https://bz.apache.org/SpamAssassin/show_bug.cgi?id=8424
Kent Oyer <[email protected]> changed: What |Removed |Added ---------------------------------------------------------------------------- CC| |[email protected] --- Comment #2 from Kent Oyer <[email protected]> --- Created attachment 6093 --> https://bz.apache.org/SpamAssassin/attachment.cgi?id=6093&action=edit proposed patch Yes I remember seeing this bug years ago. I'm surprised it hasn't been fixed already. The attached patch implements your suggestion and all tests pass. We're in the middle of a new release so we'll need at least 3 committers to vote yes in order to get this in the new release. I vote yes. Note: the upcoming release (4.0.3) implements a new "handler" framework and the PDF handler parses the full PDF object structure rather than scanning line by line as the plugin does. As a result, it can detect images in PDFs even when the content is compressed or encrypted (but not password protected), and pass them to the image handler for OCR. The PDF handler will be disabled by default and and does not replace or interfere with the PDFInfo plugin. Existing rules such as GMD_PDF_SQUARE still require PDFInfo and are not affected by it. To take advantage of the handler, you'll need to enable it and write rules that use it. -- You are receiving this mail because: You are the assignee for the bug.
