https://bugs.kde.org/show_bug.cgi?id=425148
--- Comment #7 from jpror...@gmail.com --- (In reply to Albert Astals Cid from comment #6) > > To be fair we support upgrading from such old versions, one could argue that > the annotation timestamp being changed is even correct since now it's inside > the PDF and previously wasn't, but i guess one could argue that this is a > technicality and that the annotation was really added in the past to the > users point of view. Being the user, I naturally favor the second interpretation. ;) In that line of thinking, the annotation is a data object created by the user. That it was once stored in an an okular xml file and is now stored in the pdf itself is really of little importance to the user. It's the data that I created that I care about. In the current implementation, it's not just the timestamp that I would lose. Since the new annotation element is created by me, I'm the owner, the old object may in fact have a different owner listed in the xml. That owner information is also lost by not migrating the metadata and this seems a even more significant than the time stamp. Thanks for the additional thoughts and support in working on the migration. I'm still in the process of getting a pristine okular to build in my build env so that I can test out the proposed fix with my data. At the moment, most of the build dependencies succeed (78/84). I'm running into an issue with polkit-qt-1. -- You are receiving this mail because: You are watching all bug changes.
