernel version, glibc version, etc, etc.


More information about the mono-bugs mailing list