On Thu, Jan 26, 2012 at 05:05:15PM +0200, Zeeshan Ali (Khattak) wrote: > On Thu, Jan 26, 2012 at 11:19 AM, Christophe Fergeau > <cferg...@redhat.com> wrote: > > On Thu, Jan 26, 2012 at 08:18:38AM +0100, Philipp Hahn wrote: > >> Your changed version only has the same behaviour, if the user-passed-in > >> function iter_func() never changes it->next, which you can't guarentee > >> here. > >> You need to keep the "next" copy. > > > > Yes, the for loop was changed to a while loop recently exactly for that > > purpose: > > I don't get it, your commit clearly introduced a hang in Boxes and my > change fixes it back while also simplifying the code slightly. So I > don't at all buy the 'while' being more 'reliable'.
The callback gets passed 'it', and the callback used in gvir_config_object_delete_child can unlink/free 'it', so things won't work as expected if you try to get it->next after calling the callback. Using a while loop and getting it->next before calling the callback avoids this problem. Christophe
pgpHWeV6BTMn2.pgp
Description: PGP signature
-- libvir-list mailing list libvir-list@redhat.com https://www.redhat.com/mailman/listinfo/libvir-list
