On 14/12/2010, at 3:38 AM, Daniel P. Berrange wrote: > This raises the question of how the tools depending on libvirt > will be distributed. eg virt-viewer. If they're separate, then > we need to make sure the tools continue working, after any update > of libvirt without changing paths manually. > > IMHO, rather than just distributing a libvirt installer, we should > distribute a 'virt tools' installer on virt-tools.org which has > a complete bundle of everything that has been ported to Win32, > libvirt core, language bindings and apps. The installer would of > course allow you to pick & choose which bits to install. Yeah, that's not a bad idea too. One of the interesting things that turned up from this exercise so far, is that having a shortcut to virsh.exe in the windows install menu doesn't work (tried it). Virsh (on windows) needs to be given a connection string in order to do anything. Without it, it just displays an error message then exits. Using a shortcut at the moment means the user sees a black dos prompt screen show up for a fraction of a second (not long enough to read the text) then disappear. Not real practical, so disabled the shortcut creation. Thinking what we'd probably want to do for virsh is have a dialog or something that lets the user assemble a connection string, then launches virsh using that. Someone probably has a better idea than that though. :) -- libvir-list mailing list libvir-list@xxxxxxxxxx https://www.redhat.com/mailman/listinfo/libvir-list