[Denemo-devel] Transpose

[Richard Shann]

When we were getting transpose working we wanted to apply the script to
every note is a selection. This raised the issue that we were executing
the entire script - including defining all the functions - for every
note, instead of just defining them once & then executing them for each
note.
I seem to remember thinking about this & proposing something.
The same issue has just arisen with tweaking the positions of things
that LilyPond may fail to place optimally, for a different reason.
We want scripts to be able to place their definitions somewhere where
they can be instantiated without being obliged to execute the script
itself. (For example, if I load a .denemo file that has a D.C al fine
marking in it, and I want to tweak its position I want to have the tweak
script available even though I haven't inserted the D.C al fine myself -
i.e. I have not run the script that the tweak relates to.)

My question is: did I already implement something? Or did I just think
about it, or did I write up a proposal of how to do it? If so, where?

I remember that my thoughts involved using a naming convention (e.g.
RehearsalMark::tweak) for global variables that belong to a given
command. But did I write this up somewhere? This business of losing
things is not so unusual. I came across an example when I created the
Double Barline script the other day. I called it Double Bar, and to my
surprize it promptly drew itself - there already is a double bar command
in Denemo - a built-in. It is just that it has gotten lost over time. It
is only slightly broken in fact. (We don't really want it - it is far
less flexible than a scripted one - you can't tweak it etc).

Obviously, my memory being what it is, this highlights the need for a
link from our website to design notes, proposals etc. I have been trying
to use the list of features on the About page as a log of the new
features as they are added, to avoid a desperate scrabble for release
notes at release time, but this would best be something else I guess.

Richard