Merge tests #18 and #41 currently fail with SIGABRT on the
merge-tracking branch. The stack shows that a pool cleanup is
triggering an abort(), likely due to a leaking error:
#0 0x00a42402 in __kernel_vsyscall ()
#1 0x4afa1d40 in raise () from /lib/libc.so.6
#2 0x4afa3591 in abort () from /lib/libc.so.6
#3 0x00e201df in err_abort (data=0x85dddf8)
at subversion/libsvn_subr/error.c:68
#4 0x00ca00cd in run_cleanups (cref=0x85dddd0) at memory/unix/apr_pools.c:2048
#5 0x00ca084d in apr_pool_destroy (pool=0x85dddc0)
at memory/unix/apr_pools.c:727
#6 0x00ca0838 in apr_pool_destroy (pool=0x85678b0)
at memory/unix/apr_pools.c:724
#7 0x00ca0bce in apr_pool_terminate () at memory/unix/apr_pools.c:568
#8 0x00ca279a in apr_terminate () at misc/unix/start.c:82
#9 0x4afa4939 in exit () from /lib/libc.so.6
#10 0x4af8ef34 in __libc_start_main () from /lib/libc.so.6
#11 0x0804b891 in _start ()
This can be reproduced using 'config.nice --enable-maintainer-mode'.
As I cannot reproduce these failures on trunk, I have a feeling that
they're due to the recent work on use of sub-tree merge info.
- application/pgp-signature attachment: stored
Received on Tue Jan 23 00:17:01 2007