Please attach the specific score you are having problems with and precise step by step instructions to reproduce the problem. Also say what OS you are using.
Thank you, but I've isolated the problem. It seems that the template file I was using has been somehow corrupted. I deleted and replaced the template file and that appears to have solved the issue.
I'm a little curious how this template file got corrupted. Could you post the template.
Even though the corrupted template was at fault, I would prefer musescore to handle this gracefully, for example by refusing to load the template file and producing a useful popup error message, instead of crashing upon editing a score based on that template.
I can verify that crash on Windows x86-64 on nightly dbd9e84 and 2.0.2 release when creating score with that template upon adding a 1st ending. However, I seem to be able to open that template mscz directly and then insert measures and add a 1st ending without causing a crash.
I notice in both 2.0.2 and nightly that when making a new score with that template that the Eb and Bb child scores are completely empty... So a likely explaination of the crash is that the 1st ending is trying to be added to NULL pointer Measure. Although the template file clearly has Eb and Bb child scores. I'm going to look into this a little more. Did you end up fixing that template so it is no longer corrupted? Please post that here if so.
Aside: seems like I useful template...I can't believe I haven't thought of making one like this
As a related question to any core devs out there: Are child scores for templates currently implemented behavior, or would that require a new feature request?
EDIT: I am able to create and save a mscz with multiple parts as a template and then create a new score from that template and add the 1st ending without crashing. So to answer my question, templates with parts is currently implemented behavior.
Thank you. I just fixed the problem by deleting the corrupted template file, taking one of the finished scores I had made using the template, deleting the measures and text from it, and resaving it as a new template. I have since continued to make new charts without incident.
A month ago I wrote "I can verify that crash on Windows x86-64 on nightly dbd9e84 and 2.0.2 release when creating score with that template upon adding a 1st ending."
But now I'm unable to reproduce that crash on latest self-compile master e666061 on Windows x86-64. Note the the C & Bb parts (child scores) still have no measures when I create new score from this corrupt template, so maybe there still should be a separate issue along the lines of "automatically verify template has no issues before creating score with it".
Maybe something got committed to master in the last month that fixed this crash. I don't know what, though, as I can't find anything in past month commit history that I think would have fixed this crash.
Comments
Please attach the specific score you are having problems with and precise step by step instructions to reproduce the problem. Also say what OS you are using.
Thank you, but I've isolated the problem. It seems that the template file I was using has been somehow corrupted. I deleted and replaced the template file and that appears to have solved the issue.
I'm a little curious how this template file got corrupted. Could you post the template.
Even though the corrupted template was at fault, I would prefer musescore to handle this gracefully, for example by refusing to load the template file and producing a useful popup error message, instead of crashing upon editing a score based on that template.
No problem. I had already deleted the file so I had to restore it, but after restoring it I tested it and it was still producing the same error.
I can verify that crash on Windows x86-64 on nightly dbd9e84 and 2.0.2 release when creating score with that template upon adding a 1st ending. However, I seem to be able to open that template mscz directly and then insert measures and add a 1st ending without causing a crash.
I notice in both 2.0.2 and nightly that when making a new score with that template that the Eb and Bb child scores are completely empty... So a likely explaination of the crash is that the 1st ending is trying to be added to NULL pointer Measure. Although the template file clearly has Eb and Bb child scores. I'm going to look into this a little more. Did you end up fixing that template so it is no longer corrupted? Please post that here if so.
Aside: seems like I useful template...I can't believe I haven't thought of making one like this
As a related question to any core devs out there: Are child scores for templates currently implemented behavior, or would that require a new feature request?EDIT: I am able to create and save a mscz with multiple parts as a template and then create a new score from that template and add the 1st ending without crashing. So to answer my question, templates with parts is currently implemented behavior.
Cameron Melendez, do you remember what steps you might have been taking when creating that template file?
You mean parts?
Yes. But as I remember, parts are implemented internally as child scores, right?
I don't really understand the question abotu child scores, but it's better to discuss that in the forum or maybe the developers mailing list.
To answer my question, yes indeed parts are implemented internally as child scores (see https://github.com/musescore/MuseScore/blob/master/libmscore/score.h#L3…)
Sorry, I will avoid using internal technical terminology when a user makes a forum post.
I'm sorry I'm afraid I do not remember making the template file, but I know that it functioned properly for quite a while before the problem came up.
Thank you. I just fixed the problem by deleting the corrupted template file, taking one of the finished scores I had made using the template, deleting the measures and text from it, and resaving it as a new template. I have since continued to make new charts without incident.
A month ago I wrote "I can verify that crash on Windows x86-64 on nightly dbd9e84 and 2.0.2 release when creating score with that template upon adding a 1st ending."
But now I'm unable to reproduce that crash on latest self-compile master e666061 on Windows x86-64. Note the the C & Bb parts (child scores) still have no measures when I create new score from this corrupt template, so maybe there still should be a separate issue along the lines of "automatically verify template has no issues before creating score with it".
Maybe something got committed to master in the last month that fixed this crash. I don't know what, though, as I can't find anything in past month commit history that I think would have fixed this crash.