[Removed website-discuss; it's not really their problem.]

Mark Martin wrote:
> The case at http://arc.opensolaris.org/caselog/PSARC/2009/387
> <http://arc.opensolaris.org/caselog/PSARC/2009/387/REDACTED.txt> is
> showing as largely redacted, which is an error.  I suspect it may be a
> simple flag, so would someone mind giving it a quick jiggle?
> 
> I'm curious to know what the condition is that got it into this state --
> clearly it's not content driven as that material was quite un-redacted
> just days ago.  I'm happy enough just to get it re-liberated, though.

When a case is listed as "open," the rule used to be that any files that
accidentally had special key phrases in them were excluded on a
file-by-file basis.  This was good, because there were some things (such
as the forms used for 1-pagers) that occasionally had accidental notices
in them.

That changed when we moved over to the arc.opensolaris.org
infrastructure.  The script that does the copying there doesn't just
exclude a file, it blanks out the whole case if *any* file in there has
a naughty word.  (And the list of bad words is substantially longer and
looser.)

Since then, volunteers (myself before I left Sun, I think mostly Garrett
since) have been carefully watching that list, and scrubbing the cases
manually (with vi and emacs) that wander into trouble.

The result is that simple mistakes like using the word "pro***prietary"
on an open ARC mailing list will result in having the case and all of
its materials marked as "redacted."  Anyone outside of Sun who wants to
start making cases less available can just send a single email message
to shoot a case down.

You can even trigger the mechanism by saying something like "I don't
think this should be conXXXXfidential" or "that's not a propriZZZZetary
idea."

-- 
James Carlson         42.703N 71.076W         <carlsonj at workingcode.com>

Reply via email to