fix #296034: Crash when multi-measure rests are enabled and press End

[Jojo-Schmitz]

fixes https://musescore.org/en/node/296034 and also https://musescore.org/en/node/296541

also consistenltly use lastMeasureMM() rather than last() in other places of mscore/scoreview.cpp

[MarcSabatella]

Thanks for getting this! FWIW, the downside of just using lastMeasureMM() it won't account for a frame after the last measure. Not sure if that's actually relevant in the way it's being used or not (frame won't even appear in continuous view).

[Jojo-Schmitz]

I tried with a frame on the next page and it worked

[dmitrio95]

In my tests it doesn't indeed seem to handle frame at the end correctly. Maybe we should introduce something like MeasureBase* Score::lastVisible() or lastMM() function which would return either a frame or lastMeasureMM(), depending on what exactly stands at the end of the score?

[MarcSabatella]

Such a function seems logical to me.

[Jojo-Schmitz]

test.zip
Seems to work for me with the attached score (rename to mscz).
But a lastMM() or some such might be better indeed.

[dmitrio95]

test.zip
Seems to work for me with the attached score (rename to mscz).

For me with this PR End key moves me to the last "usual" measure (MM rest spanning for 11 measures), but not to boxes in the last two pages.

[Jojo-Schmitz]

Checked again: the End key does get me to the last page of that test score.
The Page-up key only gets be back to page 3 though, no matter how ofter it gets pressed, but that is the case without the PR too, so nothing I broke ;-), the Home key does get me to page 1, and Page-down does get me to the end (if pressed repeatedly)

[MarcSabatella]

Still doesn't make sense to me that this would work with a frame all by itself on the last page.

A possibly related issue, though, that doesn't involve mmrests - the viewport moves to the last page with measures on it if there is a frame all by itself on the last page and I try adjusting its size. So somehow thr viewport updating code I think could also benefit from a lastMeasureBaseMM function or however we choose to define it.

[Jojo-Schmitz]

Sure, a lastMeasureBaseMM() function would make sense here. Just as long as it doesn't exist, this PR is better than a crash ;-)

[RobFog]

3.3.3 milestone?