Segmentation fault on drag/drop volta on last measure end

• Jan 30, 2014 - 15:17
Type
Functional
Severity
S2 - Critical
Status
closed
Project
MuseScore

mscore version: 03e0cf9

OS: Linux OpenSuSe 13.1

When dropping a line (Volta) onto the last measure of a score and the measure is a 'short' measure, a segmentation fault occurs.

1. Create test score, C major, time sig. 6/4.

2. Add 6 quarter notes to the first measure and (only) four to the second measure.

3. Delete all remaining measures leaving only the two with notes.

4. Edit the properties of the second measure to have only 4 Actual notes.

5. Open the master palette and select 'Lines'.

6. Drag a volta across to the score and drop it onto the right hand end of the final measure (where the 'end' barline is).

7. Segmentation fault:-

/home/turtle$ mscore
false
QDateTime("") QDateTime("Thu Jan 30 09:27:32 2014")
internal error: shortcut not found

cannot find shortcut
internal error: shortcut not found

cannot find shortcut
TextStyleType not found
Articulation::setSubtype: empty subtype
Articulation::setSubtype: empty subtype
setDynamicType: other
Marker::read type 0
Marker::read type 1
Marker::read type 2
Marker::read type 3
Marker::read type 4
Marker::read type 5
Marker::read type 6
loop toggled 0 - 0
QWidget::setMinimumSize: (inspector/Ms::Inspector) Negative sizes (-331,0) are not possible
setDynamicType: other
QXcbConnection: XCB error: 3 (BadWindow), sequence: 1463, resource id: 0, major code: 2 (ChangeWindowAttributes), minor code: 0
QXcbConnection: XCB error: 3 (BadWindow), sequence: 1465, resource id: 0, major code: 20 (GetProperty), minor code: 0
Segmentation fault
/home/turtle$

The segmentation fault does not happen if the line is attached to the left hand end of the final measure, nor does it happen when the measure has a full complement of notes ( six quarter notes in this case). It seems to be only on a 'short' last measure.

Attachment Size
test.mscz 1.33 KB

Comments

• Mar 25, 2014 - 12:21

I have been waiting to see 778 appear in the git master so I can update and test. To date I have not seen it, but I will continue to check and let you know after I get it.

Reply

• Mar 31, 2014 - 20:59

if a have understood,

Why having a 1st volta on the last mesure .....

you should have a third one (with 6/4 or other actual value) and then test ?

1rst volta and second volta ...?

did i get it ?

Reply

• Apr 1, 2014 - 13:55

@Lamardelmy: My music theory knowledge is insufficient to answer your comment. I only know I could and that it resulted in a segmentation fault crash. If the action of adding the volta is incorrect, then the program should detect it and give an error message - not just crash.

Still waiting for the patch to appear in the Git.

Reply

• Apr 1, 2014 - 17:31

I may be wrond, did you use
http://musescore.org/en/handbook/volta
(located in lines as I thought you should have two and you should use between tehm the second one on the second line of the first image from that page http://musescore.org/en/handbook/bar-line that means take it again ;) and instead of the mesure of "1rst volta" go to "second volta" that should be after your "reprise" bar,
It is hard for me to explain music theaory in english.... hope it helped ;)

Reply

• Jul 23, 2014 - 13:05
Status (old) fixed ⇒ active

Nightly (b94670c) / Windows 7

Of course, the practice is to release the Volta lines when the anchor appears at the beginning of the measure, and the line is then automatically placed to the fullest measure.

But, in this case
1) Stretch the line until the right hook is positioned at the opposite of the last beat (not before, but exactly above this last beat, and more to the right, of course)
2) Release

Result: crash

Like on the attachment from « Promenade »

Volta crash.jpg

Attachment Size
Volta crash.jpg 21.38 KB

Reply