Closed ARogovskyy closed 3 months ago
WRT this test file, you are correct that returning successful-ok-ignored-or-substituted is a valid implementation. The point of the 'none' case is to test the recommended behavior, but it would be good to make this more explicit somehow.
So I take it as that the course of action is not to add successful-ok-ignored-or-substituted
but instead document that this test explicitly tests this particular behavior?
Been thinking about this, and given the goal is to test things that are supposed to work and not things that are not supposed to work, I'll just remove the 'none' test.
[master 83d979b5e] Drop 'none' test (Issue #909)
I am not sure if this a bug or working as intended, just stumbled upon this while running the IPP tests under
examples/
against my ipp server.There is a test in
get-printer-attributes-suite.test
which sendsGet-Printer-Attributes
withrequested-attributes
set to the single keywordnone
. I can't seem to find any referencce ofnone
being a permitted special value for that attribute. Maybe I didn't look enough, but neither RFC 8011 nor the IANA IPP registrations page mention it.Then, it would seem like a "fake" value to instruct the printer to return no attributes (because
none
is not an attribute name), which CUPS does. But then , shouldn'tsuccessful-ok-ignored-or-substituted-attributes
be also a valid status code in this test? The RFC does explicitly allow the server to send the unsupported attributes in the Unsupported Attributes Group after all.