Re: [PATCH] Fix #305: Allow dynamics to be positioned independently

[Maximilian Albert]

Hi Neil,

sorry for the long delay in my follow-up.

>> As a user, I would normally not include it since I'm not aware of it and
>> would thus think that manual positioning of hairpins simply doesn't
>> work (or is buggy). Is there a way to avoid it (e.g., by having
>> Lilypond automatically detect when two hairpins point in different
>> directions)?
>
> I'm sure it's feasible, though it's going to increase the complexity
> somewhat due to the necessity of caching the previous direction.  I'll
> see what I can do.

That would be awesome!

> Bear in mind though that an explicit command is necessary for
> situations where the direction doesn't change (for example, the last
> dynamic in the regression test).

Fair enough. But the need of \breakDynamicSpan in such a situation
would be much more obvious to a user, I guess, and even someone not
knowing about the exact command might start looking it up in the
documentation (whereas the abovementioned behaviour simply looks like
a bug).

>> Also, is there a way to 'activate' \breakDynamicSpan
>> somehow so that it applies to all future hairpins/dynamics?
>
> It might be possible to redefine dynamic commands to send a break
> request automatically, though there is one limitation due to the way
> the engraver's coded: you can't have separate alignments for the case
> where there's an absolute dynamic directly followed by a span dynamic.
>
> For example, the following snippet wouldn't produce separate alignment
> spanners for the forte and hairpin:
>
> \relative c' {
>  c1_\f^\<
>  c1\!
> }

Hmm, okay. I suppose that in most cases people wouldn't want the
hairpin to point in a different direction anyway. Nonetheless, someone
might request it one day. So would you consider this limitation as a
bug?

Thanks again for your work!
Max

P.S.: Your patch hasn't been applied yet, has it?