Created attachment 61519 [details] Same word in Hebrew written over and over again. Last occurrence breaks mid-word. Problem description: Steps to reproduce: 1. .... Create a new Writer document. 2. .... Write a long enough sentence in Hebrew so that it spans more than one line. Current behavior: The line will break mid-word. Expected behavior: Line should break either before or after a word. Platform (if different from the browser): Arch Linux, 64-bit.
Possibly related: issue began after upgrading to LO 3.5.2, which is compiled against the new ICU 49.
can you paste in the same word here
Yeah, the word is מילים I'll attach the document too.
Created attachment 61557 [details] Document demonstrating the issue of Hebrew lines improperly broken
confirmed
http://unicode.org/versions/Unicode6.1.0/ "Two new Line_Break property values have been added. The first is for Hebrew letters: lb=HL. It is used in the definition of a new rule, LB21a, in UAX #14, for handling line breaking for Hebrew characters next to hyphens. The second, lb=CJ, allows for better customization of Japanese line breaking. Implementations of Unicode line breaking may need to be updated to correctly handle these additional line break property values." might be it
Caolan McNamara committed a patch related to this issue. It has been pushed to "master": http://cgit.freedesktop.org/libreoffice/core/commit/?id=20c24114143d6d38774b56a142fd4ae05094308e Resolves: fdo#49849 implement Unicode 6.1 hebrew line breaking rules
in the "internal-icu" case, icu 49 is only in use on master, not 3-5. So if the problem appears in a 3.5 or earlier version, it means a distro has built against a system icu 49 so if your distro has this problem point them to the above commit, and bug #49885 if you want to help out :-)