Buildbots sick (segfault in svn_cache_set_error_handler() )

All of the buildbots are currently failing. When I try to run the
basic_tests.py on my OS X Leopard system, I get:

hyrum-wrights-macbook-pro:cmdline Hyrum$ ./basic_tests.py --verbose
CMD: /Users/Hyrum/dev/svn-trunk/subversion/svnadmin/svnadmin create
svn-test-work/local_tmp/repos --bdb-txn-nosync terminated by signal 10
CMD: svnadmin create svn-test-work/local_tmp/repos
--bdb-txn-nosyncTraceback (most recent call last):
   File "./basic_tests.py", line 2371, in <module>
     svntest.main.run_tests(test_list)
   File
"/Users/Hyrum/dev/svn-trunk/subversion/tests/cmdline/svntest/main.py",
line 1433, in run_tests
     actions.setup_pristine_repository()
   File
"/Users/Hyrum/dev/svn-trunk/subversion/tests/cmdline/svntest/actions.py",
line 44, in setup_pristine_repository
     main.create_repos(main.pristine_dir)
   File
"/Users/Hyrum/dev/svn-trunk/subversion/tests/cmdline/svntest/main.py",
line 615, in create_repos
     path, *opts)
   File
"/Users/Hyrum/dev/svn-trunk/subversion/tests/cmdline/svntest/main.py",
line 322, in run_command
     None, *varargs)
   File
"/Users/Hyrum/dev/svn-trunk/subversion/tests/cmdline/svntest/main.py",
line 436, in run_command_stdin
     *varargs)
   File
"/Users/Hyrum/dev/svn-trunk/subversion/tests/cmdline/svntest/main.py",
line 416, in spawn_process
     exit_code = wait_on_pipe(kid, stdout_lines, stderr_lines)
   File
"/Users/Hyrum/dev/svn-trunk/subversion/tests/cmdline/svntest/main.py",
line 383, in wait_on_pipe
     raise SVNProcessTerminatedBySignal
svntest.main.SVNProcessTerminatedBySignal

Running the above svnadmin create command in the debugger gives:
Program received signal EXC_BAD_ACCESS, Could not access memory.
Reason: KERN_PROTECTION_FAILURE at address: 0x00000004
0x0005ffbd in svn_cache_set_error_handler (cache=0x0, handler=0x1b783
<warn_on_cache_errors>, baton=0x8229d8, pool=0x811418) at
subversion/libsvn_subr/cache.c:27
/Users/Hyrum/dev/svn-trunk/subversion/libsvn_subr/cache.c:27:1010:beg:0x5ffbd
(gdb)
No source file named address.
(gdb) backtrace
#0 0x0005ffbd in svn_cache_set_error_handler (cache=0x0,
handler=0x1b783 <warn_on_cache_errors>, baton=0x8229d8, pool=0x811418)
at subversion/libsvn_subr/cache.c:27
#1 0x0001bc6c in svn_fs_fs__initialize_caches (fs=0x8229d8,
pool=0x811418) at subversion/libsvn_fs_fs/caching.c:237
#2 0x0001ec15 in fs_create (fs=0x8229d8, path=0x812300
"svn-test-work/local_tmp/repos/db", pool=0x811418, common_pool=0x813418)
at subversion/libsvn_fs_fs/fs.c:185
#3 0x00019e56 in svn_fs_create (fs_p=0x812290, path=0x812300
"svn-test-work/local_tmp/repos/db", fs_config=0x8115e8, pool=0x811418)
at subversion/libsvn_fs/fs-loader.c:375
#4 0x0001553e in svn_repos_create (repos_p=0xbffff5fc, path=0x8115c8
"svn-test-work/local_tmp/repos", unused_1=0x0, unused_2=0x0,
config=0x811680, fs_config=0x8115e8, pool=0x811418) at
subversion/libsvn_repos/repos.c:1207
#5 0x00002567 in subcommand_create (os=0x811530, baton=0xbffff66c,
pool=0x811418) at subversion/svnadmin/main.c:553
#6 0x00005522 in main (argc=4, argv=0xbffff744) at
subversion/svnadmin/main.c:1685

So it would appear that there is a segfault somewhere in the new caching
code. Or that a caller is not using it appropriately.

-Hyrum