[tapx-dev] Broken coverage testing

[Leif Eriksen]

On Fri, 2007-08-10 at 08:44 +1000, Leif Eriksen wrote:
> OK I'll install 0.55 and 0.56 and trace the merge between the two.

I haven't gone into the diff -r between D::C 0.55<-> 0.56 too much yet,
but a few interesting things are appearing as I poke cover.

Firstly, my statement about some objects getting to D::C's
print_summary() as empty is a mistake - those empty ones are for modules
in the blib/... tree - so naturally there is no coverage of them, hence
no body in those objects.

But that brings us to the report...

File                           stmt   bran   cond    sub    pod   time
---------------------------- ------ ------ ------ ------ ------ ------
blib/lib/TAP/Base.pm            n/a    n/a    n/a    n/a    n/a
blib/lib/TAP/Harness.pm         n/a    n/a    n/a    n/a    
.../lib/TAP/Harness/Color.pm    n/a    n/a    n/a    n/a 
...TAP/Harness/Compatible.pm   91.3   95.0   86.8   93.3  100.0   70.0
blib/lib/TAP/Parser.pm          n/a    n/a    n/a    n/a    n/a    
...

see how the ones with 'n/a' are _also_ for modules in the blib tree - a
nugget to work on ! Now to see if I can work out why 0.56 started paying
attention to the blib tree in preference to lib ?

L