On 2019/12/19 21:39:20, c_sorensen wrote:
Although I am not an expert in pure/impure calculations, I think that
you are
misunderstanding this entry in the CG.
I confess, I'm finding it difficult to apprehend this.
But the rehearsal mark will potentially be placed somewhere different depending on the line breaking.
That was already the case when RehearsalMark.self-alignment-X was fixed to CENTER (and the difference in system height was more extreme). If it's a problem now, was it a problem before? Is this change an improvement? If there was/is a problem, can I prove it? Can I find a case that triggers circular evaluation? Can I find a case where a line break is poorly chosen due to the placement of rehearsal marks? I don't know. Can I play it safe by setting up a pure version of the callback that returns a constant value like before? That would be pure, but I don't understand whether it would be better.
Of course, this could all be wrong, but it's my best understanding of
how
pure/impure works.
Thanks for clarifying! :-D https://codereview.appspot.com/547340043/
