As part of #4 we switched to using a single mutex protecting the resolver caches for both reading and writing. While a mutex is pretty cheap compared to a network call we should be able to separate out the deleting from the reading so that we can restore read-write mutex usage.
@vyzo's recommendation for this #4 (comment)
As part of #4 we switched to using a single mutex protecting the resolver caches for both reading and writing. While a mutex is pretty cheap compared to a network call we should be able to separate out the deleting from the reading so that we can restore read-write mutex usage.
@vyzo's recommendation for this #4 (comment)