Apache OpenOffice (AOO) Bugzilla – Issue 34309
Numbers in Arabic are reversed at the beginning of a line
Last modified: 2013-11-15 04:21:32 UTC
When I type ARABIC 1-2 ARABIC, it displays as ARABIC 2-1 ARABIC which is true but when 1-2 or any outline numbers come in the beginning of the paragraph it displays as ARABIC 1-2 ARABIC which is clearly wrong.
Oops! I meant to say ... outline numbers come at the beginning of the paragraph it displays as ARABIC 1-2 instead of ARABIC 2-1 (for section one subsection 2)
please attach screenshots.
Created attachment 17788 [details] The first blue underlined numerals sequence is correct, the second is not.
OK, thanks - reassigning to WordProcessor component for review.
reassigned to SBA.
SBA: Looks OK in OOo 2.03 candidate (dev build 680m167). Please note that you have to set the correcr character attribute (Language Arabic for CTL) To get this working correctly. Set to Worksforme. Please comment.
SBA: Closed.
I just retested with 2.0.2-2ubuntu12.1 in my ubuntu system with CTL set to Arabic (Saudi Arabia) and the issue is still there. However, I may be wrong in counting it as a bug since I am not expert in unicode algorithm so let me clarify more. Typing an Arabic number, then space, then an Arabic number, then an Arabic word is OK but Typing an Arabic number, then hypen, then an Arabic number, then an Arabic word gives a different order for the numbers, shouldn't they be the same? Isn't both the hyphen and the space weak characters? Compare قارن 1 2 عربي 1-2 عربي In Firefox the have the same order wich what I think is correct. In OO.o they one with the hypen have what I think is the wrong order. It's worth noting that kedit and gedit also work like OO.o. So, what application is doing it properly? Firefox or the other gang? ;)
fme->munzirtaha: Of course OOo does it right ;-) To be honest, I didn't check your test case against the official rules (http://www.unicode.org/unicode/reports/tr9/) but I'm pretty confident that the icu implementation of the UBA (which is used by OOo) is correct.
SBA: Closing issue.
Created attachment 70808 [details] Examples of wrong and right numbering for right-too-left languages...
This issue should not be closed, the problem with inverted sub-numbering for right-to-left languages is still there, and manually numbering headeings is really annoying, especially if you insert a heading between two headings you'll have to re-number all following headings which is a waste of time. I attached an example screenshot showing the current wrong numbering, and right way of numbering...
Sorry for the typos... Just clarifying what i meant by sub-numbering: if you choose right-to-left typing, then choose format > bullets and numbering, and then choose numeric with all sublevels (ex: 1, 1.1, 1.1.1). After that demote one level (increase indentation), it will be numbered (1.1), type anything and then press enter. the next line will be wrongly numbered (1.2) where it should be (2.1).
You are right Ali, the bug is very clear. I will reopen the bug
ja->fme:please take a look at the latest comment
Created attachment 81925 [details] Arabic numbering problem I replicated this issue (you can find it in attached file) I'm using operating system: windows 7 oo version: AOO401m5(Build:9714) - Rev. 1524958 writer application. i added many levels to let you see the problem.