https://bugs.freedesktop.org/show_bug.cgi?id=59642
--- Comment #12 from Mike Kaganski <mikekagan...@hotmail.com> --- (In reply to comment #10) > Do you know if the LO has code to avoid this marker, or maybe handle this? Here is what I found: http://docs.libreoffice.org/tools/html/classSvStream.html#ab1d78e3df7058ca99859cb43d3b227a6 sal_Bool SvStream::StartReadingUnicodeText(rtl_TextEncoding eReadBomCharSet) If eReadBomCharSet==RTL_TEXTENCODING_DONTKNOW: read 16bit, if 0xfeff do nothing (UTF-16), if 0xfffe switch endian swapping (UTF-16), if 0xefbb or 0xbbef read another byte and check for UTF-8. If no UTF-* BOM was detected put all read bytes back. This means that if 2 bytes were read it was an UTF-16 BOM, if 3 bytes were read it was an UTF-8 BOM. There is no UTF-7, UTF-32 or UTF-EBCDIC BOM detection! If eReadBomCharSet!=RTL_TEXTENCODING_DONTKNOW: only read a BOM of that encoding and switch endian swapping if UTF-16 and 0xfffe. === The pStrm is SvStream*, so everything seems as easy as adding this just before line 316 of smdetect.cxx http://cgit.freedesktop.org/libreoffice/core/tree/starmath/source/smdetect.cxx#n316 : pStrm->StartReadingUnicodeText(RTL_TEXTENCODING_DONTKNOW); sal_uLong nBytesRead = pStrm->Read( aBuffer, nSize ); -- You are receiving this mail because: You are the assignee for the bug.
_______________________________________________ Libreoffice-bugs mailing list Libreoffice-bugs@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/libreoffice-bugs
