Force encode with code 128 A or B

[AlessandroStaffolani]

Is there a way to force the barcode type code128 to encode using either code 128 B or code 128 A? At the moment it is automatically selected, but for some scanner it is required a specific version.

[metafloor]

The only way to force use of a single subset would be to manually encode the data and use the raw option. Looking at the BWIPP code, the binary text string should include the start codeword and the codewords for the characters. The BWIPP code calculates the checkdigit and adds the stop codeword. As long as the characters you are encoding are entirely in subset A or B, it is not difficult to map the characters to codewords.

[AlessandroStaffolani]

The only way to force use of a single subset would be to manually encode the data and use the raw option. Looking at the BWIPP code, the binary text string should include the start codeword and the codewords for the characters. The BWIPP code calculates the checkdigit and adds the stop codeword. As long as the characters you are encoding are entirely in subset A or B, it is not difficult to map the characters to codewords.

Thank you for your reply, is there any example of how to perform the manual encoding?

[metafloor]

Subset-B is straightforward to encode as it is just the printable ASCII characters, space through DEL, indexed from zero. Space is ASCII 32, so we just subtract 32 to get the zero offset:

// subset-B : printable ascii
if (/[^\x20-\x7f]/.test(text)) {
    throw new RangeError("subset-B: text out of range");
}
// 0x68/104 is the subset-B start codeword
let cws = '\x68' + text.replace(/./g, ($0) => String.fromCharCode($0.charCodeAt(0)-32));

bwipjs.toBuffer({ bcid:'code128', text:cws, raw:true, alttext:text, textxalign:'center', includetext:true })
    .then((png) => {
       // `png` is a Buffer
    })

The example above assumes node-js and uses toBuffer(). Replace with toCanvas() if on the browser.

[emikelsoft]

I just tried the solution above to force Subset B with bwip-js version 4.6.0 (2025-04-20)
But it is no longer working, I am getting the following error:

Cannot read properties of undefined (reading 'get')
at $get (node_modules\bwip-js\dist\bwip-js-node.js:557:14)
at bwipp_code128 (node_modules\bwip-js\dist\bwip-js-node.js:10900:28)
at bwipp_encode (node_modules\bwip-js\dist\bwip-js-node.js:42630:5)
at _Render (node_modules\bwip-js\dist\bwip-js-node.js:284:5)
at C:\test\bwip-js-node.js:121:17

[metafloor]

Just back from a short holiday. Will investigate this tomorrow.

[emikelsoft]

Just back from a short holiday. Will investigate this tomorrow.

Could you find some time to look into this?

[metafloor]

Meant to post something about this yesterday but forgot. This has turned into the mother of all bugs. I have a test case that is currently crashing nodejs (core dump). Planning to port the test case to the browser today to see if it can be reproduced there.

[metafloor]

@terryburton: I am going to punt on this one and ignore the nodejs core dump for now. Too overwhelmed with work presently.

I can reproduce the reported bug in the browser and in the BWIPP online demo. The call stack for the exception in the browser shows that seta is undefined. Looking at the BWIPP code, I see seta get defined for encoding new and legacy but not for raw.

For reference, this is the test case I am using:

Encoder: code128
Text: ^104^033^034^035^065^066^067^016^017^018^019
Options: raw parse

The encoded text is ABCabc0123. I am testing against version 2025-04-19.

[terryburton]

Indeed, it would be better if set{a,b,c} where defined in the encounter's context and charvals were to evaporate.

I'll take a look tomorrow.

[terryburton]

Fixed with bwipp/postscriptbarcode@4aca7c7