hadoop-common-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Colin P. McCabe" <cmcc...@apache.org>
Subject Re: timsort bug in the JDK
Date Thu, 05 Mar 2015 02:36:36 GMT
Tsuyoshi Ozawa sent out an email to the common-dev list about this
recently.  It seems like the bug only bites when the number of
elements is larger than 67108864, which may limit its impact (to state
it mildly).  Also, the flawed sorting algorithm is not used on arrays
of primitives, just on arrays of Objects.  We should probably file a
JIRA to track this, though, just in case there is an impact.  And
maybe look at some of the uses of sort() in the code.

best,
Colin


On Tue, Mar 3, 2015 at 8:56 AM, Steve Loughran <stevel@hortonworks.com> wrote:
> One other late-breaking issue may we "what to do about the fact that Java 7 & 8 have
a broken sort algorithm?, which has surfaced recently<http://envisage-project.eu/proving-android-java-and-python-sorting-algorithm-is-broken-and-how-to-fix-it/>
>
> I believe some other OSS projects have tried to address this.
>
> Looking at LUCENE–6293, they weren’t clear whether it was worth the effort for a
problem that didn’t corrupt their data. I’m fairly tempted to argue the same for doing
something for 2.7, especially as a switch throughout the code base could be expensive. Except:
what if Oracle don’t ship a patch for JDK7?
>
> -Steve

Mime
View raw message