Reinhold Kainhofer <reinh...@kainhofer.com> writes: > On So., 21. Aug. 2011 02:33:06 CEST, Carl Sorensen <c_soren...@byu.edu> wrote: >> My thought for the architecture is to have two sets of properties -- the >> context set and the \once set. > > That's exactly what I was about to propose, too. override basically > sets the default, vbut a once always always takes precedence. You > don't have to worry about restoring the old value (which one? as those > examples highlight) at the end of the timestep, you just have to clear > the once value. > >> When \override x=3 is received, it's put on both the \once set and the >> context set. > > I wouldn't go that far. I would say that a once always overrides all > overrides.
Should \once \override Staff.x be able to override \override Voice.x ? Presumably not, I suppose that every context would have its own \once value. I have actually considered giving once absolute priority. The reason I have chosen against this is that if a user writes \override x=2 music \revert x then he has a reasonable expectation that music will get typeset with a value of x=2 when music itself does not contain any overrides or reverts and the override x=2 happens directly before music. So I did not want a preceding \once\override x = 3 issued by different code be able to seep beyond the following explicit \override. This is of course, a judgment call. I chose to have \override and \once\override display an immediate action, and \revert only display an immediate action if the corresponding \override is visible at the moment. This is actually a similar problem space to that of nested properties: with nested properties, several different overrides can be visible at the same time (after overriding in sequence x, x.y and x.y.z, the visible x is a composition of all three overrides), so I am focused on creating semantics that keep straight to the stack order of the input, even though revert may have effects beyond the top of the visibility stack. >> \revert means "apply the revert to the context set. If the \once set is >> not empty, also apply it to the \once set." > > \once\revert would simply set the once value to the second-to-last > value of the override stack. \set Staffgroup.x = 2 \override Staff.x = 3 \override Voice.x = 4 \once\revert Voice.x ; where is x now? \revert Staff.x ; where is x now? [wait ...] ; where is x now? Since the current semantics of the implicit revert for \once\override are based on "would simply set" and the results get quite surprising if the stack does not do the programmer of staying unchanged anywhere else, I am by now quite wary concerning sentences containing the string "\revert would simply set". -- David Kastrup _______________________________________________ lilypond-devel mailing list lilypond-devel@gnu.org https://lists.gnu.org/mailman/listinfo/lilypond-devel
