[R-sig-Geo] rgeos 0.2-1 released

pgalpern pgalpern at gmail.com
Tue Jan 17 16:18:06 CET 2012


Roger,
   Thanks to all for efforts to improve this very useful package.

   I am going to cautiously suggest that the memory leak with 
raster::rasterToPolygons(x, dissolve=TRUE) which calls 
rgeos::gUnaryUnion() is  fixed -- or has improved considerably.

Previously,  running this function on a raster of 800 000 cells (which, 
as you pointed out in another thread, is at the best of times a memory 
intensive task) quickly consumed all the available overhead memory and 
headed into disk swapping.  Now, I can run the same raster and memory is 
consumed up to what appears to be the R gc() trigger point, and then 
stabilizes (suggesting that R can now release the memory, which it 
previously could not).  I should note, however, that 7 GB was freed only 
after the R instance was closed.  Calls to gc() did not release this.  
This may be expected behaviour, not sure.

(Run under 64bit R 2.14.1).

Thanks!
Paul

-- 
Paul Galpern, PhD Candidate
Natural Resources Institute
70 Dysart Road
University of Manitoba
Winnipeg, Manitoba, Canada R3T 2M6
http://borealscape.ca



> Date: Mon, 16 Jan 2012 14:26:34 +0100 (CET)
> From: Roger Bivand<Roger.Bivand at nhh.no>
> To:r-sig-geo at r-project.org
> Subject: [R-sig-Geo] rgeos 0.2-1 released
> Message-ID:<alpine.LRH.2.02.1201161400310.21470 at reclus.nhh.no>
> Content-Type: TEXT/PLAIN; format=flowed; charset=US-ASCII
>
> Edzer Pebesma and Colin Rundel have made progress with finding and
> removing memory leaks on rgeos, so we've released 0.2-1. It would be
> useful to receive reports on changes in usability. The OSX binary for
> 0.2-1 is already on CRAN Extras, the Windows binary is diffusing across
> CRAN mirrors.
>
> Roger



More information about the R-sig-Geo mailing list