Re: 13.51 regtests

[Trevor Daniels]

Francisco Vila wrote Wednesday, February 23, 2011 6:38 PM


> 2011/2/23 Phil Holmes <address@hidden>:
> > Another change I don't understand.
>
> Although I don't understand much about regtests, this difference 
> could
> show that printing order of objects in the same layer (with the 
> same
> value of the layer property) is indeterminate, therefore you could 
> try
> again and obtain a different result. NR 5.4.7 "Visibility of 
> objects"
> under "Painting objets white".

That's exactly the cause.  Perhaps the reg test should
be changed to set 'layer explicitly on the grobs involved
so the test is reliably reproducible, like this:

\header {

 texidoc = "The whiteout command underlays a white box under a
markup.  If grobs other than staff lines are involved their
@code{layer} property should be set explicitly, since the
order of processing grobs with the same value of @code{layer}
is not reliably reproducible.
"

}
\version "2.12.0"

\paper
{
 ragged-right = ##t
}

\relative c'' {
 \override TextScript #'extra-offset = #'(2 . 3)
 \override Stem #'layer = #3
 c4-\markup  { \whiteout \pad-markup #0.5 foo } c
}