>>>>> "Georg" == Georg Baum <[EMAIL PROTECTED]> writes:
Georg> On Sunday 24 December 2006 01:02, Dov Feldstern wrote: >> Okay, this is the issue: if a paragraph's language is Hebrew, then >> the encoding is identified correctly using "use language's default >> encoding". However, if the paragraph's language is English, but you >> have some Hebrew words in it, then LyX complains about the >> encoding. Switching manually to cp1255 solves this. Perhaps auto >> identification of the encoding by language is done at the paragraph >> level, not the individual character level? Georg> Yes. The first character determines the encoding of the whole Georg> pragraph. I don't know why, maybe it is latex limitation. The inputenc documentation states: % Originally this command was only to be used in vertical mode (with % the idea that it should be only within a document when % using text from several documents to build up a composite work such % as a volume of journal articles. However, usages in certain % languages suggested that it might be preferable to allow changing % the input encoding at any time, which is what is possible now % (though that is quite computing resource intensive). This means that older inputenc versions only worked as paragraph-based, but newer ones can do better. It is not clear to me at what time the change happened, though. While looking at that, I also noticed an inputenx.sty package that may be useful... http://www.tex.ac.uk/tex-archive/macros/latex/contrib/oberdiek/ JMarc
