Re: query.str() in template query -- <blush>

[Warren Young]

Jim Wallace wrote:
> 
> It'd be cool if this abuse could be detected at compile time instead
> of runtime.

Suggestions as to how?

I wouldn't be averse to adding an exception for this.  That puts 
detection off to runtime, but it's then impossible to mistake for a 
MySQL++ bug.

I'm also willing to accept patches that allow this brokenness to work 
*as long as* they don't break something else.  That's the main reason we 
don't claim to support this: it's interfered with other things in the 
past, usually due to some C++ restriction on overloading.  Since the 
Query::template_defaults mechanism is meant for other things, I didn't 
shed a tear when learning that it can't be used to completely fill out a 
query in all situations.