Examples can (and should) be exemplary C++

[xtofl]

Obviously, the primary goal of the examples is to show how one uses the library.

However, they are also adding to the overall C++ knowledge: people using thrust are very likely to also learn the standard library algorithms, and take over the style presented (if not even copy-paste-adapt the examples).

So there would be value in making the example code exemplary, 'modern' C++.

Some instances of example code that could be modernized:

    thrust::device_vector<int> stencil(8);
    stencil[0] = 0;
    stencil[1] = 1;
...

could make use of std::initializer_list<int> to direct initialize the vector.

std::cout << "found " << (indices_end - indices.begin()) ...

Should really use std::distance(begin(indices), indices_end)

// reduce a pair of bounding boxes (a,b) to a bounding box containing a and b
struct bbox_reduction : public thrust::binary_function<bbox,bbox,bbox>

Is a poor choice of name: it tells you what function it's going to be used in (reduce), not what it does (bounding box of bounding boxes); it may be replaced by a lambda expression.

...

I think it would be a good idea to set up an 'example-wide' code review and try to exemplify that code.

(cf. also the very limited pull request NVIDIA/thrust#753)

[vrajvaghela89]

"Hi, I'd like to work on this issue. Could you assign it to me?"

[charan-003]

Hi, I'd like to work on this issue. Could you please assign it to me? I hope no one else is currently working on it. Looking at the thrust examples, I assume the changes I need to make include:

• Using std::initializer_list to directly initialize vectors (e.g., thrust::device_vector<int> stencil = {0, 1, ...};).
• Replacing manual iterator subtraction with std::distance (e.g., std::distance(indices.begin(), indices_end)).
• Replacing named functors like bbox_reduction with lambda expressions for clarity.
  and further continue with other files.
  Please let me know if I’m on the right track or if there’s anything else to consider!

[charan-003]

@bernhardmgruber @xtofl ?

[bernhardmgruber]

@charan-003 I removed the assignment of this issue to @vrajvaghela89. We are happy to receive a PR with improvements. There is no need to be assigned to this issue.

Please create small PRs with limited scope, like one PR will just improve all the use of thrust vectors etc. Another PR will move to more cuda::std algorithms etc. This way, we avoid long review times. We are looking forward to your suggestions!

[charan-003]

@bernhardmgruber thank you for the reply!!
#5670

i have created a pr :))