Re: [Jts-topo-suite-user] PreparedGeometry thread safety

[Martin D. <mtn...@te...>]

Thanks for this report. As you suggest, it seems like the simplest 
approach is to simply synchronize the build() method.

Are you able to change this and try it out?  I will commit this to SVN 
in the meantime.

On 1/14/2013 12:16 PM, Shahak Nagiel wrote:
> I saw another issue in our logs today:
>
> java.lang.NullPointerException
> at 
> com.vividsolutions.jts.index.strtree.AbstractSTRtree.build(AbstractSTRtree.java:117)
> at 
> com.vividsolutions.jts.index.strtree.AbstractSTRtree.query(AbstractSTRtree.java:255)
> at com.vividsolutions.jts.index.strtree.STRtree.query(STRtree.java:218)
> at 
> com.vividsolutions.jts.noding.MCIndexSegmentSetMutualIntersector.intersectChains(MCIndexSegmentSetMutualIntersector.intersectChains.java:100)
> at 
> com.vividsolutions.jts.noding.MCIndexSegmentSetMutualIntersector.process(MCIndexSegmentSetMutualIntersector.intersectChains.java:89)
> at 
> com.vividsolutions.jts.noding.FastSegmentSetIntersectionFinder.intersects(FastSegmentSetIntersectionFinder.java:77)
> atcom.vividsolutions.jts.noding.FastSegmentSetIntersectionFinder.intersects(FastSegmentSetIntersectionFinder.java:72)
> atcom.vividsolutions.jts.geom.prep.PreparedPolygonIntersects.intersects(PreparedPolygonIntersects.java:101)
> atcom.vividsolutions.jts.geom.prep.PreparedPolygonIntersects.intersects(PreparedPolygonIntersects.java:65)
>
> Looking at the AbstractSTRtree.build() implementation, it looks like 
> the itemBoundables must have been null to trigger that exception. 
> Since build() can only be called once--but isn't guarded against such 
> in a multi-threaded fashion, it seems like the method can simply be 
> synchronized (or, optionally, some lower-level locking mechanism for 
> the built flag)?
>
> Thanks,
> Shahak
>
>
>
> ------------------------------------------------------------------------
> *From:* Martin Davis <mtn...@gm... <mailto:mtn...@gm...>>
> *To:* "Smiley, David W." <ds...@mi... <mailto:ds...@mi...>>
> *Cc:* Shahak Nagiel <sn...@ya... <mailto:sn...@ya...>>; 
> "jts...@li... 
> <mailto:jts...@li...>" 
> <jts...@li... 
> <mailto:jts...@li...>>
> *Sent:* Wednesday, December 26, 2012 10:00 PM
> *Subject:* Re: [Jts-topo-suite-user] PreparedGeometry thread safety
>
> I have committed a changeset to make PreparedGeometry fully 
> thread-safe (I hope).  They are essentially the same as David's patch, 
> but go a bit further in terms of simplifying and clarifying the code.
>
> Shahak, are you able to build from SVN and test these out?  I don't 
> have any other way of adequately testing whether these fixes work 
> (since the issues only seem to occur under very high concurrency, and 
> I don't have a 32-core machine handy to test with 8^)  Suggestions for 
> other ways of verification are welcome...
>
> Martin
>
> On Mon, Dec 24, 2012 at 6:51 AM, Smiley, David W. <ds...@mi... 
> <mailto:ds...@mi...>> wrote:
>
>     I see Martin responded with pretty much what I was going to say.
>      I was hoping certain problematic methods would get marked as
>     deprecated as I did in my patch but alas, that didn't make it in :-(
>
>
>
>
>
>
>
>
>
>
>
>
>
>
> ------------------------------------------------------------------------------
> Master Visual Studio, SharePoint, SQL, ASP.NET, C# 2012, HTML5, CSS,
> MVC, Windows 8 Apps, JavaScript and much more. Keep your skills current
> with LearnDevNow - 3,200 step-by-step video tutorials by Microsoft
> MVPs and experts. SALE $99.99 this month only -- learn more at:
> http://p.sf.net/sfu/learnmore_122412
>
>
> _______________________________________________
> Jts-topo-suite-user mailing list
> Jts...@li...
> https://lists.sourceforge.net/lists/listinfo/jts-topo-suite-user
>
>
> No virus found in this message.
> Checked by AVG - www.avg.com <http://www.avg.com>
> Version: 2013.0.2805 / Virus Database: 2637/6000 - Release Date: 12/31/12
> Internal Virus Database is out of date.
>