Validation of extended CSV fields is not being done properly

[stsccfr]

I discovered two files in the Staffordshire database each of which contained an invalid extended CSV field, and yet the FreeREG server had accepted them. I then uploaded a baptism file, STSABCBA.CSV, which had BAPTISMS in the first header line, but whose FlexCSV fields were all for marriages (marriage_date, groom_name, etc). That file was also accepted by the server. Another consequence of the bad processing is that the date range reported for the file is wrong. The file contains a single entry dated 1608, but the date range reported by the server (in the email it sends out) is 2024-1300.

STSABCBA.CSV

I think the culprit is in the function for valid_field_definition() in lib/new_freereg_csv_update_processor.rb and is this line:

entry_fields = Freereg1CsvEntry.attribute_names

which defines the valid field names to be the entire collection of named attributes in the Freereg1CsvEntry class, which is all FlexCSV fields for all 3 event types (baptisms, marriages, burials), and anything else that happens to be defined in that class. This means that one can construct a CSV file using any random selection of the field names for the 3 events and the server will accept it. Proper validation of uploaded files is important especially given that they don't all come from a single source.

Instead of the class attribute names, I think we should be using the lists of event-specific FlexCSV field names defined in lib/freereg_options_constants.rb