On 2017/09/22 11:34:30, dak wrote:
On 2017/09/22 11:19:08, Ebe123 wrote: > Add to regression tests
I think it is a bad idea to change rgb-color here: this is a recipe
for
wide-reaching incompatibilities for the sake of limited applications.
It would
make more sense to meddle with the type of the "transparent" property,
making it
a boolean-or-number? for example.
If transparency is supposed to be made into a part of color, I think
it should
rather be optional: see <https://sourceforge.net/p/testlilyissues/issues/5162/#bd3f> for a
proposal in
that regard.
However, I think that transparency is really something that tends to
have
considerably different scope and application as color and is more
often than not
applied in manners orthogonal to color, so keeping it in a separate
property
might make for a lot less conflicts of use cases in the long run.
So [#5189] would be preferable to this. There would be way more work if done that way. Alpha transparency is integrated with color in some formats, such as SVG. Here, if using rgb-color, the opacity is optional as it defaults to 1,0. There is also a convert-ly rule for changing rgb- to rgba-. https://codereview.appspot.com/330300043/ _______________________________________________ lilypond-devel mailing list lilypond-devel@gnu.org https://lists.gnu.org/mailman/listinfo/lilypond-devel