Using \catcode to define characters gets me wierd output and error messages

Question

I am using the following code to redefine a breve in order to work with a particular truetype font I must use (Elegant Garamond) and the particular piece of code generates an error (but still works) and it generates output on the typesetted document.

\documentclass[10pt,a4paper,openany]{book}
\usepackage[utf8]{inputenc}
\usepackage[T1]{fontenc}
\usepackage[romanian]{babel}

\usepackage{eg}       %this is the definition for the font I use
\pdfmapfile{+eg.map}

\newcommand\chara{a}

\makeatletter
\def\kernbrevea#1{\kern#1\dimexpr0.47em-\strip@pt\fontdimen1\font\dimexpr-0.2em\relax\relax}
\makeatother

\begin{document}        %begining here avoids one error that \catcode line below generates

\catcode`\ă=\active
\def ă{\raisebox{.0em}\u\kernbrevea{-}{\chara}}%

text with ă character
\end{document}

Compiling the file gives me two errors:

! Missing number, treated as zero.
<to be read again> 
                   \protect 
l.17 \catcode`\ă
                 =\active
A number should have been here; I inserted `0'.
(If you can't figure out why I needed to see a number,
look up `weird error' in the index to The TeXbook.)


! Package inputenc Error: Keyboard character used is undefined
(inputenc)                in inputencoding `utf8'.

See the inputenc package documentation for explanation.
Type  H <return>  for immediate help.
 ...                                              

l.17 \catcode`\ă
                 =\active
You need to provide a definition with \DeclareInputText 
or \DeclareInputMath before using this key.

The output is an expected "text with ă character" but just before, on the first line, there is a string: "=,". As in the screencap:

Is there a way to supress it or to solve the errors I get?

I have read the suggested part in The TeXbook but I can't find the logic (well, I barely know TeX/LaTeX). Obviously my code is broken, but I can't find in what way or if there are some problems with my font...

Answer 1

You need a different structure to handle UTF-8 with pdftex (but see below for different solution):

\documentclass[10pt,a4paper,openany]{book}
\usepackage[utf8]{inputenc}
\usepackage[T1]{fontenc}
\usepackage[romanian]{babel}

%\usepackage{eg}       %this is the definition for the font I use
%\pdfmapfile{+eg.map}

\newcommand\chara{a}

\makeatletter
\def\kernbrevea#1{\kern#1\dimexpr0.47em-\strip@pt\fontdimen1\font\dimexpr-0.2em\relax\relax}
\makeatother
\DeclareUnicodeCharacter{0103}{\raisebox{.0em}\u\kernbrevea{-}{\chara}}

\begin{document}

text with ă character
\end{document}

---

However the above is still really not the correct approach : it is tying the input encoding to a specific font encoding, and as you have seen breaks if the file is encoded as latin1 or latin2 rather than UTF8. Also it does not address the standard latex markup for the character, \u{a}

So better is just to declare \u{a} to be your raisebox construct. Then \u{a} will use your definition and ă expands to \u{a} so will automatically get the new definition and so neither the catcode tricks of the previous question nor \DeclareUnicodeCharacter are needed.

\documentclass[10pt,a4paper,openany]{book}
\usepackage[utf8]{inputenc}
\usepackage[T1]{fontenc}
\usepackage[romanian]{babel}

%\usepackage{eg}       %this is the definition for the font I use
%\pdfmapfile{+eg.map}


\makeatletter
\def\kernbrevea#1{\kern#1\dimexpr0.47em-\strip@pt\fontdimen1\font\dimexpr-0.2em\relax\relax}
\makeatother

%\DeclareTextComposite{\u}{T1}{a}{160}% T1 default
\DeclareTextCompositeCommand{\u}{T1}{a}{\raisebox{.0em}\u\kernbrevea{-}{a}}
\begin{document}

text with ă character and \u{a}
\end{document}