- used HEAD instead of GET because we don't need to download the body as per this task description
- united some of the options, though it was probably a better idea to implement and commit as I go preserving option history in git...
- in real world scenario only options 2-3 are viable, because in such kind of program you usually want to keep stream of tasks decoupled from the logic that handles them
use make to build the project, the binary will be in ./build/tt
This test task has been turned into a simple cli program
option 1:
make && ./build/tt -urls=urls.txt -concurrency 1 -timeout 5
option 2/3/4 are united, use max to limit the number of OK responses while cancelling in-flight requests:
make && ./build/tt -urls=urls.txt -concurrency 3 -timeout 5 -max 2
you can find an example schema that could be used to store stats in ./postgres/schemas.sql