I pushed this earlier. I like what it does, and the test sold me. :)
Once -o works, I want to keep these things tweakable. On Mar 17, 10:35 am, Dustin <[email protected]> wrote: > On Mar 17, 4:57 am, Trond Norbye <[email protected]> wrote: > > > The issue: Memcached eats memory unnecesarily > > I downgraded the issue from medium to low (and from defect to RFE). For > > all details, see:http://code.google.com/p/memcached/issues/detail?id=14 > > > Proposed patch with test > > case:http://github.com/trondn/memcached/commit/83da85eca91c3dd9777f1479e7d... > > I had started something similar, but began by factoring out the LRU > search so that I could reuse the loop and make that function smaller. > I never got to the point where I *liked* what I was doing on the first > search, but I like what yours does. > > If nobody objects, I'd like to take this. > > There are a couple of things I'd like to do with it in the future, > though: > > 1) Make both of the depths configurable (e.g. 0 would disable a > search). > 2) Try to get these refactored into an isolated function. > > One nice thing about the depth configuration is that -M would just > need to set LRU search depth to 0.
