On 4/6/06, Han-Wen Nienhuys <[EMAIL PROTECTED]> wrote:

> They participate, in that their dimension fields are used for various
> computations.  Simply said,  a stencil is a combination of output format
> and a bbox.

Once a stencil is constructed, is it ever adjusted?  That is, once the
stencil is made, does Lilypond ever go back in and shift the contents
around, or is it treated as a single unit from then on?

> > 2.  The more I read the code and think about it, the more I think
> > stencil interpretation should be pushed to the back end and written in
> > Scheme.
>
> what exactly do you want to push back?

The way I understand it, Lilypond constructs a stencil to represent an
object, and the largest stencils (lines?  pages?) are then interpreted
to produce output.  I suggest that the back ends should get the
stencils themselves.

> > 3.  I really don't like stencil leaves being arbitrary scheme
> > expressions that produce output.  When the back end can't look inside
> > the box, it can't figure out the best way to deal with the contents.
>
> Perhaps, but it was an easy solution for what we needed: a way of
> having pluggable backends with minimum fuss.

When I changed the gsave/grestore stuff in the PostScript backend, I
was (and still am) nervous because the backend doesn't have a clue
what might be in the code fragments passed into boxes.

David


_______________________________________________
lilypond-devel mailing list
[email protected]
http://lists.gnu.org/mailman/listinfo/lilypond-devel

Reply via email to