Re: [Bug-AUCTeX] 11.81; *output* buffer is set to UTF-8

[David Kastrup]

Peter Dyballa <address@hidden> writes:

> Am 04.11.2005 um 23:41 schrieb David Kastrup:
>
>>
>>> Therefore AUCTeX should set coding-system of its
>>> *output* buffer correctly.
>>
>> What coding does your source code buffer have?
>>
>
> LaTeX file is in ISO 8859-15 (0 in mode-line), the *output* log buffer
> in UTF-8 (U in mode-line). Some German umlauts are displayed as
> \<octal code>.

Well, strictly speaking, the coding system of the output log buffer is
completely irrelevant (I remember now) and is typically in the mode of
the environment.  What _is_ relevant is the process output encoding
system from the TeX process, and that should be the same as the buffer
from which you start compilation.

preview-latex does the following contortion to get there:

          (when (featurep 'mule)
            (setq preview-coding-system
                  (with-current-buffer commandbuff
                    buffer-file-coding-system))
            (when preview-coding-system
              (setq preview-coding-system
                    (preview-buffer-recode-system
                     (coding-system-base preview-coding-system))))
            (set-process-coding-system
             process preview-coding-system))

Would you say that in general, preview-latex seems to work correctly?
If it does, then it would probably be time that we started integrating
the run buffer/error handling stuff in preview-latex and the rest of
AUCTeX.

preview-latex does a quite better job currently IIRC.  At least for
non-interactive TeX sessions, the fast location conversion hooks might
be desirable: if one converts all error locations at once, then
next-error will behave more robustly with regard to errors, and we
also could implement something like an error overview buffer, or put
error marks in the margin of the buffers.

-- 
David Kastrup, Kriemhildstr. 15, 44793 Bochum