Re: Controlling hairpin length

[David Sumbler]

On Sun, 2016-01-17 at 20:15 +0100, Thomas Morley wrote:
> 2016-01-17 19:21 GMT+01:00 David Sumbler <address@hidden>:
> 
> >
> > The only trouble is, I am getting a lot of compiler errors and warnings.
> > For each of the relevant layout blocks, I get:
> >
> > programming error: infinite rod
> > continuing, cross fingers
> >
> > and then numerous instances of:
> >
> > programming error: ignoring weird minimum distance
> > continuing, cross fingers
> >
> > In the cases of the 2nd violin and the viola, this latter message is
> > printed 50 times each for the 3rd movement alone - quite a lot for a
> > movement with only 64 bars!  And there are certainly well over 1,000 of
> > these errors for the whole quartet (score and parts).
> >
> > Have you come across this behaviour?
> 
> 
> 
> In one word: no.
> Please try to boil it down to minimal example otherwise I can't help.
> 
> Cheers,
>   Harm

After spending some hours trying to get to the bottom of this problem, I
narrowed it down to 3 different elements in my files which were causing
these hundreds of error messages when \myHairpinMinimumLength was set.
There were not hundreds of triggers in my files: but it seems that, once
triggered, the hairpin routine couldn't sort itself out for the
remainder of a movement and so kept on repeating the same message.

The first element that caused the errors was my occasional use of \invP,
defined as:
invP = \tweak stencil ##f \p

I have used this in a few places, rather than \! to terminate a hairpin.
I do this as an easy way to finish a diminuendo just before a notehead
so that the note in question does not appear to be included within the
diminuendo.

The second thing that gave rise to problems was the following bar:
d'2..*6/7 \sf\> s8\! e,8~\< |
I had done this so that there would be more space between the diminuendo
hairpin and the following crescendo hairpin.  I wanted it to be clear
that the new crescendo actually begins with the second note.

The third item that caused "infinite rod" and "weird minimum distance"
errors was the following (all 4 quartet instruments in "parallelMusic"):

    << b1\fermata {s2..\> s8\!} >> |
    << c1\fermata {s2..\> s8\!} >> |
    << e1\fermata {s2..\> s8\!} >> |
    << b1\fermata {s2..\> s8\!} >> |

My reason for using this was, again, to try to control the position of
the end of the diminuendo hairpins.

No doubt none of these 3 examples is very elegant, and they are very
likely not the best way of achieving the result I want.  But they worked
perfectly well until I used the improved hairpin length routine.  I feel
that ideally an improvement in an area where Lilypond is somewhat
unsatisfactory (so far as the user is concerned) should not at the same
time break some other aspect(s) of the program's output, so I am
pointing these out not as a criticism, but for information.

For the time being, perhaps, I shall have to go back to individual
tweaks for hairpin length, using trial and error to find the best
values.

Thanks again for your help.

David