Read Bloch. Don’t optimize until you have proven that this bit of code is causing a significant performance hit. arrayCopy can and is inlined by jit.
Chas > On Feb 16, 2018, at 5:53 AM, Stian Soiland-Reyes <st...@apache.org> wrote: > > String is still a type of Object (requiring GC handling of reference > counters), as you can do Object[] o = stringArray.clone(); without > casting. (other way not so) > > https://www.javaspecialists.eu/archive/Issue124.html says there is > hardly any difference, so I don't think we need to fight this one out > and can just go with the cleanest code :) > >> On 16 February 2018 at 11:22, sebb <seb...@gmail.com> wrote: >>> On 16 February 2018 at 10:01, Stian Soiland-Reyes <st...@apache.org> wrote: >>> I agree in general for .clone() on objects - and I'm trying to move >>> away from using .clone() in that Commons RDF fluent mutable/immutable >>> builder (See COMMONSRDF-49). >>> >>> But this is an Object[] - a native type. >> >> Huh? >> >> The code says String[], which is native as it's final. >> >> Object[] may not be native >> >>> I must admit I am not sure what is really the preferred way - I >>> thought .clone() is best as the VM can copy an array natively however >>> it wants, while with Arrays.copyOf it might have to check if length is >>> the same before doing anything clever. >> >> In the specific case of String[] it might be OK to use clone, but in >> general it's better to use the proper JVM methods and not try to >> second-guess how efficient they are. >> >>>> On 13 February 2018 at 18:58, Gilles <gil...@harfang.homelinux.org> wrote: >>>>> On Tue, 13 Feb 2018 18:40:13 +0000, sebb wrote: >>>>> >>>>>> On 13 February 2018 at 09:31, Gilles <gil...@harfang.homelinux.org> >>>>>> wrote: >>>>>> >>>>>>> On Tue, 13 Feb 2018 00:16:13 +0000 (UTC), st...@apache.org wrote: >>>>>>> >>>>>>> >>>>>>> Repository: commons-csv >>>>>>> Updated Branches: >>>>>>> refs/heads/CSV-216 637ad2d7a -> f66a83901 >>>>>>> >>>>>>> >>>>>>> CSV-216: Avoid Arrays.copyOf() >>>>>> >>>>>> >>>>>> >>>>>> Why? >>>>> >>>>> >>>>> Agreed >>>>> >>>>>>> as .clone() will do >>>>>> >>>>>> >>>>>> >>>>>> We should rather avoid using "clone()". >>>>> >>>>> >>>>> Again: why? >>>> >>>> >>>> There are many discussions about this topic on the web.[1] >>>> My point is that using "clone()" is not good advice. >>>> >>>> Gilles >>>> >>>> [1] E.g. http://vojtechruzicka.com/java-cloning-problems/ >>>> >>>> >>>>> It's not clear why Arrays.copyOf(0 is considered bad, nor why clone() >>>>> is considered bad. >>>>> >>>>> >>>>>> Gilles >>>>>> >>>>>> >>>>>>> -- at least until someone tries to do >>>>>>> .withValue(x) in an out-of-range column >>>>>>> >>>>>>> >>>>>>> Project: http://git-wip-us.apache.org/repos/asf/commons-csv/repo >>>>>>> Commit: >>>>>>> http://git-wip-us.apache.org/repos/asf/commons-csv/commit/f66a8390 >>>>>>> Tree: http://git-wip-us.apache.org/repos/asf/commons-csv/tree/f66a8390 >>>>>>> Diff: http://git-wip-us.apache.org/repos/asf/commons-csv/diff/f66a8390 >>>>>>> >>>>>>> Branch: refs/heads/CSV-216 >>>>>>> Commit: f66a83901bd026369a2e8d522bd567eb2ef3f8c0 >>>>>>> Parents: 637ad2d >>>>>>> Author: Stian Soiland-Reyes <st...@apache.org> >>>>>>> Authored: Fri Feb 9 16:49:51 2018 +0000 >>>>>>> Committer: Stian Soiland-Reyes <st...@apache.org> >>>>>>> Committed: Tue Feb 13 00:14:52 2018 +0000 >>>>>>> >>>>>>> >>>>>>> >>>>>>> ---------------------------------------------------------------------- >>>>>>> src/main/java/org/apache/commons/csv/CSVRecord.java | 4 ++-- >>>>>>> 1 file changed, 2 insertions(+), 2 deletions(-) >>>>>>> >>>>>>> >>>>>>> ---------------------------------------------------------------------- >>>>>>> >>>>>>> >>>>>>> >>>>>>> >>>>>>> >>>>>>> >>>>>>> http://git-wip-us.apache.org/repos/asf/commons-csv/blob/f66a8390/src/main/java/org/apache/commons/csv/CSVRecord.java >>>>>>> >>>>>>> >>>>>>> ---------------------------------------------------------------------- >>>>>>> diff --git a/src/main/java/org/apache/commons/csv/CSVRecord.java >>>>>>> b/src/main/java/org/apache/commons/csv/CSVRecord.java >>>>>>> index 979119f..2be5c49 100644 >>>>>>> --- a/src/main/java/org/apache/commons/csv/CSVRecord.java >>>>>>> +++ b/src/main/java/org/apache/commons/csv/CSVRecord.java >>>>>>> @@ -199,7 +199,7 @@ public class CSVRecord implements Serializable, >>>>>>> Iterable<String> { >>>>>>> public final CSVRecord immutable() { >>>>>>> if (isMutable()) { >>>>>>> // Subclass is probably CSVMutableRecord, freeze values >>>>>>> - String[] frozenValue = Arrays.copyOf(values, >>>>>>> values.length); >>>>>>> + String[] frozenValue = values.clone(); >>>>>>> return new CSVRecord(frozenValue, mapping, comment, >>>>>>> recordNumber, characterPosition); >>>>>>> } else { >>>>>>> return this; >>>>>>> @@ -260,7 +260,7 @@ public class CSVRecord implements Serializable, >>>>>>> Iterable<String> { >>>>>>> if (isMutable()) { >>>>>>> return this; >>>>>>> } >>>>>>> - String[] newValues = Arrays.copyOf(values, >>>>>>> values.length); >>>>>>> + String[] newValues = values.clone(); >>>>>>> return new CSVMutableRecord(newValues, mapping, comment, >>>>>>> recordNumber, characterPosition); >>>>>>> } >>>> >>>> >>>> >>>> --------------------------------------------------------------------- >>>> To unsubscribe, e-mail: dev-unsubscr...@commons.apache.org >>>> For additional commands, e-mail: dev-h...@commons.apache.org >>>> >>> >>> >>> >>> -- >>> Stian Soiland-Reyes >>> http://orcid.org/0000-0001-9842-9718 >>> >>> --------------------------------------------------------------------- >>> To unsubscribe, e-mail: dev-unsubscr...@commons.apache.org >>> For additional commands, e-mail: dev-h...@commons.apache.org >>> >> >> --------------------------------------------------------------------- >> To unsubscribe, e-mail: dev-unsubscr...@commons.apache.org >> For additional commands, e-mail: dev-h...@commons.apache.org >> > > > > -- > Stian Soiland-Reyes > http://orcid.org/0000-0001-9842-9718 > > --------------------------------------------------------------------- > To unsubscribe, e-mail: dev-unsubscr...@commons.apache.org > For additional commands, e-mail: dev-h...@commons.apache.org > > --------------------------------------------------------------------- To unsubscribe, e-mail: dev-unsubscr...@commons.apache.org For additional commands, e-mail: dev-h...@commons.apache.org
