Re: avoid alignment of hairpin and dynamic text

[Jean Abou Samra]

Le 01/02/2021 à 18:10, Werner LEMBERG a écrit :

> [we are actually on the wrong e-mail list, but...]
>
> > > Reason is that the order of events at a given moment is not
> > > specified.  [...]
> > Are you sure about the order of events not being guaranteed?
> Yes, see David's response to something similar syntax-wise that I
> tried some time ago:
>
>    For simultaneous events, LilyPond has no order.  That's why one
>    needs to have one phase for recording events and another for doing
>    something based on them.
>
> The problem is parallel music.
>
> The important e-mails in the thread start with
>
>    https://lists.gnu.org/archive/html/lilypond-devel/2019-07/msg00142.html
>
>
>      Werner

Indeed, you are right. Even though I don't see
a reason why events on a single note would not
be processed in order by an engraver operating
on Voice level, the expected behavior with
something like

<< { c'1\p\< c'1\f\> } { s1 s1\breakDynamicSpan } >>

is not at all clear.

Actually, maybe you could manage to make
\breakDynamicSpan a music function. Perhaps as
simple as

breakDynamicSpan = \tweak spanner-broken ##t

if the current code is updated to account for
that (I did not check in detail).

That would break some current uses such as
c'1\<\breakDynamicSpan which were clearly intended
(see input/regression/dynamics-alignment-breaker-order.ly
and commit cf3642858a2 by Reinhold). Now, it's not
like the feature has been undocumented so far…

Regards,
Jean