https://bugs.documentfoundation.org/show_bug.cgi?id=32249
--- Comment #49 from Eyal Rozenberg <[email protected]> --- (In reply to Dave Gilbert from comment #46) Indeed, the two filters are not that far apart - although they should be in principle, because we do different kind of work with Draw and with Writer. I would like to note, that we don't have to immediately make the jump from "we just can't do it" to "we do it perfectly". We could start a naive algorithm, which assumes no boxes are "real" out-of-order boxes; tries to find common baselines for boxes on a line, forms lines, applies a heuristic to when a paragraph ends (indications may include: v-spacing, last line terminate earlier, next line starts with capital etc.) - and that already gives users something workable to start from, with them fixing the rest manually. And of course, we could adopt approaches taken in other applications like Okular. Finally - some of this logic is also relevant for draw; because while in Draw we don't want all text on all pages to be one long continuous flow, we still want to be able to have larger boxes with complete paragraphs, or multiple paragraphs - which do exist in Draw (and Impress/PowerPoint, frequent sources of PDFs). (PS - this disucssion would probably be a better fit in bug 151577) -- You are receiving this mail because: You are the assignee for the bug.
