[tap-l] Uses of 1..0 without a directive

[Michael G Schwern]

There's been some question as to whether "1..0\n" with no skip directive is a
valid plan.  While reworking the Perl core tests this week I found a number of
uses.  67 to be exact.

Then I did a Google code search.
http://www.google.com/codesearch?hl=en&lr=&q=%221..0\n%22+lang%3Aperl&btnG=Search

There's a lot of uses in the wild, though take the results with a grain of
salt as there's a lot of duplicates and out of date versions.  But it's pretty
clear it's out there and it's what people have been using to mean "skip all"
for a long time.  I think we're going to have to support this one in the protocol.

I've been pondering how to specify this stuff that's only in there because TAP
writers do it in the wild.  "1..0\n" is one.  todo in the plan is another.  I
think the wording like "a TAP reader should parse it, but a TAP writer should
not use it".  We can tweak the strength of either side, but the basic idea
there are things a reader should understand but a writer should not emit
effectively codifying the idea of "be strict in what you emit, lax in what you
read".

I'm leaning towards "TAP reader may" and "TAP writer must not" which
effectively tells any TAP writers to change it ASAP while allowing a TAP
reader the choice of supporting it, but not forcing them or even particularly
encouraging it, but they can if they need to do it for backwards compat
without breaking or extending the protocol.  I also think it should be made
clear it will be removed in a future version.


-- 
29. The Irish MPs are not after "Me frosted lucky charms".
    -- The 213 Things Skippy Is No Longer Allowed To Do In The U.S. Army
           http://skippyslist.com/list/