Create Queue Library about simple-http-server HOT 4 CLOSED

If we are not going to give the queue implementation its own directory, should we put it in a utils directory, or should we just place it in the socket_tcp_manager folder for simplicity's sake?

Is the implementation just going to store a typeless pointer to the object, or should we also give it a certain size and have the queue structure actually contain the data?

from simple-http-server.

My thought is that queue, hashmap, task struct would all be in one utils folder.
This may also make the project more structured when we add the unit test code and add a run tests option to the build.

So I have seen queue structures in C implemented as a linked list. This would help to keep track of size and for later deletion. In this scenario, each node would just have a pointer to void * data.

from simple-http-server.

From an algorithms point of view, a singly linked list should be the best for the queue, as long as we just hold a pointer to the head and tail. Then insertion to the back and deletion from the front is O(1).

There is also a chance that a resizable array could utilize cache locality better, but that would require some benchmarking to see if that works out in our case.

from simple-http-server.

From an algorithms point of view, a singly linked list should be the best for the queue, as long as we just hold a pointer to the head and tail. Then insertion to the back and deletion from the front is O(1).

There is also a chance that a resizable array could utilize cache locality better, but that would require some benchmarking to see if that works out in our case.

Yes I think for our purposes the queue will be fine. We can always improve later

from simple-http-server.