Am 02.08.2005 um 19:18 schrieb D. Richard Hipp:
On Tue, 2005-08-02 at 09:30 -0400, D. Richard Hipp wrote:
On Mon, 2005-08-01 at 22:04 +0200, Jens Miltner wrote:
we get an assertion (no crash here, though) in btree.c
and the backtrace looks similar to the one scunacc provided, which
made me think the two might be related...
I am able to reproduce the bug described in ticket #1346.
It looks like this may be the same problem. I will let
everybody know as soon as I have a fix.
Patches to fix ticket #1346 are available at
http://www.sqlite.org/cvstrac/chngview?cn=2573
Please try adding these patches and see if they do not
fix the problem in the multi-threaded application.
Thanks, from a first quick test, this seems to fix our problem with
the assertion.
BTW, SQLite runs about 2x faster if you compile with
-DNDEBUG=1 which disables all of the assert() statements.
Yes, I figured that running with debug code is not optimal, but we
only link against the debugging build of sqlite in the debug version
of our app, which is why we saw the assertions.
In the release builds, we link against sqlite built with -DNDEBUG=1
(or actually, configure run without the --enable-debug option, which
I assume, will generate the -DNDEBUG=1 C option).
Thanks again for the quick fix.
</jum>
