Re: [Denemo-devel] A major re-structuring.

[Richard Shann]

This work is now in master, so those dependent on binary builds can test
from tomorrow.
The only visible change should be Check Score working in some corner
cases where it was failing, but a lot has changed under the bonnet. We
could now get rid of that annoying off-by-one in the bar numbering in
the display when there is an upbeat. We could even contemplate having
directives attached to bars, so one could say "print this bar
transparently for this layout" ...

Please report any issues ...

Richard



> >         > Sent: Thursday, March 10, 2016 10:14 AM
> >         > To: denemo <address@hidden>
> >         > Subject: [Denemo-devel] A major re-structuring.
> >         >
> >         > One of the most fundamental mistakes made when Denemo was
> >         first started was not to have a structure for a Denemo
> >         measure. Instead it is just a list of the objects in the
> >         measure. As a knock on effect, nothing can be stored with the
> >         measure that applies to the whole measure.
> >         > This has made it difficult to fix the way that information
> >         about the context (clef, timesignature ...) of an object is
> >         stored, especially when a measure has no objects. It has long
> >         been my ambition to get rid of the way the calculation of the
> >         context information is computed - a draw time. This is still
> >         causing failures in Check Score on occasions.
> >         >
> >         > As a start I have created a branch Restructure in which a
> >         staff now has a list of Measure structures, which for now just
> >         has one element, viz.
> >         > the list of objects.
> >         >
> >         > I've done a first trawl of all the code to try and make the
> >         change work properly, and I've tested (with the default build)
> >         loading files, starting new files, adding and deleting
> >         movements, staffs, measures, notes, chords, copy and paste
> >         creating snippets, pasting and following them Undo and Redo.
> >         >
> >         > It would not surprise me if I have missed something, it is
> >         perilous work. The symptoms of an error are crashes, often
> >         with memory allocation warnings. But, so far it seems good. If
> >         anyone can help testing this version that would be good!
> >         >
> >         > Richard