When calling a link to a PDF export directly without an active browser session the request will fail with a 403 HTTP status code (e.g. sending http://confluence/spaces/flyingpdf/pdfpageexport.action?pageId=1234567 to somebody via eMail). It seems to be caused by Confluence adding a ";jsessionid=..." to all request coming the first time without an active session and flyingpdf interpreting this as filename.

To clarify: This is not caused by permission restrictions for a) I made sure that anonymous has read access and is allowed to export the space and b) it also happens even if giving an administrative account to the URL (e.g. http://confluence/spaces/flyingpdf/pdfpageexport.action?pageId=1234567&os_name=abc&os_password=abc).

Rewriting ";" to "?" in Apache conf works around the issue, but that's very dirty (probably with side-effects).

Also see related question at https://answers.atlassian.com/questions/212249/pdf-retrieval-issue-via-webrequest-issue-after-upgrade-from-confluence-3-5-17-to-5-1-5