Chris Frey wrote:
> double<2>, price, // 2 digit money
Not a good example. Serious programs dealing in money do it in some
form of fixed-point arithmetic.
As for the broader implications of your proposal, I'm going to have to
cogitate on that. I can tell you right now that it shouldn't go in 1.7,
since it's a pure feature. Any other thoughts about what goes into 1.8
>>the C99 extensions...
> I don't have a copy of the spec,
Through ANSI, you can get it for $18, in PDF form:
If you go through the ISO, they want $284 for the same thing!
> C99 has long long
So the question is, does it have formalized printf format specs for that
type? And quoth ISO 9899: "ll...specifies that a following d, i, o, u,
x, or X conversion specifier applies to a long long int argument..."
So, the GNU library is doing it the standard way. For now, I'm content
to let the other platforms go hang if they can't cope with that format
> C++ hasn't officially added this to the standard yet, but this is working its
> way through the committee.
I'm betting that all C/C++ compilers that are now shipping a C99 library
for their C side make it available on the C++ side as well, even though
it isn't strictly correct.
> This was all work done at once
> while cycling through numerous compiles, so it all came out at once.
Yes, it's a lot of work to untangle them on either end. The question is
who should get stuck with the job. :)
> The guy who wrote the patches is in the best position to organize them though,
> so I agree.
Also, it helps the guy evaluating them to understand the point of each
conceptual change. Such things aren't always obvious to those who
didn't implement the patch.