I believe there is a problem with the way pdfTeX constructs the CharSet for a subsetted font. Not all fonts are affected, and it only shows up when attempting to validate for PDF/A. I’ve mentioned the problem to Thanh and others, but as there’s been nothing done, I’ll bring it up here, with an example that cannot be worked-around.

Attached are two PDF/A documents, using Hebrew text.

The main font is the “David” style font from the culmus package:   DavidCLM-Medium

hebart-david.pdf  — produced by pdfTeX

hebart-gs.pdf — produced using  dvips+GS

hebart-david fails PDF/A validation in one aspect only (see image):

    “CharSet in subset font is incomplete (font contains glyphs that are not listed)"