[Gnu-music-discuss] etf2ly.py fails

[David Chan]

Hi,

I've been trying to convert some etf files from www.freenote.org to
mudela, using etf2ly.py.  The files I have been working on are
twopt01.etf, ..., twopt15.etf, located at

 http://www.freenote.org/music/jsbach/twopt01.etf

etc.  Here are some errors I've found.


twopt03.etf
-----------
etf2ly.py dies with the following output:

> Processing `../twopt03.etf'
> parsing ...Traceback (innermost last):
>   File "../etf2ly.py", line 1004, in ?
>     e = Etf_file(f)
>   File "../etf2ly.py", line 666, in __init__
>     self.parse (name)
>   File "../etf2ly.py", line 842, in parse
>     m = self.try_Sx (l)
>   File "../etf2ly.py", line 767, in try_Sx
>     self.slurs[-1].append_entry (params)
> AttributeError: 'None' object has no attribute 'append_entry'

I don't speak python but I have an idea as to what the problem might be.
In twopt03.etf, the lowest slurnumber an Sx(n) ever has is 5:

> ^Sx(5) 15 -32256 24576 1 6 45 

I think etf2ly.py is assuming that the lowest slurnumber will be 1.

>       def try_Sx(self,l):
>               m = Sxre.match (l)
>               if m:
>                       slurno = string.atoi (m.group (1))
>                       if len (self.slurs) == slurno:

                        ^^^^^^^^^^^^^^^^^^^^^^  never true for this file?

>                               self.slurs.append (Slur (slurno))
>
>                       params = list (m.groups ()[1:])
>                       params = map (string.atoi, params)

                 ** at this point "print self.slurs" outputs "[None]" **

>                       self.slurs[-1].append_entry (params)
>
>               return m        



twopt05.etf
-----------
etf2ly.py dies with the following output:

> Processing `../twopt05.etf'
> parsing ...processing ...Traceback (innermost last):
>   File "../etf2ly.py", line 1004, in ?
>     e = Etf_file(f)
>   File "../etf2ly.py", line 666, in __init__
>     self.parse (name)
>   File "../etf2ly.py", line 861, in parse
>     m.global_measure = self.measures[mno]
      ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
> IndexError: list index out of range

This is in Etf_file.parse().  mno = 33 when the indicated line fails,
wheras len(self.measures) = 33; i.e. it is failing on measure 33, the last
measure.  Relevant-looking ^GF statements in twopt05.etf are as follows:

> ^GF(1,31) 0 0 0 0 0 
> ^GF(1,31) 0 0 31 0 0 
> ^GF(1,32) 0 0 0 0 0 
> ^GF(1,32) 0 0 32 0 0 
> ^GF(1,33) 0 0 0 0 0 
  ^^^^^^^^^^^^^^^^^^^   this looks dodgy
> ^GF(2,1) 3 0 33 0 0 
[...]
> ^GF(2,31) 3 0 63 0 0 
> ^GF(2,31) 0 0 0 0 0 
> ^GF(2,32) 3 0 64 0 0 
> ^GF(2,32) 0 0 0 0 0 

Note that ^GF(1,33) is only 1 line long, and there is no ^GF(2,33).  In
fact, the piece is only 32 bars long.

Similarly,
        twopt09.etf (34 bars long) has a 1-line ^GF(1,35)
        twopt14.etf (20 bars long) has 1-line ^GF(1,21) and ^GF(1,22),
                but also 1-line ^GF(2,21) and ^GF(2,22).

In all three cases, if I delete the dodgy looking lines then the
conversion succeeds.



Perhaps Finale doesn't stick to the spec entirely.  Any chance of a
workround in the converter, Han-Wen?  :-)

Many thanks,
-- 
David