#1704 Notation Editor, clef causes cursor position to be inaccurate

[Lorenzo]

I thought this happened just on first measures of a segment but that's just because typically, unless you insert clef changes, that's where a clef is by default.

This becomes particularly evident if you use the keyboard to move around (left, right) and to select (left, right with SHIFT modifier), as you end up selecting notes which don't match the cursor position.

Left right arrow key behaviour is consistent in editing windows and moves backward / forward of the currently selected time (this also works in the matrix editor), so the visual inconsistency makes it very hard to edit these measures.

Ideally (because they are not currently selectable via keyboard - i.e. SHIFT + left/right will not select a clef), they should be 'visually ignored' by the cursor.

Attached is a short screengrab video which hopefully helps explain the problem (keyboard presses are shown via screenkey).

[Ted Felix]

Confirmed.

With nothing selected and the cursor at the far left, the problem is really obvious. Hold down shift and press right arrow. The cursor jumps over the clef and selects the first note. Shift+right again and the cursor jumps over the time signature and the first note and selects the second note. Shift+right and the cursor jumps over the second and third notes and selects the third. Finally the fourth shift+right does what we expect.

In a bar without a clef or time signature, right arrow works as expected.

[Philip Leishman]

I made a change for [#1672] which causes this problem. I am not really sure why I made the change :(
I have changed back - please merge

[Philip Leishman]

As I feared - this change breaks the fix for bug 1672.
Still needs more work.
Maybe wait with the merge

[Philip Leishman]

OK - I think I have a version that works for both cases.
The cursor does not jump to the current segment if the playback position is outside that segment but inside the segment the cursor is layout accurate.
See latest version of the merge request.

[Ted Felix]

Pushed as [c006f7]. Please test latest git.

It has been a long time since I've tested [#1672], but a little regression testing based on the various test cases described there indicates the original 1672 fix is still working.

[Lorenzo]

I have tested with the latest git and this seems to be fixed now. Thanks!
To 'stress test' a bit I inserted notes and weirdly placed clefs as well as with quarter, 1/8 and 1/16 notes, including overlapping.