Slurs lost on exchange voices (if no note in other voice at anchor point)

• Dec 23, 2014 - 17:14
Type
Functional
Severity
S3 - Major
Status
closed
Project
MuseScore

Ubuntu 14.04, GIT commit: bf97799

1) open attached score
2) select all
3) edit / voices / exchange 1-2

Result: first slur remains, but second is lost. Actually, "lost" is a good word. It is apparently still in the score, but no longer attached properly. Certain operations performed on the score afterward yields console errors about "zero slur", although I don't know exactly how to reproduce that.

Comments

• Dec 23, 2014 - 17:49

This is off topic marc but I was wondering whether your simply forgot to attach the score, or whether something went wrong with the website. The latter happened in the past, hence my question.

Reply

• Dec 23, 2014 - 19:21

Ok, Marc, I can easily reproduce this issue.
I am trying to locate. At the moment, works on August 21, but fails on August 28. Yet a little while, and I may clarify.

Reply

• Dec 23, 2014 - 19:32

So, after investigation, this issue appears on August 22

- Works from this Nightly: https://github.com/musescore/MuseScore/pull/1198
August 22 3d172d7.jpg
- Failure with this one: https://github.com/musescore/MuseScore/pull/1196
August 22 b4dbb5a failure.jpg

So, the cause is most likely a side effect of these commits (the one or other, or both?)

- https://github.com/musescore/MuseScore/commit/24facb5918b87c0cc8878b3b4…

- https://github.com/musescore/MuseScore/commit/f2ea47bc3a4ea3fc22a8311b4…

Attachment Size
August 22 3d172d7.jpg 10.66 KB
August 22 b4dbb5a failure.jpg 10.77 KB

Reply

• Dec 23, 2014 - 20:25

"narrowing it down" is an English expression for the sort of thing you do so well - we initially have a problem where we don't know where to begin looking, but you make the problem "narrower" by restricting the number of places we need to look :-)

Reply

• Dec 24, 2014 - 00:28

The problem appears to be that in cmdExchangeVoice, we need to go through the spanners and fix the track info for all and/or slur segments. However, the details of how this all works - including making sure the slur segments match the slurs - is pretty tricky and not something I am very familiar with, so I will leave it to someone else.

Reply