Remove invalid syntax for $mode when using a country/language code

[jakejackson1]

The syntax en_GB is invalid for the $mode. Based on the PDF 1.6 specification, the format should follow RFC 3066 en-GB (the code is used in the /Lang entry).

RFC 3066 has been revised multiple times, and looks to fall under the BCP 47 language tag standard now. The PR also adds a link to documentation on the BCP 47 language tag format.

References

• Page 865 of the PDF 1.6 Spec: https://opensource.adobe.com/dc-acrobat-sdk-docs/pdfstandards/pdfreference1.6.pdf
• IETF language tag: https://en.wikipedia.org/wiki/IETF_language_tag
• RFC 3066: https://www.rfc-editor.org/rfc/rfc3066
• BCP 47: https://datatracker.ietf.org/doc/bcp47/
• Line 10 of mPDF's LanguageToFont class: https://github.com/mpdf/mpdf/blob/development/src/Language/LanguageToFont.php#L10
• Lines 333-337 of MetadataWriter: https://github.com/mpdf/mpdf/blob/development/src/Writer/MetadataWriter.php#L333-L337

[finwe]

The question is, couldn't mPDF just process en_GB to en-GB internally?

[jakejackson1]

The question is, couldn't mPDF just process en_GB to en-GB internally?

That was my first thought: just do a quick str_replace from underscore to hyphens. But after considering the complexity of language and font processing in mPDF, I didn't want to risk altering functionality in the library without 100% understanding the potential side effects. It was much easier doing a preliminary investigation and updating the docs to better reflect the intended input.