Hi!
On Apr 20, 2010, at 7:40 AM, Joe Daly wrote:
> Id rather see a chunk removed then regressions ran, and if it is a
> significant regression file a bug. I suspect there are chunks that can be
> removed that will show no regression and some that will, if they get mixed
> together you dont have a baseline.
I certainly agree with this.
Right now I am disconnecting the heap to myisam conversion. What have I found?
1) Needed to fix the variables testing.
2) I've found four bugs in the join optimizer for when table conversion was
failing.
3) One big bug in INSERT/SELECT where rows would be cutoff if failure of disk
occurred.
4) One subselect bug in UNION where data truncation occurs because... HEAP to
MyISAM didn't happen correctly.
Fun, Fun. I should be finished with the tree today.
Cheers,
-Brian
_______________________________________________
Mailing list: https://launchpad.net/~drizzle-discuss
Post to : [email protected]
Unsubscribe : https://launchpad.net/~drizzle-discuss
More help : https://help.launchpad.net/ListHelp
