On 07/ 6/10 09:26 AM, Dermot McCluskey wrote:
Sarah,
A couple more nits/queries:
- is the value of disk_prop/dev_size in the same format as size/val? ie
an integer followed by the units, like "20GB"?
Yes.
- the definition of <swap> sub-elements looks unusual:
<!ELEMENT swap (size|zvol|slice)*>
Can there really be multiple <size> sub-elements?
No, actually this should just be:
<!ELEMENT swap (size|zvol)>
thanks,
sarah
Dermot
On 07/06/10 14:37, Sarah Jelinek wrote:
Hi Dermot,
On 07/ 5/10 12:02 PM, Dermot McCluskey wrote:
Hi Sarah,
Regarding the following line in target.dtd:
<!ATTLIST target_device keyword (boot_disk) "boot_disk">
Does this imply there is only one allowed value for keyword, which
is "boot_disk", and that this is also the default value? If so,
what is the
point of this attribute? Once we load defaults (eg "xmllint
--dtdattr")
then target_device will *always* have a keyword attribute whose value
is "boot_disk".
We may have additional keywords in the future so making it a
attribute seemed reasonable. As for the default, yes I should not set
that, because you are right we will always have this set when we load
the defaults.
Also, regarding this line:
<!ELEMENT disk ((disk_name|disk_prop|disk_keyword)*, partition*,
slice*)>
what is the "(disk_name|disk_prop|disk_keyword)*" portion saying? I
think
it says you can have any *one* of those 3 sub-elements, but you can
repeat
the choice multiple times, thereby allowing you to have all 3 or any
combination thereof. If those 3 elements are mutually exclusive
(?), but also
optional, then perhaps you should have:
<!ELEMENT disk ((disk_name|disk_prop|disk_keyword)?, partition*,
slice*)>
You are correct, they are mutually exclusive so I will change this.
Also, there now appear to be two separate ways to specify that the
"boot_disk"
is to be the install target:
<target>
<target_device keyword="boot_disk"/>
</target>
or
<target>
<target_device>
<disk>
<disk_keyword key="boot_disk"/>
</disk>
</target>
</target>
Is that intended?
No. I intend for it to be only a disk_keyword. I will remove the
target_device keyword.
thanks,
sarah
***
- Dermot
On 07/03/10 00:12, Sarah Jelinek wrote:
Hi All,
I updated the AI/DC schemas and the design document to reflect the
comments received in the initial design review. If you want to see
the specific changes, please use the .odt file and set the show
changes and you will see the edits.
The areas I modified were:
-Target dtd. This is based on feedback plus discussions about the
use cases for AI. You will see a high level disk object now, that
encapsulates the partitions, slices or other parts. The new AI
schema supports mutliple disks and pools. And, the associated
components.
-Execution dtd. We are no longer embedded software in the
checkpoint elements. The software spec now has a name which much
match the checkpoint name to differentiate different software data
for multiple transfer checkpoints.
-Software spec dtd. This now has elements and attributes for
facets, and better publisher/origin definition.
-I removed the extraneous definitions that were in text for each of
the schemas. The schemas themselves should be sufficient.
-I added the dtd's as well as example instance documents to the
gate for reference.
Please review and send comments. I would like to close this out by
COB 7/9/10.
Thank you,
sarah
_______________________________________________
caiman-discuss mailing list
[email protected]
http://mail.opensolaris.org/mailman/listinfo/caiman-discuss
_______________________________________________
caiman-discuss mailing list
[email protected]
http://mail.opensolaris.org/mailman/listinfo/caiman-discuss
_______________________________________________
caiman-discuss mailing list
[email protected]
http://mail.opensolaris.org/mailman/listinfo/caiman-discuss
