Apache OpenOffice (AOO) Bugzilla – Issue 2725
Imported czech .doc documents are displayed wrong in word processor
Last modified: 2003-04-09 13:54:45 UTC
First, I should notice I am new to OpenOffice.org, so if I commit this in a bad place, please excuse me. The word processor is not able to import czech written .doc files. These files are in win1250 coding. AFAIK, the problem is not in the coding, but in the font management. I've read somewhere on the StarOffice web pages that the word processor component somehow decides, which fonts it will and which fonts it will not use. The problem is, that the fonts the word processor decides to use do not work in Czech. It prints great, without any mistake, using postscript printer and its internal fonts, it prints great even imported documents written in fonts that I know for sure that are not in the printer. But, it doesn't print them well on the screen. Problem is with the hyphen sign (the one of middle length) and the czech signs of quoations (the bottom one, nines), which are displayed as question marks, but printed correctly on the printer. I thought first I would just find the right fonts to use, but I could not make the word processor use the fonts installed in the X font server, resp. it used only some of them and all incorrectly. Then I played with the fonts in the Adobe font software to find out what is happening there, but it all behaved totally chaoticaly for me (it always did not display some characters, like z with acorn etc., but always printed them well). But what I do not understand at all is the fact, that I do not have any problems like this with the spreadsheet component. It uses all the fonts installed in X font server, it imports all the .xls files with all the czech characters and displays them correctly. .doc document consisting of the problematic characters is included. Thank you for any answer. I use RedHat 7.2, with locale set to Czech, ISO-8859-2 fonts from Adobe and URW, KDE 2.2.0 and oo.org 641b.
Created attachment 883 [details] .doc containing one line of problematic characters
Reassigned to Michael.
MRU->CP: looks like a problem with the font replacement. The font "Bookman" is replaced by a font, which is not capable of such czech characters. When formatting the text with "Times" or "Thorndale", everything looks fine.
Doesn't seem to me like font replacement problem. The - sign, the indent dot and the quoation marks are not displayed correctly in any font, including the times font, whereas the diacritics is displayed well in times.
I have one more hint. I found out that the only fonts that work well for me are the fonts ArialCE and TimesNewRomanCE taken from the Windows 2000. The problem is that in standard Windows instalation, these fonts are the only ones that DO NOT WORK correctly. They have bad metrics, coding and whatever else. So it is quite surprising that they work so great in the oo.org word processor :-) I made some screenshots for you to see what I mean: 1. this is, how it should look like (and how it looks like using the windows fonts): http://www.crs-net.cz/oo/timesCE-good.png 2. this is, how it looks with every other font: http://www.crs-net.cz/oo/times-bad.png You can download the fonts (TimesNewRomanCE and ArialCE) here: http://www.crs-net.cz/oo/fonts.tar.gz Thank you for your care.
Hi Ondrej, please provide a list of fonts that you use (i.e. the output of the xlsfonts command). What version of OOo do you use ? (goto menu "Help -> About OpenOffice.org" type Ctrl s-d-t and note the version and build number) Where are the fonts installed on your system ? Can you copy them to <your_ooo_path>/share/fonts/truetype/ directory, does that change anything ?
1. oo.org version = 641 h (build 7454) 2. output of xlsfonts is in the included file. 3. Fonts are installed in the standard RH7.2 font paths (/usr/share/fonts/ /usr/share/X11/fonts/8859-2/) We made some more experiments with the fonts and oo.org, here are the results: 1. Type1 fonts: a) Fonts with their own encoding vector are not supported by oo.org. b) Fonts with Adobe standard encoding are supported, but some characters (ccaron, Ccaron, caccute, Caccute, ...) are not displayed, but correctly printed on PS3 device. 2. TrueType fonts: a) Fonts with their own encoding vector (i.e. symbol encoding) are not supported by oo.org. b) Standard TT fonts must follow these rules to work correctly with oo.org: 1. The font must present itself as unicode font. (It doesn't really have to be unicode font, it is sufficent that it claims it.) 2. The font must claim that it contains only codepage Win1250. 3. All characters must have all three mappings: Apple standard, Macintosh Roman and Microsoft unicode. Example: Euro sign must be 8364 Apple standard character, 219 Macintosh Roman character and 8364 Microsoft unicode character. If the Apple and Macintosh mapping is missing, then the character is not displayed at all, but is correctly printed on PS3 device. Fonts that do not follow the rules 1-3 are either not supported by oo.org at all, or some characters are not displayed (and have big metrics problems). By the way, are there plans for pair kerning (kerning, that is included in the TT font file) implentation? ondrej
Created attachment 1027 [details] Output of xlsfonts
Please update to the current OOo 641 C (641q build id 7565). We have done a lot of fixes for fonts (which may or may not help), especially for using fonts found in the filesystem (and finding them if they are hidden behind a fontserver). Please give it try and post your findings
works for me with OOo 1.0
closed