When exporting to DXF (for example, to create a marker/cut file in different software), DXF works pretty well since the addition of passmarks.
However, any text added to labels in pattern pieces, are converted to lines, and are then rendered unreadable by the receiving/importing software. This then leads to manual work where you have to designate which fabric and how many pieces need to be laid on the marker. For volume work this is too much work and thus not practical in real-world usage of Valentina.
The text right now looks very "nice". It doesn't however, need to look nice. Real, sloppy, un-anti-aliased looking text would suffice and would probably be easier to convert to MTEXT in DXF.
This way, valentina could provide easily usable data to other systems which posess capabilities valentina doesn't have (yet! ;))
Attached is also an example of how text is usually detected (this from optitex)
^ mtext definition
Note for myself: This task will require redesign in two places: Class VTextGraphicsItem should start probably using QGraphicsTextItem for showing text. Redesign part that draw text on layout detail.