Am Freitag, den 29.03.2019, 15:30 +0000 schrieb Guenter Milde: > I want to bypass "auto" in order to keep its functionality as-is: > > * Currently, the "auto" setting basically¹ means > > use a mix of fixed-width 8-bit encodings
It means: select the most appropriate encoding for the language(s).
> For documents with languages using the Latin, Greek, Cyrillic, and
> Hebrew script, this allows to encode all supportedt letters even in
> cases where utf8 would break something.
>
> ¹ For some languages it will be a mix of fixed-width and variable-
> width
> encodings (10 out of the 82 languages with Babel support use
> utf8 or legacy variable-width encodings.)
>
> * When changing Encodings in lib/languages, "auto" becomes basically
> an
> alias for "utf8", because "utf8" will be the new default for all²
> languages.
>
> ² Only "Japanese (platex)" currently requires "utf8-platex".
Which is in line with the above definition.
> * With my proposal, the "auto" setting is kept available as-is but
> not
> used for new documents (unless set in a custom template).
So it is basically deprecated.
> To make clear that this is now a legacy setting, it may be renamed.
> (We could also rename "Encoding" in lib/languages to something like
> "LegacyInputEncoding".)
My proposal, instead, is to add a new tag in languages,
FixedWithEncoding, which holds the old encodings for things such as
listings.
> The proposed format conversion ("auto" to current buffer default
> encoding) will break the export tests for
> autotests/export/latex/CJK/en-de-el-ru-zh_CN_language-default.lyx and
> autotests/export/latex/en-ar_language-default.lyx. Try compiling
> after changing the inputencoding to "Western European (ISO 8859-15)"
> manually.
As said, I will consider that and fix it, if it breaks.
> If going ahead with a file format change, please do the changes in a
> feature branch, because reverting a file format change is not nice.
We will just add another file format change, if everything fails.
> > In any case, just painting over the language's default encoding
> > everywhere is not a sane solution.
>
> In my opinion, using "utf8" instead of the language's legacy default
> encoding for all languages is simple and sane.
As argued, I disagree.
Jürgen
>
> Günter
>
signature.asc
Description: This is a digitally signed message part
