Opening 1.3 score with Beam Middle set on full measure rest crashes
A little investigation shows the problem stems from the Oboe 3 part, measure 63. The first eighth rest and all the following rests have the "beam middle" property set, which is kind of odd. This is also why, BTW, the eighth note on the downbeat is missing its flag when the score is loaded into 1.3.
If you go to 1.3, select all the rests in that passage, then double click the "no beam" icon in the Beams palettes, you can save the file and it fixes both problems. That is, the flag reappears in 1.3, and 2.0 won't crash on load.
Not sure what you mean about the time signature. I don't see any issues in either 1.3 or 2.0. Do you perhaps mean the text element in the title box that is designed to look like a time signature? It's true that the manual positioning of the text elements within the title box was not preserved well. That's worth submitting as a separate issue.
Hmm, I think I somehow accidentally edited my response when above I meant to post a new one.
I added the the paragraph about the time signature, so see above for that. I also thought I added a paragraph saying that I didn't post the previous response to say this isn't a bug. I just wanted to record useful info for whomever ends up fixing it.
I would observe this bug won't happen very often, since setting that property would be a very senseless thing to do. Was this a score you created from scratch, or did it start off as an import from MusicXML? If the latter, can you post the original? We should see if the problem with the score was created during import or if was the result of something you or whomever created the score did.
Also, I have now ascertained that attempting to create this situation - beam middle set on full measure rest - crashes 2.0 in exactly the same spot, same reason:
1) create score with four measures in 2/4
2) add a single eighth note to beat one of the first measure
3) click the eighth rest in bar 1, shift click last full reasure rest
4) double click Beam Middle icon in Beams toolbar
This score was created with an XML import in 1.3. I modify some things and save it in 1.3 . I attach the xml file to this post. There is others bugs in XML import. Slurs are not correctly imported in 2.0 (in 1.3 work and also in finale). I put an new issue this afternoon.
For the key signature i create an other bug issue soon. But prior i want to isolate it in a small musescore file.
Expected in 2.0 and rigth in 1.3
In 2.0
Sorry, i make a mistake previously telling you that it's a key signature issue, it's a clefs issue
In 2.0 clefs are on the right in the measure, i can remove this bug removing time signature, putting clefs and then add time signature.
I don't think you can simply prevent setting beam properties on rests. That's how you create beams *over* rests (eg, quaver, rest, quaver in 6/8). Full measure rests, now maybe those shouldn't have beam properties.
So, in fact, I haven't prevented beaming over rests, I have prevented the crash, although the additional clef problems are still present, I imagine that is a completely separate issue.
How do I resubmit this patch, since Travis failed with an apparently entirely unrelated issue.
I manually fixed this bug. At least there is no crash anymore.
It seems that MuseScore does not handle some combinations of beamed notes + rests.
Eventually i will prepare some regression tests for this.
Comments
I have no idea if it can be useful. Contrary to the usual: I saved with 1.3 as XML; Open with dfab91 and saved. ?
Thank's but i post this bug to have a fix in 2.0.
I know it's possible to work around this issue ;-).
Robert
A little investigation shows the problem stems from the Oboe 3 part, measure 63. The first eighth rest and all the following rests have the "beam middle" property set, which is kind of odd. This is also why, BTW, the eighth note on the downbeat is missing its flag when the score is loaded into 1.3.
If you go to 1.3, select all the rests in that passage, then double click the "no beam" icon in the Beams palettes, you can save the file and it fixes both problems. That is, the flag reappears in 1.3, and 2.0 won't crash on load.
Not sure what you mean about the time signature. I don't see any issues in either 1.3 or 2.0. Do you perhaps mean the text element in the title box that is designed to look like a time signature? It's true that the manual positioning of the text elements within the title box was not preserved well. That's worth submitting as a separate issue.
Yes with this, it's work. :-)
a fix in 2.0 may be necessary to avoid crash on another score.
PS : Strange position of the time signature on first measure.
Hmm, I think I somehow accidentally edited my response when above I meant to post a new one.
I added the the paragraph about the time signature, so see above for that. I also thought I added a paragraph saying that I didn't post the previous response to say this isn't a bug. I just wanted to record useful info for whomever ends up fixing it.
I would observe this bug won't happen very often, since setting that property would be a very senseless thing to do. Was this a score you created from scratch, or did it start off as an import from MusicXML? If the latter, can you post the original? We should see if the problem with the score was created during import or if was the result of something you or whomever created the score did.
Also, I have now ascertained that attempting to create this situation - beam middle set on full measure rest - crashes 2.0 in exactly the same spot, same reason:
1) create score with four measures in 2/4
2) add a single eighth note to beat one of the first measure
3) click the eighth rest in bar 1, shift click last full reasure rest
4) double click Beam Middle icon in Beams toolbar
Hello,
This score was created with an XML import in 1.3. I modify some things and save it in 1.3 . I attach the xml file to this post. There is others bugs in XML import. Slurs are not correctly imported in 2.0 (in 1.3 work and also in finale). I put an new issue this afternoon.
For the key signature i create an other bug issue soon. But prior i want to isolate it in a small musescore file.
Expected in 2.0 and rigth in 1.3
In 2.0
Sorry, i make a mistake previously telling you that it's a key signature issue, it's a clefs issue
In 2.0 clefs are on the right in the measure, i can remove this bug removing time signature, putting clefs and then add time signature.
I put a new issue also for that in afternoon
Thank's
Robert
This patch should prevent the crash reported.
https://github.com/musescore/MuseScore/pull/665
It stops MS trying to add a beam property to a rest, or trying to lay out a rest with a beam property.
I don't think you can simply prevent setting beam properties on rests. That's how you create beams *over* rests (eg, quaver, rest, quaver in 6/8). Full measure rests, now maybe those shouldn't have beam properties.
So, in fact, I haven't prevented beaming over rests, I have prevented the crash, although the additional clef problems are still present, I imagine that is a completely separate issue.
How do I resubmit this patch, since Travis failed with an apparently entirely unrelated issue.
Look again: Travis is happy with your PR meanwhile
I manually fixed this bug. At least there is no crash anymore.
It seems that MuseScore does not handle some combinations of beamed notes + rests.
Eventually i will prepare some regression tests for this.
Where is that fix? I don't see any corresponding commit.
Ah, there it is: guthub-musescore-musescore-51d8cbdc
Automatically closed -- issue fixed for 2 weeks with no activity.