POE Failures

[Andy Armstrong]

On 12 Nov 2007, at 07:51, Ovid wrote:

> I just noticed in the POE failures that raw output looks like this:
>
>  prove -vl t/30_loops/10_select/all_errors.t
>  t/30_loops/10_select/all_errors......
>  1..0 # most of these should move into other test files
>   No subtests run
>
>  Test Summary Report
>  -------------------
>  t/30_loops/10_select/all_errors.t (Wstat: 0 Tests: 0 Failed: 0)
>    Parse errors: No plan found in TAP output
>  Files=1, Tests=0,  0 wallclock secs ( 0.01 usr  0.01 sys +  0.03  
> cusr  0.01 csys =  0.06 CPU)
>  Result: NOTESTS
>
> Why is this a failure?  Hadn't we agreed that 1..0 is enough to  
> specifiy 'skip all'?  While it's not officially part of the TAP  
> spec, the '1.0 # SKIP' message is redundant and probably suboptimal  
> (what happens when someone does '1..13 # SKIP', for example?).
>
> Was this change never implemented?  We've been bitten by this  
> before.  From the Changes file, I see this:
>
>        - Relaxed strict parsing of skip plan for pre v13 TAP.


The change was to allow things like 1..0 # skipping everything - i.e.  
in pre v13 TAP we don't require the strict "# SKIP".

I think we talked about relaxing the requirement altogether but it  
never happened.

I explained the problem to Chris Williams in Vienna and apparently POE  
1.000 is fixed but unreleased for unrelated reasons.

-- 
Andy Armstrong, Hexten