On 2017-01-14, Richard Heck wrote:
> On 01/14/2017 12:11 AM, Scott Kostyshak wrote:
>> On Sat, Jan 14, 2017 at 05:56:21PM +1300, gordon cooper wrote:
>> Thanks for reporting this problem. Next time it would be helpful if you
>> sent an example file. I went ahead and created an example file and I can
>> reproduce the same problem you saw. The example file is attached. If you
>> open LyX in the terminal, you see the following message when compiling:
>> Error 84 returned from iconv when converting from UCS-4LE to
>> ISO-8859-15: Invalid or incomplete multibyte or wide character
>> It does not give a clue where the error comes from, and indeed the LyX
>> GUI is not helpful at all in indicating where the error might be.
>> I'm moving this thread to the lyx-devel list now. I think you wanted to
>> just give lyx-users a warning in case they come across the problem, and
>> I think potential discussion of how to improve LyX to handle this case
>> should proceed on lyx-devel.
> This kind of error does seem to crop up quite a bit. It'd be nice if we
> could say something more informative.
> Could we at least say what the character was? Or report its location in
> the file?
With an 8-bit input encoding or "forced" characters like °, LyX reports
<LyX Warning: uncodable character '°'> in the source pane and also
the character in question in the Exort Error dialogue.
Reporting the location in the file would be a bonus:
With just one "exotic" character in the document, a search would be possible
(the only problem is, that drag and drop from the Error dialogue does not
work).
With characters that only fail in verbatim context, a search in a larger
document is cumbersome.
With "inputenc utf8", LyX only reports uncodable characters for "forced"
conversion. LyX has currently no correct idea about which characters fail to
be converted by [utf8]{inputenc} in verbatim context and the LaTeX error
"TeX capacity exceeded" does not help either.
Günter
