On Saturday, 1 October 2016 12:31:41 UTC-7, Brian Foley wrote: > > On Sat, Oct 01, 2016 at 06:43:46PM +0000, Robin H. Johnson wrote: > What sort of environment are you building in? Would you mind sending me a > copy of the pylint output to look at? (privately if you like.) > Started a new thread for that.
> > It doesn't need a specific PCI slot, just not to conflict with other > > stuff. > Thanks. NB this is a patch we wouldn't be able to apply to 2.15 anyway, > because > it introduces an RPC change, and for stable releases we want to guarantee > that > running different point releases on different nodes in a cluster won't > cause > compatibility problems. Otherwise rolling upgrades or downgrades > potentially > get a lot more difficult. > Still reasonable to get it into 2.16+2.17? For 2.15, where I need it now, I'm just going to ship it as an optional off-by-default patch for Gentoo. > ACK, but the qga_info string doesn't make a huge amount of sense on its > own. > To understand what it means you need to find where it's used, and at that > point > since it's only used in the one place, you may as well just inline it. > Will change in the next rev. > kvm_cmd.extend([ > "-chardev", "socket,path=%s,server,nowait,id=qga0" % qga_path, > "-device", "virtio-serial,id=qga0,bus=pci.0,addr=%#x" % qga_addr, > "-device", > "virtserialport,chardev=qga0,name=org.qemu.guest_agent.0", > ]) > That last line is a min of 72 characters, before extra indenting, which then pushes it over the 80-char limit. > >
