Re: [boost] boost.log's text_multifile_backend is not performant on windows - perhaps it should be documented?

A quick correction. I noticed my performance numbers were not super reliable - my log output directory was on my c drive, which is a bad idea as OS reads/writes will have a greater impact on numbers. I moved the output to a separate, faster drive. There with the cache keyed on path, throughput was 94K, while with cache keyed on the string representation of the path, throughput was 110K. Tweaking further, I could key the cache just on channel name (based on my logging output structure), and that got me to 120K. So while still a good improvement, not quite as dramatic as I first thought.