> Michael Shapiro writes:
> > > sfwnv-x{small}104: xmllint --valid ipmievd.xml
> > > ipmievd.xml:92: element stability: validity error : Value "Uncommitted"
> > > for attribute value of stability is not among the enumerated set
> > > <stability value='Uncommitted' />
> [...]
> > The stability values used by smf(5) are the attributes(5) values, which
> > are the stable tokens that pre-date the recent decision to use a different
> > set of tokens in our internal ARC paperwork. You should use External
> > or Unstable instead.
>
> The new stability attributes approved by the ARC are *NOT* merely
> internal paperwork. They're new *public* attributes that supersede
> the previous ones and are intended to be used in all documentation
> that refers to the stability of public interfaces.
>
> See PSARC 2005/220 for details.
>
> Given that we'll be using these new attributes in future man pages and
> that attributes(5) will be updated to have them, it would be good to
> have SMF accept the new values rather than forcing project teams to
> step backwards into the previous taxonomy.
I understand where you're coming from, but I disagree with the outcome.
The previous attributes tokens were themselves Stable. They were
further encoded into multiple pieces of software which was also made
Stable. Changing them actually violates the existing definitions, an issue
which I raised in that ARC case but was lost amidst the zillion emails on
that case discussing many other concerns. I've been busy with other things,
but I plan to re-open the discussion on the stability case so we can discuss
the specific issue of impact on our software that I raised and its impact,
and then come to an appropriate concensus.
-Mike
--
Mike Shapiro, Solaris Kernel Development. blogs.sun.com/mws/
