We already have lots of that image format logic in the glance/common/format_inspector.py which is utilized in the location.py (and would have covered these should we have implemented the staging as location).
That already limits our VMDK support to 'monolithicSparse' and does the inspection directly rather than relying qemu-img. I think we should utilize that and add any additional inspections there rather than creating yet another block of validation code in the plugins.
@Dan ref #33:
We already have lots of that image format logic in the glance/ common/ format_ inspector. py which is utilized in the location.py (and would have covered these should we have implemented the staging as location).
That already limits our VMDK support to 'monolithicSparse' and does the inspection directly rather than relying qemu-img. I think we should utilize that and add any additional inspections there rather than creating yet another block of validation code in the plugins.