MuseScore 3.6 crashes on open of a file with start repeat in continuous view

• Jan 14, 2021 - 16:50
Reported version
3.6
Priority
P0 - Critical
Type
Functional
Frequency
Many
Severity
S2 - Critical
Reproducibility
Always
Status
closed
Regression
Yes
Workaround
Yes
Project

Install MuseScore 3.6 from https://github.com/musescore/MuseScore/releases/tag/v3.6
(I used the MSI)

Open any of the attached files.
Observe crash.

Video of crash: https://www.youtube.com/watch?v=LnmjtHtCrCI&feature=youtu.be


Comments

Severity S1 - Blocker S2 - Critical

I can confirm the crashes with these files in particular. Issue in both cases seems to have to with what appears to be invalid content that crashes when we try to decide if it's a clef or not. I think this is connected to the start repeat in bar 21 - it causes the previous measure to not have an end barline, and crash happens in a place where we aren't prepared to handle a measure with no end barline.

In reply to by Marc Sabatella

Regression Yes No

I have this issue as well with a big file (c. 250 KB). Start repeat is around 12 measures in to the piece. Just updated to 3.6 and this issue affected me for the first time. Tried creating a piece in 3.6 and adding start repeats but the bug didn’t affect me so presumably it only affects files going from pre-3.6 to 3.6 versions.

If you could fix this ASAP, that would be amazing! This is a file which means a lot to me and to basically have it lost forever, especially because I was planning on working on it soon, would be heartbreaking.

Workaround No Yes

Workaround (not for the faint at heart): Extract the mscx from the mscz (like using 7Zip), put it into a plain text editor (like WordPad, or NotePad++) and you'd see something like this:

<?xml version="1.0" encoding="UTF-8"?>
<museScore version="3.02">
  <programVersion>3.6.0</programVersion>
  <programRevision>10dee08</programRevision>
  <Score>
    <layoutMode>line</layoutMode>
    <LayerTag id="0" tag="default"></LayerTag>
    <currentLayer>0</currentLayer>
    <Division>480</Division>
 

Delete just this one line:

    <layoutMode>line</layoutMode>

Save the file, move it back into the mssz (or just open it directly in MuseScore, if the score didn't contain any images)

I'm sorry, but I can't explain the workaround simpler than I did. And I did mention that it isn't something for the 'faint at heart' ;-)
It isn't rocket science either though, and if you can't afford to wait for a fix and don't want to fall back to 3.5.2, this is what you'd need to do (or get done by someone)

But here's another and easier workaround: get the 3.5.2 PortableApp and 'install' it, open the score with that, switch to page view, save, quit.
Now 3.6 should open it without a crash

I don't think the issue is with the continuous view but rather with a file that has instruments broken-down into Parts. I have a file created with 3.5 with continuous view and it opens fine, but another file with continuos view and instrument Parts keeps crashing.

See title - the issue is specifically about files in continuous view that contain start repeats. Parts shouldn't be relevant. Or if you have a file with no start repeat that crashes, it's a different bug, and we'd need you to attach the score so we can investigate.

Could you also re-download an older version (3.5.x), open the file in said older version just to remove the repeats, then delete the older version and re-install 3.6.0 before re-adding the repeats? I tried uninstalling and reinstalling 3.6.0 last night before I knew what the issue was so I’m no stranger to the process. Obviously you would have to avoid opening any files already opened in 3.6.0 (I worked on a few other files with no issues in the new update last night). My only concern is whether or not the fact that I merely tried to open the affected file in 3.6.0 would make it incomparable with 3.5.x versions.

I’d certainly rather do something like that, which I’m familiar with, than do something more complicated though!

That score does have repeat barlines (here for example a start repeat in measure 10), having them at the start is not needed for 3.6 to crash on them in continuous view

I had the same problem. Every time I tried to open MuseScore 3.6, it would crash even before opening the startup center. I tried reinstalling it, and it works fine now. So, I don't know what could have happened. Possibly an incomplete installation? I uninstalled the previous MuseScore version too. So, I don't know if doing that fixes it. Glad it kept my soundfonts though.

In reply to by Jojo-Schmitz

I'm still confused. What should I do to open recent files of mine? I don't have a problem uploading other scores with repeats in them. Only the most recent scores i have been working on. I tried downloading 3.5 back but i t just downloads 3.6 instead.

You can use one the the numerous workarounds mentioined here to fix them yourself. The easiest is probably to grab and unpack ('install') a 3.5.2 PortableApp (if you're on Windows)

In reply to by Jojo-Schmitz

That PortableApp workaround worked for me. I think, like some of the other posters, that it has to do with repeats, not parts. I haven't created parts for my score, but I have got two repeats, one with first and second time bars. None of the other scores I've opened have repeats, and all of them open OK.

Of course it matters. Would you let a very large number of users suffer the inconvenience of this issue, for eample, for several weeks, or a few months, as if nothing had happened?
This is a use case (score in Continuous Display, and with repeats) so frequent and basic hat one is entitled to expect it to work without any failure from a "major" program such as MuseScore.

Doesn't make it a blocker here in the issue tracker and now chaning any of the properties of this issue though is not going to change anything at all. So, no, it doesn't matter at all

It should have made it a release blocker though (bit that's not what this severity field is about), I 100% argee with you here. It is just too late for that...

My comment was not a request to change the status of this issue. It's an internal kitchen, it's secondary. By referring to "blocker", I meant that it becomes urgent that 3.6.1 be published. So it's no longer "When it's ready", it should be " Right now and now"!

(redirected from https://musescore.org/en/comment/1056768#comment-1056768)

I had a score that did the same! I reported it on a different thread, unaware that what I encountered was unrelated to that issue and instead related to this issue. (You can check that thread if you want to examine my score.)

Luckily, Jojo-Schmitz was kind enough to fix my score in Musescore 3.5.2 to be transferable to 3.6.0. When I downloaded the score, it was in Page View as I expected. I was able to switch it to Continuous View just fine. However, when I made a change and saved the file in Continuous View, then tried to reopen the file later, it crashed again. I redownloaded the file and tried making a change and saving it in Page View, and it works with no issues.

For reference, one of my other scores is set to Continuous View, and it made the transfer between versions just fine, even opening in Continuous View by default. However, this score just seems to have issues with being saved in Continuous View.

EDIT: I noticed that one of the instruments in this score doesn't play at all unless you click on one of its notes first, and then it will play. However, this instrument is also no longer obeying any dynamics markings or velocity changes that I have placed onto it before or after the update, always at the default volume instead. Perhaps this could be a corrupted instrument that partly led to the original file crashing? (It's the Soprano Saxophone, if you have my score.)

Oh, so it was! I had no idea, since the "Check for Update" button in the Help bar still said that I was up to date with 3.6.0 just 10 minutes ago. I looked up the new version online, downloaded and installed it, and the crashing is fixed now. Thank you all for quickly addressing these issues!

(The other issues I mention at the end of my previous post still exist, and I may make a separate thread for them. I found a way to fix the dynamics issue, but it involves changing a setting that causes undesirable outcomes elsewhere in the score, which wasn't necessary before. It may have been an intentional change with unintended consequences, or it might be more case-specific than I'm realizing.)

Fix version
3.6.1