Mathew Heard
Mathew Heard
But that's supposed to be fixed in 2.18.2? Now that I (appear to) have replication I can do some more testing. If there's anything you can think of let me...
So if I am understanding correctly theres a race between the start of the scheduler that normally occurs and the start of a scheduler on "ALTER TABLE"? The theoretical validation...
#7156 related?
Sounds like the same as #7156 which itself is a re-post of #6141 and #6224
Despite the name Cursors used in this library do not appear to be declared cursors, instead they are an API to stream results.
I think 1 should be fixed I think 2 is fine due to r->main not being NULL. I'm not sure I see the issue with 3 (you see thats does...
You are right re; 2.