Comments on pages 29 - 42.

Josh Simons


Page 29

"New images are created to mark a boundary between the packages installed in
different images." Sounds like a circular definition.

This paragraph: "After you have completed the installation of the OPenSolaris 2009.06..." is unclear. What seems to be missing is a statement that software management tasks may
only be performed on images. Perhaps this would be clearer:

Software management tasks are performed are performed on images. Since a complete OpenSolaris 2009.06 installation creates a full image, it not necessary in the usual case to explicitly create an image to perform software management tasks. However, if one wants to provide logical separation beween different software applications as during
        zone creation process, an image must be explicitly created.

Page 30

Figure 4-1 might be clearer if you labeled the images as full or partial.

The switches to image-create aren't shown correctly. It should be either:

image-create [-R] [-FfPUz] ... (dashes on the inside of the brackets)
or
image-create [-FfPRUz] ... (not sure why -R was called out separately?)

Page 31

Example 4-1

Suggest changing last sentence to: "The -p option is used to specify the
        location of the package repository."   (You mention earlier that 
http://pkg.example.com
        is the repository in this example.)

Put a backslash continuation character after -p to should this is one command line?

Displaying image properties. "pkg property" shows properties...but for which image since none is specified? Is it really the case that this command does not take an image/directory as an argument? [I see on page 34 that this is explained...it would be clearer to introduce
the idea of a current image before describing these commands.]


Example 4-2

Output of command says "lush-content-cache-on-success" .... the leading "f" has been omitted.

Page 32

        set-property applies to which image? (none specified)

        Example 4-3

                "pkg property" command should be preceded by "$ " prompt as in 
other
                examples.

Same comment for the "pkg -R set-property" and "pkg property" commands that
                follow.

Page 33

This discussion includes the concept of "a current image". How is that defined?

        Example 4-5

        typo: pfexeec --> pfexec

        (also , that line is duplicated...remove one?)

This seems to equate the concept of a clone with an image which is confusing. (An image-update results in a clone being updated.) [This is partially explained on page
        35...would be better if explained earlier.]

Page 39

        In the Note remove "section." after "page 41"

        Define "key attribute".

        Last paragraph says the table shows type of actions and their
        "standard attributes"? What is a standard attribute?

Page 40

In the table, the link and hardlink descriptions are written more clearly than the
        directory and file entries. How about:

        The directory action represents a file system directory.
        path -- the file system path where the directory is installed.

        The file action represents an ordinary file.

        Typo in "Driver" section: TThe --> The

        "Driver"...Capitalize "The driver files must be installed..."

        "Driver" payload reference is obscure. How about just:

"The driver action represents a device driver. Note that the driver files must be
        installed separately using File actions."

[This seems error-prone, especially if the argument to this action is "usually the file name of the driver binary." In such cases, might we expect people will forget to specify the appropriate File action with the same argument? Is this an error case
        we catch and flag?]

"Depend" -- with no key attribute, how is the dependency to another package expressed? Same question for Set, Group, and User. There is no required attribute (which is how I interpret "key attribute"?) [If the table really is meant to show "standard attributes" as described on Page 39, I would expect some argument to be described for these actions.]

Page 42

        "ISA"  --- we support SPARC as well, not just x86 platforms.

        Define "ARC case" for non-Sun readers?

        Human-readable written twice here with em dashes.

Same comment on ISA in its second mention on this page. (Why is it mentioned a 2nd time and included in an opensolaris attribute table when it isn't an opensolaris
        attribute (i.e. not prefixed by opensolaris)?)

        




                







_______________________________________________
pkg-discuss mailing list
[email protected]
http://mail.opensolaris.org/mailman/listinfo/pkg-discuss

Reply via email to