Musescore 4.0.2 file Corrupt and will no longer open

• Apr 11, 2023 - 12:13

The file spontaneously kicked me out and closed the application and when I went back in it said the file was corrupt and when I opened it anyway nothing came up. The error details were 'Error=XML_ERROR_EMPTY_DOCUMENT ErrorID=13 (0xd) Line number=0' This is really frustrating so any help would be welcome, as I've already attempted many fixed to no avail.

Attachment Size
Screenshot (69).png 24.66 KB
Something in D minor.mscz 41.35 KB

Comments

In reply to by OptimusChrist

I was able to open the included mscx file with 3.6.2 without any problems.
Attached is the file saved with 3.6.2. Only the text formatting from the title needs to be redone.
I hope it helps.

By the way: if you use Ctrl-K to enter the chords instead of staff text, they can be played - I modified the first measures that way.

Attachment Size
Something in D minor_3.6.2.mscz 28.49 KB

In reply to by OptimusChrist

MuS 4.x is quite new and still a bit unstable.
For serious work you should use the older version 3.6.2. You can copy the 3.6.2 scores into the MuS 4 folder and open them there too. Then you have the improved sound and engraving, but for writing 3.6.2 seems to be much more stable.
Maybe in half a year the bugs will be mostly gone. Test it from time to time.
And: make regular backups!

In reply to by HildeK

Strongly disagree. There are literally hundreds of known bugs in MuseScore 4 fixed in MuseScore 4. Certainly there are bugs in MuseScore 4 too, but in general, if you value stability and the ability to benefit from fixes and improvements and wish to be able to continue to enjoy those advantages going forward, you should absolutely be using MuseScore 4. Only people with very specific need for particular MuseScore 3 features not available in MuseScore 4 with no good workaround should resort to continuing to use MuseScore 3.

In reply to by Marc Sabatella

But when I see the almost daily requests for help because of corrupt files here and hours or days of work broken, I can't agree with you. Sure, some of it is also due to operator error.
Since 3.6.2 has been around, I have had no crashes from scores I created myself. Only occasionally it happens with those that I had tried to repair from already corruptet MuS 4.x mscx import files.
I don't deny that 3.6.2 still has a lot of minor bugs, but they rarely have dramatic effects. Some things just do not work as expected ...

In reply to by HildeK

4.0.2 is much more thorough about reporting corruption, which is why you see people asking about it more now. The vast majority of corrupted files reported in 4.,0.2 were actually corrupted long ago, either by earlier 4.0 builds (due to bugs already fixed) or in MuseScore 3 or earlier (due to bugs that still exist and are still corrupting files every time you use it). The 3.6.2 corruption detection capabilities were much more limited until now, so the majority of corruptions were never reported before.

Now, the issue discussed here is a different type of corruption - one in which the ZIP file itself is messed up. There reports are not very common, but check this forum and you'll see there are hundreds of such reports with MuseScore 3, and so far only a small handful with MuseScore 4. So far no one has figured out the cause of these, but they are statistically much more likely to happen with Msuecore 3 - probably by an order of magnitude or more.

In reply to by Jojo-Schmitz

That’s what I’m talking about here - scores like the one above that are broken as ZIP files (bad header, or good header followed by a bunch of zeroes). Those are the problems that prevent a score from loading at all and actually cause lost work. My sense is that this is much less common in MU4. Not that this is due to any specific fixes, probably just different versions of the libraries used for ZIP creation.

Ordinary score corruptions - too many/few beats in a measure - are far less serious. But here it’s a s I said - tough to get accurate data because 4.0.2 is far more aggressive about reporting problems that previously went undetected. My gut sense from following the various issues and PR’s is that the original 4.0 made these corruptions more common due to a couple of specific bugs bugs now fixed, and that at this point, once we factor out newly reported cases of corruption introduced in earlier versions, the actual incidence of new occurrences of these score corruptions is roughly unchanged over the past several years (since the big Fraction refactor, perhaps).

Do you still have an unanswered question? Please log in first to post your question.