I have submitted a fix (https://gerrit.libreoffice.org/#/c/15340/) that requires someone knowledgeable with redline and .doc exporting to review.
The marker used in Word to identify table cell division was being included in the track-changes deletion from the following cell if the deletion was the first node in the cell. This would cause the two table cells to merge into one when changes were accepted by Word. (Problem not seen in LibreOffice itself, only in Word.) I don't understand this code - I just figured out something that fixed my problem and tried to make it as targeted as possible. I'm not really sure of the implications - it just didn't seem to do anything bad to my sample document.... The biggest problem is that this is called at the end of EVERY cell, regardless of whether it is touching a redline. I couldn't figure out how to recognize when that occurs. So the possibility of introducing a regression is fairly high... Several testcases are attached to the bug report. It is super easy to reproduce the problem. Simply open the testcase in LibreOffice and save. Then in MSWord, accept all of the tracked changes. Thanks, Justin _______________________________________________ LibreOffice mailing list [email protected] http://lists.freedesktop.org/mailman/listinfo/libreoffice
