As a user of the nio library when writing tests I might want to run an asynchronous function within a synchronous function and wait for the result. This function acts as a synchronous wrapper around the asynchronous function.
This is either an alternative to #369 or a complement. It lets me use the normal busted it function and write code like this:
it('Does something, function()
local tree = nio.tests.with_async_context(adapter.discover_positions, tempfile)
local result = tree:to_list()
assert.are.same(expected, result)
end)
Here the function adapter.discover_positions is asynchronous, so we wrap it in nio.tests.with_async_context to provide it with an asynchronous context. The resulting call is synchronous and can be used in tests. We can pass arguments to the wrapped function and we return its result to the caller.
Why not use nio.tests.it instad? nio.tests.it runs the entire tests in an asynchronous context, which messes up error reporting, at least when used with the real busted. And it turns all assertion failures into errors when they should be test failures instead.
rcarriga
commented
6 months ago
As a user of the nio library when writing tests I might want to run an asynchronous function within a synchronous function and wait for the result. This function acts as a synchronous wrapper around the asynchronous function.
This is either an alternative to #369 or a complement. It lets me use the normal busted
itfunction and write code like this:Here the function
adapter.discover_positionsis asynchronous, so we wrap it innio.tests.with_async_contextto provide it with an asynchronous context. The resulting call is synchronous and can be used in tests. We can pass arguments to the wrapped function and we return its result to the caller.Why not use
nio.tests.itinstad?nio.tests.itruns the entire tests in an asynchronous context, which messes up error reporting, at least when used with the real busted. And it turns all assertion failures into errors when they should be test failures instead.