On Sun, 21 Dec 2003 23:55:22 -0800 David Hampton <[EMAIL PROTECTED]> wrote:
> > Of course it is possible, but it is not trivial. The aim is to > > compile gnucash with libxml2, but the existing parser relies on a > > bug of libxml1. I don't think you can use the parser the same way as > > it works now with libxml2. > > Please explain further. What bug does gnucash rely on? I've been > using libxml2 in the gnome2 branch without any noticeable problem. > I am glad to hear that. I thought that libxml2 guesses differently from libxml1. Because gnucash passes strings in the locale encoding to libxml, libxml1 assumes a latin1 string while libxml2 assumes the string to be utf-8. If in the gnome2-branch all strings are handled in UTF-8 internally, libxml2 will guess correctly. But I wonder how it will handle old files that were created under an EUC-JP encoding. Libxml1 had encoded the EUC-JP multi-bytes in single latin1 pieces. Now when I open the same file with libxml2 how will it handle putting together the decoded latin1 pieces to clean multi-byte EUC-JP? I will try to find some time to compile the gnome2 branch and have a look how it works with multi-byte encodings. Reinke _______________________________________________ gnucash-devel mailing list [EMAIL PROTECTED] http://www.gnucash.org/cgi-bin/mailman/listinfo/gnucash-devel
