On Mon, Feb 22, 2021 at 06:42:00PM +0100, Paolo Bonzini wrote: > On 22/02/21 15:57, Markus Armbruster wrote: > > * The block layer's pseudo-protocol "json:" (which can get embedded in > > image headers) > > If it gets embedded in image headers, I don't think we'll be able to > deprecate it ever. We'd need to keep a converter for old images, at which > point it's simpler to keep the extensions. Even if we can only use a standard JSON parser for QMP + QGA, that's a already a significant net win long term IMHO. Both of those are security critical components and also areas where we might want different language impls as we increasingly use a multi-process model, so avoiding use of extensins is good. Even for the block layer, we don't neccessarily need to keep compat at runtime. It could be sufficient to have an extended deprecation period and then provide an offline helper script to re-write the qcow2 backing store field to use " instead of ' ....assuming we actually get real world pushback from people who really have used ' - we don't know if there are such people yet. We can do deprecation in a multi stage process - deprecation for everything, then block it for QMP + QGA after 2 cycles, while still allowing it for qcow2, and eventually block for qcow2 3 years later or something like that. IOW, I wouldn't give up on trying to deprecate it. Regards, Daniel -- |: https://berrange.com -o- https://www.flickr.com/photos/dberrange :| |: https://libvirt.org -o- https://fstop138.berrange.com :| |: https://entangle-photo.org -o- https://www.instagram.com/dberrange :|
