https://bugs.documentfoundation.org/show_bug.cgi?id=167246
V Stuart Foote <[email protected]> changed: What |Removed |Added ---------------------------------------------------------------------------- See Also| |https://bugs.documentfounda | |tion.org/show_bug.cgi?id=89 | |727 Severity|minor |enhancement CC| |[email protected], | |[email protected], | |[email protected] --- Comment #2 from V Stuart Foote <[email protected]> --- Actually we do the correct thing. The pdfium based insert filter is minimalistic. It only parses the first page of a PDF into a BMP (resolving to a system PDFIMPORT_RESOLUTION_DPI variable). When embedded the BMP generated is both the preview and the source image--size recorded to ODF archive pictures directory *doubles* as two copies of the image are held. When an image is inserted as linked, the "preview" is still generated, but the ODF does record the path to source of the PDF. The resulting ODF is 1/2 the size, just a single copy of the BMP recorded to archive as .PNG at the resolution specified. Guess the request might be to bypass recording a preview (or use the thumbnail view logic of the start center for the preview) and only record the BMP as .PNG when the PDF is embedded. And then for the linked PDF image source--refactoring would be needed to reparse the source PDF on reload and only then might a very small archive result (the XML and a low res thumbnail actually present in the archive). Of course I'd still love to see progress on a skia SVG canvas for the pdfium to write out vector renderings of the PDF as a alternative to handling as BMP image. -- You are receiving this mail because: You are the assignee for the bug.
