Re: [Opensync-users] opensync+isync breaks fast syncing, generates dupes

[Simon J. <si...@jo...>]

Michael Bell <mic...@cm...> writes:

> Hi Simon,
>
> Simon Josefsson schrieb:
>
>> Hi again.  I'm not sure from the latest discussions whether my original
>> problem is supposed to be fixed or not, but I just tried using the
>> latest svn version of opensync, file-sync, syncml, vformat (but
>> libsyncml 0.4.6 because libsyncml svn breaks directly with smlSafeFree
>> errors) and I still get duplicates when syncing syncml-obex-client to
>> file-sync.  Slightly trimmed output as below.  Should I expect any
>> change in this yet?
>
> The problem changed a little bit. My actual environment is libsyncml
> rev 383 and all other stuff is at rev 3179. I get no smlSafeFree
> errors. So if you get such errors then I need your traces or at
> minimum the last 20-100 lines of each log file (Thread-*). I can sync
> and the syncml part works very well.

Hi Michael.  I'm attaching the zip'ed logs in this mail.  It dies pretty
fast, here is the output:

jas@mocca:/tmp/log$ time syncml-obex-client -b $MAC $CHANNEL --slow-sync text/x-vcard Contacts --wbxml --identifier "PC Suite"
/home/jas/src/osync/libsyncml/libsyncml/sml_support.c:441:E:smlSafeFree: Assertion "*address" failed
Aborted

real	0m0.027s
user	0m0.012s
sys	0m0.012s
jas@mocca:/tmp/log$ 

A gdb backtrace is:

(gdb) bt
#0  0xffffe410 in __kernel_vsyscall ()
#1  0xb7b03f15 in raise () from /lib/i686/cmov/libc.so.6
#2  0xb7b05891 in abort () from /lib/i686/cmov/libc.so.6
#3  0xb7efd855 in smlSafeFree () from /usr/local/lib/libsyncml.so.0
#4  0xb7efd88e in smlSafeCFree () from /usr/local/lib/libsyncml.so.0
#5  0xb7ef3269 in smlNotificationSend () from /usr/local/lib/libsyncml.so.0
#6  0x0804c440 in main ()
(gdb) 

I'll have to figure out how to build with debug information...

> The problem is that the plugin signals SLOW-SYNC during connect but
> OpenSync or msynctool does not react. I don't know which part of the
> software should check for such an event. Perhaps Daniel can clarify
> this. If I know where I have to search then I can try to find the bug.
>
> So your problem is still some kind of open. We fixed the first bug and
> found another one :(

Ok, no problem, and thanks for explaining.  Forcing slow-syncing makes
things work for me, although I suspect it is causing unnecessarily many
merge dupes with the evolution plugin.  But I can track down and report
those separately.

/Simon