[branch 1.2] [trunk] copying straight chords into tuplet brackets causes corruption

Confirmed using r.2179 nightly and 0.9.5 stable (Windows XP)

Doesn't crash in r.2597 but it creates a corrupt measure (not enough beats in the measure).

I check Sibelius and when you try to copy something starting in a tuplet, you have a very nice dialog saying "You can't copy this passage as it intersects a tuplet". I guess we could do something like this?

While checking this, I tried something else

1. Create new score in 4/4
2. In first measure, create an half note and 4 8th rests
3. In second measure, create a tuplet of 3 quarters
4. Select the first half note and copy it
5. Select the last 8th rests and paste
6. The result looks "good" but if you right click on the new 8th note, you have access to a tuplet menu
7. If you select remove tuplet from the menu, you destroy the timing

Attached is a file to reproduce the problem. Please note that there is no tuplet menu element in this file, since save or load removed it.

For the record, Sibelius allows such a manipulation. The tuplet is not removed and contains a dotted quarter notes + a 8th rest + quarter. So once the previous measure is full (with an 8th), the remaining duration (the dotted quarter) is included in the tuplet by reducing its size by 1/3.

The tuplet menu that Lasconic reports in comment #3 above goes away after a reload. Apart from the tuplet menu, the current behavior seems fine to me. Sibelius' behavior is fine too.

Pasting a tuplet accross a barline is not easy and lead to weird behavior and possible time corrupition.
Sibelius displays a message "You can't paste here as one of the tuplet would cross a barline"

Latest revisions do not crash or corrupt the score so i changed status to "normal". The behaviour should be enhanced by showing a dialog explaining the user what did not work.

Copying a 3 quarter tuplets on the last quarter of the same tuplets corrupts the score in r2923.
See http://screenr.com/IdM

I can reproduce #3 as well. See http://screenr.com/HxM
I put it back to critical. We can decide to postpone it if neceassary.

You are right. It does not work for copied ranges. I tested with copying only the half note which is a simpler case which works. The current "paste" code for selected ranges cannot be fixed easily so it may be better to postphone the fix.

Reactivating this critical bug report to see whether it still applies to the trunk.

The original bug is fixed on branch 1.2. It was fixed as part of
issue #14648: Copying from within a tuplet bracket to the same or another or to a normal rest produces strange results .
For the trunk a patch has been proposed at time of writing, but not been accepted yet.

The issue of comment #3 works fine in branch 1.2. The result is:
NOT FOUND: 01
If you start off with a chord, the result is this:
NOT FOUND: 02
The eighth note in the second bar offers a tuplet menu. If you delete the tuplet and undo, you get this:
NOT FOUND: 03

In the trunk, like in the branch, it works fine with a single note, but with a chord you get this:
NOT FOUND: 04

Actually, you don't even need to make the example so difficult.
Take a bunch of 1/8 triplets. Then copy a straight 1/8 chord from elsewhere onto one of the triplets.
You get this in branch 1.2 and trunk (the difference is that the trunk shows the "3" for the first two notes):
NOT FOUND: 01
NOT FOUND: 02
I should mention: The processing is in Score::pasteStaff (in cmd.cpp).

"Straight" ;-)

There is now a dialog telling you that you cannot paste into a tuplet.