[prev in list] [next in list] [prev in thread] [next in thread] 

List:       lyx-devel
Subject:    Re: iconv errors and deurl.py
From:       Guenter Milde <milde () users ! berlios ! de>
Date:       2009-03-31 8:47:41
Message-ID: gqslbd$3f5$1 () ger ! gmane ! org
[Download RAW message or body]

On 2009-03-27, John McCabe-Dansted wrote:
> On Fri, Mar 27, 2009 at 8:43 PM, Guenter Milde wrote:
>> Setting Document>Settings>Language>Encoding to UTF8 (ucs enhanced/utf8x)
>> solves this particular problem.

> Would this be a sensible default?

No, because 
a) it does not solve all cases
b) the usc package is said to be experimental and no longer maintained.

But using one of the Unicode encodings is a good way to locate the problem.

>> IMO, it is correct not to use the LaTeX transcription from
>> "unicodesymbols" for ERT, as ERT is "raw LaTeX".

> One option would be to use the transcription when pasting into the
> ERT. Although then LyX would have to know which characters were valid
> in the current encoding, and handle the case of the encoding of a
> document being changed in someway.

This would go a wide way from the simple "ERT is raw source: LyX uses it
verbatim, you use it at your own risk" philosophy.

>> LyX should at least show it in a pop-up (similar to LaTeX errors), so
>> that the user has immediate feedback on why there is no reaction. Even
>> better: parse it and report the location.

> Would a python script that generated a LaTeX style error log be helpful?
> E.g., like
>   http://wiki.lyx.org/Tools/LyX-GrammarChecker
> but in python?

> I can write C/C++, but python sounds more maintainable in this case.

Why not a patch for LyX's C++ sources?

Günter

[prev in list] [next in list] [prev in thread] [next in thread]