> ich bekomme folgenden Fehler beim Import einer musicxml-Datei: > NameError: name 'mxl_note' is not defined. Did you mean: 'Xml_node'?
Please send me (privately) a PDF file that shows how the MusicXML file should actually look like (i.e., how the original version in Capella looks like). Furthermore I ask to check how key changes are stored in Capella – the script `CapToMusic.py` (that converted this report's MusicXML input) inserts a zero-duration bar if the key changes; this is neither musically correct (since the measure counter becomes wrong), nor technically (zero-duration rests are invalid, as described in a previous e-mail). Maybe there is a newer version of `CapToMusic.py`? Otherwise I suggest to write a bug report to Capella... * * * See https://gitlab.com/lilypond/lilypond/-/merge_requests/2704 for fixes that make `musicxml2y` finish without aborting. Irrespective of that, the script cannot convert the original input file – in its current state – to something useful. First of all, there are glaring bugs, for example, bar 59 has a 2/2 time signature but only three quarter beats. There are many other bars with similar problems in the various voices of the piece, i.e., the length of the contained notes or rests in a measure don't fit the length as specified by the current time signature. This is true in particular for suppressed staves; `musicxml2ly` expects that even such voices are musically valid (think of producing parts, for example); it's not sufficient to have a good 'visual' result of the full score. Another peculiarity is that time signatures are repeated again and again in the MusicXML data after a line break. This is not what should be done in MusicXML. I've stopped having a closer look after seeing all these problems. I hope that Capella has a tool to validate its input to catch such bugs. Maybe conversion is possible after fixing them.