Less durable queue

[jakepearson]

Would it be possible to create a version of the queue that would not be durable for testing purposes? Maybe you could pass an optional parameter to set it to in memory on mode.

[ztellman]

What are you trying to test? You could simply set the :fsync-intervalto be arbitrarily high, but it will still fsync when swapping out slab files.

[jakepearson]

The idea was to isolate queues between tests, I would like to be sure that one test's queue state cannot effect another. I can use a uuid for a queue name for each test, but over time I would end up with a large directory. If I can keep the queue in memory then I won't need to make sure to use a uuid or cleanup a named queue between test runs.

[ztellman]

I can provide a delete-queues function, if that helps, but I suggest actually exercising the code in your tests is a good idea. Skipping over writing to disk would also skip over serializing the tasks, which is an important thing to validate.

[jakepearson]

That sounds great. Thanks.

[vedang]

This issue can be safely closed at the delete! functionality has been added to the code via 37e31d3 9 years ago.