ledermann / datev Goto Github PK

Hi 👋

Thanks for the work on this gem! Since I'm updating this gem as a dependency right now I would love to see a CHANGELOG.md or similar to immediately see what changed. Maybe you could consider introducing this? Thanks!

The classes Datev::AccountHeader, Datev::BookingHeader and Datev::ContactHeader use Time.now.utc for Erzeugt am. I suggest that the timestamp is evaluated at runtime and not when evaluating the class. Otherwise it uses the timestamp when the server starts and not when generating the export.

You can fix this with a proc in each subclass of Header, for example:

self.default_attributes = proc do
    { ... }
end

And in the Datev::Base class the code below could be used to prevent breaking existing code:

attr_writer :default_attributes

def default_attributes
  if @default_attributes.is_a?(Proc)
    @default_attributes.call
  else
    @default_attributes
  end
end

What do you think?

We are a happy user of your gem and are in the process for the DATEV "Buchungsstapel"-certification.

The Feedback we got is:

• Versionsnummer must be 700 instead of 720:

  datev/lib/datev/base/header/booking_header.rb

  Line 5 in ac75ac0

  'Versionsnummer' => 720,

• Formatversion must be 9 to 7:

  datev/lib/datev/base/header/booking_header.rb

  Line 8 in ac75ac0

  'Formatversion' => 7,

Feedback in German:

Die Versionsnummer (Feld 2) des Headers beträgt zur Zeit 700. Sie haben 720 aufgezeichnet. Diese Nummer ist ungültig.
Die Formatversion (Feld 5) ist beim Buchungsstapel 9, wenn der Header die Versionsnummer 700 hat.

It's possible to overwrite the defaults in the constructor like

    Datev::BookingExport.new(
      "Versionsnummer" => 700,
      "Formatversion" => 9,
      ....
    )

but I question, whether the defaults should be changed?

The headers did not match the datev pattern, shown in the picture below.

I try to add the five missing headers and prepare a pull request.

We are running into issues when trying to import data via the ASCII Import of the 8.26 "Datev Rechnungswesen compact".
I noticed that the exported data is defined for version 7.2 so I guess the gem currently does not support the newest Datev version.

According to Datev Spec V.5.0

• values of type text must be quoted in the output csv
• if they do not contain a value - quotes must be present
• if the value of a text column contains a double quote it must be doubled

As reported by Datev-Prüfprogramm V5.0. According to the spec, the field "Kreditoren-Skonto 3" does not exist.

This is supposedly due to LF being used as the line separator instead of CRLF.

Another feedback we got from the DATEV-review

In den Belegfeldern sind folgende Zeichen zulässig:
Ziffern: 0 1 2...9
Großbuchstaben: A B C...Z
Kleinbuchstaben: a b c...z
Sonderzeichen: $ & % * + - /

Do you think it would make sense to validate or sanitize the input in the gem directly?
My proposal would be to throw an exception if the input contains invalid chars and leave the sanitization (Umlaut-Replacement and the like) to the caller.