Apache OpenOffice (AOO) Bugzilla – Issue 32785
Excel Export BIFF5: missing character encoding
Last modified: 2013-08-07 15:13:10 UTC
I have a XLS file. If I save it from OOo and open with OOo 111 on Windows machine, the diacritics is gone ('?' takes place instead of dacritical symbols), although font seems unchanged. No way to change the behavior. If I export the file as SXC, no problems.
Created attachment 17040 [details] Original XLS doc
Created attachment 17041 [details] Saved from OOo 680_m49
Created attachment 17042 [details] exported from OOo680 as SXC
Hi, the underlying problem is the original xls file format. It's Excel 5 and therefore not fully supported. To avoid this problem just save in OOo format or use Excel 97/2000/XP. @Daniel As discussed for fixing with target Later Frank
accepted DR->DR: todo: nobody has implemented the CODEPAGE record in BIFF5...
title adjusted
*** Issue 42933 has been marked as a duplicate of this issue. ***
*** Issue 54059 has been marked as a duplicate of this issue. ***
*** Issue 58726 has been marked as a duplicate of this issue. ***
@dr,fst: duplicates and voted, we should reconsider the target !
new target
*** Issue 70824 has been marked as a duplicate of this issue. ***
Dear developers, is there any chance of fixing this issue in v2.2? This is nasty bug, that can cause loss of work for the full work day (this has happened twice with my users).
Daniel, it turns out that we have customers who do use MSO95 so they need export filter that correctly handles encoding. Any chance for 2.3? Thanks a lot for your attention. WBR, Kirill Palagin.
I will see if there is some time to fix this in 2.3
Daniel, any luck?
*** Issue 80872 has been marked as a duplicate of this issue. ***
Maybe for 2.4?
change target from 2.x to 3.x according to http://wiki.services.openoffice.org/wiki/Target_3x
given the number of duplicates and votes set target to 3.0. sorry for the noise, choosing 3.x seems to be an oversight of mine.
Daniel, are we on track for 3.0 with this issue?
I am too busy to do it for myself. Maybe there is a volunteer to provide a patch?
There is an article on the Microsoft server referring to the BIFF5 file creation: http://support.microsoft.com/kb/150447 There is a demonstration code available for downloading there. On the first look, the code seems to be available for free to use in any application. (It's a bit strange as the Microsoft seems to be supporting Open Source, but it looks that way. But, perhaps, I'm mistaken, somehow)
I have got yet another point. The OpenOffice does read the Excel 95 files correctly. So, the correct translation from BIFF5 is implemented already. The problem is just in the export to the BIFF5. It should be relatively simple to use the import code page to create a proper export code page. It's quite possible no other changes will be necesary.
The imported codepage is not stored anywhere. Additionally, the Excel95 export filter has to work for all files, also for new files and files imported from other formats such as ODS and Excel97+ which use Unicode and thus do not contain a codepage. The codepage has to be received from the operating system depending on the current locale, and this has to work on all supported platforms (Windows, Linux, Solaris, Mac...)
Please see http://www.openoffice.org/issues/show_bug.cgi?id=9709#desc43 for a way to get system encoding.
*** Issue 61401 has been marked as a duplicate of this issue. ***
*** Issue 102886 has been marked as a duplicate of this issue. ***
*** Issue 106869 has been marked as a duplicate of this issue. ***
Hello tuharsky, I stumbled upon this issue during my TCM test for the Germanophone version of OOO320_m9. It looks O.K. under Debian SID AMD64, but as I can neither read nor understand the text in the file, I may be wrong ... ;) Could have a look at this issue to confirm, if it is fixed, please? If it is fixed, it would be nice, if you could close this issue :) Have fun Thomas.
I'll test with 3.2 stable. However, with 3.1.1 the problem persists.
Hello tuharsky, *, just as a reminder: Have you tested it in the meanwhile, like you wrote on Jan 18th 2010? If not: Would you be so kind to test it with 3.2 or OOO320m17, please? TIA Thomas.
With OOo3.2m12(@Ubuntu10.4) bug still present. Should there be any difference with m17?
Created attachment 69473 [details] OOo3.2-good (original)
Created attachment 69474 [details] OOo3.2-bad (saved)
Ubuntu 9.10, DEV300m77 - save file ihriská.xls(Original XLS doc) as excel5 crashes OOo. - open file ihriská.xls - save as - format excel 5 - crash - no crash report created
For the crash in m77, I created a separate issue (issue 111745), because the fix for the crash doesn't resolve the encoding problem.
Hello tuharsky, *, sorry for the delay. Too busy since my last entry, sorry ... :( You wrote <quote> ------- Additional comments from tuharsky Mon May 17 06:30:09 +0000 2010 ------- With OOo3.2m12(@Ubuntu10.4) bug still present. Should there be any difference with m17? </quote> I have not read the release notes (neither for m12 nor for m17), but there are always changes between the release ... ;) But as both are outdated, it would be nice, if you could test it again with OOO320m18 ... ;) HTH Thomas. P.S.: I still cannot reproduce it with the Germanophone version of OOO320m18 under Debian SID/Experimental AMD64 ... :(
thackert, apparently you cannot reproduce because all German letters are in Latin-1 codepage? See http://en.wikipedia.org/wiki/ISO/IEC_8859-1
Of course it's still here. Also, is OpenOffice abandoned now?
Good news: OO is alive and well. Bad news: This bug, for writing a format that's a decade out of date, is unlikely to be fixed. More bad news: the @ openoffice.org email addresses are going away very soon. If you use one here, log in and click on the Preferences link at the top of the page, and change it. BZ is smart: you'll still have all your old issues and comments properly linked.