We recently upgraded from OJS 3.1.1-4 to 3.2.1-1, but we’re not sure exactly when the problem began.
When using the direct URL for downloading a document, old PDF files created with PDF 1.3 are downloading as .txt documents. These would be downloads listed in page metadata as “citation_pdf_url.” The .txt document begins like this:
%PDF-1.3 %‚„œ” 167 0 obj <</Linearized 1/L 12398451/O 170/E 65690/N 28/T 12395063/H [ 916 670]>> endobj xref 167 31
Download from PDF viewer
The same older files DO download as PDFs from the viewer, but without a name.
These are PDF documents that have worked successfully in OJS 2 and 3, going back about a decade. There are no PHP errors associate with the issue. We are not having any issues with documents created with PDF 1.4+.
Is OJS 3.2.x having trouble “reading” the metadata of these older PDF files.
Indiana University Libraries