Invalid language escape sequences in text strings not detected

[a20dev]

verapdf-arlington 1.29.166 does not detect invalid language escape sequences in text strings.

esc1.pdf

[bdoubrov]

Strictly speaking, UTF8 encoded string would be recognized as PDFDocEncoding in PDF 1.7 or earlier. But it would not result in any syntax errors.

[bdoubrov]

@a20dev Sorry, I realized I might have misinterpreted the issue. Do you mean that some of the language escape sequences might correspond to the codes that are not defined by ISO 639-1(2) / 3166-1 ? These codes are subject to dated revisions, and, for example, PDF 2.0 does not even fix the year for 3166-1. So, any such validation would need to be updated each time the new revision is published. I don't think this is a good idea

[a20dev]

In esc1.pdf (attached above), the TU entry of field dictionary 54 has five characters (fr-FR) instead of four characters (frFR) between the two U+001B characters.

[a20dev]

This syntax differs from the syntax of the Lang entries, so this mistake is easy to make.

[bdoubrov]

Makes sense! We'll try to add this check to the veraPDF Arlington implementation

[MaximPlusov]

We added the warning for this case.