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

[bugzilla-daemon at bugzilla.ximian.com]

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	2006-04-06 18:26:11.000000000 -0400
+++ shadow/78035.tmp.13902	2006-04-09 16:59:51.000000000 -0400
@@ -52,6 +52,81 @@
 ------- Additional Comments From vargaz at gmail.com  2006-04-06 18:26 -------
 I can't repro this, and neither can the buildbot, which is running
 a clean build:
 
 http://mono.ximian.com:8008/x86-64-head-mono/builds/3/test%20runtime/0
 
+
+------- Additional Comments From matt at use.net  2006-04-09 16:59 -------
+This appears to happen due to memory corruption, which will cause 
+the test to pass or fail depending on the memory layout as 
+influenced by glibc version, gcc version, optimization levels, 
+kernel revision, etc. Here is the relevant valgrind output:
+
+==28738== Source and destination overlap in strncpy(0x128618DC, 
+0x128618DE, 6)
+==28738==    at 0x11B1C17A: strncpy 
+(in /usr/lib/valgrind/vgpreload_memcheck.so)
+==28738==    by 0x49041C: canonicalize_path (image.c:115)
+==28738==    by 0x49119F: mono_image_open_full (image.c:924)
+==28738==    by 0x48F3F3: get_runtimes_from_exe (domain.c:1415)
+==28738==    by 0x48F546: mono_init_internal (domain.c:526)
+==28738==    by 0x4EEAA0: mini_init (mini.c:10737)
+==28738==    by 0x418953: mono_main (driver.c:876)
+==28738==    by 0x1264CCFF: __libc_start_main 
+(in /lib/libc-2.3.5.so)
+==28738==    by 0x418249: ??? (start.S:113)
+==28738==
+==28738== Source and destination overlap in strncpy(0x128632FC, 
+0x128632FE, 6)
+==28738==    at 0x11B1C17A: strncpy 
+(in /usr/lib/valgrind/vgpreload_memcheck.so)
+==28738==    by 0x49041C: canonicalize_path (image.c:115)
+==28738==    by 0x491CB0: do_mono_image_open (image.c:803)
+==28738==    by 0x491232: mono_image_open_full (image.c:943)
+==28738==    by 0x48F3F3: get_runtimes_from_exe (domain.c:1415)
+==28738==    by 0x48F546: mono_init_internal (domain.c:526)
+==28738==    by 0x4EEAA0: mini_init (mini.c:10737)
+==28738==    by 0x418953: mono_main (driver.c:876)
+==28738==    by 0x1264CCFF: __libc_start_main 
+(in /lib/libc-2.3.5.so)
+==28738==    by 0x418249: ??? (start.S:113)
+
+
+This one may also be relevant, but I don't think it's related to the 
+tests failing on my machine:
+==6793== Thread 1:
+==6793== Invalid read of size 8
+==6793==    at 0x4BAC60: mono_bitset_sub (monobitset.c:627)
+==6793==    by 0x49BB27: mono_unload_interface_ids (class.c:1459)
+==6793==    by 0x490F30: mono_image_close (image.c:1121)
+==6793==    by 0x48C7E2: mono_assembly_close (assembly.c:2192)
+==6793==    by 0x48ED27: mono_domain_free (domain.c:977)
+==6793==    by 0x4EF013: mini_cleanup (mini.c:10962)
+==6793==    by 0x4195CE: mono_main (driver.c:976)
+==6793==    by 0x5518CFF: (below main) (in /lib/libc-2.3.5.so)
+==6793==  Address 0x57E1F40 is 0 bytes after a block of size 24 
+alloc'd
+==6793==    at 0x4A19777: calloc (vg_replace_malloc.c:260)
+==6793==    by 0x4E52866: g_malloc0 
+(in /usr/lib/libglib-2.0.so.0.800.6)
+==6793==    by 0x4BA47C: mono_bitset_new (monobitset.c:50)
+==6793==    by 0x49BBC4: mono_get_unique_iid (class.c:1500)
+==6793==    by 0x49BF5A: mono_class_setup_mono_type (class.c:2552)
+==6793==    by 0x49D459: mono_class_create_from_typedef 
+(class.c:2795)
+==6793==    by 0x49CF7A: mono_class_get (class.c:3672)
+==6793==    by 0x49D224: mono_class_get_full (class.c:3697)
+==6793==    by 0x494EB8: mono_metadata_interfaces_from_typedef_full 
+(metadata.c:2758)
+==6793==    by 0x49D77A: mono_class_create_from_typedef 
+(class.c:2798)
+==6793==    by 0x49CF7A: mono_class_get (class.c:3672)
+==6793==    by 0x49C8D3: mono_class_from_name (class.c:3902)
+
+
+Also worth noting is that all the other tests appear to have these 
+same valgrind warnings, but they are randomly lucky that the 
+corruption doesn't cause them to crash/fail.
+
+Please let me know if more testing or information would be useful in 
+helping fix this bug.