\documentclass{article}
\usepackage{parskip}
\usepackage{booktabs}

\usepackage{lmodern}
\usepackage{textcomp}
\usepackage[LGR,T1]{fontenc}

% UTF8 input encoding for pdfTeX or LuaTeX in 8-bit compatibility mode:
% (XeTeX in compatibility mode would require xetex-inputenc.sty, which is not
% not at CTAN but on https://github.com/wspr/xetex-inputenc)
\usepackage[utf8]{luainputenc}

\usepackage[pdfencoding=auto,colorlinks=true]{hyperref}
\usepackage{bookmark}
% hyperrefs PU encoding supports Greek LICR macros
\DeclareTextCommand{\ensuregreek}{PU}[1]{#1}

\usepackage[normalize-symbols] % comment option out to test error reporting
	   {textalpha}

\begin{document}

\title{The \emph{textalpha} package}
\author{Günter Milde}
\maketitle

Standard 8-bit TeX fonts do not support the Greek script. If Greek
characters shall be used with 8-bit TeX, setup of a Greek text font and an
explcit font-encoding switch is required.

With the \emph{textalpha} package, you can easily write a single Greek
symbol (like \textPsi{} or \textmu{}) or a
\textlambda\textomicron\textgamma\textomicron\textvarsigma{} in non-Greek
text as well as ISO-conforming formulas with upright constants (like
\textpi): $A = \mbox{\textpi} r^2$ vs. $A = \pi r^2$. Input is possible via
LICR macros (\verb|\textalpha| \ldots \verb|\textOmega|)
or Unicode literals\footnote{%
Requires \emph{\href{http://www.ctan.org/pkg/greek-inputenc}{greek-inputenc}}
and the utf8 option or XeTeX/LuaTeX.}.

\tableofcontents

\section{Usage}

Load this package in the preamble of your document with
\begin{verbatim}
  \usepackage{textalpha}
\end{verbatim}
If it is loaded after the setup of Unicode fonts with
\href{http://www.ctan.org/pkg/fontspec}{\emph{fontspec}} package under
LuaTeX and XeTeX, it provides a compatible interface for Greek in text mode.

If you want to merge symbol variants, use the
\texttt{normalize-symbols} option:%
\footnote{This option is ignored with Unicode fonts.}
\begin{verbatim}
  \usepackage[normalize-symbols]{textalpha}
\end{verbatim}

See the source of this document \url{textalpha-doc.tex} for a setup and
usage example, the literate source of the package
\href{textalpha.sty.html}{textalpha.sty} for the implementation, and
\emph{Font setup for Greek with XeTeX/LuaTeX}
(\href{euenc-doc.pdf}{euenc-doc.pdf}) for special feature under XeTeX/LuaTeX.


\subsection{option \texttt{normalize-symbols}}

This option is ignored with Unicode fonts.

Mathematical notation uses variant shapes of some Greek letters as
additional symbols.  There are separate code points for the symbol variants
in Unicode but not in the LGR font encoding used for Greek in 8-bit TeX.
TeX supports some of the variant shape symbols in mathematical mode
($\theta|\vartheta, \phi|\varphi, \pi|\varpi, \rho|\varrho,
\epsilon|\varepsilon$).

The variations have no syntactic meaning in Greek text and text fonts may
use the variant shapes in place of the “regular” ones as a stylistic choice.
However, some Greek texts in Unicode encoding use these GREEK
... SYMBOL characters in place of the corresponding GREEK LETTER ...
characters.

The \texttt{normalize-symbols} option merges letters and symbols to Greek
letters. This way, text copied from external sources can be compiled without
errors even if it contains GREEK SYMBOL characters in place of GREEK LETTERS:
\begin{quote}
  This text uses both variants for beta (β|ϐ), theta (θ|ϑ), phi (φ|ϕ), pi
  (π|ϖ), kappa (κ|ϰ), rho (ρ|ϱ), Theta (Θ|ϴ), and epsilon (ε|ϵ) in the LaTeX
  source.
\end{quote}

\textbf{Attention}: Do not use this option in cases where the distinction
between the symbol variants may be important (e.g. in a mathematical or
scientific context). Try the \emph{alphabeta} package with the
respective characters in math mode or use XeTeX/LuaTeX with Unicode fonts in
these cases.

\section{Limitations}

Because the internal font encoding switch interferes with other work behind
the scenes, kerning, diacritics and up/down-casing show problems if Greek
letters are used without explicit change of the font encoding.

These problems can be avoided by use of \emph{babel} and the correct
language setting (greek), an explicit font encoding switch (e.g.
wrapping the Greek text in \verb|\ensuregreek|%
\footnote{The \texttt{\textbackslash ensuregreek} macro ensures the argument
  is set in a font encoding supporting Greek.
  % This can be used to fix these
  % problems without adverse side-effects if the active font encoding is
  % already LGR or Unicode.
  },
or XeTeX/LuaTeX with Unicode fonts.


\subsection{Diacritics}

Composition of diacritics (like \verb|\accdasia\acctonos| or \verb|\<\'|)
fails in other font encodings. Long names (like \verb|\accdasiaoxia|) work.

With LGR, pre-composed glyphs are chosen if available. In other font
encodings, accent macros do not select pre-composed characters. (The
difference is a sub-optimal placement of the accent and becomes obvious if
you drag-and-drop text from the PDF version of this document.):

\begin{quote}
  \ensuregreek{\<'a \accdasia\acctonos a \accdasiaoxia a
  \accdasiaoxia\textalpha} (LGR) vs. \accdasiaoxia\textalpha{} (T1).
\end{quote}

According to Greek typographical tradition, diacritics (except the
dialytika) are placed before capital letters in Titlecase and dropped in
UPPERCASE:
%
\begin{quote}
  \ensuregreek{%
    \<{\textalpha} \>{\textepsilon} \"'{\textiota} \`>\texteta{}
    \'<{\textomicron} \~<{\textupsilon} \~>{\textomega}
    \quad
    \<{\textAlpha} \>{\textEpsilon} \"{\textIota} \`>\textEta{}
    \'<{\textOmicron} \~<{\textUpsilon} \~>{\textOmega}
    \quad
    \MakeUppercase{%
      \<{\textalpha} \>{\textepsilon} \"'{\textiota} \`\>\texteta{}
      \'<{\textomicron} \~<{\textupsilon} \~>{\textomega}
    }
  }
\end{quote}
%
This fails for accent macros if the active font encoding is not LGR or
Unicode. Pre-composed literal Unicode characters are handled correctly:
\begin{quote}
  \ensuregreek{\'\textAlpha/Ά}  (LICR/Unicode in LGR) vs.
  \'\textAlpha/Ά (LICR/Unicode in T1).
\end{quote}

The dialytika marks a \emph{hiatus} (break-up of a diphthong). It must be
present in UPPERCASE even where it is redundant in lowercase (the hiatus can
also be marked by an accent on the first character of a diphthong). The
auto-hiatus feature works in LGR and Unicode font encodings only:
\begin{quote}
  \acctonos\textalpha\textupsilon{}, \acctonos\textepsilon\textiota{} $\mapsto$
  \MakeUppercase{\ensuregreek{
    \acctonos\textalpha\textupsilon{}, \acctonos\textepsilon\textiota{}
  }} (LGR) vs.
  \MakeUppercase{
    \acctonos\textalpha\textupsilon{}, \acctonos\textepsilon\textiota{}%
  } (T1)
\end{quote}

Currently, the second vowel of the diphthong must be given as macro, not
Unicode literal if the auto-hiatus feature should work:
\begin{quote}
\ensuregreek{ἀ\textupsilon{}πνία} $\mapsto$
\ensuregreek{\MakeUppercase{\ensuregreek{ἀ\textupsilon{}πνία}}} (LGR) vs.
\ensuregreek{\MakeUppercase{\ensuregreek{ἀυπνία}}} (T1).
\end{quote}

\subsection{Kerning}

No kerning occurs between Greek characters in non-Greek text due to the
internal font encoding switch: compare
\ensuregreek{\textAlpha\textUpsilon\textAlpha} (LGR) to
\textAlpha\textUpsilon\textAlpha (T1).

Compiling with LuaTeX provides kerning also over font encoding boundaries.

\section{Test and Examples}

\subsection{Greek alphabet}

Greek letters via Latin transcription in LGR font encoding:

\begin{quote}
  \ensuregreek{A B G D E Z H J I K L M N X O P R S T U F Q Y W}\\
  \ensuregreek{a b g d e z h j i k l m n x o p r sv c t u f q y w}
\end{quote}

Greek letters via default macros in other font encoding (here T1):

\begin{quote}
  \textAlpha{} \textBeta{} \textGamma{} \textDelta{} \textEpsilon{}
  \textZeta{} \textEta{} \textTheta{} \textIota{} \textKappa{}
  \textLambda{} \textMu{} \textNu{} \textXi{} \textOmicron{} \textPi{}
  \textRho{} \textSigma{} \textTau{} \textUpsilon{} \textPhi{}
  \textChi{} \textPsi{} \textOmega{}
  \\
  \textalpha{} \textbeta{} \textgamma{} \textdelta{} \textepsilon{}
  \textzeta{} \texteta{} \texttheta{} \textiota{} \textkappa{}
  \textlambda{} \textmu{} \textnu{} \textxi{} \textomicron{} \textpi{}
  \textrho{} \textsigma{} \textvarsigma{} \texttau{} \textupsilon{}
  \textphi{} \textchi{} \textpsi{} \textomega{}
\end{quote}

Archaic Greek letters and Greek punctuation

\begin{quote}
  \textDigamma
  \textanoteleia
  \textQoppa
  \textSampi
  \textStigma
  \textnumeralsigngreek
  \texterotimatiko
  \\
  \textdigamma
  \textkoppa
  \textqoppa
  \textsampi
  \textstigma
  \textnumeralsignlowergreek
  \textvarstigma
\end{quote}

Diacritics

\begin{quote}
  Symbol macros:%
  \footnote{Composite diacritics require wrapping in
  \texttt{\textbackslash ensuregreek}.}
  \"{} \'{} \`{} \~{} \<{} \>{} \u{} \={}
  \ensuregreek{\"~{} \"'{} \"`{} \<~{} \<`{} \<'{} \>~{} \>'{} \>`{}}

  Named macros:
  \accdialytika{}
  \acctonos{}
  \accvaria{}
  \accperispomeni{}
  \accdasia{}
  \accpsili{}
  \ypogegrammeni{}
  \prosgegrammeni{}
  %
  \accdialytikaperispomeni{}
  \accdialytikatonos{}
  \accdialytikavaria{}
  \accdasiaperispomeni{}
  \accdasiavaria{}
  \accdasiaoxia{}
  \accpsiliperispomeni{}
  \accpsilioxia{}
  \accpsilivaria{}
  %
  \accinvertedbrevebelow{} % == \textsubarch{}
  \accbrevebelow{}


\end{quote}

Accent macros can start with \verb|\a| instead of \verb|\| when the
short form is redefined, e.\,g. inside a \emph{tabbing} environment.
This also works for the new-defined Dasia and Psili shortcuts:
\begin{quote}
  \begin{tabbing}
       COL1\quad \= COL2\quad \= COL3\quad \= COL4\quad \\
       COL1 \>          \> COL3 \\
       Viele    \> Gr\a"u\ss e
       \> \greekscript \a<\textalpha{}
       \> \greekscript \a>\textomega
  \end{tabbing}
\end{quote}


\subsection{Greek Unicode characters in non-Greek text}

With the \emph{textalpha} package,
\href{http://www.ctan.org/pkg/greek-inputenc}{greek-inputenc} and input
encoding \texttt{utf8}, Greek Unicode characters can be used in text with
any font encoding. See Tables \ref{tab:greek-and-coptic} and
\ref{tab:greek-extended}.

Kerning is preserved if the font encoding is LGR: \ensuregreek{AΫA}

\begin{table}[tbp]
\centerline{
\begin{tabular}{rrrrrrrrrrrrrrrrr}
\toprule
    & 0 & 1 & 2 & 3 & 4 & 5 & 6 & 7 & 8 & 9 & A & B & C & D & E & F\\
\midrule
370 & * & * & * & * & ʹ & ͵ & * & * &   &   & ͺ & * & * & * & ; &  \\
380 &   &   &   &   & ΄ & ΅ & Ά & · & Έ & Ή & Ί &   & Ό &   & Ύ & Ώ\\
390 & ΐ & Α & Β & Γ & Δ & Ε & Ζ & Η & Θ & Ι & Κ & Λ & Μ & Ν & Ξ & Ο\\
3A0 & Π & Ρ &   & Σ & Τ & Υ & Φ & Χ & Ψ & Ω & Ϊ & Ϋ & ά & έ & ή & ί\\
3B0 & ΰ & α & β & γ & δ & ε & ζ & η & θ & ι & κ & λ & μ & ν & ξ & ο\\
3C0 & π & ρ & ς & σ & τ & υ & φ & χ & ψ & ω & ϊ & ϋ & ό & ύ & ώ &  \\
3D0 & * & * & * & * & * & * & * & * & Ϙ & ϙ & Ϛ & ϛ & Ϝ & ϝ & * & ϟ\\
3E0 & Ϡ & ϡ & * & * & * & * & * & * & * & * & * & * & * & * & * & *\\
3F0 & * & * & * & * & * & * & * & * & * & * & * & * & * & * & * & *\\
\bottomrule
\end{tabular}
} % end centerline
\caption{Greek and Coptic Unicode Block, input as literal Unicode
   characters in T1 font encoding (legend: * glyph missing in LGR).}
\label{tab:greek-and-coptic}
\end{table}


\begin{table}[tbp]
\centerline{
\begin{tabular}{rrrrrrrrrrrrrrrrr}
\toprule
     & 0 & 1 & 2 & 3 & 4 & 5 & 6 & 7 & 8 & 9 & A & B & C & D & E & F\\
\midrule
1F00 & ἀ & ἁ & ἂ & ἃ & ἄ & ἅ & ἆ & ἇ & Ἀ & Ἁ & Ἂ & Ἃ & Ἄ & Ἅ & Ἆ & Ἇ\\
1F10 & ἐ & ἑ & ἒ & ἓ & ἔ & ἕ &   &   & Ἐ & Ἑ & Ἒ & Ἓ & Ἔ & Ἕ &   &  \\
1F20 & ἠ & ἡ & ἢ & ἣ & ἤ & ἥ & ἦ & ἧ & Ἠ & Ἡ & Ἢ & Ἣ & Ἤ & Ἥ & Ἦ & Ἧ\\
1F30 & ἰ & ἱ & ἲ & ἳ & ἴ & ἵ & ἶ & ἷ & Ἰ & Ἱ & Ἲ & Ἳ & Ἴ & Ἵ & Ἶ & Ἷ\\
1F40 & ὀ & ὁ & ὂ & ὃ & ὄ & ὅ &   &   & Ὀ & Ὁ & Ὂ & Ὃ & Ὄ & Ὅ &   &  \\
1F50 & ὐ & ὑ & ὒ & ὓ & ὔ & ὕ & ὖ & ὗ &   & Ὑ &   & Ὓ &   & Ὕ &   & Ὗ\\
1F60 & ὠ & ὡ & ὢ & ὣ & ὤ & ὥ & ὦ & ὧ & Ὠ & Ὡ & Ὢ & Ὣ & Ὤ & Ὥ & Ὦ & Ὧ\\
1F70 & ὰ & ά & ὲ & έ & ὴ & ή & ὶ & ί & ὸ & ό & ὺ & ύ & ὼ & ώ &   &  \\
1F80 & ᾀ & ᾁ & ᾂ & ᾃ & ᾄ & ᾅ & ᾆ & ᾇ & ᾈ & ᾉ & ᾊ & ᾋ & ᾌ & ᾍ & ᾎ & ᾏ\\
1F90 & ᾐ & ᾑ & ᾒ & ᾓ & ᾔ & ᾕ & ᾖ & ᾗ & ᾘ & ᾙ & ᾚ & ᾛ & ᾜ & ᾝ & ᾞ & ᾟ\\
1FA0 & ᾠ & ᾡ & ᾢ & ᾣ & ᾤ & ᾥ & ᾦ & ᾧ & ᾨ & ᾩ & ᾪ & ᾫ & ᾬ & ᾭ & ᾮ & ᾯ\\
1FB0 & ᾰ & ᾱ & ᾲ & ᾳ & ᾴ &   & ᾶ & ᾷ & Ᾰ & Ᾱ & Ὰ & Ά & ᾼ & ᾽ & ι & ᾿\\
1FC0 & ῀ & ῁ & ῂ & ῃ & ῄ &   & ῆ & ῇ & Ὲ & Έ & Ὴ & Ή & ῌ & ῍ & ῎ & ῏\\
1FD0 & ῐ & ῑ & ῒ & ΐ &   &   & ῖ & ῗ & Ῐ & Ῑ & Ὶ & Ί &   & ῝ & ῞ & ῟\\
1FE0 & ῠ & ῡ & ῢ & ΰ & ῤ & ῥ & ῦ & ῧ & Ῠ & Ῡ & Ὺ & Ύ & Ῥ & ῭ & ΅ & `\\
1FF0 &   &   & ῲ & ῳ & ῴ &   & ῶ & ῷ & Ὸ & Ό & Ὼ & Ώ & ῼ & ´ & ῾ &  \\
\bottomrule
\end{tabular}
} % end centerline
\caption{Greek Extended Unicode Block, input as literal Unicode
characters in T1 font encoding.}
\label{tab:greek-extended}
\end{table}

Combined Diacritics work ᾅ, diacritics (except diaeresis) are dropped with
MakeUppercase (μαΐστρος $\mapsto$ \MakeUppercase{μαΐστρος}).

\subsection{PDF strings}

With \emph{textalpha} and
\emph{\href{http://www.ctan.org/pkg/greek-inputenc}{greek-inputenc}}, there
are two options to get Greek letters in PDF strings: LICR macros and literal
Unicode input.

\subsubsection{\textlambda\textomicron\textgamma\textomicron\textvarsigma{},
         λογος and \ensuregreek{logos}}

The subsection title above uses: LICR macros, Unicode input and the LGR
transcription for the Greek word \ensuregreek{logos}. Check the table of
contents in the PDF viewer: LICR macros and Unicode literals work fine, the
Latin transcription remains Latin in the PDF metadata.

\end{document}