Bug 62628

Summary: FILEOPEN: xls file character encoding mismatch
Product: LibreOffice Reporter: Vladislav <vladisslav>
Component: SpreadsheetAssignee: Not Assigned <libreoffice-bugs>
Status: NEW --- QA Contact:
Severity: normal    
Priority: medium CC: sasha.libreoffice, vladisslav
Version: 4.1.0.0.alpha0+ Master   
Hardware: All   
OS: All   
Whiteboard:
i915 platform: i915 features:
Attachments: A test file
Libreoffice screenshot
Openoffice.org screenshot
Excel 2003 opens the test file just fine
LibreOffice git

Description Vladislav 2013-03-22 10:17:01 UTC
Created attachment 76897 [details]
A test file

Incorrect character encoding.

How to reproduce:
1. Open attached file.

What expected:
Russian letters in cells.

Actual result:
East-european letters instead of russian ones.
Comment 1 Vladislav 2013-03-22 10:17:49 UTC
Created attachment 76898 [details]
Libreoffice screenshot
Comment 2 Vladislav 2013-03-22 10:18:25 UTC
Created attachment 76899 [details]
Openoffice.org screenshot
Comment 3 A (Andy) 2013-03-30 09:22:11 UTC
I can see it with LO 4.0.1.2 (Win7 Home, 64bit), but I do not have OO and therefore could not test it with OO.  I am wondering why only parts of the Russian text seem to show this mistake.  But if you open this file with MSO 2007 then you get the same result.

@Sasha: Do you maybe have an explanation for this or/and can you confirm this, too?
Comment 4 Vladislav 2013-03-30 14:20:50 UTC
Created attachment 77231 [details]
Excel 2003 opens the test file just fine
Comment 5 Vladislav 2013-03-30 14:30:52 UTC
(In reply to comment #3)
> I can see it with LO 4.0.1.2 (Win7 Home, 64bit), but I do not have OO and
> therefore could not test it with OO.  I am wondering why only parts of the
> Russian text seem to show this mistake.  But if you open this file with MSO
> 2007 then you get the same result.
> 
> @Sasha: Do you maybe have an explanation for this or/and can you confirm
> this, too?

I do not have access to a machine with MSO 2007, but MSO 2003 opens the test file just fine.
A screenshot of MSO 2003 is included.
Maybe MSO 2007 has the same bug too :-)
I'll test the latest git (today or tomorrow) and include more screenshots.
Comment 6 Vladislav 2013-03-31 06:34:54 UTC
Created attachment 77240 [details]
LibreOffice git

Yesterday git still shows the bug.
Comment 7 sasha.libreoffice 2013-04-01 06:34:31 UTC
reproduced in 4.0.1 on Fedora 64 bit and Win XP 32 bit

This file is in format of Excel 95. It contains Russian text not as unicode, but as ANSI Windows cp 1251 encoding. It is one-byte encoding. And looks differently on different locales. On Russian Windows it looks correctly, on another - not.

IMHO this file generated by some old program. If it is 1c then needed upgrading to new version. New version has no such problem.

Possible solution: use Excel to re-save in format of 2003 or more new.

Use of freedesktop.org services, including Bugzilla, is subject to our Code of Conduct. How we collect and use information is described in our Privacy Policy.