What do you think about simply adding a configuration option to control the CSV delimiter? You could then subclass from the csv.excel dialect and override the delimiter attribute? We'd leave it as comma by default. Let's call the option "csv_delimiter".
What do you think about simply adding a configuration option to control the CSV delimiter? You could then subclass from the csv.excel dialect and override the delimiter attribute? We'd leave it as comma by default. Let's call the option "csv_delimiter".