PDF exports with arabic numerals instead of regular numbers

I'm pretty sure that upstream bug was not the correct one, that bug report is about Arabic being exported to PDF incorrectly. This bug report is about Hebrew being exported incorrectly and showing up as Arabic.

Chris

I faced the same problem (arabic numbers to calc tables when i export as pdf in OO 2.4 in 2.6.24.19 generic on i386) but with an english text. The problem got solved when i unchecked the PDF/A-1 option in the export as pdf dialog. hope this helps.

Ok, I tested this on a document that didn't contain any Hebrew and it still happens, like jarlaxl lamat wrote in his comment. Also, unchecking the PDF/A-1 option prevents this problem for me too.

So the exact bug is:
Arabic numerals when exporting to PDF with PDF/A-1 enabled, a Writer document that contains a table pasted from Calc. Only if Calc _is still open_ when the export is being done.

There is an upstream bug report but it sounds like this might actually be an ooo-build problem from what has been mentioned in the upstream bug report.

I had a "related" problem: in a text in English with several formulas, some of the numbers in the formulas become some weird characters (I'm an ignorant in this, I believe there were Hebrew, but I'm not sure) when I export the document to PDF.
The solution I found was to only check "create PDF form" selecting "PDF" in the "submit format", in the "General" tab when the export dialog appears (I'm using the File->Export as PDF, and no the button on the toolbar). I also check "Export Bookmarks" and "Export automatically inserted blank pages" but that shouldn't change the result.
Good Luck !

Same problem with me. I have a Writer document with several formulas inserted in. Looks fine in Writer, but when I convert to PDF some of the characters randomly show up as unusual characters. Sometimes if I export again, it seems as if some of them fix, and others get screwed up.

I have the same problem with a text document with mathematical formulae added using OpenOffice Maths after exporting it as PDF. I have a workaround by using 'Print to PDF' which comes with Ubuntu. Press Ctrl+P in OpenOffice and the document should be "printed" as a PDF file stored in /home/[your own username]/PDF folder.

This should all have been pre-installed, but if not here are the instructions:

1. In control centre (type gnome-control-center in terminal), go to 'Personal' section and click on 'Sessions'. Under 'Startup programs', make sure 'Print queue applet' is enabled.

2. Go to 'System' section and click 'Services', make sure 'Printer services (cupsys)' is enabled.

3. Go to 'Hardware section' and click on 'Printing'. See if 'Print to PDF' is under Local Printers. If not, click 'New Printer', select 'Print to PDF file' then 'Forward', select 'Generic' then 'Forward', select 'PDF file generator' then 'Forward', change the names as you wish then 'Apply'.

I've been having the a similar problem but with Impress. I need to export to PDF to show my slides on the Universities lecture theater computers because they do not have OpenOffice installed (just some weired product call Microsoft Office or something like that). Anyway, in the PDF's regular numerals are randomly replaced by arabic-script numerals, both within text boxes and within equations. Very strangely objects copied and pasted from one page to another may have arabic script on one page and regular on another. There seems to be no rhyme or reason - I am not using any other non-western scripts anywhere in the docs. I tried the 'Print to PDF' workaround suggested by tpcchu but that gives me a second, presumably unrelated problem with pages being randomly either landscape or rotated to portrait - extreme annoyance. I have a whole term of lectures planned using OpenOffice impress, I would really like suggestions/solutions ASAP!

Thanks

James

Screenshot is available in this forum thread:
http://user.services.openoffice.org/en/forum/viewtopic.php?f=12&t=8737

I also have this problem.

Using Intrepid Alpha 6 on x64.
OO version 2.4.1-8ubuntu

This is only a problem of numerals which appear within equations not in general text.
None of the "solutions" work for me, although I do find that which numerals are incorrect does seem to erratically change.
The print to pdf option is not a good solution - the resulting pdf does not contain the bookmarks.
 Attached pdf shows problems. (for example subscripts in equation 4)

This discussion of the same issue concludes that the bug is not present in the openoffice official release only in the Ubuntu packages: http://user.services.openoffice.org/en/forum/viewtopic.php?f=7&t=7796&st=0&sk=t&sd=a&start=10

Unfortunately I cannot easily test this as I am using 64bit.
I might give it a go on a virtual machine - which would be a solution for me for the short term.

I have been aware of this bug for some time and keep seeing reports which conclude it has been fixed... it is not fixed.
I see this as a major issue, I am pretty certain that I have already submitted one technical journal paper with these misprints in them!

Thanks, Jack.

Today i recognized this buggy behavior (numbers in wrong eventually hebraic -- i don't know -- font in exported pdf) with newest openoffice from ubuntu. I had to create PDF from a document with heavy formula use.

My workaround with best result: Print to generic postscript printer, choose print to file and create pdf from this output.

Some month ago i continued my work on a windows box and forgot to report a bug for this reason.

Architecture: amd64
Source: openoffice.org
Version: 1:2.4.1-9ubuntu2 from intrepid repositories

btw. test.pdf (http://launchpadlibrarian.net/17778579/test.pdf) lets acroread go nuts here on amd64.

Acroread complains about invalid color space

Sorry if more information isn't needed. I've seen this bug, too--the version where strange numerals appear in English files that contain formulas. Easily reproducible using the numbers.odt file available on the openoffice bug page. This is in 64-bit Intrepid and Hardy. However, I can verify that the bug did not seem to be present when I installed OOO300m9 in Hardy from the .deb available on the openoffice.org site.

Unchecking the various PDF options in the export dialog didn't change the behavior of the bug for me.

As someone who creates lots of PDF's from files with formulas, I'll say thanks in advance for your work on this!

I add to Scott Thatcher, that I have the same situation (64bit, jaunty). Playing with PDF/X-a1 settings just changes where the wrong characters show up, whether on embedded calc sheet, native formulas or in tables. Sometimes part of the formula has numbers and the other part has these wrong characters (some non-arabic numbers as seen in test.pdf).

I can confirm this bug with gentoo linux and the app-office/openoffice package.
As a workaround I suggest printing the document to a .ps file via the print menu, and using ps2pdf to convert it to a pdf. That worked for me.

2.6.28-14-generic #46-Ubuntu 64-bit
I can correct the file by editing calc tables changing some bit, mostly after 1 or 2 edits the problem is gone, but it is nagging. It ofte happens when the document is previously edited by oo windows version...

Having the same problem.

English document containing MANY formula and for each export to PDF from ooo im random getting Arabic numerals in place of english digits.
None of the above fix's help.
Print to .ps is fine but then cups-pdf is only printing the first page for me. If anyone can suggest a work around...
If i reinstall ooo not from rep would that help.

ubuntu 9.04 64bit
ooo 3.01

Do you still have this issue with the new version in Ubuntu 9.04 with the ppa at https://launchpad.net/~openoffice-pkgs/+archive/ppa or Ubuntu Karmic?

Just finished install ooo updates and yeah problem is still there.

Seems to be less frequent though and im getting what looks like numerals from another language now.

My work "around" has been to print to file in ooo .ps then open that with the standard viewer (eye of gnome or whatever) then print to file again but this time you get the option to print to file as pdf instead of .ps

cheers

Jaunty [NOT WORKING]

Linux ubuntu 2.6.31-9-generic #29-Ubuntu SMP Sun Aug 30 17:39:26 UTC 2009 x86_64 GNU/Linux

openoffice.org:
  Installiert: 1:3.1.1-1ubuntu1~jaunty1
  Kandidat: 1:3.1.1-1ubuntu1~jaunty1
  Versions-Tabelle:
 *** 1:3.1.1-1ubuntu1~jaunty1 0
        500 http://ppa.launchpad.net jaunty/main Packages
        100 /var/lib/dpkg/status
     1:3.0.1-9ubuntu3 0
        500 ftp://192.168.0.100 jaunty/main Packages

Karmic Alpha5 [NOT WORKING]

Linux ubuntu 2.6.31-9-generic #29-Ubuntu SMP Sun Aug 30 17:39:26 UTC 2009 x86_64 GNU/Linux

openoffice.org:
  Installiert: (keine)
  Kandidat: 1:3.1.1-1ubuntu1
  Versions-Tabelle:
     1:3.1.1-1ubuntu1 0
        500 http://archive.ubuntu.com karmic/main Packages

Karmic Daily [SEEMS TO WORK]

Linux achim-laptop 2.6.31-10-generic #31-Ubuntu SMP Thu Sep 10 14:14:28 UTC 2009 i686 GNU/Linux

openoffice.org:
  Installiert: (keine)
  Kandidat: 1:3.1.1-1ubuntu2
  Versions-Tabelle:
     1:3.1.1-1ubuntu2 0
        500 http://de.archive.ubuntu.com karmic/main Packages

As it seems this problem is still present but only on x86_64 platform. I wasn't able to reproduce this problem on my x86 system. The package version of the x86 system is not the same as of the x86_64 system, so I am not 100% if it is a x86_64 problem.

Regards
Achim

Installed the ppa and haven't seen the bug since then on 64-bit Jaunty.

I have the same type of problem on Ubuntu 9.04 (64-bit) with the Ubuntu OpenOffice (3.0.1). Upon PDF export, some numeric characters in Math equations are replaced, especially in sub and superscripts.

Using a colleague's OpenOffice 3.1.0 on Windows XP, no PDF export problem was encountered and the math looks normal. For now, I will use this method to generate PDF, while waiting that hopefully the problem will be solved in Ubuntu OpenOffice 3.1.1 which appears to be included in the upcoming Ubuntu 9.10.

While looking for a solution to the problem, I had the impression that it is mostly 64-bit Linux users who encounter this problem.

See also:
http://qa.openoffice.org/issues/show_bug.cgi?id=87669

I am running 64-bit Jaunty, installed the ppa as described above (OO3.1 for Linux) and the problem vanished.

> I am running 64-bit Jaunty, installed the ppa as described above (OO3.1
> for Linux) and the problem vanished.

That is not entirely true, with complex documents the problem is still
present.

Take a look at comment #22

Regards
Achim

Achim, I suggest that we exchange one or 2 documents with proven problems, going to try out one on my 32-bit system without ppa, it was the same a few days ago...I will get back to you, not today, I am crowded with work...
Cheers/Antonio

Oliveira:
> Achim, I suggest that we exchange one or 2 documents with proven problems, going to try out one on my 32-bit system without ppa, it was the same a few days ago...I will get back to you, not today, I am crowded with work...
> Cheers/Antonio

Okay, so I have a new (simple) Test Dokument. As you can see the
PDF-File has still strange characters.

I think if you embed a spreadsheet the result can be even worse, as you
can see in comment #22.

openoffice.org:
  Installiert: 1:3.1.1-1ubuntu1~jaunty1
  Kandidat: 1:3.1.1-1ubuntu1~jaunty1
  Versions-Tabelle:
 *** 1:3.1.1-1ubuntu1~jaunty1 0
        500 http://ppa.launchpad.net jaunty/main Packages
        100 /var/lib/dpkg/status
     1:3.0.1-9ubuntu3 0
        500 ftp://192.168.0.100 jaunty/main Packages

$ uname -m
x86_64

Regards
Achim

Achim,

Just tested your document on karmic alpha 6 with the current official karmic OO.org package which is 1:3.1.1-2ubuntu1 (karmic)
I am on 64bit.

Firstly let me just be clear that if I download test.odt and export as pdf I get erroneous characters (please see attached file "testexport.pdf").

However because I have been messing around with mozplugger, the first time I tried it opened test.odt readonly and I had a perfect export of the pdf. This seems to be repeatable - if I change the permissions so that I do not have write permission before I open test.odt in OO I can export the pdf perfectly (see attached file "testexport_ro.pdf").

Please can you see if this happens for you too. Perhaps it was a fluke.

uname -a:
2.6.31-10-generic #35-Ubuntu SMP Tue Sep 22 17:33:14 UTC 2009 x86_64 GNU/Linux

second attachment.

I am not able to reproduce, what you have seen. If I open the file with read only permission I still get wrong characters.

The interesting part is, that the document from you, that you opened with read only permission is smaller in size.

36,1 KiB vs. 39.5KiB

I also noticed that there are less fonts embeded in the smaller document.

Regards
Achim

Hi

This was what I got on 64-bit, you were right, it still does the trick...

try this one...it exports cleanly here...

Oliveira:

> try this one...it exports cleanly here...
>
> ** Attachment added: "trash.odt"

That file works, but isn't a good test case, there is too less
information on the page.

If I have documents with 10+ pages the chance is very high that there
are wrong characters.

Regards
Achim

I exported to pdf the
test2.odt listed in
<https://bugs.launchpad.net/openoffice/+bug/244353/comments/28>
in (U)OOo 3.1:
$ cat /usr/lib/openoffice/program/versionrc
[Version]
AllLanguages=en-US
BuildVersion=openoffice.org-core 1:3.1.1-5ubuntu1, Thu Oct 22 20:13:41
UTC 2009
buildid=310m19(Build:9420)
ExtensionUpdateURL=http://updateext.services.openoffice.org/ProductUpdateService/check.Update
OOOBaseVersion=3.1
ProductBuildid=9420
ProductMajor=310
ProductMinor=19
ProductSource=OOO310
UpdateID=OpenOffice.org_3_en-US
UpdateURL=
UpdateUserAgent=<PRODUCT> (${buildid}; ${_OS}; ${_ARCH};
BundledLanguages=${AllLanguages})
Vendor=Debian and Ubuntu

and got garbage fonts in the output. See: Test2UOOo.pdf

I then opened the doc
in standard OOo 3.1:
$ cat /opt/openoffice.org3/program/versionrc
[Version]
AllLanguages=en-US
buildid=310m19(Build:9420)
ExtensionUpdateURL=http://updateext.services.openoffice.org/ProductUpdateService/check.Update
OOOBaseVersion=3.1
ProductBuildid=9420
ProductMajor=310
ProductMinor=19
ProductSource=OOO310
UpdateID=OpenOffice.org_3_en-US
UpdateURL=http://update32.services.openoffice.org/ProductUpdateService/check.Update?pkgfmt=deb
UpdateUserAgent=<PRODUCT> (${buildid}; ${_OS}; ${_ARCH};
BundledLanguages=${AllLanguages})

and the export was also wrong, see: Test2standardOOo.pdf

But then tested from standard OOo (s)OOo again and the export was fine, see: Test2standardOOo2.pdf

I've since tried to reproduce in (U)OOo and (s)OOo and have not been able to do so. I notice in the Test2UOOo.pdf and Test2standardOOo.pdf files that DejaVuSeriff-italic font is included. The pdf files without issue do not include these. So perhaps it's a DejaVu font issue?

Also note that printing to a cups-pdf printer and/or to a ps file produces a correct pdf. But given that the bug is not reliably reproducable this should only be taken as a "maybe"?

First off: I'm using Ubuntu's OpenOffice, installed from repositories, fully up to date... I experience the same issue as many others here have reported with embedded OpenOffice.org Math formulas, but I don't appear to have any issues with embedded tables. The problem seems to be completely random, as I tried exporting the same document several times in a row, both with the 'Export Directly as PDF' button on the toolbar, and the 'File > Export as PDF' menu option... both methods produce errors, though the latter seems to produce far fewer errors when selecting the "PDF/A-1a" option (which turns off the default option of 'Create PDF Form, of type FDF').

When I try exporting the odt file listed in comment #1, the file exports normally, and as I would have expected it to. However, when exporting the test document in comment #28, it produces erroneous characters, though not very many (see attachment: Test2.pdf).

andrew@BEC-LIN:~$ cat /usr/lib/openoffice/program/versionrc
[Version]
AllLanguages=en-US
BuildVersion=openoffice.org-core 1:3.1.1-5ubuntu1, Thu Oct 22 20:13:41 UTC 2009
buildid=310m19(Build:9420)
ExtensionUpdateURL=http://updateext.services.openoffice.org/ProductUpdateService/check.Update
OOOBaseVersion=3.1
ProductBuildid=9420
ProductMajor=310
ProductMinor=19
ProductSource=OOO310
UpdateID=OpenOffice.org_3_en-US
UpdateURL=
UpdateUserAgent=<PRODUCT> (${buildid}; ${_OS}; ${_ARCH}; BundledLanguages=${AllLanguages})
Vendor=Debian and Ubuntu

The error seems to happen more often in very long documents, so as an example, I'm uploading a copy of a lab report I wrote this past semester, which includes embedded tables, pictures, and formula boxes. When exporting on either of my computers, random numbers from different formula boxes get messed up, and the location and severity seem to vary randomly as well. The pdf's that im attaching of that report were all made within a few moments of each-other, using the "Export to PDF" button on the toolbar.

Note how, on the previous PDF export, the bulk of errors are on page 6 of the report, yet on the second export, which is attached, (exported less than 30 seconds later) the errors are on pages 3 and 8

@AFarris01

I have experienced this problem persistently for around a year now on various versions of OO but I have not seen it the last few times that I have exported pdfs.

It looks like we have exactly the same package installed for OO (on ubuntu 9.10 I presume)

jacko@mpc540:~$ cat /usr/lib/openoffice/program/versionrc
[Version]
AllLanguages=en-US
BuildVersion=openoffice.org-core 1:3.1.1-5ubuntu1, Thu Oct 22 20:13:41 UTC 2009
buildid=310m19(Build:9420)
ExtensionUpdateURL=http://updateext.services.openoffice.org/ProductUpdateService/check.Update
OOOBaseVersion=3.1
ProductBuildid=9420
ProductMajor=310
ProductMinor=19
ProductSource=OOO310
UpdateID=OpenOffice.org_3_en-US
UpdateURL=
UpdateUserAgent=<PRODUCT> (${buildid}; ${_OS}; ${_ARCH}; BundledLanguages=${AllLanguages})
Vendor=Debian and Ubuntu
jacko@mpc540:~$

I have downloaded your odt and exported and it looks fine to me (see attached) can you spot any errors?? I exported three times and it was fine on all occasions (I too have seen the errors moving around the document)

I am racking my brain to work out if I have changed my config but I cannot think of anything significant.

Just noticed that the file size of my pdf was much bigger than yours so I checked my pdf export options (attached).
When I choose "JPEG compression" or "Reduce Image Quality" I get formula errors in the pdf - although they do not look the same as yours, rather than Arabic characters I get a rectangle character appearing in numerous places.

Try replicating my settings and seeing if you still have the same problem.

@jackocleebrown

You're correct, I'm on ubuntu 9.10. Also, I checked the file...didn't see any errors! very odd indeed... I don't think it's a config issue, because I'm getting the same results on a fresh install of ubuntu on my laptop, as well as a long-standing upgraded one on my desktop.

What are your settings under 'File > Export as PDF"? It seems, to me anyway, that with the "PDF/A-1a" option selected, I get far fewer errors on the output documents than the defualt, which is 'Create PDF Form: FDF'

@AFarris01

Replied above (#43) with the settings I used.

This issue is not just an (U)OOo or Novell/go-oo issue. Please see standard OOo upstream bug:
http://www.openoffice.org/issues/show_bug.cgi?id=105912
[Formula symbols export weirdly to PDF when several objects in document]

I tested w/standard OOo & attached a copy of my PDF to the report.

That bug is left as an open issue from:
http://www.openoffice.org/issues/show_bug.cgi?id=107358
[Issue 107358 blocks: 87669 99999 100007 101715 105912 87669,99999,100007,101715,105912]

@jackocleebrown

that was weird... the last reply I saw from you was your long one... our replies must have crossed paths.

About the settings:
My "Export to PDF" defaulted to using the "JPEG Compression", and a quality of 90%. I tried a variety of combinations on the quality and format settings that seem to have been affecting it to see the different results, and here's what I came up with:

90% JPEG compression, FDF style PDF -- Significant Errors
90% JPEG compression, A-1a style PDF -- Minor Errors/No Errors (varies)
90% JPEG compression, FDF style PDF, clicking 'lossless' then back to 90% -- No Errors/Minor Errors (varies)
90% JPEG compression, A-1a style PDF, clicking 'lossless' then back to 90% -- No Errors/Significant Errors (varies)
80% JPEG compression, FDF style PDF -- Significant Errors
80% JPEG compression, A-1a style PDF -- Significant Errors ** This one generated corrupted tables. I had never seen this before this test.
50% JPEG compression, FDF style PDF -- Minor Errors
50% JPEG compression, A-1a style PDF -- Minor Errors/No errors (varies)
10% JPEG compression, FDF style PDF -- Significant Errors
10% JPEG compression, A-1a style PDF -- Minor Errors
Lossless compression, FDF style PDF -- Minor Errors
Lossless compression, A-1a style PDF -- Significant Errors

I've attached all my test exports for others to view... All were made using my report document, as it was uploaded earlier.

I've added some updated testing results on upstream http://www.openoffice.org/issues/show_bug.cgi?id=105912. Tested/testing on OOo 3.2RC1 (OOO320_m8). So far I've not been able to produce a corrupted PDF using that version on linux, WinXP, or Win2K.

Econtered this today. The document was originally a word file (.doc) Added some rows and colums from another open OO spreadsheet and then tried to export as PDF. Numbers but not text in the spreadsheet objects appeared in wrong (appears to be hebrew) font. Tried various fonts, saved doc as ODT rather than DOC, tried again, and again etc. Sometimes the problem seemed fixed with a change of font if saving closing and reopening the file before exporting. The problem appears random in than some times a save or change of font appears to fix the problem only to findanother table has reverted to the erroneous font in the pdf.

This seems to affect the 64bit version
3.0.1 OOO300m15 (build:9379)
openoffice.org-core 1.3.0.1-9ubuntu3.2 Fri Feb 19 23:33:57 UTC 2010

but not the 32 bit
OpenOffice.org 3.1.1
OOO310m19 (build:9420)
openoffice.org-core 1:3.1.1-5ubuntu1.1 Fri Feb 19 23:52:09 UTC 2010

as transfering the .odt file to another machine and creating the pdf with the latter version was ok.

<rant>
I hate OpenOffice, every stupid misfeature in MS Office doggedly imitated and then some more of their own. If I didn't need to use it for all the stupid .doc and .ppt file idiots send me I would never touch it. What's wrong with plain text.

Noam Lewis, this issue is unreproducible using LibreOffice Writer via the Terminal:

cd ~/Desktop && wget https://bugs.launchpad.net/ubuntu/+source/openoffice.org/+bug/244353/+attachment/313739/+files/test.odt && unoconv --listener && unoconv -f pdf test.odt

wait a few seconds then type:

acroread test.odt

looks as it does in Writer. Does this work for you?

lsb_release -rd
Description: Ubuntu 11.04
Release: 11.04

apt-cache policy libreoffice-writer
libreoffice-writer:
  Installed: 1:3.3.2-1ubuntu5
  Candidate: 1:3.3.2-1ubuntu5
  Version table:
 *** 1:3.3.2-1ubuntu5 0
        500 http://us.archive.ubuntu.com/ubuntu/ natty-updates/main i386 Packages
        500 http://us.archive.ubuntu.com/ubuntu/ natty-proposed/main i386 Packages
        100 /var/lib/dpkg/status
     1:3.3.2-1ubuntu4 0
        500 http://us.archive.ubuntu.com/ubuntu/ natty/main i386 Packages

apt-cache policy unoconv
unoconv:
  Installed: 0.3-6
  Candidate: 0.3-6
  Version table:
 *** 0.3-6 0
        500 http://us.archive.ubuntu.com/ubuntu/ natty/universe i386 Packages
        100 /var/lib/dpkg/status

apt-cache policy acroread
acroread:
  Installed: 9.4.2-0natty1
  Candidate: 9.4.2-0natty1
  Version table:
 *** 9.4.2-0natty1 0
        500 http://archive.canonical.com/ubuntu/ natty/partner i386 Packages
        100 /var/lib/dpkg/status

[This is an automated message.]
There are no new official OpenOffice.org releases in Ubuntu packaging anymore => Won't Fix

If the problem persists, please mark this bug as "also affects project Libreoffice" or "also affects distribution Libreoffice (Ubuntu)" if that has not happened already.

Please leave references to upstream OpenOffice.org bugs in place to allow cross pollination.

closing as invalid (expired) because of missing reply