[GEM-dev] [pd-gem:bugs] #209 pix_image magick leaks memory

"IOhannes m zmölnig" via GEM-dev gem-dev at lists.iem.at
Wed Jul 9 10:10:50 CEST 2014


indeed, there is a memleak in the MagickCore backend (using the C-api).

there is *no* memleak in the Magick++ backend (using the C++-api).

both backends can be used to provide the gem_imageMAGICK.so plugin (Magick++ is preferred automatically)

i'll have a look into it


---

** [bugs:#209] pix_image magick leaks memory**

**Status:** open
**OS:** any
**Release:** 0.93-git
**Created:** Tue Apr 02, 2013 03:55 PM UTC by Claude Heiland-Allen
**Last Updated:** Wed Jul 09, 2014 07:42 AM UTC
**Owner:** IOhannes m zmölnig

I have a patch that loads two new images every frame with two pix_image]--[pix_texture].  Only the latest two images are needed for each frame.  I'm also using pix_image "thread 0" so that the images are loaded before the current frame is displayed, and pix_texture "rectangle 0, quality 2".

After some time Pd is using all my memory (I have 8GB ram and no swap) and the system becomes unresponsive.

Using valgrind I see there is a severe leak, probably in Magick loader (log attached).  Same problem whether I use png or tif.

Versions:

Pd version 0.42-6
compiled 08:43:13 Mar 19 2012

GEM: ver: 0.93.git 867c5a7
GEM: compiled: Mar 26 2013

libmagickcore5:
  Installed: 8:6.7.7.10-5
  Candidate: 8:6.7.7.10-5
  Version table:
 *** 8:6.7.7.10-5 0
        500 http://ftp.uk.debian.org/debian/ wheezy/main amd64 Packages
        100 /var/lib/dpkg/status



---

Sent from sourceforge.net because gem-dev at lists.iem.at is subscribed to https://sourceforge.net/p/pd-gem/bugs/

To unsubscribe from further messages, a project admin can change settings at https://sourceforge.net/p/pd-gem/admin/bugs/options.  Or, if this is a mailing list, you can unsubscribe from the mailing list.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.puredata.info/pipermail/gem-dev/attachments/20140709/c8f23e81/attachment.html>


More information about the GEM-dev mailing list