Possible Regression in Code128 (GS1 FNC1)

[dmk2014]

Hello,

We are using this library with great success for various symbologies - thanks very much for your work!

Description

We have run into an issue encoding GS1 data using the Code128 symbology in the latest version (3.1.5 at time of writing). The library uses Convert.ToChar(200) to represent FNC1 (or È). Sample data to encode (including both the leading FNC1 and an FNC1 after the variable length batch number application identifier): È10222333È15100701

Version 2.4.0:

Version 3.1.5:

From a brief review of the code it looks like the behaviour changed in 8c9d351 where the lookup tables were modified. Other characters are impacted in a similar way, but the relevant change for FNC1 was:

Before:
C128_Code.Rows.Add(new object[] { "102", FNC1, FNC1, FNC1, "11110101110" });

After:
entry(102, "FNC1", "FNC1", "FNC1", "11110101110");

Attempted Solution

Modified the lookup table entry as suggested in #179 to:
entry(FNC1.ToString(), FNC1.ToString(), FNC1.ToString(), "11110101110");

This results in:

Summary

• Am I correct in my assumption that this is a regression or was the change as intended?
• The attempted solution above is somewhat naive
  • Has not been tested beyond basic encoding checks with the sample app
  • Unsure if it introduces side effects elsewhere in the encoding
  • Doesn't account for other changes to the lookup table (FNC2, FNC3, FNC4)

[dmk2014]

There is also a Java implementation of barcodelib - https://github.com/barnhill/barcode-java/.

Its Code128 implementation adds FNC1 to the lookup table as barcodelib did in versions prior to v3. See Code128.java#L264:

this.C128_Code.add(new Entry ( "102", getChar(200)/*FNC1*/, getChar(200)/*FNC1*/, getChar(200)/*FNC1*/, "11110101110" ));