Bug#119531: latex/3568: inclusion wish for latin10.def

[Frank (bugs reply) <latex-bugs@latex-project.org>]

Hilmar,

 > Down here in the Debian Bug Tracking system we've got the request to
 > include latin10.def to support input encoding for the Romanian
 > language according to ISO8859-16. Additionally to the input encoding
 > it defines 2 letters (s and t with comma below), which are not part
 > of the EC-fonts. Char #164 can probably replaced by \texteuro . The
 > complete bug report is available on http://bugs.debian.org/119531 .
 > Please give us a short statement if you refuse the inclusion.

i have no principal objection in including latin10 support in core LaTeX, on
the contrary. however, the presented file is _not_ a proper inputencoding
definition file and i'm grateful if you do not _in this form_ include files in
the debian tetex packages.

general comments:

input encoding declarations should only refer (in the second argument) to
LICRs (LaTeX Internal Character Representations), eg

 > \DeclareInputText{170}{\ooalign{S\crcr\hidewidth\raise-.31ex\hbox{\scriptsize,}\hidewidth}}

is a nogo

similarly \euro is wrong as it should be \texteuro which is  the official LICR
(the fact that some package uses different a name is irrelevant (just like the
fact that i also prefer typing \euro to \texteuro :-)


LICRs that are not provided by the kernel but by, say, textcomp, should be
properly set up as

\ProvideTextCommandDefault{\texteuro}
   {\TextSymbolUnavailable\texteuro}

thus, \textdegree and perhaps others should be handled in this way

as for

 > \DeclareInputText{170}{\ooalign{S\crcr\hidewidth\raise-.31ex\hbox{\scriptsize,}\hidewidth}}
 > \DeclareInputText{186}{\ooalign{s\crcr\hidewidth\raise-.31ex\hbox{\scriptsize,}\hidewidth}}
 > \DeclareInputText{222}{\ooalign{T\crcr\hidewidth\raise-.31ex\hbox{\scriptsize,}\hidewidth}}
 > \DeclareInputText{254}{\ooalign{t\crcr\hidewidth\raise-.31ex\hbox{\scriptsize,}\hidewidth}}

i would like to get to a decision what the LICR name should be, something like
\C{T} is out of question for various reasons.

questions:

 - what are the unicode names? if any?
 - are there similar chars with comman-under-accent?

it might be that something like \textundercomma (or something equally
horrible:-) might be the best but suggestions are welcome

whatever the final name the latin10.def should go

\providecommand\textundercomma[1]{....}

offering that accent withought compromising the general interface

assuming i get such a file i'm happy to include it in the upcoming release

good night
frank

ps what is meant by:

%%
%% Latin10 is also coming with support for the German double quotations.
%% You have to use babel with a language that support those quotations,
%% German and Romanian come now in my mind...
%%

pps:

 > %% The comma below accent for S, s, T and t doesn't look good
 > %% for large characters. A solution would be to include internal
 > %% support for comma below in the same way like for the dot below,
 > %% so \C{t} to create the t comma below, etc.

that might be possible after we have agreed on the LICR name, nevertheless a
\providecommand is probably in order and it might be the best to start out in
this way