Re: Doc bug: beams

[Carl Sorensen]

On 9/27/09 12:47 PM, "Nicolas Sceaux" <address@hidden> wrote:

> Le 27 sept. 09 à 18:25, Carl Sorensen a écrit :
>>>   \time 4/4
>>>   \repeat unfold 8 a8
>>>   %% group 8th notes by 2
>>>   \overrideBeamSettings #'Score #'(4 . 4) #'end #'(((1 . 8) . (2 2 2
>>> 2)))
>> 
>> This is a feasible beamSettings override, but it's probably not
>> desirable.
>> 
>> The new autobeaming code requires that you set a *complete* set of
>> beam
>> rules for any time signature.  This is not a complete set; it
>> doesn't have
>> default behavior (but if no default behavior is in the beaming rule,
>> it will
>> default to grouping at beatLength).
> 
> You mean that a beam setting override must contain a (* . (...)) rule?
> Even if a (* . (...)) rule exists in the default beamSettings?

Yes, that is correct.  There is only one beaming rule that applies per time
signature.  It is an alist that contains all beaming information.  That way,
you don't have to go around figuring which rules to revert; you just replace
the old rule with the new rule.

> 
> I would have naively expected that a beam rule override would change
> some particular beaming behavior, and that when it does not apply the
> rules
> defined in beamSettings would be used.

That's the old-style autobeam beahavior -- a bunch of individual rules that
need to be managed individually.  That caused lots of challenges, which is
why we changed it.

Under the new style, when you define autobeaming rules for a time signature,
you define *all* the rules that apply, so there's no need to go search
through the defaults and revert rules that you don't want to have apply any
more.

> 
> That should appear clearly in the doc (but maybe I have missed it).

Thank you for reading this with new eyes.  The docs clearly need a better
explanation.  I'll clean it up.

Thanks,

Carl