Make test-output-distance regtest contain span bars

Make test-output-distance regtest contain span bars

Since span bars don't turn up in the metrics, the regtests don't show
them.  To make catastrophic span bar failures show up at least
somewhere, the ever-changing test-output-distance test is made to
include them.

[lemzwerg, 2013-02-27 19:15:27]

LGTM.

[thomasmorley65, 2013-02-28 00:45:57]

LGTM

One suggestion:

https://codereview.appspot.com/7400057/diff/1/input/regression/test-output-di...
File input/regression/test-output-distance.ly (right):

https://codereview.appspot.com/7400057/diff/1/input/regression/test-output-di...
input/regression/test-output-distance.ly:20: 
How about setting \bar "|."
Might be a stronger flag.

[dak, 2013-02-28 05:57:45]

On 2013/02/28 00:45:57, thomasmorley65 wrote:
> LGTM
> 
> One suggestion:
> 
>
https://codereview.appspot.com/7400057/diff/1/input/regression/test-output-di...
> File input/regression/test-output-distance.ly (right):
> 
>
https://codereview.appspot.com/7400057/diff/1/input/regression/test-output-di...
> input/regression/test-output-distance.ly:20: 
> How about setting \bar "|."
> Might be a stronger flag.

I think we have a special span bar at the start, and a default span bar at the
end.  \bar "|." would leave us only with special span bars unless one expands
the whole example by another bar.  Neither option makes me really enthused, but
I am willing to let myself be convinced.

[Keith, 2013-03-02 07:30:11]

LGTM.

On 2013/02/28 05:57:45, dak wrote:
 
> I think we have a special span bar at the start,
The bar at the start is a SystemStartBar, which was not affected by the recent
span-bar bug.  You might want an extra measure, and a final bar, then.  Also,
some note in the texidoc saying that this test should have span bars.

Of course, it would be better if the regression tests that concern span bars
could have an override so that the extents of the span-bars as printed are
written into the .signature files.  Unfortunately, I could not find a way to
make this happen.  I like to blame everything on the pure-from-neighbor patch,
but here I cannot: even before that patch SpanBars had X/Y-extents unrelated to
what was actually printed.

Another more satisfying option, if possible, would be to finish the TODO in
stencil.scm:found-grob suggesting "use stencil extent if available" but I don't
know how to add the correct offsets to the stencil extents to give an accurate
output in the .signature file.

[dak, 2013-03-05 14:58:45]

On 2013/03/02 07:30:11, Keith wrote:
> LGTM.
> 
> On 2013/02/28 05:57:45, dak wrote:
>  
> > I think we have a special span bar at the start,
> The bar at the start is a SystemStartBar, which was not affected by the recent
> span-bar bug.  You might want an extra measure, and a final bar, then.  Also,
> some note in the texidoc saying that this test should have span bars.

I think that this issue can be made to explode into arbitrary overkill.  I'd
rather just commit it as-is and leave further improvements to whoever considers
them really pressing.  There is at least one span bar now, and a multimeasure
rest (I seem to remember that those were somewhat weak in metrics as well, but
whether or not this is the case, this is about the most simple content for
another staff anyway).