[
https://issues.apache.org/jira/browse/MATH-183?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12565203#action_12565203
]
Luc Maisonobe commented on MATH-183:
------------------------------------
I have looked again at the synchronisation issues.
The one in ResizableDoubleArray was straightforward and fixed.
The ones in SummaryStatistics are more difficult. They are due to the various
setters and getters for implementations (set*Impl and get*Impl) which are
synchronized, whereas other methods using these implementations are not
synchronized. In fact, there are more issues than the two appearing in the
report: if we fix these two, a bunch of other problems appear, which should
also be fixed. At the end, almost all methods end up being synchronized. This
seems not to be the objective of this class but rather the objective of the
SynchronizedSummaryStatistics.
So I would suggest moving the synchronization on the setters/getters from the
SummaryStatistics class to the SynchronizedSummaryStatistics. derived class.
What do you think ?
> Findbugs Report
> ---------------
>
> Key: MATH-183
> URL: https://issues.apache.org/jira/browse/MATH-183
> Project: Commons Math
> Issue Type: Bug
> Reporter: Sebb
> Attachments: MathFindbugs.csv
>
>
> The attachment (to follow) is a summary of a Findbugs run.
> It's possible that the use of == for comparing floats is intended; if so
> perhaps it should be commented.
> I think all the other bug reports are valid, though of course the ones
> relating to the exposure of internal implementation may be ignored
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
