here is the end of an strace:

read(18, "\253\1\0\0 \2\0\f", 8)        = 8
write(2, "memo-file-conduit-Message: MemoF"...,
50memo-file-conduit-Message: MemoFile Conduit v 0.9
) = 50
write(2, "memo-file-conduit-Message: PreSy"...,
35memo-file-conduit-Message: PreSync
) = 35
brk(0x815d000)                          = 0x815d000
select(19, NULL, [18], NULL, {10, 0})   = 1 (out [18], left {10, 0})
write(18, "\0019\0\0\0\n", 6)           = 6
nanosleep({0, 16000}, NULL)             = 0
select(19, NULL, [18], NULL, {10, 0})   = 1 (out [18], left {10, 0})
write(18, "\33\1 \6\3\0\0\0\377\377", 10) = 10
nanosleep({0, 20000}, NULL)             = 0
select(19, [18], NULL, NULL, {10, 0})   = 1 (in [18], left {10, 0})
read(18, "\0019\0\0\1$", 6)             = 6
select(19, [18], NULL, NULL, {10, 0})   = 1 (in [18], left {10, 0})
read(18, "\233\1\0\0\240\0\1\34\1\32\0\17Unfiled\0\0\0\0\0\0\0\0"...,
292) = 292
write(2, "memo-file-conduit-Message: load_"...,
40memo-file-conduit-Message: load_records
) = 40
open("/home/james/Memos", O_RDONLY|O_NONBLOCK|O_LARGEFILE|O_DIRECTORY)
= 20
fstat64(20, {st_mode=S_IFDIR|0755, st_size=4096, ...}) = 0
fcntl64(20, F_SETFD, FD_CLOEXEC)        = 0
write(2, "memo-file-conduit-Message: load_"...,
43memo-file-conduit-Message: load_categories
) = 43
open("/home/james/Memos/.categories", O_RDONLY) = -1 ENOENT (No such
file or directory)
write(2, "memo-file-conduit-Message: no ca"...,
53memo-file-conduit-Message: no categories, no records
) = 53
close(20)                               = 0
write(2, "gpilotd-Message: Performing Slow"..., 49gpilotd-Message:
Performing Slow Synchronization
) = 49
select(19, NULL, [18], NULL, {10, 0})   = 1 (out [18], left {10, 0})
write(18, "\1:\0\0\0\f", 6)             = 6
nanosleep({0, 16000}, NULL)             = 0
select(19, NULL, [18], NULL, {10, 0})   = 1 (out [18], left {10, 0})
write(18, " \1!\10\3\0\0\0\0\0\377\377", 12) = 12
nanosleep({0, 22000}, NULL)             = 0
select(19, [18], NULL, NULL, {10, 0})   = 1 (in [18], left {10, 0})
read(18, "\1:\0\0\0\331", 6)            = 6
select(19, [18], NULL, NULL, {10, 0})   = 1 (in [18], left {10, 0})
read(18, "\240\1\0\0 \323\0o`\35\0\0\0\311\0\2Board meeting\n\nn"...,
217) = 217
--- SIGSEGV (Segmentation fault) ---
write(3, "\33\0\2\0\0\0\0\0 N\2\0\0\0\0\0+E\1\0", 20) = 20
read(3, "\1\2b\0\0\0\0\0q\222#\3\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0"...,
32) = 32
fork()                                  = 1940
wait4(1940, [WIFEXITED(s) && WEXITSTATUS(s) == 0], 0, NULL) = 1940
--- SIGCHLD (Child exited) ---
sigreturn()                             = ? (mask now [SEGV RTMIN])
_exit(1)                                = ?

*** Bug 114794 has been marked as a duplicate of this bug. ***

Other than bug 114794, which I just marked as a duplicate of this bug,
this appears to be a unique stack trace, according to the
simple-dup-finder.  I'm adding the bugsquad keyword and marking as new.

*** Bug 114987 has been marked as a duplicate of this bug. ***

Looks like I forgot to mark priority->high as per bugsquad guidelines
with new stack trace crasher bugs.  I'm doing that now.

*** Bug 116616 has been marked as a duplicate of this bug. ***

Created attachment 19582 [details] [review]
patch against gnome-pilot-2.0.10 distribution, to fix memo-file conduit bug

The patch I've just submitted fixes the reported memo-file bug.
It should be applied to the stock gnome-pilots-2.0.10 
distribution.

The problem actually lay in gpilotd code, that doesn't seem
to be used by the other gnome-pilot-conduits conduit code,
AFAICS.

The gnome-pilot-conduit-standard-abs.c code hadn't been
updated to use the newer gtk marshallers, so when the
gtk code called the marshal (when the gtk signal fired)
the marshaller would grab nonsense of the stack and call
a bogus handler.

The fix affects gpilotd/gnome-pilot-conduit-standard-abs.c
and gpilotd/gpmarshal.[c,h,list]

Changes were made to memo_file_conduit.c in v.2.0.10, to try to
prevent the memo-file bug from crashing gpilotd.  At least some
of these changes were unnecessary, and had the side effect
of making it impossible to run memo-file-conduit unless the
memo-file categories already existed locally.  I'm attaching
a patch by Nathan Owens, against the gnome-pilots-conduits-2.0.10
distributed memo_file_conduit.c that reverses these changes.

I believe both these patches should go into 2.0.11

Created attachment 19586 [details] [review]
patch to memo_file_conduit 2.0.10, to allow memo-file to run before local records exist. partial reverse of 2.0.9--2.0.10 patch.

this patch solves the memo-file sync problems for me against
gnome-pilot*-2.0.10.

I put in the gnome-pilot one and modified the conduit one to leave in error
checking for load_records but still work on a brand new setup.

*** Bug 122572 has been marked as a duplicate of this bug. ***

*** Bug 115319 has been marked as a duplicate of this bug. ***