On Fri, Jun 13, 2014 at 04:49:18PM +0800, Hu Tao wrote: > On Fri, Jun 13, 2014 at 10:18:44AM +0200, Paolo Bonzini wrote: > > Il 13/06/2014 10:03, Michael S. Tsirkin ha scritto: > > >>>> > Two points I would like to see addressed before it's going upstream: > > >>>> > - please validate and test the pci branch > > >>>> > there have been lots of changes > > >>>> > - please work to drop the SignedRange: for all > > >>>> > current purposes existing Range is sufficient, > > >>>> > instead of a list and sort re-implementation, it is > > >>>> > better to use g_list_insert_sorted > > >>> > > >>> I see. Thanks for your work! > > >Could you confirm you'll be able to fix these by Monday? > > >I'd like to send the pull request then. > > > > I think SignedRange can be dealt with after the merge. > > > > Paolo > > I agree. Using existing Range makes string output visitor don't support > negative values which it does support before SignedRange.
Are there any users though? I think it's more a bug than a feature. Also we do want to get rid of rewriting g_list_insert_sorted though, that's for sure. > Besides SignedRange, tests show that there are several problems with the > pci tree(core dumps, unexpecting exit of qemu). > > Hu OK pls share this info, also pls bisect if possible. Thanks! -- MST
