[PD-dev] potential memory leak in g_graph.c

[Ivica Bukvic]

Many thanks for the clarification, Miller.

The challenge I am trying to figure out is a memory leak report that
Jonathan shared with me. See:

https://sourceforge.net/tracker/?func=detail&aid=3605235&group_id=55736&atid=478070

Basically deleting and creating the same object (in this case through
script as that is obviously quicker than doing it manually) keeps pd (or
pd-l2ork in this case) growing in terms of its memory footprint. I was
hoping that closing the patch would free that up but it doesn't...

Hence, my little adventure into figuring out why the said call. I am sure
there are others as even disabling that call doesn't help much in curbing
the growing memory footprint (at least over here).


On Mon, Feb 25, 2013 at 11:17 PM, Miller Puckette <msp at ucsd.edu> wrote:

> I don't remember why it's there (and perhaps it's no longer necessary) -
> but I suspect it had something to do with a subtle problem when deleting
> a GOP whose interior got mad wieh there was no rtext for the owning object.
>
> The rtext eventualy goes away when the window is closed - but perhaps it's
> a problem if there's a self-editing patch that creates and deletes
> 1000s of objects in an open window?
>
> cheers
> Miller
>
>
> On Mon, Feb 25, 2013 at 10:36:46PM -0500, Ivica Ico Bukvic wrote:
> > Does anyone know why does the following exist inside the
> > glist_delete function inside g_graph.c:
> >
> >             /* if we're a drawing command, erase all scalars now,
> > before deleting
> >             it; we'll redraw them once it's deleted below. */
> >         if (drawcommand)
> > canvas_redrawallfortemplate(template_findbyname(canvas_makebindsym(
> >                 glist_getcanvas(x)->gl_name)), 2);
> >         if (glist_isvisible(canvas))
> >             gobj_vis(y, x, 0);
> >   -->      if (x->gl_editor && (ob = pd_checkobject(&y->g_pd)))
> >   -->          rtext_new(x, ob);
> >         if (x->gl_list == y) {
> >
> > That rtext is never freed after that so unless I am missing
> > something I see no reason why we would want to do this. Any ideas?
> >
> > --
> > Ivica Ico Bukvic, D.M.A
> > Composition, Music Technology
> > Director, DISIS Interactive Sound & Intermedia Studio
> > Director, L2Ork Linux Laptop Orchestra
> > Head, ICAT IMPACT Studio
> > Virginia Tech
> > Department of Music
> > Blacksburg, VA 24061-0240
> > (540) 231-6139
> > (540) 231-5034 (fax)
> > disis.music.vt.edu
> > l2ork.music.vt.edu
> > ico.bukvic.net
> >
>
> > _______________________________________________
> > Pd-dev mailing list
> > Pd-dev at iem.at
> > http://lists.puredata.info/listinfo/pd-dev
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.puredata.info/pipermail/pd-dev/attachments/20130225/61947518/attachment-0001.htm>