On Mon, Feb 14, 2005 at 09:21:11AM -0700, Warren Young wrote:
> I'm beginning to think that it's time that all the legacy exception
> stuff in MySQL++ should die. It was created in a time long ago, when
> not all C++ compilers supported Standard C++ exceptions. Now that we've
> moved the bar up so high w.r.t. compilers that we support, is this
> feature even necessary any more?
I'm all for this.
I think exceptions should be arranged in a heirarchy anyway, to give the
user the choice in how to deal with them, and the old mysql++ exceptions
don't give the user that choice.
> - I could, in good conscience, accept Chris Frey's Row::operator at()
I don't think this is a conscience issue. :-) C++ programmers should be
aware that any C++-oriented exceptions can occur, and the at() function
is part of the std::vector<> class.
mysql++ uses the new operator, and makes use of stl internally. These
facts mean that mysql++ in general inherits the exception characteristics
of the standard libraries. Nothing we can do about that except wrapping
all the library code in try/catch, and nobody wants that....
It's good that someone is thinking of backward compatibility though...
I wasn't :-)