Apache OpenOffice (AOO) Bugzilla – Issue 17503
Patch to enable rtf filter handle local language files
Last modified: 2013-08-07 14:41:36 UTC
The patch will resolve the following two issues. 1. In rich text format, the face name of a font specification in encoded with the codepage of the font itself. Current rtf filter does not hadle it well. 2. The style names are not localized in rtf files. It should be translated to the localized style names.
Created attachment 8068 [details] Enables localized rtf handling
HI->AMA: Please check the attachment.
Created attachment 8270 [details] EXamples generated by Wordpad in Windows 98 local versions
Please have a look.
I haven't applied or looked at the example rtf's yet. But on a brief look at the patch I'm not so sure about the use of SetEncoding in svxrtf.cxx, that would set the fontencoding 'currently in use' when each fontdefinition's F_CHARSET is parsed inside the fonttable definition, not when that font gets used which is where that information is useful. Or am I missing something, the encoding of the fontnames themselves ? I like the standard names lookup, similiar to the winword one, that's good stuff, I'd like to put that right in for 2.0. Looking at the whole SetScriptAttr method itself I'm a little concerned at the code before its patched and its assumptions about LOW and HIGH chars. I think I'll have to look through the examples to get a feel for what it should be doing. I wonder if the whole if else should just be the else branch on its own.
Woops, didn't actually read the top of the issue which explains what the charset business was all about, at least my guess as to fontnames being in their own encoding was correct. That's familiar from what we've seen in the winword filter, so a green light on that one, good stuff.
Integrated to 2.0 in limerickfilterteam08. Interestingly wordpad can open the .rtf files, but word 2003 cannot it seems.
reopen to reassign
cmc->mru: localized fontnames working in limerickfilterteam08
Hi, all Is it possible to change the target milestone to 1.1.1? We Japanese team have been applying his patch and provide as `japanese build' for user's convinence. We regard his patch quite important. Regards, maho
Opps sorry, I changed target milestone without discussion. Backouted.
You could write a new bug for 1.1.1 to support the native charset for fontnames alone. This patch does a few more things than just that and for 1.1.1 you want to keep things as massively simple as possible.
I'm as bad, I didn't mean to set the target of this either :-)
Checked fix in CWS filterteam08.
Verified. Fix will be available for OO 2.0.
Closed. Fixed in current snapshot.
The fix in only partially included in current 680 codeline. Especially the standard name lookup is omitted. Are there any reason? And recently I found out the fix for issue 35653 interfares decoding of Japanese languege RTF. As my patch in issue #20264 is included in current codeline, sw now generates local encoded RTF. Therefore unless these problems are fixed, sw even cannot read Japanese RTF generated by itself. I will post an additional patch right next.
Created attachment 22010 [details] Additional patch for 680 codeline
Please do not open such issues, if ther's ONE problem with ONE certain Language. The issue itself is fixed. So please file a new issue with the required information. That is a VERY MUCH better issue handling! Thanks for your patience.
Closed.