PDF convertion to PDF/A-1a with ConvertErrorAction.Delete parameter

What exaktly will be deleted if ConvertErrorAction.Delete is used during conversion to PDF/A-1a?
Example:
pdfDocument.convert(log, PdfFormat.PDF_A_1A, ConvertErrorAction.Delete);

The documentation states “You can also specify the action for the elements which can not be converted using the ConvertErrorAction enumeration” and it’s difficult to understand the consequences of using that option

Can images or text in the resulting PDF be deleted or is it internal constructions that is not allowed accoring to the PDF/A-1a?
Are there specific clauses or groups of clause in the PDF/A-1a specification that is considered during the deletion?

@010101
Judging by the code:
If ConvertErrorAction.Delete is specified when calling the Convert method, then in some situations certain objects will be deleted/changed from the converted document.
(as opposed to ConvertErrorAction.None, when these objects remain but the result will obviously not fully comply with the standard).
The corresponding entry is writed into the log file.

Such objects include:

  • EmbeddedFiles and JavaScript names in documents Catalog
  • Additional actions for ‘Widget’ and ‘Field’ dictionaries
  • Non-symbolic TrueType font that contains Encoding entry with Differences array which is prohibited for these kinds of fonts. (another font is created).
  • The transparency is prohibited