Hi Harald.
Thanks for your continued work on this.
I'm wondering... does your solution support self-joins? Do you have
tests for that? If not, it would be fantastic if you could add that.
If you have enough time to help with this early this week, we can get
it into the next 3.2 release.
At this point, I'm looking over your code, trying to understand what's going on.
Thanks!
Patrick Earl
On Mon, Mar 28, 2011 at 2:09 AM, Harald Mueller <[email protected]> wrote:
> Hi Patrick, all,
>
> In my long text from yesterday, there are (at least) two errors:
>
> ----
> In section 2., about 3 lines below (||-1), there is a "not" missing:
>
> Other member expressions (with simple types or component types) are NOT of
> relevance to this discussion, except where explicitly noted.
>
> ----
> In section 8., the third of the "Four examples" is simply wrong! The correct
> argument is:
>
> * When we encounter a.B.C.Q == null (which would become a_B_C.Q IS NULL on
> the database), we do not know anything: This condition becoming true could be
> the result of a missing a.B, or a missing a.B.C, or actually that a.B.C.Q is
> null. So we cannot add anything to the "definitely != null" member expression
> set.
>
> Sorry for the glitches.
>
> Best regards
> Harald
>
> --
> NEU: FreePhone - kostenlos mobil telefonieren und surfen!
> Jetzt informieren: http://www.gmx.net/de/go/freephone
>
