Reason to use pragma once and inclusion guarding marco together? about diffusion-maps HOT 2 CLOSED

I guess using #pragma once and #include guards together is from a habit I picked up several years ago. I learnt somewhere that using #pragma once could decrease compilation time, so I started to use it (although the gain should be marginal for a project of this scale). At the same time, #pragma once is non-standard, so I use a regular #include guard as a backup. I guess I'll stick with regular #include guards in the future.

The Python API will pretty much be what's stated in the proposal with some minor changes. Pytest will be used to test the whole algorithm on simple datasets to see if the result makes sense – or, at the very least, doesn't return an array of NaNs. They should be pushed to the repository within a few days.

from diffusion-maps.

I guess using #pragma once and #include guards together is from a habit I picked up several years ago. I learnt somewhere that using #pragma once could decrease compilation time, so I started to use it (although the gain should be marginal for a project of this scale). At the same time, #pragma once is non-standard, so I use a regular #include guard as a backup. I guess I'll stick with regular #include guards in the future.

Either is OK. Having said that, I have not encountered a modern compiler that does not make #pragma once work perfectly for a while.

The Python API will pretty much be what's stated in the proposal with some minor changes. Pytest will be used to test the whole algorithm on simple datasets to see if the result makes sense – or, at the very least, doesn't return an array of NaNs. They should be pushed to the repository within a few days.

I look forward to it and will come back to take a look when I have time. Feel free to tag me if you want me to see anything.

from diffusion-maps.