Am 19.06.2012 11:20, schrieb Daniel P. Berrange: > On Tue, Jun 19, 2012 at 09:39:34AM +0200, Kevin Wolf wrote: >> Am 18.06.2012 20:31, schrieb Anthony Liguori: >>> Irrespective of anything else, I think it's safe to say the experiment of >>> "rich >>> errors" has been a failure. We still have way too many places using >>> error_report. >>> >>> As I mentioned in another thread, I think we should: >>> >>> 1) Introduce a GENERIC_ERROR QError type. It could have a 'domain' and a >>> 'msg' >>> field. >>> >>> 2) Focus on converting users of error_report over to use propagated Error >>> objects. >>> >>> We shouldn't/can't change existing QError users. We also shouldn't >>> consider >>> changing the wire protocol. But for new error users, we should/can relax >>> the >>> reported errors. >>> >>> We need a clear support policy on whether the contents of 'msg' are stable >>> or >>> not too. >> >> Another point that you used to bring up in earlier discussions is >> translated error messages. If we start returning error messages that are >> meant to displayed to the user, should we get your gettext patches >> applied which you did for the GTK backend? libvirt would then have to >> pay attention to start qemu with the same locale as the client has. > > You can't really start the VM in the same locale as the client app, > because there's no persistent 1:N relationship between libvirt clients > and VMs - it is M:N, so you can't choose a single VM. In addition there > is a bunch of work that libvirt does against VMs in contexts that have > no associated client. You just have to have 1 system wide locale for > all QEMU VMs on a host and libvirt.
Good point. So if we ever needed it, we would have to introduce a monitor command to switch. But in most cases client and server locale should be the same anyway, so I think we can ignore that part for the start. Kevin
