[Patch] New VHD support... now with added dynamics

[leilei]

I've accidentally made a differencing VHD to a differencing VHD and have broke the VHDs in the process.

oops

[shermanp]

I've accidentally made a differencing VHD to a differencing VHD and have broke the VHDs in the process.

oops

Differencing VHD's are fragile unfortunately. Microsoft created the newer VHDX format for a reason...

Having said that, it should be possible to chain them. The biggest gotcha is accidentally changing the parent after you created the child image.

If the following happens when chaining differencing VHD's, please open an issue https://github.com/shermp/MiniVHD/issues/new/choose:

• You create a differencing VHD from another differencing VHD, and something breaks IF YOU HAVE NOT OPENED/MODIFIED THE PARENT DIFFERENCING VHD.
• The parent VHD (differencing or otherwise) breaks in any way.

[leilei]

I've also had some differencing VHDs not boot because it thinks the parent was modified - which they were never modified (and set to read only after they were created+finalized anyway)

The only thing I can think of about this is... possible daylight savings disreprancies? Parent was 2-18-2021, diff vhd was 2-20-2021...

[shermanp]

I've also had some differencing VHDs not boot because it thinks the parent was modified - which they were never modified (and set to read only after they were created+finalized anyway)

The only thing I can think of about this is... possible daylight savings disreprancies? Parent was 2-18-2021, diff vhd was 2-20-2021...

That... wouldn't actually be surprising

The timestamp check may simply be too fragile, and I may have to remove it. Windows itself doesn't bother setting the timestamp, so MiniVHD may have to follow suit. Either that, or I do what Windows does when creating differencing VHD's and set the timestamp value to zero.

VHD really has no good way of verifying the integrity of the parent unfortunately.