On Mon, Mar 12, 2018 at 04:03:59PM +0100, Andrea Bolognani wrote:
On Mon, 2018-03-12 at 14:58 +0000, Daniel P. Berrangé wrote:
On Mon, Mar 12, 2018 at 03:53:09PM +0100, Andrea Bolognani wrote:
On Mon, 2018-03-12 at 14:46 +0000, Daniel P. Berrangé wrote:
The build system doesn't get refreshed automatically after this change, so it's still trying to generate virkeycodetable_rfb.h and failing because the current version of keymap-gen doesn't know about it.
Perhaps it would be better if I just made 'keymap-gen' continue to accept 'rfb' as a deprecated alias for 'qnum'
That would probably be nice regardless, but it wouldn't really help here - unless we decided to stick with the old name in libvirt too. Are you advocating for doing just that?
No, just take your patch as is. Just fix keymap-gen, so we make the (harmless) error message disappear for people with existing checkouts that have built the old filename.
But that still wouldn't work seamlessly, would it? The existing build system would try to generate virkeycodetable_rfb.h, and succeed, but the code won't compile because it's looking for virkeycodetable_qnum.h and related symbols now.
The resulting virkeycodetable_rfb.h would never be used - it would only be generated because the automagic .deps files automake creates still list the old filename. In fact on that basis I think we can just ignore the failure - it is self-correcting - automake triggers re-generation of the old filename, and that's non-fatal because we actally use the new filename. So just push your current patch as is Reviewed-by: Daniel P. Berrangé <berrange@redhat.com> 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 :|