multimeasure rest at beginning

[David Bobroff]

I sent in a bug report a few days ago.  This is a repeat report.  I'm
sending a repeat to be sure that my first one didn't get lost in the
shuffle or something as I have seen no response to it yet.  In this
example:

\version "2.5.24"

\score {
  \relative c {
    \set Score.skipBars = ##t
    R1*2
  }
}

I get this console output:

GNU LilyPond 2.5.24
Processing `test.ly'
Parsing...
Interpreting music... [1]
Preprocessing graphical objects...
Calculating line breaks...
programming error: Multi_measure_rest::get_rods (): I am not spanned!
continuing, cross fingers
[2]
Calculating page breaks...
Layout output to `test.ps'...
Converting to `test.pdf'...

The resulting PDF simply contains a clef, default time signature and a
staff.  If I comment out the skipBars line I get a completely blank
first measure followed by a normal single bar rest.  This seems to be
affecting only multimeasure rests at the beginnings of \score {} blocks.
This behavior occurs with current CVS (ChangeLog 1.3589)

-David