PDF attached via XML file - ojs 3.5.0-3 error

[simgiallorosso]

Hi everyone, up until version 3.5.0-2, when we imported a volume/article with native XML, the PDF attached to the XML was called via a URL.
With version 3.5.0-3, however, this is no longer allowed; the PDF must be attached directly to the XML.

this is the error received:

The process failed. Check below for errors/warnings.
Errors occured:
Generic Items

• Filter (Native XML issue import) supports input schema(plugins/importexport/native/native.xsd) - string given

Generic Items Id: 1

• Line 2 Column 1: Extra content at the end of the document

• Line -1 Column 0: The document has no document element.

Validation errors:

• Extra content at the end of the document

• The document has no document element.

We have modified the code in the xml, but we have another error

The only problem is the size of the PDFs, which still blocks the process. The cdata tag is where you tell me to include the PDF, but then when importing it, the system says the PDF is too large to be uploaded this way.

The process failed. Check below for errors/warnings.
Errors occured:
Generic Items

• Filter (Native XML issue import) supports input schema(plugins/importexport/native/native.xsd) - string given

Generic Items Id: 1

• Line 51 Column 10000042: CData section too big found

• Line 51 Column 25171211: Sequence ‘]]>’ not allowed in content

• Line 51 Column 25171213: internal error: detected an error in element content

• Line -1 Column 0: The document has no document element.

Validation errors:

• CData section too big found

• Sequence ‘]]>’ not allowed in content

• internal error: detected an error in element content

• The document has no document element.

Why was this changed with the version change?
It was much better to attach the PDF that was called via a URL.

[asmecher]

Hi @simgiallorosso,

The change to restrict web-based uploads to embedded content only was done for security reasons (see your other thread). If you really need the previous behaviour, you could remove the check there – but this is a security risk, albeit low.

Your other option would be to continue to specify hrefs, but run the import using the command-line tools, where that is still permitted.

How big are your PDFs? We’ve worked with lots of embed-based imports, and never hit that limit. There is a mode to allow bigger embeds in the underlying XML toolset, but I would want to tread carefully on removing resource limits.

Regards,
Alec Smecher
Public Knowledge Project Team

[simgiallorosso]

20 mega byte is the file size