Lars Gullik Bjønnes wrote: > Angus Leeming <[EMAIL PROTECTED]> writes: > > | Jean-Marc Lasgouttes wrote: > | > Angus> In Python, you can write: unicode(string,'latex+latin1') > | > Angus> http://aspn.activestate.com/ASPN/Cookbook/Python/Recipe/252124 > | > | > Angus> so I guess that we could define a similar converter and pop up > | > Angus> an error box if a character cannot be converted. > | > | > Yes, if the number of characters we want to translated is small > | > enough. > | > | I'd hope that we could either cut and paste a character array or, > | better, use libiconv to do the work for us. > | > | #include <iconv.h> > | > | iconv_t iconv_open(const char *tocode, const char *fromcode); > | > | size_t iconv(iconv_t cd, > | char **inbuf, size_t *inbytesleft, > | char **outbuf, size_t *outbytesleft); > | > | int iconv_close(iconv_t cd); > | > | In other words, use ICU to create a UTF-8 version of the latex file and > | then pass this through iconv to generate a latin-1 encoding. > > That will work "just like that" in a lot of cases, but remmeber that > our .lyx file does not really have a defined norm.
Surely, it will always work with all "new" LyX files created with the unicoded LyX 1.5? .lyx files saved by LyX 1.5 will be UTF-8 encoded, right? Assuming that this is the case, the problem will be one of converting pre-LyX 1.5 files to LyX 1.5 format. Ie, a lyx2lyx problem. I'd anticipate that lyx2lyx would assume that the .lyx file it receives as input was encoded in the system locale unless told otherwise. It should probably emit a warning about this and abort if it comes across illegal characters. Similarly, it should be possible to specify the encoding of the input file implicitly. Just thinking out loud... -- Angus
