Regression between lily 1.4.4 and 1.4.5--7

[Felix Braun]

Hi there,

seem to have run into a bug that has slipped into lily 1.4.5 and has not
been fixed until 1.4.7. I am typesetting an orchestral score but lily
aborts with the following error as soon as I include more than one
instrument in the same StaffGroup (the individual parts all work fine):

GNU LilyPond 1.4.5
Now processing: `Sommernachtstraum.ly'
Parsing...
Interpreting 
music...[8][16][24][32][40][48][56][64][72][80][88][96][104][112][120][128][132]
Preprocessing elements...
Calculating column positions... 
[3][6][9][12][15][18][21][24][27][30][33][36][39][42][45][48][51][54][57][60][63][66][69][72][75][78][81][84][87][90][93][96][99][102][105][108][111][114][117][120][123][126][129]
paper output to `Sommernachtstraum.tex'...
programming error: Improbable offset for translation: setting to zero 
(Continuing; cross thumbs)
[ last line repeated 37 times ]
ERROR: In procedure ly-number->string in expression (map ly-number->string l):
ERROR: Wrong type argument in position 1: +#.#

Sometimes lily also segfaults although I haven't figured out which
combination of instruments causes the segfault, when I try to typeset the
whole score it always segfaults. Unfortunaly I am unable to do further
debugging myself because I am severly RAM-challenged (and don't have all
that much time) but I'd be glad to help debugging if you could direct my
steps a bit. The score that I am having trouble with is available on
mutopiaproject.org (Mendelssohn-Sommernachtstraum) everything works fine
with lily 1.4.4.

Thanks for your help
Felix

-- 
Felix Braun
1910, rue Wellington
Montreal PQ
Canada H3K 1W3
Tel: ++1-514-933 60 58