Hallo Sergey,
> I wonder why the error should be produced at all? The UPDATE
> statement in the bug example updates non-indexed column, so from
> user POV, there is no reason for the UPDATE to fail with an error.
This is indeed what I was looking into when I got side-tracked by
some other bugs. I expect to take a look at it by Monday, presumably
tomorrow. The good news is that IIRC, DUPE KEY is thrown right
away, for the second row. Once I know where it happens, I hope
to write a test that's smaller than 6 MB, as well! ;)
> <non-verified-speculation>
Very possible -- this meshes well with what I've seen so far.
thanks,
Tatjana
--
Tatjana A. Nurnberg, Software Developer * MySQL AB, http://www.mysql.de/ * EST
IRC: Azundris * Skype: azundris * MySQL-SIP: 4550 * eMail: azundris@stripped
.de: Firmensitz: MySQL GmbH, Radlkoferstr. 2, D-81373 Muenchen
Geschaeftsfuehrer: Hans von Bell, Kaj Arnoe - HRB Muenchen 162140
