On Thu, Jan 03, 2013 at 04:12:54PM -0200, Leandro Dorileo wrote:
> Hi,
>
> On Thu, Jan 03, 2013 at 05:14:12PM +0900, Carsten Haitzler wrote:
> > On Tue, 27 Nov 2012 11:16:45 +0900 Cedric BAIL <cedric.b...@free.fr> said:
> >
> > > Hi,
> > >
> > > On Mon, Nov 26, 2012 at 11:35 PM, Leandro Dorileo
> > > <dori...@profusion.mobi> wrote:
> > > > Hi Cedric,
> > > >
> > > > On Mon, Nov 26, 2012 at 05:23:15PM +0900, Cedric BAIL wrote:
> > > >> Yop,
> > > >>
> > > >> On Wed, Nov 21, 2012 at 2:42 AM, Leandro Dorileo
> > > >> <dori...@profusion.mobi>
> > > >> wrote:
> > > >> > What happens if a group has 2 name statements("double named")?
> > > >> > edje will fail to free the groups cache while shuting down due
> > > >> > the collection directory entry hash having two entries with
> > > >> > different names but pointing to the same object - the second try
> > > >> > to free the part mem pool will fail - since it was freed before -
> > > >> > and issue an abort.
> > > >> >
> > > >> > Attached goes a patch to fix the issue.
> > > >>
> > > >> Just got back from vacation and reading your patch. Wouldn't
> > > >> current_de->entry be set to something != NULL, before the parse_str(0)
> > > >> call ? If that's the case, it would be a way better patch in my
> > > >> opinion.
> > > >
> > > > I'm not sure if I correctly understand your suggestion but
> > > > current_de->entry *is* something != NULL. What the patch does is try to
> > > > find if current_de is already in in the edje_file->collection.
> > >
> > > My understanding is that if the group is already in the hash,
> > > current_de->entry should already be defined to something before the
> > > parse_str(0) call. So instead of walking all the hash (that may be
> > > slow), we could just check at the entry point of the function if
> > > current_de->entry != NULL. I am just not sure how this work with
> > > inheritance. So that's why you get a question there :-)
> > >
> > > > What currently happens is that the same current_de is been added twice
> > > > case
> > > > a group is "double named" - two name statements for the same group. This
> > > > behavior doesn't look correct to me.
> > >
> > > I agree on that.
> >
> > and the conclusion to this saga is????
>
>
> I just got sometime to review the issue now, cedric was right and I came up
> with
> a simpler solution without the hash iteration I proposed initially, I'm
> commiting it
> as soon as possible.
in rev 82093.
--
Leandro Dorileo
ProFUSION embedded systems
http://profusion.mobi
------------------------------------------------------------------------------
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. ON SALE this month only -- learn more at:
http://p.sf.net/sfu/learnmore_122712
_______________________________________________
enlightenment-devel mailing list
enlightenment-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/enlightenment-devel
