Hi. FTR: I don't agree with such changes when they obfuscate the code for no measurable gain (until proven otherwise[1]). If the class is used properly, the "manual" copying is a minute part of usage duration; if not used properly (e.g. calling "create()" instead of "withCache()" on an existing instance), the "saved" copying time will be marginal wrt the lost time due to repeating the other (likely much lengthier) computations.
Regards, Gilles [1] Through JMH benchmarks _on the use-case at hand_. Le dim. 20 juin 2021 à 09:35, GitBox <g...@apache.org> a écrit : > > > arturobernalg commented on a change in pull request #99: > URL: https://github.com/apache/commons-numbers/pull/99#discussion_r654891918 > > > > ########## > File path: > commons-numbers-combinatorics/src/main/java/org/apache/commons/numbers/combinatorics/LogFactorial.java > ########## > @@ -62,9 +62,7 @@ private LogFactorial(int numValues, > } > > // Copy available values. > - for (int i = beginCopy; i < endCopy; i++) { > - logFactorials[i] = cache[i]; > - } > + if (endCopy - 2 >= 0) System.arraycopy(cache, beginCopy, > logFactorials, beginCopy, endCopy - 2); > > Review comment: > True. Changed. TY > > ########## > File path: > commons-numbers-combinatorics/src/main/java/org/apache/commons/numbers/combinatorics/FactorialDouble.java > ########## > @@ -64,9 +64,7 @@ private FactorialDouble(int numValues, > } > > // Copy available values. > - for (int i = beginCopy; i < endCopy; i++) { > - factorialsDouble[i] = cache[i]; > - } > + if (endCopy - 2 >= 0) System.arraycopy(cache, beginCopy, > factorialsDouble, beginCopy, endCopy - 2); > > Review comment: > True. Changed. TY > --------------------------------------------------------------------- To unsubscribe, e-mail: dev-unsubscr...@commons.apache.org For additional commands, e-mail: dev-h...@commons.apache.org
