Alternative polymetric representation

[Arle Lommel]

(Sorry for the lengthy post, but a couple of issues came up when I tried to 
work with an older file I did some time ago.)

About 18 months ago Reinhold helped me figure out how to display time 
signatures like (3+3+4)/16 in which there is a complex upper portion of the 
time signature over a simple lower portion. Reinhold provided a fairly complex 
fix at that time, but with updating from 2.12.2 to 2.13.18, it runs into some 
problems with beat grouping. In particular running the convert does a 
reasonable job except for this error:

> Not smart enough to convert beatGrouping. 
>    beatGrouping with a specified context must now be accomplished with
>    \overrideBeamSettings.
> Please refer to the manual for details, and update manually.

OK. Fair enough, I need to look at the current manual, but there are some 
problems with the documentation online that make this problematic.

If I go to here:

http://lilypond.org/doc/v2.13//Documentation/web/manuals.html

and search, it does not search the 2.13 documentation, but the 2.12 
documentation (where \overrideBeamSettings is, for obvious reasons, not found).

So I modify the search string to change the version number to 2.13. Not a big 
deal, just a minor annoyance, but it only brings up one hit:

http://lilypond.org/doc/v2.13/Documentation/ce/lily-547b3574.ly

Which returns a 404. (And the Google hit doesn't have a cache to look at!)

So the question is where can I find the documentation for 
\overrideBeamSettings? I'd like to consult that *before* I come here and get a 
RTFM response, but in this case it seems that TFM doesn't exist.

For what it's worth, what I'm trying to get right is the equivalent in function 
to 
http://lilypond.org/doc/v2.13//Documentation/snippets/rhythms#changing-time-signatures-inside-a-polymetric-section-using-_005cscaledurations
 , but the display is quite a bit different in appearance and cleaner, e.,g.,

That was the output from 2.12.2. If I convert it and convert to PDF, I get the 
following, which is (obviously) not what I want:

(I'm not sure why the first line only is doubled.)

At the time I first got Reinhold's solution, I believe he said that this time 
signature notation would make it into the trunk for 2.13, but I'm not entirely 
certain I recall that right.

So my question is if there is now a preferred way to accomplish this sort of 
time signature in 2.13.x with autobeaming (and hopefully no unexplained 
duplication of the first line of the *monophonic* score)

If anyone wants to see the entire scores in question, here is the link to the 
original Lilypond file for 2.12.2, which worked under that version:

http://dl.dropbox.com/u/223919/lilypond/compound.ly

And here is the PDF:

http://dl.dropbox.com/u/223919/lilypond/compound.pdf

Here is the link to the output from the convert function (2.13.18):

http://dl.dropbox.com/u/223919/lilypond/compound2.ly

And the PDF it makes:

http://dl.dropbox.com/u/223919/lilypond/compound2.pdf

Any guidance on this would be welcome since this is digging into some pretty 
obscure code.

Thanks in advance for any insights,

Arle Lommel