Hi Jack,

Thank you for the review. I have responded to comments on files I modified/added:

ai_dtd:
------
How come target has * (which means zero or more)?

Because, targets are not required to be specified for AI.
ai_manifest.xml:
---------------

I'd like to change the Driver Update section. Issues are outdated packages, an incorrect searchall entry (already fixed per other's comments), and some comments (which I admittedly wrote) that need to be clarified or updated. I would also like the <search_all/> at the bottom not to be commented.

Below is what I propose. I have attached a diff to help show the changes since there are more than a few. For the diff, I left searchall where it was before to aid in showing diffs; however, I move it to the top as follows:

<add_drivers>
<!--
Driver Updates: This section is for adding driver packages to the
            boot environment before the installation takes place.  The
            installer can then access all devices on the system.  The
packages installed in the boot environment will also be installed
            on the target.

            A <searchall> entry performs a search for devices which are
            missing their drivers.  A repository publisher and location
            may be specified, and that repository and its database will
            be used.  If no publisher and location is specified, the
            configured repositories will be used.
            (See pkg publisher command.)  If <addall> is specified as
            "true", then drivers the database says are third-party drivers
            will be added like all others; otherwise third-party drivers
            will not be added.

<search_all addall="true">
<source>
<publisher name="opensolaris.org">
<origin name="http://pkg.opensolaris.org/release";
</publisher>
</source>
<search_all/>

<software> entries are user-provided specifications of packages
needed in order to perform the install. types are P5I, SVR4, DU. A <software_data> action of "noinstall" inhibits adding to target.

P5I: A pkg(5) P5I file, full path is in source/publisher/origin.
            Path may be to a local file or an http or ftp specification.
<software>
<software_data type="P5I"/>
<source>
<publisher>
<origin
name="http://pkg.opensolaris.org/release/p5i/0/driver/firewire.p5i"/>
</publisher>
</source>
</software>

SVR4: An SVR4 package spec. source/publisher/origin corresponds to the directory where the package is located. Name refers to the
            package name subdirectory (for expanded package) or datastream
            file.

<software>
<software_data type="SVR4">
<name>my_disk_driver.d"/>
</software_data>
<source>
<publisher>
<origin name="/export/package_dir"/>
</publisher>
</source>
</software>

            DU: An ITU (Install Time Update) or Driver Update image.
source/publisher/origin refers to the path just above the image's
            DU directory (if expanded) or the name of the .iso image.  All
            packages in the image will be added.

<software>
<software_data type="DU"/>
<source>
<publisher>
<origin name="/export/duimages/mydriver.iso"/>
</publisher>
</source>
</software>
-->

<search_all/>
</add_drivers>

I will take a look at your suggestions and change the comments in the ai_manifest.xml. In some places you refer to 'searchall' and it is search_all. However, in general the rewording seems reasonable to me.
default.xml
-----------

Copyright had 2008 removed.  Was that intentional?
Well.. it is a new file from a pure content view. Not sure it was intentional. Do you think it should be included?


I plan on adding back <add_drivers> with some bugfixes I'm working on. Do you want to add it back in a comment now instead, as is in the current default.xml? If not, please let me know where in the file you would want me to put it.

software.dtd:
-------------

83: A facet is an option...

Fixed.

89: I'm wondering about the value of the facet set attribute. If it wasn't there, then if the facet was present it would be "true". To temporarily disable the facet, one could just comment it out. This is easier than setting an attribute to false.

No, this doesn't actually work with the way facets are defined. A facet, if defined is set to true or false. So, if we only have the facet name, and assume it it true, the lack of that name doesn't mean false. That needs to be explicitly set.

thanks,
sarah
_______________________________________________
caiman-discuss mailing list
[email protected]
http://mail.opensolaris.org/mailman/listinfo/caiman-discuss

Reply via email to