"Douglas Gregor" <doug.gregor@gmail.com> wrote in message news:c9d9f5edd926cc1d7f2ab2228d9754bd@cs.indiana.edu...

On Jul 31, 2005, at 12:00 PM, Douglas Gregor wrote:

...

31 failures in 1 library: serialization (31)

|serialization| test_class_info_load_text_warchive_dll: vc-6_5 test_const_load_fail1: borland-5_6_4 test_const_load_fail1_nvp: borland-5_6_4 test_const_load_fail2: borland-5_6_4 test_const_load_fail2_nvp: borland-5_6_4 test_const_save_fail1: vc-6_5 vc-6_5-stlport vc-7_0 test_const_save_fail2: vc-6_5 vc-6_5-stlport vc-7_0 test_const_save_fail3: vc-6_5 vc-6_5-stlport vc-7_0 test_demo_fast_archive_dll: vc-6_5 test_demo_pimpl_dll: vc-6_5 vc-7_0 test_demo_portable_archive_dll: borland-5_6_4 intel-win32-8_1 intel-win32-9_0 test_demo_shared_ptr_dll: vc-6_5 vc-7_0 test_map_binary_archive_dll: vc-6_5 vc-7_0 test_map_text_archive_dll: vc-6_5 vc-7_0 test_map_text_warchive_dll: vc-6_5 vc-7_0 test_mult_archive_types_dll: vc-6_5 test_non_intrusive_binary_archive: vc-8_0 test_shared_ptr_132_xml_archive_dll: intel-win32-8_1

I've gone through and marked up all of these, except this one strange regression:

http://tinyurl.com/ajprh

Looking at the output of the test, I don't even seeing an error. To make things more interesting, the other vc-8_0 tester isn't having any problem. Ideas, anyone?

Different builds of the compiler? Because Microsoft is issuing monthly community updates, it is really easy for testers to diverge as to what build they are running. --Beman