Tom Mueller (pkg-discuss) wrote:
Shawn,
Changes look good except for one point: is the policy value needed on
unset-policy? Maybe this is a cut/paste typo.
I appreciate that you provided better consistency between the
set-property, set-publisher and set-policy commands, but for the record,
I maintain that the -v option adds unnecessary complexity to the command
interface, and that if it is really deemed necessary, then it would be
needed for set-property/set-publisher too (as part of an "improve
scripting interface" wad). Also, -v is already used for "verbose" in
other subcommands.
There is established precedent for using the same option letter, but
with different meaning for different subcommands (see CLIP). Or, as
another example, image-create takes a --variant option, which obviously
implies that -v (if it was accepted) has a different meaning for that
subcommand.
However, when taking a step back and looking at the overall system, I'm
questioning whether the degree of helpfulness that this delineation
provides is sufficient to overcome the conceptual complexity that this
adds. More is not always better; sometimes it is just more. This feature
could have been implemented with properties (some of which have boolean
values) and by adding publisher-specific properties (which we
effectively already have via the set-publisher command). No new
subcommands to learn or document - just some additional property values.
These doubts are very subjective, so if I'm the only one having these
doubts, feel free to ignore them.
I believe it is better than the current morass that can't decide whether
to call image attributes properties, attributes, or policies, which is
really quite confusing IMO.
From a technical perspective, it is also preferable, as policy
management may expand in the future.
Cheers,
--
Shawn Walker
_______________________________________________
pkg-discuss mailing list
[email protected]
http://mail.opensolaris.org/mailman/listinfo/pkg-discuss
