(Some?) text styles imported with wrong (Y?) offsets from 2.x scores

• Mar 16, 2019 - 17:24
Reported version
3.0
Type
Development
Frequency
Many
Severity
S3 - Major
Reproducibility
Always
Status
duplicate
Regression
Yes
Workaround
No
Project

At least some of the text styles found in a 2.x score are imported with wrong offsets. I noticed wrong Y offsets, as most X offsets are 0. For instance:

Subtitle has a 9mm Y offset in my 2.x scores and it is imported with 10sp (note the change of unit too!)
Lyrics odd/even lines 7.5sp => 0
Dynamics 7.5sp => 4sp
Tempo -4sp => -2sp
and so on...

A small test score from ver. 2.3.2 is attached.

I noticed that, at last, some of these styles are split between above and below position, but this does not seem a valid reason to ignore the original setting:
- practically all of them have a default / prevalent position (lyrics and dynamics are usually below and tempo is usually above) which should be read from the score, even if the other (above position for dynamics and lyrics, below for tempo) is set to some default value;
OR
- if the value set in the score is negative, obviously the intended element position was above the stave and the above value should be set to the score value with the below value set to some value; if the value is positive, the opposite.

So, there seems to be plenty of ways to deal with values from before the above/below split and it should be done, rather than filling the score with arbitrary values.

EDIT: There seems to be some issues with X offsets too: measure number was X -1sp / Y -2sp in ver. 2.x score and becomes X 0sp / Y -2sp (Y is correct, but X is reset)

Attachment Size
TextStyleImportTest.mscz 41.93 KB