[L2Ork-dev] Delayed pd_unbind

[Albert Graef]

Hi Ico,

On Thu, Nov 14, 2013 at 3:04 AM, Ivica Ico Bukvic <ico at vt.edu> wrote:
> So, yes, delayed unbind is critical in ensuring that the example I provided
> earlier does not cause a segfault (the same holds true for all iemgui
> objects, not just the external provided in the example).

Thanks for the example. Yes, vanilla pd segfaults on that one, as does
pd-l2ork with my suggested patch. So the solution I proposed to get
rid of the memory leaks doesn't work. But still the other parts of my
patch remain valid. My latest version of the patch is attached below.

Note that without any of my patches, the latest pd-l2ork still
segfaults on the boom example if you click `boom` *twice*. So it
appears that your solution doesn't completely solve the problems with
the example either. However, after applying the latest version of my
patch I can click `boom` as often as I want without making pd-l2ork
segfault. The iem_r-help.pd help patch in the distribution also
continues to work as advertised. Can you confirm this?

So I'd say that my latest version of the patch does improve the
situation. It also solves the segfaults I had in pd-faust. The memory
leaks I described previously are still there, but they are not any
worse than before. ;-)

I still have some ideas for a better solution that would also get rid
of the memory leaks, but that will take some time to implement. So I'd
be happy if you could commit the patch as it stands now, so that we
get rid of the segfaults in pd-faust and the boom example.

Albert

-- 
Dr. Albert Gr"af
Dept. of Music-Informatics, University of Mainz, Germany
Email:  aggraef at gmail.com
WWW:    https://plus.google.com/+AlbertGraef
-------------- next part --------------
A non-text attachment was scrubbed...
Name: m_pd.patch
Type: application/x-octet-stream
Size: 3453 bytes
Desc: not available
URL: <http://disis.music.vt.edu/pipermail/l2ork-dev/attachments/20131114/868740fa/attachment-0001.bin>