RTF file containing Japanese characters created in MS-Word can not be opened correctly. Wrong encoding. This error appears in buth Windows and Linux (ubuntu).
Please attach a document.
Created attachment 56596 [details]
Sample file created with MS-Word 2007
I confirm the error with LibO 3.5 Beta2 on Windows XP (with east-asian signs installed) and Windows Vista 64.
I works fine with LibO 3.4.5 (tested on Windows Vista 64).
This is a regression.
In RTFDocumentImpl::resolveChars we read in one code unit at a time but in a stateful encoding like ms932 we need all chars. we can't convert half the input sequence at a time
Reproduced, will have a look at it soon. We already handle multiple \u tokens in one go, something similar could be done for \' tokens.
Fixed in master: http://cgit.freedesktop.org/libreoffice/core/commit/?id=d7baacd81bbcfaa35b7fbf9981fa3fa7c9fb1cb4
Miklos Vajna committed a patch related to this issue.
It has been pushed to "libreoffice-3-5":
fdo#45543 fix RTF import of ms932-encoded characters
It will be available in LibreOffice 3.5.2.
Works well in LO 3.5.2. Closing. Thank you.
Thank you very much !