jtamate marked 9 inline comments as done. jtamate added a comment.
> If anyone attempts this, please name the struct and its members, don't use QPair ;-) > But no, that cannot possibly be faster. QVariant has lots of overhead itself. I've tried, just before reading your comment :-) Three tests: fill the structure, compare two structures and read 3 values. AnotherV2 (If someone finds a better name, it will be welcome). 0.00041 msecs per iteration (total: 55, iterations: 131072) 0.00022 msecs per iteration (total: 59, iterations: 262144) 0.00048 msecs per iteration (total: 64, iterations: 131072) QPair+QVariant: 0.00056 msecs per iteration (total: 74, iterations: 131072) 0.00020 msecs per iteration (total: 55, iterations: 262144) 0.00049 msecs per iteration (total: 65, iterations: 131072) INLINE COMMENTS > dfaure wrote in udsentry_benchmark.cpp:619 > This relies on insert being called in ascending "field" order, for > lower_bound to work. > But that is not necessarily the case in kioslaves, so you'd have to insert at > "index" here, instead of appending. Yes, it was badly done. Changed the fill order to detect this problems. REPOSITORY R241 KIO REVISION DETAIL https://phabricator.kde.org/D12659 To: jtamate, dfaure, #frameworks Cc: bruns, michaelh