Markup and markup-list protocol

[Nicolas Sceaux]

Le 17 févr. 2010 à 07:27, Boris Shingarov a écrit :

> Another change I was thinking about, is to redesign the API contracts
> around interpret-markup-list. [...]

Boris,

The patch you submitted seems to me as a hack making the \score markup
command behaving differently from other markup commands, and breaking
the current protocol in which the distinction between markup/stencil
and markup-list/stencil-list is clearly made.

For instance, currently we have
  \column and \column-lines
  \justified and \justified-lines
so in that context the natural solution for the \score issue would be
  \score making one stencil for all systems
and e.g.
  \score-lines making one stencil per systems

So the patch you proposed cannot be accepted (afaic) in its current form.


However, I think you're right when you say that this protocol should 
probably be redesigned, in order to tackle several issues:

- we indeed certainly want to get rid off the dual definitions of \column
vs \column-lines, etc, ie. merge the markup and markup-list things, so that
"the right thing" happens without having to care about it;

- the current internal representation of markups (bare lists of markup
commands and their arguments) is not very expressive.  By using a more
complexe data type, some extra information may be added to the markups.
For instance, you cannot add attach a \label inside a piece of text,
instead you have to stop the \markup(lines) block, add a label, and start
a new \markup(lines) block.  An other example: you cannot change the
after-title-spacing &co settings attached to a particular toplevel markup;

- the point you made about keeping a context across stencil interpretations
is interesting.

Nicolas