* James Kanze:
> On May 13, 9:18 am, "Alf P. Steinbach" <al...@[EMAIL PROTECTED]
> wrote:
>> * Pawel_Iks:
>
> Just a couple of nits (and the correction of a typo), but...
>
>> <code>
>> class A
>> {
>> private:
>> size_t myN;
>> std::vector<int> myElements;
>>
>> size_t indexFor( size_t i, size_t j ) const
>> {
>> return myN*i + j;
>> }
>
>> public:
>> A( size_t n ): myN( n ), myElements( n*n ) {}
>
>> int at( size_t i, size_t j ) const
>> {
>> return myElements.at( indexFor( i + j ) );
>
> You certainly meant "indexFor( i, j )"
Yes.
>. As written, it won't
> compile (and wouldn't do the right thing if it did).
>
> Also, I very much question the wisdom of using at() here.
That's the design of std::vector::at.
It may be unwise (or not!) that std::vector::at doesn't terminate the
program instead of throwing, but what it does that is absolutely not
unwise is to guarantee to /fail/ on an invalid index value, which
guarantee operator[] doesn't offer -- all operator[] has is UB.
Especially for a newbie it's therefore a good idea to use at(), and a
bad idea to use operator[], unless one knows that one's standard
library implementation guarantees failure also for operator[] -- but
in general one won't know that, and in particular, a newbie won't.
> If
> there's an error in the indexing, you want an assertion failure,
> not an exception. Which is what you'll get with any reasonable
> implementation of std::vector<>::operator[]. But of course,
> even that is only partially right, since something like 20, 0
> will pass even if myN is 10. You really need to use assert in
> indexFor, i.e.:
>
> size_t indexFor( size_t i, size_t j ) const
> {
> assert( i < myN && j < myN ) ;
> return myN*i + j;
> }
Ideally, yes, agreed.
I congratulate myself on still being able to write off-the-cuff code
that is so easily modified, so maintainable. ;-)
Cheers,
- Alf
--
A: Because it messes up the order in which people normally read text.
Q: Why is it such a bad thing?
A: Top-posting.
Q: What is the most annoying thing on usenet and in e-mail?
|
