Jim Wallace wrote:
> I'm thinking the deadlock test may not be a useful for mysql++. I've
> altered xaction.cpp to test errnum(), which is much simpler. Here's the
Argh... :) I was going to suggest essentially this very thing, but
decided against it because I didn't want to add an intentional
programming bug to xaction.cpp, because I didn't remember that it
already had one. I figured that if we had to add buggy code, it'd be
better to segregate it into a new example.
And the kicker is that I'm reading this message of yours now because I'd
just finished checking the new example in (with extensive changes of
course, because that's the way I am) and so fired up the email program
to reply to your previous message.
Since we've both gone to all this effort, I say we keep it, unless
someone else complains that it's too ugly to live. The deadlock example