some progress here: if I replace german Umlauts and other special characters inside the data fields in the exported html from roboform, I'm able to import the logins step by step. By now I reached a number of 800 imported logins.
Because not only german Umlauts (ä,ö,ü,ß) are reasons for the "nothing to import"-condition, I had to import step by step, and found datasets from roboform with "graphical dots" (Unicode+2022) as password representation and others. One dataset of a login starts with
<DIV class="floatdiv orph">
so I get closer when I choose a subset of my data and and, if it failes to import, removed the last half of the subset, repeating with removing the last half of the subset from the not properly imported subset and so on.
(example: choose aprox. 1000 lines of the html, tried to import, failed, removed the last 500 lines, tried to import, removed the last 250 lines from the remaining 500 lines from second try an so on)
regards