Hey all,My GSoC project is quite close to fixing beam subdivision with custom beat structures, base moments and even nested tuplets. Currently, my implementation fails only one regression test: beam-subdivide-tuplets.ly. It's a sextuplet expecting to subdivided once halfway, while my code currently subdivides it twice at the thirds since the tuplet factor automatically simplifies. To pass that reg test and retain the old behavior of subdividing tuplets with non-prime numerators, I would just need a way to get the unsimplified numbers.
I tried inspecting the code responsible for handling the tuplet fraction. IIUC, currently, the only use for the unsimplified version of the tuplet fraction is to set the printed text above tuplet brackets (TupletNumber.text). With that being the only use, only the string text returned by the Scheme function from inputting the unsimplified tuplet ratio is recorded on the Scheme side.
Does anyone know how the process of parsing tuplet fractions goes? At what location in the codebase does the original tuplet numerator and denominator become a Moment? Where the TupletNumber.text function is invoked to store in Scheme? Once I understand, then the rest of the journey shouldn't be too hard for me.
After I am able to pass this 1 remaining regression test, I will create a merge request of my project. My code is always at https://gitlab.com/ljyip/lilypond. There will be more to come after that as I am way ahead of schedule.
-- - Jason Yip
OpenPGP_0xB69A3DD87D22F506.asc
Description: OpenPGP public key
OpenPGP_signature
Description: OpenPGP digital signature
