[caiman-discuss] Requesting a Code Review for 5265 distro-constructor should auto-generate image hashes

[Jean McCormack]

ValidatorModule.py:
lines 141-148: Prior to your change, if count was 0 it would return 
FALSE. Now it returns TRUE. Does this effect other callers?

Jean


Glenn Lagasse wrote:
> Hi All,
>
> Could I get at least two reviewers for this change please?  It's not
> going in to 2009.06 but I'd like to get some feedback from a few people.
>
> With this change, the distribution constructor can now generate
> cryptographic hashes for the images that it creates.  This is done by
> specifying the hashes you want (must be supported by the digest(1)
> command which I validate using a Validator method) in the manifest used
> to create the images.  The design is flexible in that you
> create/recreate hashes as a separate finalizer step (so if you change
> your mind and want different hash types you can get them without having
> to start the run over from scratch).
>
> I've tested this code using the following methods:
>
> 1) Sample manifest with proper values validates cleanly
> 2) Sample manifest with duplicate hash types fails validation
> 3) Sample manifest with checksums enabled but no hashes fails validation
> 4) Sample manifest with proper values creates hashes
> 5) Sample manifest with proper values creates hashes and can be resumed
> 6) Sample manifest without checksums validates cleanly
> 7) Sample manifest without checksums but specified finalizer script
>    runs ok but displays warnings when the checksums finalizer script is
>    run without having specified any checksums to generate
> 8) Sample manifest with checksum type not supported by digest command
>    fails appropriately
> 9) Sample manifest, resumed after initial checksum generation, with
>    different checksums works as expected (ie it generates the proper
>    checksums as now specified in the manifest
>
> Webrev:
>
> http://cr.opensolaris.org/~glagasse/slim_5265
>
> Bug:
>
> http://defect.opensolaris.org/bz/show_bug.cgi?id=5265
>
> While there isn't any rush to review this since I've no place to
> integrate it yet I would appreciate comments no later than one week from
> tomorrow.
>
> Thanks!
>
>