Each time a record with a large text field is written to a file with active journalization, the current journal file becomes corrupted.
As we cannot predict the requested field length, we define such fields with a length of 100 bytes delimited and turn on "extended records". We must always lie about the length in such cases. PxPlus has no problems with file I/O to the data file. So the 100 bytes are a signal for the developer, that there may me much more bytes.
We expect the journal file to be written as properly as the data file. How can we avoid the problem?