Comment 18 for bug 492391

Revision history for this message
Stewart Smith (stewart) wrote : Re: [Bug 492391] Re: Random randgen test failures

On Tue, 20 Apr 2010 18:37:13 -0000, Brian Aker <email address hidden> wrote:
> On Apr 20, 2010, at 11:24 AM, Lee Bieber wrote:
>
> > drizzled: field/varstring.cc:90: virtual int
> Field_varstring::store(const char*, uint32_t, const CHARSET_INFO*):
> Assertion `!table || (table->write_set == __null || isWriteSet())'
> failed.
>
> Bitfield issues. We have no compressive test on those at the moment.
> Engines are not safe to use them as a means for testing anything but
> BLOBs.

The API is a little obscure, but it should work. This is how we do
partial row fetch/set (to save disk reads or network hops). Anything
hitting these asserts should be considered a bug and be fixed.

If using the Field classes to do read row for this, it's pretty simple
to get it going right.

--
Stewart Smith